1. Artificial Intelligence as a Game Changer

AI is no longer just a buzzword. In 2025, it will further revolutionize marketing by analyzing data in real-time, predicting customer behavior more precisely, and creating personalized experiences automatically.

AI-Powered MarTech Applications:

• Predictive Analytics: AI analyzes historical data to predict trends and customer preferences, enabling companies to act more targeted.
• Automated Content Creation: Generative AI produces personalized content for social media, emails, and websites.
• Conversational Marketing: AI-powered chatbots optimize customer interaction around the clock.
• Personalized Advertising: Dynamic ads adapt to user behavior in real-time.
• Optimization of the Customer Journey: AI identifies bottlenecks and increases the efficiency of marketing measures.

2. Personalization vs. Data Protection: A Balancing Act

Customers expect personalized content, while the protection of their data is becoming increasingly important. The GDPR and the end of third-party cookies require new strategies.

Strategies for Success:

• Use First-Party Data: Companies increasingly rely on their own customer data instead of third-party cookies.
• Collect Zero-Party Data: Customers consciously share information through surveys and interactions.
• Data Protection-Friendly Personalization: AI groups customers into anonymous segments instead of creating individual profiles.
• Privacy-Enhancing Technologies (PETs): Technologies like Differential Privacy enable data protection-compliant analyses.
• Improve Consent Management: Transparent communication and flexible opt-in options strengthen customer trust.

3. The Fusion of MarTech and AdTech

Previously separate systems are growing together to enable a unified customer approach. This development increases the efficiency and ROI of marketing campaigns.

Important Developments:

• Data-Driven Advertising Automation: Integration of MarTech and AdTech data to optimize campaigns.
• Omnichannel Marketing: Unified customer communication across social media, email, websites, and digital advertising.
• Customer Data Platforms (CDPs): Centralized customer data for consistent and personalized campaigns.
• Programmatic Advertising: Automated buying and selling of advertising space in real-time.
• Attribution Modeling: Optimized budget allocation through more precise channel evaluation.

Conclusion: Actively Shape the Future

MarTech will become more intelligent and data protection-oriented in 2025. Companies that focus early on AI-powered analyses, data protection-compliant personalization, and the integration of MarTech and AdTech will secure long-term competitive advantages. However, it is crucial to use these technologies responsibly to maintain customer trust and ensure sustainable success.

1. In-House Focus: Insights from Leading Brands

This year's MarTech Summit stood out by featuring numerous brand-side speakers. Hearing directly from major advertisers about their in-house strategies offered unique perspectives on challenges and opportunities in MarTech. Many brands shared how they are optimizing internal processes to enhance agility and minimize reliance on external partners.

2. High-Quality Networking in an Open Atmosphere

In addition to insightful panels and presentations, the event provided excellent networking opportunities. The open-minded and friendly atmosphere encouraged meaningful conversations between brands, tech providers, and industry experts. It was a refreshing contrast to more sales-driven events, fostering real knowledge exchange rather than just pitching solutions.

3. AI Readiness Starts with Data Quality

One of the strongest themes throughout the summit was AI readiness. Before fully leveraging AI, companies recognize the need to acquire, store, and manage high-quality data. This was reaffirmed during multiple panel discussions, where audience polls consistently ranked data quality as the top priority. Many brands are currently focused on structuring their data in a way that ensures effective AI implementation in the near future.

Final Thoughts

The Martech Summit successfully brought together industry leaders to discuss both strategic and operational MarTech developments. With a clear emphasis on in-house transformation, data quality, and AI readiness, the event provided valuable insights for brands looking to future-proof their marketing efforts.

But what are the differences between the three now-available open-source solutions, and what limitations exist?

We have been active as a Marketing Mix Modeling provider for over five years and continuously work with clients to increase the efficiency of their marketing activities.

We have taken a close look at the various open-source solutions for you and compared them with our self-developed MMM method, Scope.

Key Insights

1. Hierarchical Modeling

With Meridian from Google, it is now possible to use data from different geographic regions when training the model, even with an open-source solution. Training on granular data helps a model tremendously to determine the effect of media activities more accurately and reliably. This gives you a much more valid basis for making decisions to optimize your media activities, provided you have the necessary data from multiple regions.

2. Flexible Media Efficiency Over Time

Two of the open-source solutions assume by definition that the effect of a media variable is constant over time. On the one hand, this makes sense because the amount of data available in most MMM projects is only sufficient to determine the average effect of a media variable over the entire period under consideration with reasonable certainty. On the other hand, this does not correspond to reality, in which it is crucial for the media effect when and how media is played out in what form.

3. User Interface for Use Without Coding

Without coding knowledge, I cannot cope with any of the freely available open-source solutions. All three solutions are provided in the form of code that you can download and then execute to get the results.

Conclusion:

The ability to model hierarchically on regions is another big step forward for the capabilities of the available open-source solutions. The prerequisite for use remains coding knowledge or at least the ability to execute code and deal with error messages.

Furthermore, the release of Meridian by Google shows that Google has also recognized that the previously available digital attribution methods are no longer up-to-date and sufficient to evaluate the impact of marketing and media activities. The importance of Marketing Mix Modeling as the most important standard tool in the marketing measurement toolkit is underlined once again by this step.

How MMT can help you further?

The data available for Marketing Mix Modeling always remains the same, regardless of the solution used. We have developed a platform where you can easily use all existing open-source solutions without coding knowledge to train meaningful models, and we also provide our own, even more powerful method that eliminates some of the current limitations of the open-source solutions.

Without much effort, this allows you to take a comparative look at the results from the different solutions. Which model comes to which conclusions and why? This look at the different model results gives the user a deep understanding of how the instrument, Marketing Mix Modeling, and the effect of your media activities work. This deep understanding will enable you to make the right decisions for more media efficiency. Depending on your needs, you can either be the user of the platform yourself or leave the use to one of our MMM experts.

Marketing measurement encompasses the methods, tools, and processes used to evaluate the effectiveness of marketing activities. It's about quantifying the impact of your marketing efforts. Often, the focus is on the impact of various marketing channels and campaigns due to data availability. However, marketing measurement also includes quantifying impact at a granular level, such as evaluating the effectiveness of individual ad placements, as seen in many attribution models.

Why is Marketing Measurement Important?

Anyone investing in marketing should know the return on their investment. Only by understanding the impact of your marketing activities in detail can you optimize effectively and get the most out of your campaigns. Marketing measurement is a crucial task that every company should take seriously – the higher the advertising spend, the more critical this task becomes.

The primary goal of marketing measurement is clear: increase the efficiency of advertising spending and get the maximum return from every euro invested.

Key Components of Marketing Measurement

Modern and effective marketing measurement in 2024 relies on the combination of three methods that complement each other:

Marketing Mix Modeling (MMM): Quantifies the relationships between marketing activities and key business KPIs (such as sales, revenue, web visits, etc.), even beyond marketing activities. This provides an ideal basis for optimizing marketing activities and valuable support for internal communication with stakeholders beyond marketing.

Experimentation: Allows you to examine cause-and-effect relationships that cannot be clarified by marketing mix modeling due to the data situation. For example, when newly emerging marketing channels have never been used in the past or only at a very low level.

Always-on Measurement and Attribution: With very granular data, the statements from marketing mix modeling can also be questioned and additional questions answered that marketing mix modeling cannot answer, such as, "Which display creative has a better click-through rate?" "With which SEA creative do we generate more revenue?".

If you want to focus on one method first, you should start with marketing mix modeling, as it is the connecting element between the three methods and offers the best potential to make your marketing activities more efficient directly through the use of a method. Depending on the business model and needs, you can then use the other two methods in addition to leverage further potential.

Challenges in Marketing Measurement

The basis for value-creating marketing measurement is data that arises around marketing activities. Data that comes from different sources, must be merged from different systems and must be continuously updated. The ever more diverse mix of marketing activities means that the number of sources continues to increase. Nevertheless, stricter data protection regulations must be observed and alternatives found to understand the surfing behavior of users in the future despite cookie regulation.

It is important to ensure that the resources spent on marketing measurement are in a healthy relationship to the total marketing investment. Every euro you invest in measuring and analyzing marketing activities must also pay off in the form of increased marketing efficiency.

The Future of Marketing Measurement

Modern and particularly value-creating marketing measurement will be heavily influenced by technical innovation in order to be able to use it more cost-effectively and effectively. The market of software-as-a-service providers will consolidate. At the same time, new players will keep emerging who will try to stand out with hype topics and set themselves apart from the competition. As an advertiser, I can best prepare for this and benefit from market developments if I collect and store the data surrounding my own marketing activities as granularly as possible and over longer periods of time.

Conclusion

Marketing measurement is an important task within marketing to ensure high efficiency of advertising expenditure. The higher the annual advertising expenditure, the more important it is to establish marketing measurement processes in the company. In general, it is important to ensure that the use of resources for marketing measurement is in a healthy relationship to advertising expenditure. The best starting point for value-creating and effective marketing measurement is the marketing mix modeling method.

Jochen Schweizer mydays offers a wide range of experiences and gift ideas, relying on optimal marketing in a dynamic competitive environment. The collaboration with MMT enables a deeper understanding of the performance of media activities and the integration of these insights into media planning. A key focus is on the interplay between upper- and lower-funnel activities. The goal is to derive clear recommendations for action that support both short-term and long-term business success.

Jochen Schweizer mydays anticipates valuable impulses from the collaboration with MMT, leading to a continuous increase in marketing efficiency. Comprehensive initial modeling and regular updates conducted within the MMM module ensure maximum transparency and foster trust in the insights gained—an essential prerequisite for taking action based on these findings. Additionally, Jochen Schweizer mydays expects valuable insights beyond media activities through the holistic consideration of influencing factors.

"With MMT, we have found a competent partner who provides us with valuable insights into our media activities. Thanks to their transparent approach and ongoing collaboration, we can trace the modeling results in detail through Mercury’s MMM module and have the flexibility to bring the marketing mix modeling processes in-house via self-service at any time."

Jessica Bielmeier
Team Lead Paid Media, Jochen Schweizer mydays Group

"Jochen Schweizer mydays, like us, has a strong hands-on mentality. This makes the collaboration particularly enjoyable because valuable insights can be quickly operationalized, and efficiency improvements are realized early in the partnership. With Mercury’s self-service MMM module, we can react flexibly and efficiently to new challenges with every model update."

Torben Seebrandt
Director Data & Intelligence, MMT

About the Jochen Schweizer mydays Group

The companies of the Jochen Schweizer mydays Group offer experiences, travel adventures, and gifts anytime, anywhere, and for a wide variety of occasions. Whether thrilling adventures, impressive outdoor activities, or relaxing experiences for quality time with loved ones—the team has made it their mission to work alongside thousands of experience partners to provide the perfect experience for every customer. Since October 2017, the two leading experience providers, Jochen Schweizer and mydays, have been united under the umbrella of the Jochen Schweizer mydays Group as a member of ProSiebenSat.1 Media SE.

Press Contact:

Jochen Schweizer mydays Group

Laura Werner

presse@jochen-schweizer.de

++++++

About Mercury Media Technology GmbH & Co. KG

Since 2016, Mercury Media Technology has been an independent and growing MarTech company based in Hamburg. Agencies, advertisers, and publishers across Europe are among its clients. Over 30 experts in software and data engineering, data intelligence, and media develop data-driven solutions to maximize the efficiency of media operations. The "Mercury" Media Operations platform, which is part of the Marketing Resource Management segment, enables data-driven media management. From the self-service Marketing Mix Modeling module, which provides actionable insights for media planning, to strategic and tactical planning, booking, and automated dashboards in the integrated reporting center, Mercury supports the entire media workflow. Data pipelines are in place with relevant technical systems, including delivery systems such as ad servers and DSPs, all social platforms, website tracking tools, and validation systems. Additionally, all external and internal data is cleanly prepared and structured via the "Bridge" data connector.

The company, headquartered in Hamburg, is managed by Tobias Irmer, Gunnar Neumann, and Andreas Sand.

More information at mercurymediatechnology.com

Press Contact:

Mercury Media Technology GmbH & Co. KG

Klostertor 1

20097 Hamburg

Michelle Tejada

hello@mercurymediatechnology.com

In this article, we offer a perspective to answer these questions and address a possible distinction between different formats of in-house media buying.

Content

• Technical in-housing

• Hybrid in-housing

• Full in-housing

• Full in-housing: a utopia?

At the first glance, in-house media buying looks like an enticing concept for advertising companies. After all, it promises full control over the company’s data and budgets as well as independence from agencies and other intermediaries. Indeed, 73% of brands have already brought at least some parts of their digital marketing back under their control. But inhousing is not a “one size fits all” approach: There are different models, which vary in suitability depending on the size and focus of the advertising company. How brands organize their media buying and what role an agency plays as a partner in the process depends largely on internal expertise, financial and time resources, and the ad tech setup. First of all, a distinction must be made between the different formats of in-house media buying: technical in-housing, hybrid in-housing, and full in-housing.

Technical in-housing

Technical inhousing is when companies license technical infrastructures themselves, e.g. for media buying and/or full campaign management. For example, they can license ad servers, DSPs, or verification services themselves or set up a Facebook advertising account independently. Companies take control over technical systems and conclude direct contracts with corresponding service providers without using an agency as an intermediary. However, the agency then works operationally in the licensed system & receives appropriate accesses.

The major advantage of technical inhousing is that advertisers can potentially save money, as they do not need a partner for the complex technical processes and can enter into direct negotiations with providers. In addition, their own data infrastructure can be connected directly based on their resources. This breaks down data silos, reduces dependencies and complexity, and gives companies direct access to all campaign information. Advertisers can also manage the access rights of service providers to different platforms. Accordingly, they gain a significant advantage in control and can continue their work without any extra efforts or major data leaks, e.g. after changing the agency.

However, advertisers need to either have the appropriate expertise for this endeavor or take time to build it up first, or they can call in external consulting. Relevant capacity should also be available in Purchasing and IT. This concerns the initial setup, but it’s comparatively manageable. Ongoing control is also possible with a small number of own employees.

Any company that can muster these resources is capable of implementing technical inhousing. For many advertisers, therefore, taking media buying into their own hands is a good starting point.

Hybrid in-housing

In hybrid inhousing, advertisers also work together with agencies. For example, they take over individual process steps or media types and hand over everything else to an agency. For example, the company can take control over negotiations with publishers and thus remove the purchasing component from the agency's tasks. Or it can take care of the digital advertising and commission a service provider with the analog channels. It is even conceivable to take over individual sub-steps, such as budget allocation among various channels with subsequent handover to the agency for operational implementation. Overall, many forms of hybrid collaboration can be found on the market.

One argument in favor of this is that advertisers can flexibly take over individual aspects of the entire media process, depending on their available resources, thus gaining more independence and knowledge. They can leverage their strengths and competencies, and work closely internally.

However, this requires agencies willing to embrace this model - and there are still too few of those.

Companies that decide in favor of hybrid inhousing should carefully consider in advance which components of media buying they could manage internally in the long run - even in case of headcount fluctuations. 

Full in-housing

The most complex format of self-directed media buying is full inhousing. In this case, companies work together autonomously and without agency support. The company handles all aspects of media buying internally - from technical infrastructure setup and strategic planning to implementation and evaluations.

The advantages are clear: advertisers have full control over their budgets, data, and content. Brand sovereignty remains 100% within the company and there is no dependence on other service providers, except for walled gardens.

However, this entails a high investment of time and money to provide personnel, develop know-how, and build up infrastructure. In-house experts are also in high demand in the labor market and therefore difficult to retain permanently.

Therefore, full inhousing is initially particularly suitable for smaller companies that only use a few advertising channels. Large advertisers are faced with an almost insurmountable task when aiming for full inhousing. Here we recommend a step-by-step approach based on the intermediate solutions described above. Of course, it is also important to have some flexibility and, depending on the progress, the process should be accelerated or decelerated. In addition, recourse should be taken to practice and expert consulting can be engaged throughout the process. 

Full in-housing: a utopia?

Thesis: Large companies that want to - and also have to - be present in all channels cannot sustainably afford fully in-house media buying. Not only do they have to train and retain the necessary talent, but they also have to constantly respond to changes in the market. Excellently trained and positioned teams also need someone to drive innovation. So it's less a question of initial resources and more a question of long-term cost efficiency. Small companies, on the other hand, that regard media buying as a manageable expense, eventually reach the limits of full inhousing along with their growth - after all, every company wants to develop. So for many advertisers, hybrid inhousing is the model of the future.

And yet, there are still too few agencies embracing this format and offering it to the customer as well as it is an opportunity. 

The renowned fashion retailer Peek&Cloppenburg Hamburg has been collaborating with Mercury Media Technology (MMT) since early 2024 to optimize its media mix and sustainably improve footfall and ad-driven revenue. To achieve this, Peek&Cloppenburg Hamburg leverages Mercury's Marketing Mix Modelling module, which precisely evaluates the impact of individual marketing activities across specific sales regions and provides recommendations for greater marketing efficiency.

Peek&Cloppenburg Hamburg has been a key player in the German fashion retail sector for decades. With numerous brick-and-mortar stores primarily in northern and eastern Germany and its online shop VAN GRAAF, the company offers customers a carefully curated, high-quality range of textiles and both physical and digital shopping experiences. Its more than 100 years of success demonstrate that P&C Hamburg has consistently made the right decisions to maintain its market presence over the long term.

Driven by macroeconomic factors, the textile trade in Germany—and P&C Hamburg—faces significant challenges, making an efficient marketing mix essential. To address these challenges, Peek&Cloppenburg Hamburg sought a marketing measurement partner and, since March 2024, has relied on MMT. Based on the positive experience of the first six months, the company plans to continue and expand this partnership over the coming years.

"With Mercury's Marketing Mix Modelling module, we can precisely determine which media activities contribute to revenue increases, both overall and in specific regions. We can also identify the optimal combination of media activities for each region to maximize incremental revenue across the board"

Torben Seebrandt
Director Data & Intelligence, MMT

"With MMT, we have found a reliable partner to help us use our advertising budgets more efficiently. The open and personal collaboration, combined with a high degree of flexibility, has truly impressed us"

Katja Hünnekens
Director Media/Online/PR, Peek&Cloppenburg Hamburg

About Peek&Cloppenburg KG Hamburg

The Peek&Cloppenburg KG Hamburg Group* is one of the leading companies in the textile retail sector, currently operating a total of 41 stores in Germany and Europe. In locations such as Poland, Switzerland, Hungary, Latvia, and the Czech Republic, as well as through its online shop, the company operates exclusively under the internationally recognized name VAN GRAAF. Through its VANGRAAF.COM platform, which features over 500 fashion labels, Peek&Cloppenburg KG Hamburg continuously integrates a growing number of partner products. The marketplace, along with the VAN GRAAF app, is a cornerstone of its omnichannel strategy, seamlessly combining virtual and in-store shopping experiences.

*There are two independent companies named Peek&Cloppenburg, headquartered in Düsseldorf and Hamburg, respectively. Our client is the Peek&Cloppenburg KG Hamburg Group. For more information about the Hamburg-based group, visit: https://www.peek-und-cloppenburg.de/de/unternehmen/

+++++

Press Contact:
Peek&Cloppenburg KG
Mönckebergstraße 8
20095 Hamburg
Katja Hünnekens
Katja.Huennekens@peek-und-cloppenburg.de

About Mercury Media Technology GmbH & Co. KG

Founded in 2016, Mercury Media Technology is an independent and growing MarTech company based in Hamburg. Its clients across Europe include agencies, advertisers, and publishers. Over 30 experts in software and data engineering, data intelligence, and media develop data-driven solutions aimed at maximizing the efficiency of media operations.

The "Mercury" media operations platform, which also falls under the broader category of marketing resource management, enables data-driven media management. From the self-service Marketing Mix Modelling module, which delivers actionable insights for media planning, to strategic and tactical planning, booking, and automated dashboards in the integrated reporting center, Mercury supports the entire media workflow.

The platform includes data pipelines to relevant technical systems, including delivery systems like ad servers and DSPs, all social platforms, website tracking tools, and validation systems. All external and internal data is also cleanly prepared and structured via the "Bridge" data connector.

The company is headquartered in Hamburg and led by managing directors Tobias Irmer, Gunnar Neumann, and Andreas Sand. For more information, visit: mercurymediatechnology.com

Press Contact:
Mercury Media Technology GmbH & Co. KG
Klostertor 1
20097 Hamburg
Michelle Tejada
hello@mercurymediatechnology.com

From deep dives into new methodologies to thought-provoking use cases, the conference left me with countless takeaways, and I wanted to share some of the key insights and reflections from my time there.

LLM-Based Conversational Agents in Action

It was fascinating to see how companies like Feedly , Stepstone , and Lufthansa are not just experimenting with large language models (LLMs), but actively integrating them to tackle real-world problems.

What stood out was their focus on developing robust evaluation metrics and feedback mechanisms , ensuring that these systems are continuously improving. A key takeaway here is the critical role of human-in-the-loop systems , which empower LLM-based tools to adapt and deliver value in meaningful and measurable ways.

These efforts go well beyond simple chatbots, showing just how versatile and impactful LLMs can be when applied strategically.

Beyond Chatbots: Innovative Generative AI Use Cases

Generative AI was a huge theme at the conference, and the use cases presented really opened my eyes to the diverse ways organizations are leveraging this technology:

Team Internet Group is transforming the way domain name searches are conducted by using Generative AI to make the process smarter and more intuitive.

Campana Schott is working on AI applications in healthcare, demonstrating how Generative AI can provide tangible solutions to complex medical challenges.

Zeiss is using Generative AI to tackle duplicate data detection with impressive precision, addressing a persistent problem in data-intensive industries.

These examples reinforced how Generative AI isn’t just a buzzword—it’s driving real revenue and tackling challenges across sectors I hadn’t even considered before.

DATEV’s AI Playground

One of the most exciting revelations came from DATEV, the company behind the payroll processing systems we often rely on. They’ve built an AI Playground , a space for experimentation that includes operational tools, beta tests, and pilot projects.

By embracing state-of-the-art LLMs, DATEV is tackling both straightforward and complex challenges. It was inspiring to see how they’ve created a structured environment for continuous testing and iteration, proving that a culture of experimentation is critical to staying ahead in this fast-moving field.

A Deep Dive into Reinforcement Learning from Human Feedback (RLHF)

For those of us fascinated by the mechanics of AI, the session on Reinforcement Learning from Human Feedback (RLHF) was one of the most insightful. It explored key techniques like Proximal Policy Optimization (PPO) and Direct Policy Optimization (DPO) and their importance in advancing Generative AI models.

RLHF stood out to me as a game-changing methodology—it shows how human feedback shapes AI systems to make them smarter and more aligned to real-world goals. This nuanced approach highlighted where the future of AI is heading: systems that learn and adapt in a way that feels increasingly “human.”

Traditional Machine Learning Techniques Still Hold Their Ground

Amid the buzz around Generative AI, it was refreshing to see that traditional machine learning techniques are still at the forefront of solving real-world problems. Two sessions were particularly memorable:

Causality in Machine Learning: Steffen Wagner delivered a thought-provoking session on causal inference, including the use of Double Machine Learning (Double ML) , to understand deeper cause-and-effect relationships. It was a reminder that foundational methodologies are just as critical as newer ones.

Video Analytics with Computer Vision: Merantix Momentum showcased how combining computer vision techniques like keypoint detection algorithms with models such as XGBoost can create effective solutions for video analysis challenges.

Hearing these examples reminded me that a hybrid approach—taking the best of both traditional ML and emerging AI techniques—often yields the most impactful results.

Personal Connections and Reflections

One of the most rewarding aspects of Machine Learning Week wasn’t just the sessions, but the personal connections I made during coffee breaks and lunches. Engaging in conversations with industry leaders, researchers, and fellow practitioners gave me fresh perspectives on the challenges we face as professionals in this space.

These discussions not only sparked meaningful reflections but also left me inspired to take action. I came away with new ideas and strategies that I’m excited to explore and implement at Mercury Media Technology .

Looking Ahead

Attending Machine Learning Week 2024 provided me with a diversified and balanced view of AI implementation in the real world. Whether it was diving into cutting-edge developments in LLMs, exploring innovative AI use cases, or revisiting traditional ML techniques, the event left me with a renewed sense of excitement about the possibilities ahead.

As AI continues to evolve, I’m inspired to apply these insights to create solutions that drive real value for our teams and clients. The future is full of opportunities, and I’m excited to help Mercury Media Technology stay at the forefront of innovation.

Content:

• What is Marketing Mix Modeling (MMM)?

• Why should marketers use Marketing Mix Modeling?

• What are the limitations of Marketing Mix Modeling?

• How to build a Marketing Mix Model?

What is Marketing Mix Modeling (MMM)?

Marketing Mix Modeling (MMM) is a powerful tool used to measure and analyze the effectiveness of marketing activities. It helps companies understand how their marketing investments influence their target KPIs, identify which marketing activities are most effective, and accurately predict future performance. MMM enables marketers to optimize their marketing mix in order to get the maximum return on their marketing investments.

Marketing Mix Modeling (MMM) is a data-driven statistical analysis technique. Data scientists develop advanced multivariate statistical models that analyze the contribution of numerous drivers to target KPIs.

To put it simply, Marketing Mix Modeling is an essential tool for marketers looking to get the most out of their campaigns. It helps better understand how different elements of the marketing strategy such as advertising, social media, and promotions are working together with external factors, e.g. weather, seasonality, or competitors' activity.

By measuring the impact of each marketing activity on the ROI, businesses can make smarter decisions which marketing activities to invest in and which ones to cut. The goal is to forecast the impact of future campaigns, reallocate marketing budgets and optimize the media mix and promotional tactics in order to maximize the target KPI, e.g. sales revenue or profit.

Which models are used in the MMM?

Linear regression model

The linear regression model is the simplest form of analysis for marketing mix modelling. It helps to analyse the relationship between two variables - for example, how turnover (dependent variable) changes depending on marketing expenditure (independent variable).

Independent variables are factors such as the level of marketing spend, which are considered as influencing variables. Dependent variables are the results that are influenced by these variables, such as sales.

The model describes the relationship between the variables by drawing a straight line through the data points. This line illustrates the relationship between the independent variable and the dependent variable. A steeply rising line indicates that higher marketing expenditure tends to lead to higher sales. A flat line, on the other hand, shows that marketing expenditure only has a minor influence on sales.

The linear regression model is particularly suitable for analysing the direct influence of a single marketing measure on a business result. It is particularly useful in scenarios in which a linear relationship between the variables is expected.

Multiple regression model

The multiple regression model extends the linear regression approach by analysing the influence of several independent variables simultaneously on one or more dependent variables. This makes it possible to understand the combined effect of different marketing activities such as TV advertising, digital campaigns or promotions on sales or other KPIs. This model provides a more comprehensive overview of the marketing mix and is particularly suitable for more complex scenarios in which several factors influence the result.

Bayesian model

The Bayesian model incorporates prior knowledge or assumptions into the analysis and updates them as new data becomes available. This gives the model particular strength in dealing with uncertainty and variability in the marketing data. This method allows for more flexible and dynamic modelling, especially when the available data is sparse or uncertain. The model is ideal when prior knowledge can be incorporated into the modeling or when complex and uncertain environments exist where traditional methods may reach their limits.

Hierarchical model

The hierarchical model, also known as the multi-level model, analyses data that is structured at different levels, such as regions, branches or customer segments. It takes into account variations at each of these levels, providing detailed and segmented insights. This model is particularly useful for organisations that want to understand the impact of marketing activities in different markets or segments and how they differ and interact with each other at different levels.

Which target KPIs could be optimized with MMM?

Marketing Mix Modeling enables the optimization of a variety of Key Performance Indicators (KPIs) that are directly related to the effectiveness of marketing activities. By analysing and adjusting these KPIs, companies can precisely align their marketing strategies and achieve better business results.

The following KPIs can be optimized by MMM:

• Sales revenue
• Conversions
• Transactions
• App installs or Downloads
• Registrations
• Leads
• Customer Lifetime Value
• Brand Awareness
• ROI (Return of Investment)
• Customer Acquisition Cost

Why should marketers use Marketing Mix Modeling?

Marketing Mix Modeling is an essential tool for marketers who like taking data-driven decisions based on valuable insights, aiming to refine the marketing strategy for optimal results.

With an MMM solution, marketers can identify the most effective channels that contribute to sales, ROI, or other target KPIs. Based on this, they can make better decisions about media budget allocation and refine their marketing mix in order to maximize the ROI.

With this technology, marketers can also gain a clear understanding of the effectiveness of their campaigns. As a result, they can use the insights provided by these solutions when building new or refining existing campaigns to ensure their success. Additionally, marketers can also measure the short-term and long-term effects that campaigns generate.

MMM also helps understand the impact of external factors such as the macro- and microeconomic environment, competition, and seasonality on the target KPI.

What are the benefits of Marketing Mix Modeling?

• Develop an understanding of what influences the target KPI
• Optimize current and future marketing campaigns
• Allocate budgets efficiently across the media mix

What are the limitations of Marketing Mix Modeling?

While MMM is an immensely useful tool for marketers, it comes with a few limitations.

Firstly, the accuracy of the results heavily depends on the data quality and completeness of the dataset used. If the data is inaccurate or incomplete, the output of the analysis might be distorted. The time series length of the data set is also important. To get a high-quality model, data from the last few months is not enough. Depending on the model, an average of three years of data may be sufficient.

It's also important to take all (or at least most) relevant factors into account. For example, if we are modeling sales of ice cream and don't take the weather into account, the model's accuracy will always be inferior.

Additionally, modeling techniques are only as good as the software and algorithms used for analysis. If these are not updated frequently, marketers may see subpar performance compared to models using more advanced methods driven by recent machine learning developments, or sophisticated approaches like bayesian modeling.

How to build a Marketing Mix Model?

Step 1 - Define the objectives: The first step of Marketing Mix Modeling is defining the objectives relevant to your business. You should consider what success will look like in terms of profitability and sales.

Step 2 - Collect data: Before you get started with Marketing Mix Modeling, you need to collect enough data to measure the outcomes accurately. This data should include sales figures, competitor activity, media variables (both contacts e.g. reach or impressions, and spend), promotions, and external factors like weather, inflation rate, and holidays as well. Using this data, you'll be able to identify trends and create insights about your own business performance.

Step 3 - Analyze data: Now it's time for the analysis of all the data collected previously. This process usually involves running different regression algorithms such as simple linear models or multiple regressions in order to isolate the impact of each component of the marketing mix on sales or profitability in specific markets or regions.

Step 4 - Determine the impact of activities: Using the results of the analysis, marketers can now test various hypotheses related to their campaigns without having to make any actual changes. This process helps determine which activities are effective and which ones are not, so the marketers can refine their budgets even further with optimization techniques to maximize the effectiveness and ROI across channels

Step 5 - Generate insights: Marketing Mix Modeling allows marketers to generate high-level insights into how they can optimize their activities in order to generate more leads or increase conversions at a lower cost per acquisition (CPA).

Step 6 - Monitor & adjust strategies: After analyzing the impact of individual factors on the target KPI, it’s important for marketers who do this type of modeling to follow up on its performance by monitoring campaigns against corresponding goals and metrics over time on a regular basis so they know whether their strategies are achieving the desired results or need adjustment as conditions change.

Step 7 - Reassess the objectives: Finally, once the results of these strategies have been tracked over a certain period of time, one should reassess their original goals and objectives based on recent performance metrics so that accurate forecasts of marketing ROI and growth possibilities can be made going forward to support executives and stakeholders in the decision-making process.

What are the most frequently asked questions about marketing mix modeling?

How is the model created?
Historical data on the target KPIs is combined with the relevant influencing factors to create a modeling data set. Each entry in the data set corresponds to an experiment with measured values for the target KPIs and the influencing factors (the more granular the data, the more experiments the data set contains, the more reliably the algorithm learns the cause-effect relationships. Variables are analyzed). Mathematical models are trained on the basis of the data set, which learn the cause-effect relationships.

What results does an MMM deliver?
A marketing mix model (MMM) quantifies the impact of marketing activities on KPIs such as sales or market share and optimizes budget allocation. It analyzes marginal benefit curves to show when additional spending in a channel is less effective. It also calculates the ROI per channel and helps to identify the most efficient ones. The MMM makes recommendations for optimal budget usage and simulates different budget scenarios to maximize revenue and profit. It provides data-based insights for future campaigns and budget decisions.

How often should the model be updated?
Updates are valuable because they can determine whether or not actions taken have led to greater media efficiency. As a standard, it is recommended to start with quarterly updates. However, depending on the brand, business model and data situation, it may make sense to update the models more frequently (e.g. monthly, weekly)

What data is required?
Historical data is required for each week for at least two years retrospectively. Data is required for at least one target KPI and for the relevant influencing factors, in particular, of course, media activities, but also sales activities (discounts, promotions), for example. Depending on the brand, business model and target KPI, the relevant influencing factors can vary greatly. In addition, the more detailed the data available, the more reliable and valuable the models trained on it will be. For example, it is better if the data is available per day and not per week. It is even better if the data is available per day and per federal state, for example.

What are the costs?
The costs vary depending on the scope of an MMM project. A key driver is usually the number of different brands and countries to be modeled. As a rule of thumb, an MMM project with quarterly updates that is to be carried out for one brand and one country can be expected to cost around €25-50k per year.

Which industries benefit most from marketing mix modeling?
Ultimately, many industries that have media spend across multiple channels can effectively use MMMs to optimize their marketing approach. However, the benefits increase with increasing complexity, i.e. the more budget, channels or markets are involved.

Here you can find more information on: The most frequently asked questions about Marketing Mix Modeling.

Conclusion

Marketing mix modeling is a powerful tool for measuring and optimizing the performance of marketing activities. With efficient marketing mix modeling, marketers can gain valuable insights into their advertising campaigns and make informed decisions based on the data. The potential impact of marketing mix modeling is huge, and it is a must-have tool for any marketer who wants to maximize the value of their marketing campaigns.

&Beyond with the media agencies Infinity Media and Vertitas Media stand for progress and innovation in the industry. They are open to new challenges and are not afraid of change. The curiosity firmly rooted in their DNA drives &Beyond to find creative and intelligent solutions for their clients. They are proactive and always up to date with the latest trends and technologies.

Media agencies still use Excel to manage their clients' campaigns. As this error-prone method cannot be the basis for data-driven processes, &Beyond set out to find a future-oriented MarTech solution that would support them in streamlining and automating media processes, starting with strategic planning, through tactical detailed planning and booking, to reporting. They came across the Hamburg-based tech provider Mercury Media Technology and opted for the media operations platform “Mercury”. From January 2025, all &Beyond agencies will work together across clients in Mercury, making communication faster and easier. Workflows will become more efficient through standardization and automation, as content templates can be used and data is automatically transferred to the billing system. This gives &Beyond full transparency over its purchasing and sales processes. 

"With &Beyond, we have gained our first, but certainly not our last client in the Spanish market. Talks are already underway with other agencies. We are particularly pleased that &Beyond is an agency that is constantly looking for innovations and wants to develop further. We are happy to support them as an equal partner."

Gunnar Neumann
Managing Director, MMT

"At &Beyond, we believe that all advertisers, regardless of their size, deserve personalized services and full access to the best professionals and technologies in the industry. With Mercury, MMT has developed innovative software for media agencies that helps us to offer our clients the best possible service. As an agency that is always on the lookout for innovation, we find it very important that the software is constantly being developed to keep up with the times."

Rai Pérez
Chief Innovation Officer, &Beyond

About &Beyond Content Media Data Group 

Founded in 2008, it exceeded 80 million euros in turnover in 2023. Formed by the agencies Veritas Media and Infinity Media, it currently works for more than 50 clients and has 70 professionals in its offices in Barcelona and Madrid. 

Leading the Agency Scope prepared by Scopen with Veritas Media in first place and Infinity Media in second based on customer ratings. 

Grupo &Beyond is part of Local Planet, the largest global network of independent media agencies, bringing together 62 agencies in 85 countries with a turnover of 17.2 Bn and 12,000 professionals. 

The &Beyond Group is led by Albert Gost as Executive Chairman, Enrique de la Torre as CEO and Laia Regués completes the Executive Committee in her position as CFO. 

Press contact:
Calle de El Españoleto 17 1 Planta
28010 Madrid España
+34917941270
Miguel.justribo@andbeyond.es

About Mercury Media Technology GmbH & Co. KG

Mercury Media Technology has been an independent, growing MarTech company based in Hamburg since 2016. Its clients include agencies, advertisers and publishers across Europe. Over 30 experts from the fields of software and data engineering, data intelligence and media develop data-driven solutions with the aim of maximizing the efficiency of media operations. The Media Operations Platform “Mercury”, which can be counted as part of the wider Marketing Resource Management segment, enables data-driven media management. Starting with the self-service marketing mix modeling module, which provides actionable insights for media planning, through strategic and tactical planning and booking to automated dashboards in the integrated reporting center, Mercury supports the entire media workflow. There are data pipelines to relevant technical systems including delivery systems such as ad servers and DSPs, all social platforms, website tracking tools and validation systems. All external and internal data is also available in a clean, structured format via the “Bridge” data connector.

The managing directors of the Hamburg-based company are Tobias Irmer, Gunnar Neumann and Andreas Sand.

More information at mercurymediatechnology.com

Press contact:
Mercury Media Technology GmbH & Co. KG
Klostertor 1
20097 Hamburg
Vivian Reifschneider
hello@mercurymediatechnology.com

DMEXCO 2024 in Cologne proved yet again why it’s the must-attend global event for digital marketing and tech experts. This year, however, something felt refreshingly different – a shift from glamour to in-depth content , and that made a huge impact on us at MMT! 

Key Themes, Meaningful Networking, and Direct Impact on MMT

The agenda highlighted critical areas like AI in marketing, advanced automation, and sustainability . But what truly stood out was the practicality and depth of conversations. These haven’t stayed buzzwords – they’ve evolved into actionable strategies that MMT can directly leverage in our projects moving forward.

One of the most noticeable and welcome changes was the less flashy, more substantive approach. For MMT, this shift was a game-changer . The space created by this focus on content allowed us to engage in more impactful discussions . We connected on a deeper level with innovative companies from all over the digital ecosystem. The conversations weren’t just surface-level networking – they were strategic and offered concrete opportunities for collaboration and growth. Several big leads are now in the pipeline, and the feedback we received on our own developments has been incredibly positive.

Despite the more toned-down vibe, DMEXCO continues to be a global leader for innovation – one of the best places for companies like MMT to understand the market’s future direction. From AI advancements to cutting-edge automation tools , the event gave us insights that directly influence MMT’s strategic planning .

"From my perspective, DMEXCO has truly found its place, even at this large scale. It has established itself as a genuine global beacon for top speakers, leading companies, and rising stars."

Gunnar Neumann
Managing Director

Inspiration, Team Strength, and Our Takeaway

We appreciated the relaxed atmosphere that allowed for high-quality, stress-free conversations . With less overcrowding and more breathing room, real connections were made. Plus, the perfect sunny weather added an extra layer of enjoyment! 

The event also worked wonders for our internal team, reinforcing our camaraderie and shared vision . We walked away feeling more unified and motivated , with renewed energy and an even clearer focus for what’s ahead at MMT.

"I leave DMEXCO feeling energized and motivated. The chance to connect with industry peers, make new contacts, and spend meaningful time with the team is truly inspiring."

Torben Seebrandt
Director Data & Intelligence

Takeaway: DMEXCO 2024 Reaffirms MMT’s Strategic Direction

DMEXCO isn’t about chasing size or festival vibes – it’s about substance . This year’s focus on deep connections, genuine insights, and actionable takeaways perfectly complements who we are as a company. We’ve returned to MMT full of fresh ideas, valuable leads , and new opportunities in sight. DMEXCO isn’t just an event for us – it’s a key catalyst for our continued growth .

Missed the Chance to Speak with Us at DMEXCO?

We’d love to connect! If you didn’t catch us at DMEXCO 2024 or have any questions about how MMT can help your business thrive, don't hesitate to reach out.

With this objective in mind, Benjamin Moser and Ahmad Hoteit knew from the outset that they needed a modern software solution that would enable robust media processes and automate them in a meaningful way, as well as bring together all relevant data and thus provide the basis for developing innovative campaigns. With the Media Operations Platform “Mercury”, pilot suisse has, after an intensive evaluation process, found a software that enables data-driven media management. Both online and classic campaigns can be managed. There are data pipelines to relevant technical systems including delivery systems such as ad servers and DSPs, all social platforms, website tracking tools and validation systems. All external and internal data is also available in a clean, structured format via the “Bridge” data connector. The clean data basis is the foundation for meaningful, automated dashboards in the integrated Reporting Center. 

"We are delighted to have gained another customer in the Swiss market with pilot suisse. We feel well equipped and encouraged to meet the unique challenges of the Swiss market with our Media Operations Platform “Mercury”. We want to gain a foothold in the Swiss market and are looking forward to talks with other agencies.” "

Gunnar Neumann
Managing Director

"With our many years of agency experience, we know that data and tech are not an end in themselves, but tools that make it possible to develop new and innovative solutions. Modern software that makes us fit for the future and enables automation and collaboration with our customers and the market is therefore key. We have evaluated various tools and see MMT as the partner that makes this possible with its “Mercury” software.”

Benjamin Moser
founder and managing partner

About pilot suisse AG

pilot suisse is a new media agency founded in 2024 for the Swiss market for all those who don't want to stand still. For people and brands who want to make a difference and are not satisfied with the status quo. With pilot suisse, the two founders and managing partners Benjamin Moser and Ahmad Hoteit, who are deeply rooted in the Swiss media industry, want to offer Swiss advertisers more effective campaigns with the best and individual consulting services. As long-standing managers at one of Switzerland's leading communications agencies and now part of the owner-managed pilot agency group and the Local Planet network, both have the market knowledge and the network to realize this goal for their clients - simply, effectively, agilely and with short communication channels. With a determined challenger mentality, they are always on the lookout for gaps for their customers and identify opportunities that enable sustainable success. Together, they work on the relevant goals and find ways to move brands forward. Based in Zurich, the agency is ready to tackle the unique challenges of the Swiss market with a determined challenger mentality.

+++++

Press contact:

pilot suisse AG

Giesshübelstrasse 106

8045 Zurich

Benjamin Moser

info@pilot-suisse.ch

About Mercury Media Technology GmbH & Co. KG

Mercury Media Technology has been an independent, growing MarTech company based in Hamburg since 2016. Its clients include agencies, advertisers and publishers across Europe. Over 30 experts from the fields of software and data engineering, data intelligence and media develop data-driven solutions with the aim of maximizing the efficiency of media operations. The Media Operations Platform “Mercury”, which can be counted as part of the wider Marketing Resource Management segment, enables data-driven media management. Starting with the self-service marketing mix modeling module, which provides actionable insights for media planning, through strategic and tactical planning and booking to automated dashboards in the integrated reporting center, Mercury supports the entire media workflow. There are data pipelines to relevant technical systems including delivery systems such as ad servers and DSPs, all social platforms, website tracking tools and validation systems. All external and internal data is also available in a clean, structured format via the “Bridge” data connector.

The managing directors of the Hamburg-based company are Tobias Irmer, Gunnar Neumann and Andreas Sand.

More information at mercurymediatechnology.com

+++++

Press contact:

Mercury Media Technology GmbH & Co. KG

Klostertor 1

20097 Hamburg

Vivian Reifschneider

hello@mercurymediatechnology.com

Composability, i.e. the structure of software in individually usable components, is becoming increasingly important in order to be able to implement individual processes flexibly and efficiently. The software developed in this way meets existing applications in the company, which leads to an increased need for integration.¹ 

"Our Media Operations Platform Mercury has a modular & API-first structure. (e.g. “Self-Service MMM”, “Strategic Planning”, “Projects”, “Data Connectors/Data Connection Service” or our “Buying Module”. Our customers only license the components they need within their workflow. Of course, we integrate with the customer's existing systems & with relevant peripheral systems on the market via API. "

Gunnar Neumann
Managing Director

Artificial intelligence: turbo booster for MarTech

The market potential of MarTech will increase from €7 billion to €15.5 billion between 2023 and 2027 with an average growth rate of +18.8%. AI as a “turbo booster” in the direction of “real-time automation” and analytics will provide a further boost to growth.²

"We can clearly see that the market is approaching us much more strongly than a few years ago and that there is a great need and openness for external solutions like ours. On the one hand, the use of our system is an ideal basis for using AI-based solutions in a targeted manner, and on the other hand, we are also working on solutions within the platform in which AI can support our customers in the long term. For example, we are already using AI in the analytics area of our MMM module and are working on gradually integrating even more intelligence into media planning and the planning process. Theoretically, use cases are possible in which media planning and optimization suggestions can be generated automatically. We discuss the practical applicability & potentials closely with our customers on an ongoing basis."

Gunnar Neumann
Managing Director

Unified data management: standardizing and using data

Even in 2024, only just under 25% of companies state that they have already fully unified and aggregated their available data: from merging data to breaking down data silos and organizing master data.

"Merging and aggregating data into a clean data basis is a key pillar of our Media Operations Platform. ID-based & without complex naming conventions, we merge target values, planned, booked, realized & billed values via the native marketing & media process. Our customers can access these via a “bridge” with their own BI solutions or view meaningful dashboards directly within the platform. Data-based decisions are ideally enabled by Mercury. "

Gunnar Neumann
Managing Director

Causal AI: understanding cause-and-effect relationships

The fundamental problem with most AI models is that the model learns from training data and, in this case, the data can hardly be used to answer counterintuitive questions regarding cause-and-effect relationships. Causal models solve this problem via causal interference.

“We are continuously working on optimizing our marketing mix model: Starting with a regression analysis via a Bayesian MMM, we are now working on a causal model that better incorporates the effects of the individual media channels on each other. When developing and applying AI models, it is crucial for us to uncover the correct cause-and-effect relationships using causal discovery and to validate them using causal inference. Only if we consider the correct causal relationships between the various influencing factors in our MMM projects, for example, during model development can we keep our performance promise to our customers and increase media efficiency.”

Torben Seebrandt
Director Data Intelligence

Every company is unique. However, we believe that the experience we have gained in terms of our culture, structure and way of working can be helpful to other companies. In this series of articles, we would like to give you an insight into how we at Mercury Media Technology (MMT) manage the balancing act between mobile working and maintaining the MMT identity despite hybrid working methods.

To begin with, we would like to give you an overview of the composition of our international team as well as our corporate culture and values. From there, the following articles will take a closer look at the different teams and the work they do.

Table of Contents

• Our Diverse and International Team

• Our Team Composition

• Our Culture and Values

• Conclusion

Our Diverse and International Team

MMT consists of an organically growing team of currently 30 people, whose diversity strengthens our company. Our team currently includes 9 different nationalities and 17 different languages, which is why we have chosen English as our common corporate language. This diversity of cultural and professional backgrounds allows us to bring different perspectives and innovative solutions to the table. Each team member brings a unique set of experiences and ideas that enrich the way we work and help us respond creatively to challenges.

These different approaches and ways of thinking not only foster creativity but also an open and dynamic work environment where we are constantly learning from each other. As a result, we are better able to adapt to change and discover new opportunities. For this reason, the diversity of our team is a key factor in our success and continued growth.

As a tech company, we know that our industry tends to be male-dominated, and we still have a gender imbalance within the company. Therefore, it is important to us to address this issue and make working at MMT attractive to all genders and to meet their different needs. At MMT, everyone should have the same opportunities, regardless of gender, because we are convinced: Diversity leads to better ideas, innovations and ultimately to our joint success.

Our Team Composition

At MMT, there are three teams that contribute to the overall success of MMT with their specialized skills and tasks: the Engineering Team, the Data Intelligence Team, and the Business Team.

The Business Team, which currently consists of seven people, acts as the interface between the client and the two technical teams. In addition to the traditional goals of lead generation, customer acquisition, and revenue growth, the Business Team includes product specialists responsible for implementation and onboarding. This team is responsible for customer care, support, and customer expansion.

The Engineering Team currently consists of 14 members and is divided into the Software Engineering Team, which programs the core functionality of our software, and the Cloud Infrastructure Team, which ensures the operation of our servers and IT infrastructure.

The Data Intelligence Team is comprised of 10 people and is divided into two teams: Data Engineering and Data Analytics. The Data Engineering team develops API connectors to collect marketing (or performance) data and make it available in our software, while the Data Analytics team develops dashboards and reports, handles marketing mix modeling, and integrates AI.

Our Culture and Values

At MMT, we define ourselves by our shared curiosity and vision. This not only helps us deliver great products that benefit our customers but also ensures that they will continue to get the best in the future. We value treating each other with respect and acting responsibly as individuals and as a company. We are optimistic about new challenges and see ourselves as passionate, creative innovators working together on a product that helps other people overcome their challenges.

The physical distance of working remotely makes it difficult to experience the company's vibrant culture. In response, we rely on a combination of in-person and hybrid events. In this way, we want to strengthen the sense of togetherness and bring the MMT culture to life by enabling both remote and onsite team members to participate. For example, our regular game nights are held as hybrid events, and links for online participation are provided for each in-person meeting. We send onboarding kits and goodies to team members' homes and accommodate the availability of remote colleagues for team events.

The diverse cultural backgrounds of our team are an asset, but also require special attention. To promote mutual understanding and appreciation, we regularly organize cultural events and offer language training in various languages. We believe that openness, interest and respect for other cultures are prerequisites for working together. These activities help create an inclusive work environment where all team members feel understood and valued.

Flat hierarchies & positive discussion culture

At MMT, we make it a priority to ensure that all team members, whether junior or team leader, can contribute their ideas and be heard. A flat hierarchical structure and a positive discussion culture are important to us, as they allow for direct and open exchange across all levels.

This ensures that no valuable ideas are lost. Our goal is to create a work environment that allows everyone to actively contribute to our products and help shape innovation right from the start.

Work-life balance & flexibility

Promoting a healthy work-life balance is a key component of our culture at MMT. Originally started in an office environment, the Corona pandemic in 2020 required a rapid adaptation to remote working. This challenge led to the development of a hybrid way of working that combines the flexibility of working from home with the benefits of face-to-face collaboration in the office. This adaptation supports our efforts to maintain work-life balance, which is critical to the well-being and satisfaction of our team members.

For this reason, we have implemented flexible work schedules that allow all team members to balance their professional responsibilities with family obligations and personal interests. This is especially valuable for colleagues with children or extensive family commitments. To support this flexibility, we rely on our communication tool, Slack. With the status feature and a dedicated channel, each team member can transparently communicate their availability. We also use Google Calendar, which makes it easy for everyone to enter their vacation days, in-office days, and out-of-office days. With these tools, we ensure that everyone on the team is aware of their presence and absence, which ensures effective collaboration - regardless of where they work. Through this flexibility, we strive to provide a supportive and understanding environment that respects and encourages the individual needs of our employees and does not create conflicts between family and work.

Team orientation & cooperation

Teamwork is an essential part of our culture. We emphasize the importance of shared success and encourage collaboration beyond individual teams. For example, in the technical team, we work extensively with the pairing method, which strengthens knowledge transfer and problem-solving skills. In the business team, sales responsibility is shared so that all members are involved in the sales process, not just the dedicated sales team.

Appreciation & Feedback Culture

Last but not least, recognition and appreciation are at the heart of our corporate culture at MMT. It is important to us that every team member receives recognition and appreciation. This is why we rely on monthly feedback meetings, which not only serve to record and acknowledge joint successes and progress, but also provide a space to openly discuss feedback and suggestions for improvement. This practice not only makes it possible to recognize commitment and performance, but also promotes an environment in which everyone feels valued and motivated. In addition, there is a development meeting twice a year in which employees and their team leads set goals and track their achievement. This is not just about quantitative targets, but also about personal development.

In addition to regular feedback and review meetings within the individual teams, we also organize a monthly meeting, our "All MMT", which is attended by all teams. This meeting is used to provide updates from all teams, present new product features, celebrate successes and reflect on important milestones.

Conclusion

In this article, we have provided a basic overview of working at MMT, from our diversity and team structure to our culture and way of working.

At MMT, the diversity of our employees not only enhances our solutions, but also fosters a dynamic working environment. Flat hierarchies and a positive discussion culture encourage an open exchange of ideas. Flexible working hours and our focus on work-life balance support the well-being of our team members. Team-oriented collaboration and a strong feedback culture ensure that everyone feels valued and motivated to contribute to our collective success. These factors are essential to MMT's continued growth and accomplishments. We hope to provide some valuable insights for companies facing similar challenges.

As an experienced partner, Twins Digital is available to both independent and well-known agency networks when rapid support is required. With a growing direct client base, Twins Digital was looking for a future-proof, sustainable tool that was perfectly tailored to the needs of media agencies. Mercury offers particular added value in this respect: The platform not only covers all media processes holistically, from strategic and tactical planning to buying and reporting, but also makes it possible to offer own inventories to buyers. Mercury thus proves to be a comprehensive all-in-one solution for Twins Digital and actively supports them in achieving outstanding marketing results for their partners.

Gunnar Neumann, CEO at MMT: "We are very pleased to have won Twins Digital, an innovative media agency, as a new customer. In addition to the optimization of the media plan creation, our focus here is on the use of the inventory module, through which Twins Digital can manage and offer its own media inventories. We are convinced that Mercury will efficiently help Twins Digital with both current and upcoming challenges."

Franziska Honold, Group Head Media Solutions at Twins Digital said, "We are excited to take our Media Operations to the next level with Mercury's support. Our selection process was thorough; it included extensive research as well as face-to-face meetings. MMT impressed us from the start with their professional demeanor and technical expertise. After several productive conversations and an extensive testing period, we were confident that MMT's platform was a perfect fit for our goals."

About Twins Digital

Twins Digital acts as a media service desk specializing in media operations and management, analytics, tech solutions and strategic consulting. They offer a comprehensive range of services from media planning to buying, the implementation of complex campaigns to data-driven analytics. With a focus on transparency and effectiveness, Twins Digital helps well-known companies such as Siemens, Disney and GLS Bank reach their target audiences both locally and in international markets.

About Mercury Media Technology GmbH & Co. KG

Mercury Media Technology is an independent, growing MarTech company based in Hamburg since 2016. Across Europe, its clients include agencies,advertisers and publishers. More than 30 experts from the fields of software and data engineering, data intelligence and media develop data-driven solutions with the aim of maximizing the efficiency of media operations. The Media Operations Platform "Mercury" enables data-driven media management from strategic and tactical planning to booking and automated reporting. The integrated self-service marketing mix modeling module delivers actionable insights for media planning. The "Bridge" data connector enables all relevant internal and external data to be automatically integrated, prepared and structured to create a clean data infrastructure. All data is visualized in meaningful dashboards. Reporting is possible at the push of a button.

The managing directors of the Hamburg-based company are Tobias Irmer, Gunnar Neumann and Andreas Sand.

More information at mercurymediatechnology.com

+++++

Press contact:

Mercury Media Technology GmbH & Co. KG

Vivian Reifschneider

hello@mercurymediatechnology.com

The convergence of data analytics, machine learning, and Generative AI has given rise to a new era of marketing intelligence. AI-driven solutions are not just augmenting human capabilities but are fundamentally reshaping the way marketers understand, reach, and engage with their audiences. This shift towards AI integration is significantly reducing manual, repetitive tasks, enabling marketers to focus more on strategic planning and creative aspects. AI is paving the way for more automation in marketing processes, from data analysis to customer interaction, thereby allowing marketers to dedicate their expertise to crafting more innovative and impactful marketing strategies.

Content

• AI Elevates Marketing

• AI Shaking Up Marketing Mix Modeling

• AI Reinventing Media Planning

• AI Optimizing Media Operation Workflows

• Conclusion

Inspired by that and with my experience as a Data Scientist developing industrial-grade data products, I present this article that explores the latest AI trends in marketing for 2024, diving into specific use cases where AI is not just a tool but a game-changer. The fusion of AI with marketing is not merely an enhancement of existing practices but a complete overhaul, heralding a new age of data-driven, AI-powered marketing strategies.

AI Elevates Marketing

1. Elevating Predictive Analytics to New Heights

Sample Use Case: Strategic Insights for Optimizing Advertisements and Data Analysis

AI's ability to forecast user behavior is a game-changer. This is not mere data collection; it's an advanced, predictive understanding of consumer behaviors. Imagine AI algorithms that don't just process but predict user actions, offering real-time insights to fine-tune advertisement strategies for maximized engagement and ROI.

2. Marketing Campaign Optimization

Sample Use Case: Channel, Message, and Timing Predictions Powered by AI

Predictive analytics has become a cornerstone in marketing campaign optimization, e.g. in the form of marketing mix modeling (MMM). By analyzing past campaign data, AI can predict which marketing channels, messages, and timings will be most effective. This approach ensures that marketing efforts are not wasted on unresponsive audiences or ineffective channels, leading to more successful and cost-efficient campaigns.

3. Personalization at Scale

Sample Use Case: Targeted Messaging and Content

In the realm of personalization, AI in 2024 will offer tools that redefine targeting precision. This goes beyond traditional demographic targeting to a more nuanced, behavior-driven approach. AI will enable you to craft messages and content that resonate on a personal level with each segment of your audience, enhancing the effectiveness of your campaigns and ensuring a higher degree of customer engagement.

4. Ad Targeting and Marketing Automation

Sample Use Case: Enhanced Chatbots and Automated Marketing

AI's role in ad targeting and automation is more sophisticated than ever. Improved algorithms enable precise ad targeting, ensuring that marketing messages reach the most receptive audiences. Additionally, AI-driven marketing automation tools are streamlining campaign management, and optimizing the workflow of media operations by reducing manual tasks thereby helping marketing managers to focus on strategy. Advanced chatbots are providing real-time customer engagement and support.

5. Smarter Content Generation

Sample Use Case: AI-assisted Content Creation

Gone are the days of generic content. In 2024, AI assists in generating smarter, more personalized content. This technology is not only about automating content creation; it's about enhancing its relevance and appeal to specific audience segments. Coupled with conversational AI, this leads to more engaging and effective marketing communications.

AI Shaking Up Marketing Mix Modeling

Let's talk about Marketing Mix Modeling (MMM) – it's getting a serious facelift thanks to AI. Remember the old-school MMM? Lots of historical data, linear regression models, that sort of thing. Well, AI’s stepping in with its machine learning smarts, and it’s changing the game. Now, we're getting a much clearer picture of how different parts of our marketing efforts play together and what that means for sales and ROI. It's like having a high-powered microscope to see the nitty-gritty of what works and what doesn’t.

Here's the cool part: AI's not just sitting back and analyzing past data; it's predicting the future! Think about it – AI models can now forecast how different marketing moves will impact sales. That means you can be smarter about where you put your money. And it's not just about looking ahead. AI lets you tweak your marketing mix on the fly, using the latest data. Plus, it’s great at figuring out how different channels work together, so you can create marketing strategies that really gel across all platforms.

AI Reinventing Media Planning

Now, let's dive into media planning. It used to be a lot of guesswork and gut feelings, right? But AI’s stepping up to make things a lot more precise. The whole process of picking the best media platforms for advertising is getting a major upgrade. AI gives you the lowdown on your audience with such detail; it’s like having a crystal ball. This means you can segment your audience like a pro and hit them with messages that really resonate.

And here’s where it gets even better – AI is turbocharging programmatic media buying. It’s all about making smart, real-time bids for ad placements, tailored to how your audience behaves and engages. Plus, AI’s not just playing matchmaker between ads and audience; it's also predicting how well your campaigns will do across different channels. This helps you to be smart with your budget, putting your dollars where they’ll make the most impact. And the cherry on top? AI can whip up media plans that are tailor-made for your campaign goals and audience preferences. It's like having a custom suit – fits perfectly, looks great, and gets results.

AI Optimizing Media Operation Workflows

AI is revolutionizing the way marketers manage their workflows, transforming the landscape of media operations in 2024. Imagine AI not just as a tool for analysis but as an active assistant in the marketing process. It's like having a smart colleague who points out anomalies, suggests tasks, and provides strategic advice on your media plans. This integration of AI into daily marketing operations is a game-changer. It's streamlining the mundane, often error-prone manual tasks, allowing marketing professionals to focus their energies on crafting winning strategies.

With AI's advanced capabilities in marketing mix modeling (MMM) and media planning, marketers are now equipped to make more informed decisions. AI algorithms can sift through vast amounts of data, identifying patterns and insights that might be missed by the human eye. This level of analysis helps in fine-tuning media plans, ensuring that each campaign is not only effective but also cost-efficient. The result is a dramatic reduction in the 'monkey work,' liberating marketers from the repetitive tasks that used to consume much of their time. Now, with AI handling the heavy lifting of data processing and analysis, marketing professionals can devote more time to strategic thinking and creative problem-solving.

This shift towards a more AI-centric approach in marketing is not just about efficiency; it's about empowering marketers to be more innovative, more responsive, and more strategic in their roles. The future of marketing, powered by AI, is not just about working harder; it's about working smarter, with technology as a trusted partner in the quest for excellence.

Conclusion

As we wrap up this exploration into the AI-driven future of marketing in 2024, it's clear that we're not just talking about incremental changes or slight improvements. We're witnessing a fundamental shift in the entire marketing paradigm. AI, with its predictive analytics, personalized content creation, and sophisticated media planning, is not just a tool in the marketer's toolkit; it has become the architect of a new marketing era.

As marketing professionals, this new era offers an exhilarating opportunity. It's a chance to redefine how we connect with our audiences, to make our marketing efforts more relevant, more personal, and, ultimately, more successful. The future of marketing is here, and it's powered by AI. It's an exciting time to be in this field, and the possibilities are as limitless as the technology driving them. Let's embrace this AI-powered future and shape marketing strategies that are not just effective but truly transformative.

In this article, we will answer the following questions:

• What is ROI (Return on Investment)?

  • What does ROI mean in the context of marketing?

  • How is ROMI (Return on Marketing Investment) calculated?

  • Which metrics and key figures are relevant for determining ROMI?

  • How can ROMI be interpreted?

• What role does ROMI play in budget allocation and decision making in marketing?

  • Why is it important to keep an eye on the ROI of the campaign?

• What are the challenges in calculating ROMI and what are possible solutions?

• What technologies and tools are available to support ROMI analysis and optimization?

What is ROI (Return on Investment)?

ROI is the abbreviation for Return on Investment. It is also often referred to as capital profitability or return on equity. ROI shows the relationship between investment and profit. It is a metric for the economic success of an investment and, thus, an interesting and important figure at many points within a company, including in marketing.

What does ROI mean in the context of marketing?

When ROI is used as a KPI in marketing, it is often referred to as ROMI, which stands for Return on Marketing Investment. ROMI is the metric that measures the profitability of marketing activities. It indicates the profit or sales growth a company achieves in relation to the marketing expenses invested. ROMI is usually expressed as a percentage and indicates how much profit is made for every euro, dollar, etc., invested. Marketers can thus deduce whether the costs invested are in a satisfactory relationship to the benefit. 

Which metrics and key figures are relevant for determining ROMI?

Determining the return on marketing investment (ROMI) requires the consideration of a variety of metrics and key figures in order to obtain a comprehensive picture of marketing effectiveness. Here are some relevant metrics and key figures that are important for the ROMI calculation:

• Total turnover: The total turnover generated by marketing activities is a key figure for the ROMI calculation.
• Marketing spend: The total spend on marketing activities, including media budget, content creation, staff and service providers, as well as other costs, must be recorded.
• Profit: The profit that is directly attributable to the marketing efforts must be determined.
• Conversion rate: The conversion rate indicates how many customers became paying customers as a result of the marketing activities. This can be measured for various stages of the sales funnel.
• Customer Lifetime Value (CLV): The CLV metric shows the estimated total revenue that a customer will generate over their entire relationship with the company.
• Customer Acquisition Cost (CAC): CAC indicates how much it costs to acquire a new customer by dividing the marketing spend by the number of customers acquired.
• Customer Retention Rate: The rate at which existing customers return and buy again is important to evaluate long-term success.
• Market share gain: The increase in market share due to marketing activities is another important indicator.
• ROI by marketing channel: It is crucial to determine ROI for different marketing channels separately to understand which channels are most profitable.
• Click-through rate (CTR): For digital advertising, CTR is relevant as it shows how many users have clicked on ads.
• Cost per click (CPC): This is relevant for pay-per-click advertising and indicates how much each click on an ad costs.
• Cost per conversion (CPC): The cost to convert a user into a paying customer is crucial for profitability.
• Lead generation rate: For B2B companies, the number of leads generated is important for calculating ROMI.
• Engagement: For companies that use social media or content marketing, key figures such as likes, comments and shares are relevant.

The exact metrics and key figures that are included in the ROMI calculation can vary depending on the industry, objectives and marketing strategy. The choice of relevant metrics is crucial in order to measure the effectiveness of marketing efforts and determine ROMI.

How is ROMI (Return on Marketing Investment) calculated?

Return on Marketing Investment (ROMI) is calculated to measure the effectiveness of marketing activities and to understand how much profit has been made in relation to marketing spend. ROMI is usually calculated using the simplified formula:

The calculation of ROMI is explained step by step below

Calculate profit from marketing: This value represents the total profit generated by marketing activities. All profits that are directly attributable to marketing efforts are taken into account. To calculate the profit, a marketing mix modeling tool should be used.

Calculate marketing expenditure: The total spend on marketing activities, including media budget, content creation, personnel and service providers, as well as other costs, must be recorded.

Subtract marketing spend from profits: Marketing expenses are subtracted from profits to arrive at the net profit from marketing activities.

Divide by marketing expenses: The net profit from step 3 is divided by the marketing spend. This gives the ROMI value, which is often expressed as a percentage.

A positive ROMI indicates that marketing efforts are profitable. For example, a ROMI value of 100% means that for every euro or dollar spent, a profit of one euro or dollar was achieved.

How can the ROMI be interpreted?

In theory, any marketing investment that achieves a ROMI of over 100% pays off. At this value, the profits are sufficient to cover the total costs, which means that there are no losses. In general, the higher the ROI, the more positive the valuation.

It is important to note that the interpretation of ROMI depends on various factors. Depending on the industry, company, objective, target group, the marketing channel used and the specific campaign, a different value can be considered positive. A ROMI of 100% may be good for one company but not enough for another. However, a positive ROMI score is an indication that marketing efforts have helped to generate profits and should serve as an incentive to optimize and expand the marketing strategy.

Why is ROMI important?

Data-based methods are crucial to achieving marketing goals effectively and efficiently and making the best use of limited budgets. For this reason, return on marketing investment has become an indispensable metric. ROMI plays an essential role in measuring, comparing and optimizing marketing activities. When applied correctly, it enables a reliable determination of marketing success.

What role does ROMI play in budget allocation and marketing decision-making?

Return on marketing investment plays a crucial role in budget allocation and marketing decision making. Here are some key roles that ROMI plays in this context:

Optimizing budget allocation: ROMI helps marketers determine which marketing activities are the most profitable. By analyzing the ROMI of different channels and campaigns, resources can be allocated more effectively to achieve the best results.

Identification of high and low performing channels: By identifying ROMI, companies can determine which marketing activities are particularly effective and which are less profitable. This allows resources to be redirected from less successful areas to channels that are performing well.

Better decision making: ROMI provides hard data on which to base marketing decisions. It enables the selection of strategies and tactics that will deliver the best results, rather than relying on assumptions or intuition.

Setting targets and performance indicators: ROMI helps to set realistic targets and performance indicators for marketing activities. These targets can be set based on expected ROMI values, improving transparency and alignment of efforts towards profitability.

Cost control: ROMI supports the monitoring and control of marketing costs. Companies can ensure that their marketing budgets are in line with expected ROMI targets, leading to better financial planning.

Proof of the value of marketing: ROMI provides quantifiable proof of the value of marketing. This is particularly important to increase management and board confidence in marketing efforts.

Overall, ROMI plays a central role in rationalizing and increasing the efficiency of the marketing budget. By using ROMI data, companies can maximize the profitability of their marketing activities and ensure that the budget is used optimally. This enables better decision making and long-term success in marketing.

Why is it important to keep an eye on the ROI of the campaign?

The return on investment allows the profitability of different campaigns to be expressed in concrete figures, making it possible to compare them. Even individual phases of the campaigns can be specifically analyzed, which makes it easier to carry out optimizations during the ongoing advertising campaign.
If the right conclusions are drawn from the data analysis, immediate improvements can be made. Based on the collected data, it is possible to gradually optimize campaigns and make well-founded decisions. In this way, campaigns can be better optimized without compromising the invested budget.

What are best practices for calculating and optimizing ROMI?

Calculating and optimizing ROMI can present marketers with challenges. Many adjustments can be made to increase ROMI, which can vary greatly depending on the company, marketing strategy and campaign. In order for marketers to identify where there is potential for optimization, the basis is always good performance monitoring and in-depth analysis.

What challenges are there when calculating ROMI and what are possible solutions?

Calculating the return on marketing investment can be associated with various challenges.

     1. Data quality and availability:

• Challenge: The data needed to calculate ROMI can be incomplete or inaccurate.
• Solution: Companies should invest in improving data collection and quality. Implementation of analytics tools such as marketing mix modeling and regular review of data sources are critical.

     2. Darketing budget allocation:

• Challenge: Accurately allocating marketing costs to specific campaigns or channels can be difficult.
• Solution: Using budget allocation methods and spend attribution tools can help track marketing costs more accurately.

     3. Attribution and multichannel marketing:

• Challenge: With multichannel marketing campaigns, attributing sales to specific marketing activities is complex.
• Solution: Using attribution models that take into account the contribution of different channels can help to more accurately assess the impact of marketing on overall sales.

     4. Time lags:

• Challenge: There can be a time lag between marketing activities and the resulting sales or profits.
• Solution: To account for this lag, companies can perform the ROMI calculation over longer periods of time and analyze the impact of marketing activities on the entire sales cycle. The ad-stock effect, i.e. the delayed advertising impact, should also be included as a factor in calculations.

     5. Seasonal fluctuations:

• Challenge: Some industries are subject to seasonal fluctuations that can influence the ROMI calculation.
• Solution: By adjusting ROMI to seasonal trends and taking these patterns into account in budget planning, companies can mitigate the impact of seasonal fluctuations.

It is important to note that solutions to these challenges may vary by company, industry and objectives. Careful data collection, analysis and customization of the ROMI calculation to a company's specific circumstances can help produce more accurate and meaningful results.

What technologies and tools are available to support ROMI analysis?

The Marketing Mix Modeling (MMM) is an advanced method for improving return on marketing investment, as it provides detailed insights into the impact of different marketing activities on overall success. An MMM helps with attribution and budget optimization by evaluating the contribution of each marketing channel or campaign to revenue generation. With these insights, companies can better allocate their budgets and invest resources in the channels that generate the highest ROMI. Additionally, cross-channel optimization is possible. Companies can use MMM to understand how different marketing channels interact with each other. This enables fine-tuning of cross-channel strategies to increase overall ROMI. MMM can incorporate external factors such as economic conditions, competition, and seasonal fluctuations into the ROMI calculation, enabling a more accurate assessment of marketing effectiveness. Through continuous marketing budget allocation based on MMM results, companies can improve their ROMI by ensuring that their expenditures match expected returns.

Marketing Mix Modeling is a powerful tool to enhance ROMI by assisting companies in optimizing their marketing efforts and boosting the return on their marketing investments. It allows for a data-driven approach to budget allocation and marketing strategy, ultimately leading to improved financial performance and competitiveness.

Conclusion: ROMI makes campaigns assessable and comparable

Return on Marketing Investment (ROMI) is a crucial concept in marketing that helps companies measure the effectiveness of their marketing activities and ensure that the marketing budget is being used in the best possible way. In this article, we have addressed the important aspects of ROMI and highlighted some key insights:

• ROMI provides a clear financial evaluation of the profitability of marketing efforts. A ROMI above 100% indicates that the marketing investments are generating profits, while values below 100% suggest that costs exceed revenues.
• The calculation of ROMI requires accurate and relevant data. Companies need to establish a reliable data basis for a meaningful ROMI assessment.
• ROMI allows for comparisons across different marketing campaigns and strategies. This comparability helps identify the most effective marketing measures and optimize future campaigns.
• Marketing Mix Modeling is a valuable tool that supports the optimization of ROMI by providing deep insights into the efficiency of different marketing channels and activities.

For some time now, there has been a renewed interest in MMM for media planning, as they are very well suited to map the effectiveness of all relevant media channels and are not limited to the digital domain. However, they are usually only used on a national level for one brand, although many brands would also be interested in planning media for different regions or products. Using a traditional method, however, would involve an enormous amount of work to create individual models for each region or product. Other modern methods, on the other hand, are associated with poor interpretability.

Therefore, Bayesian Hierarchical Marketing Mix Modeling (BHMMM) is the appropriate approach with some methodological and practical advantages to optimize media planning for different subgroups of a brand such as regions or products.

In this article, we will answer the following questions:

• What is the difference between a BHMMM and traditional methods?

• What methodological and practical advantages does it bring with It?

• How to use it to optimize regional media planning?

What is Bayesian Marketing Mix Modeling?

The concept of marketing mix modeling is further developed in a Bayesian approach by using Bayesian statistics, which is based on the use of probabilities. This methodology makes it possible to incorporate prior knowledge and combine it with the data in the modeling process, which leads to more robust results and allows statements to be made regarding their certainty when interpreting them. An extension is Bayesian Hierarchical Marketing Mix Modeling, which allows the data for a large number of subgroups to be used in one model.

One of the methodological advantages of a BHMMM is that in one modeling run, the analysis of all subgroups happens together. For example, if the model is run for the different regions in which a brand is on the market, the regions form a larger system in which they learn from each other. A general media effect is thus assumed and general trends are also mapped, but individual characteristics of the regions are also taken into account so that a media channel can have a differently strong effect depending on the region. In addition, this also makes it possible to analyze regions for which data is only available over a shorter period of time because the procedure uses information from other regions to validate the influences of various factors. Furthermore, a hierarchical approach automatically uses a larger data set compared to a national model, i.e. a larger sample from which to learn. This leads to robust models and, by definition, information on their uncertainty is possible for all model parameters resulting from the Bayesian approach. Another advantage of the BMMM is that the parameters that reflect the time-delayed media effect (adstock effect) and the saturation of a media channel are also determined directly in the modeling process. The determination of these factors is associated with a time-consuming iterative process when using classic MMM methods, which is why a Bayesian approach contributes to a significant simplification in this point. Moreover, in a BMMM an extension is possible so that a changing media effect over time is taken into account. In contrast, classical methods provide a fixed point estimate for the influence of the respective variables. This makes the BHMMM a very comprehensive and flexible approach.

For media planning, Bayesian Media Mix Modeling has some practical advantages, as all model results are available for the subgroups, e.g. for the different regions. Thus, it can be seen, for example, if in region A the influence of out-of-home campaigns is comparatively high and in another region perhaps digital campaigns work particularly well. Based on the model results, media planning is then possible to adapt to the different regions. Implications for campaign planning can therefore be derived and the effects compared by means of forecasts for all the regions included. From this, it becomes obvious which of the plans promises the greatest effect on the target key figure (e.g. sales, website traffic, advertising perception, etc.). Furthermore, an optimization of the budget allocation to the different media channels is also possible for the regions, so that the different effectiveness depending on the region is also taken into account in the distribution of the budget.

What recommendations for action can be derived from Bayesian Hierarchical Media Mix Modeling?

Using a client example, we will now show how a BHMMM concretely supports regional media planning. For one of our clients at MMT, we implemented a project for 50 cities in which the brand is active. The aim was to provide the media agency in charge with recommendations for more efficient planning for each region and to compare different campaign plans based on the resulting model. In addition to media use, various factors were included to explain sales in the different cities. On the one hand, these were general influences such as seasonality, weather, and holidays. Furthermore, customer-specific influencing factors such as fees, new product launches, and an indicator of brand interest were taken into account.

The descriptive preliminary analysis already revealed differences in sales growth and, for example, in the reaction of sales to external factors such as the introduction of fees. These findings already provide a basis for the later result interpretation of the MMM. From a methodological point of view, a Bayesian approach was considered useful for this project, because the multitude of regions can be mapped well, and finally results from the MMM can be interpreted well. At the end of the project, it was confirmed that valuable strategic insights for regional media planning could be derived from the model results.

For all regions, it was shown how strongly the various influencing factors contributed to sales in the period under consideration. The consideration of the media contribution per region gives first indications of which channel in which region contributes how strongly to sales. From this, the return on investment or the cost per order can be calculated, which allows a deeper comparison of how efficient the different channels are in the respective regions. This is the basis for later decisions regarding the budget allocation per region and channel.

Conclusion

With hierarchical Bayesian media mix modeling, the general suitability of a media channel for different regions can be compared. This allows recommendations for the budget allocation per media channel and region, as it shows in which regions the use is particularly favorable or costly. Furthermore, concrete implications for the weekly media deployment per channel and region can be derived and compared by means of forecasts for different campaign scenarios in order to make a well-founded decision for optimized media planning.

With the Media Operations Platform "Mercury", MMT provides ANKOMM with a comprehensive SaaS solution for increasing media efficiency. The media management platform automates - at a sensible point - the entire workflow from planning, execution, purchasing, optimisation to reporting of media campaigns. This reduces errors and increases quality. Mercury transparently maps the media process and simplifies the collaboration between media agencies and their advertising clients through more efficient workflows.

Gunnar Neumann, CEO at MMT: "We are very pleased to have won ANKOMM as an agile, exciting client with a great team that wants to take their internal media processes and also their clients to the next level with the help of our Media Operation Platform."

Andreas Hoffmann, CEO at ANKOMM: "What sets ANKOMM apart is that we act quickly, flexibly and smartly for our customers. For this, processes have to be right. MMT helps us to set up our media process cleanly and efficiently with its Media Operations Platform. This also helps us in the cooperation with our customers, who appreciate transparency."

-----------------------------------------------------------------------------------------------------------------------------------------------------------

About ANKOMM

ANKOMM is a full-service agency based in Hamburg. For the past 10 years, we have been able to convince renowned clients such as BMW, Metro and Sixt with our knowledge, commitment and creativity. We are particularly specialised in these three areas: Campaign, Production and Media.

About Mercury Media Technology GmbH & Co. KG

Mercury Media Technology has been working for agencies and advertisers across Europe since 2016 as an independent, growing company based in Hamburg. More than 30 experts from the fields of software and data engineering, data intelligence and media develop data-driven solutions with the aim of maximising the efficiency of media operations. The Media Operations Platform "Mercury" enables data-driven media management from strategic planning to booking and automated reporting. The integrated self-service marketing mix modelling module delivers actionable insights for media planning. The data connector enables all relevant internal and external data to be automatically integrated, prepared and structured to create a clean data infrastructure. All data is visualised in meaningful dashboards. Reporting is possible at the push of a button.

The managing directors of the Hamburg-based company are Tobias Irmer, Gunnar Neumann and Andreas Sand.

More information at mercurymediatechnology.com

+++++

Press contact:

Mercury Media Technology GmbH & Co. KG

Vivian Reifschneider

hallo@mercurymediatechnology.com

In this article, we will talk about:

• What is Hierachical Data?

• How traditional MMM’s model Hierarchical Data?

• Hierarchical Modeling using the Bayesian Approach?

Firstly, what is Hierarchical Data?

Before we unravel the magic that Bayesian MMM does on Hierarchical data, let’s quickly understand what hierarchical data is and why it's a treasure trove for marketers. Imagine a multinational company selling multiple products in different regions. The sales data can be structured at various levels – country, state, city, and product categories. This multi-level structure is hierarchical data.

In marketing, consider a global brand's advertising data. At the top level, you might have different regions like North America, Europe, and Asia. Within each region, there are individual countries. Each country may be split into different states or provinces, and each state might be further divided into different cities or towns. For each city, you might track different types of advertising spend, like television, radio, print, or online. So, the data is organized hierarchically - from regions down to specific types of ad spend in individual cities. Each level of this hierarchy is part of the larger whole, and data at each level can provide different insights for the brand's marketing strategy.

For marketers and advertising agencies, this data is gold. The ability to analyze and understand the interplay between different levels of hierarchy can reveal insights about customer preferences, regional trends, and product performances. But, here's the catch: traditional MMM struggles to accurately capture the relationships in such data.

How traditional MMM’s model Hierarchical Data ?

Within the realm of traditional machine learning based marketing mix models, two primary approaches emerge: the pooled and the unpooled models.

Pooled Model (Complete Pooling):

In a pooled model, we don't differentiate between the different hierarchical groups. We treat all data as if it comes from a single group. This is akin to aggregating all the data and running a single regression model on it.

Building one comprehensive model is often the most straightforward approach: all samples are aggregated, and distinctions between different groups are disregarded.

However, this approach, especially when using simple models like linear regression, might overlook nuances in the data, a phenomenon known as underfitting. More complex "black-box" methods, such as gradient boosting, might detect and learn from the different sub-datasets on their own, offering potentially better accuracy. But this comes at the expense of interpretability, making it challenging to understand the model's underlying mechanics and decisions.

Here's the visualization of the pooled model :

• The dots represent sales data for three regions (A, B, and C) against different advertising spends.
• The black line is the regression line for the pooled model, which is fitted to all the data across regions.
  

You can notice that while the black line may provide a general trend, it doesn't seem to perfectly capture the individual trends for each region.

Unpooled Model:

In contrast, we create a separate model for each group. So, if we have data for 3 regions, we'll run 3 separate regression models, one for each region. These models are referred to as unpooled models.

While each model specializes in a specific subset of the data, collectively, they aim to provide a comprehensive understanding of the entire dataset. The primary advantage is that these models can capture specific trends and nuances within each subset.

Here's the visualization for the unpooled models:

• The dots still represent the sales data for the three regions (A, B, and C) against different advertising spends.
• The colored lines are the regression lines for each of the regions:
• Red line represents the regression for Region A
• Green line represents the regression for Region B
• Blue line represents the regression for Region C

You can observe that the individual regression lines fit their respective regional data better than the pooled model did. This is the advantage of the unpooled approach: it can capture nuances and variations specific to each group.

However, there are challenges:

• The need to fit multiple models, which can be computationally intensive.
• The potential for overfitting, especially when the datasets for individual groups are small.
  

To summarize: 

• Pooled Model: Provides a general trend across all data but might miss nuances in individual groups.
• Unpooled Model: Captures the trends within individual groups more accurately, but you end up with multiple models.
  

Enter Bayesian MMM with its magic wand.

Hierarchical Modeling using the Bayesian Approach

Now, let's discuss the hierarchical Bayesian approach.

While both pooled and unpooled methods have their advantages, they also have drawbacks:

• The pooled approach might be too generalized and miss out on nuances in the data.
• The unpooled approach can capture nuances but might overfit to individual groups, especially if some groups have limited data, and there will be too many models to build and analyze, which is unrealistic.

Hierarchical Bayesian Modeling offers a middle ground. Hierarchical models, often implemented using Bayesian techniques, strike a balance between pooled and unpooled models. Here's a simple explanation:

Partial Pooling: Hierarchical models allow for "partial pooling", meaning they share information across groups (like the pooled model) but also allow for group-specific effects (like the unpooled models). 

Here's how:

Information Sharing: When analyzing sales of a new product, some regions might lack data. Bayesian MMM cleverly uses information from well-represented regions to improve predictions in data-scarce areas. This technique is invaluable for diverse businesses with varying regional characteristics.

Managing Complexity: While traditional models can stumble over intricate hierarchical data, Bayesian MMM thrives on it. It adeptly captures interactions, such as a local promotion's impact on broader sales, while also recognizing ongoing trends and seasonal shifts.

Leveraging History and Expertise: The Bayesian method integrates past data and expert insights. This is crucial when navigating hierarchical structures, ensuring decisions are grounded in both historical context and specialized knowledge.

Balanced Modeling: Bayesian MMM operates on the idea that each region or group, while unique, is part of a larger shared pattern. If a region's data suggests a significant deviation, the model adjusts. This balance between individuality and shared trends ensures both specificity and broad applicability.

In essence, hierarchical Bayesian modeling blends the strengths of both pooled and unpooled approaches. It allows for individual group differences while also benefiting from the shared information across groups. In the context of marketing mix models, this can lead to more robust insights into the effects of different marketing levers across various segments or regions.

Illustrative Example: The Coffee Shop Chain

Imagine you're the marketing manager for "BeanStreet," a thriving coffee shop chain with locations spread across various neighborhoods, cities, and states.

You're planning a new advertising campaign and want to know which types of ads (TV, social media, radio) work best in different areas, and how external factors such as weather conditions and local events affect sales.

The Challenge

Your data is hierarchical; you have coffee shops (level 1) nested within neighborhoods (level 2), which are part of cities (level 3), which belong to different states (level 4). The challenge is understanding the interplay between advertising mediums and external factors across these levels to optimize your marketing strategy.

Enter Bayesian MMM

You decide to employ Bayesian Marketing Mix Modeling to tackle this challenge.

Making Informed Decisions

With these insights from Bayesian MMM, you make a series of data-driven decisions:

• Allocate more budget to social media advertising in State A and to TV advertising in State B.
• Launch special hot coffee promotions in City X during winter and cold brew campaigns in City Y during summer.
• Collaborate with event organizers in Neighborhood Z to offer special discounts or limited-edition menu items during the art festival.

This example demonstrates how Bayesian Marketing Mix Modeling’s adeptness at handling hierarchical data empowers BeanStreet to make highly targeted and effective marketing decisions that account for variances at each level of hierarchy – from state down to individual neighborhoods.

-

The Way Forward

For marketers and advertising agencies, hierarchical data is like a well of insights waiting to be tapped. Bayesian Hierarchical Marketing Mix Modelling (BHMMM), with its ability to gracefully navigate through the complexities of hierarchical data, is the rope that will help you draw water from this well.

Whether you are trying to optimize advertising campaigns or looking to better understand regional trends, Bayesian Hierarchical Marketing Mix Modelling (BHMMM) offers a powerful and adaptable tool for turning data into insights and insights into action.

So, put on your magic glasses and dive into the world of hierarchical data with Bayesian Hierarchical Marketing Mix Modelling (BHMMM) – where data tells stories and insights drive success.

In this article, we will answer the following questions:

• What is Bayesian Marketing Mix Modeling?

• What are the advantages of Bayesian MMM over traditional approaches?

• What are the additional features that Bayesian MMM possesses and how are they useful for marketers?

Let’s dive in with first understanding what Bayesian MMM is! But before that, let’s recall what Marketing Mix Modeling is!

Marketing Mix Modeling: A Primer

Marketing Mix Modeling is like the magnifying glass for marketers. It helps them examine the effectiveness of different marketing channels in their "marketing mix" – from print ads, television commercials, and online marketing to price variations and other business factors. By quantifying the impact of each marketing channel on sales, marketers can strategically allocate resources to maximize return on investment (ROI).

Bayesian Marketing Mix Modeling (MMM) steps in to enhance the traditional MMM. Think of it as swapping your magnifying glass for a microscope, providing deeper insights into your marketing efforts.

What is Bayesian Marketing Mix Modeling?

The Bayesian MMM is a really useful tool for marketers. What it does, essentially, is two things. Firstly, it measures how effective different marketing channels are by giving us an average of their impact. So, we get to know, on average, how much each marketing method helps in making sales.

But that's not all. The second thing it does is predict a range of possible effects from these channels. This means it helps us understand what the smallest and largest impacts could be from each marketing method, giving us a fuller picture of what might happen with our marketing efforts.

Now, you might be wondering what's behind all this. Well, it's a mathematical method called Bayesian statistics. This method uses probabilities to solve statistical problems. It's different from other statistical methods because it allows us to incorporate what we already know or believe into the analysis. This is really important because, in real life, we often have some knowledge that we want to combine with new data to make the best decisions.

Bayesian MMM resolves the challenge of stubborn prior knowledge conflicting with new data by providing a framework that combines both sources of information in a coherent manner. Rather than dismissing or ignoring prior knowledge, Bayesian MMM allows for its integration with new data, updating and refining our understanding based on the evidence at hand. Bayesian MMM ensures that decisions and strategies are grounded in a balanced combination of historical knowledge and current data, leading to more informed and effective marketing outcomes.

This process mirrors how humans learn and evolve their understanding. When we encounter new information, we don't discard our existing knowledge completely. Instead, we integrate the new insights with what we already know, updating and refining our understanding. It's a continuous cycle of accumulating knowledge and adapting our perspectives based on the latest evidence. Bayesian MMM follows a similar pattern by incorporating prior knowledge into the analysis and allowing it to be refined and updated with new data.

How Bayesian MMM outshines Traditional Machine Learning-Based MMM

1. Embracing Uncertainty

Uncertainty is an inevitable part of the marketing landscape. Instead of sidestepping it, Bayesian MMM harnesses uncertainty. Traditional MMM delivers 'point estimates,' or single-valued predictions, of marketing impacts. In contrast, Bayesian MMM provides a spectrum of potential impacts, offering a more comprehensive perspective. This allows marketers to evaluate different scenarios and understand the probability of each outcome. By embracing this uncertainty, companies can create more robust marketing strategies that are prepared for a range of possibilities, reducing the risk of being caught off guard.

2. Valuing Past Experiences

Just as we draw lessons from past experiences, so does Bayesian MMM. It integrates 'prior knowledge' or historical insights into the model, improving prediction accuracy—a key aspect often overlooked by traditional machine learning models. By incorporating this historical data, Bayesian MMM effectively learns from the past, refining its predictions and accounting for trends and patterns that have emerged over time. This not only enhances the accuracy of forecasts but also provides a more reliable basis for making marketing decisions that are in sync with market dynamics.

3. Clearer Interpretation

Bayesian MMM serves a rational framework for decision-making under uncertainty, resulting in highly interpretable models. It provides not only the outcome but also the reasoning behind the outcome, facilitating informed decision-making. This transparency is critical in building trust and understanding among stakeholders, as it helps them to understand the logic behind the predictions. Furthermore, by providing the reasoning, it empowers decision-makers to fine-tune strategies based on concrete insights, fostering more targeted and effective marketing campaigns.

4. Superior Performance with Limited Data

While data is vital for machine learning models, collecting it in vast quantities can be challenging. This is where Bayesian MMM shines—it can deliver robust results even with smaller datasets, making it a potent tool for marketers working with limited data. This is particularly beneficial for startups and small businesses, which may not have access to large datasets. Additionally, Bayesian MMM’s ability to extract meaningful insights from limited data means that companies can make well-informed decisions faster, without having to wait for large amounts of data to accumulate. This agility is crucial in the fast-paced world of marketing.

5. Robustness

Bayesian MMM stands out for its robustness compared to Machine Learning based MMM models, particularly in handling diverse and imperfect data. While Machine Learning models often require large and clean datasets to function effectively, Bayesian MMM can work efficiently even with limited or missing data by integrating historical information and prior knowledge. This integration not only compensates for data imperfections but also adds a layer of contextual understanding that Machine Learning models might miss. Furthermore, Bayesian MMM is less sensitive to outliers and avoids overfitting, which can be a challenge for Machine Learning models in sparse data scenarios. Consequently, marketers find Bayesian MMM to be a more reliable and flexible option, especially when data quality and quantity are constrained, as it provides meaningful insights that are grounded in both data and domain knowledge.

Illustrative Example

Let's take a look at a real-world example to demonstrate the different analytical capabilities between Traditional MMM and Bayesian MMM. Imagine a beverage company, "Quench Corp," that has just launched EnerBoost. The company's marketing team has devised a multi-channel advertising campaign, including TV commercials, social media, billboards, and collaborations with fitness influencers.

Quench Corp's marketing team is eager to know how each channel has contributed to the sales during the first quarter after the launch, to better allocate their budget for the next quarter.

Using Traditional MMM:

After feeding the data into a traditional MMM, the marketing team concludes:

• TV commercials have contributed to 40% of sales.
• Social media ads contributed 30%.
• Billboards contributed 15%.
• Fitness influencer collaborations contributed 15%.

With this information, the marketing team decides to allocate the budget according to the contributions.

Using Bayesian MMM:

Bayesian MMM, on the other hand, takes a different approach. Besides using the current data, the team knows that fitness trends fluctuate and there was a fitness expo during the first quarter, which could have temporarily boosted the impact of fitness influencer collaborations.

The Bayesian MMM estimates the contributions to be:

• TV commercials: 40%, but with a range of 35% to 45%.
• Social media ads: 30%, but with a range of 25% to 35%.
• Billboards: 15%, but with a range of 10% to 20%.
• Fitness influencer collaborations: 15%, but with a range of 5% to 25%.

The Bayesian MMM considers uncertainty and the fitness expo's temporary effect, revealing that fitness influencer collaborations could range from 5% to 25%, which is quite uncertain in nature, since the range is quite huge.

Decision Making:

Using the traditional MMM, the marketing team might have been tempted to significantly increase the budget for fitness influencer collaborations. However, the Bayesian MMM reveals that the actual contribution could be much lower when the temporary fitness expo effect wears off.

Armed with this nuanced information, the marketing team makes a more informed decision to moderately increase the budget for influencer collaborations but also puts an emphasis on social media and TV commercials, which have a more stable impact.

What MMT has in offer ?

MMT boasts a comprehensive suite of MMM solutions catering to diverse needs, both in self-service as well as personalized-service mode. For those seeking a rapid snapshot of the performance of their marketing strategies, MMT provides a Machine Learning-based MMM solution. This option is ideal for clients who need a quick overview and actionable insights into how their marketing campaigns have been faring. On the other hand, for more detailed and comprehensive analysis, our advanced Bayesian MMM solution comes into play. This advanced option digs into historical data and clients prior knowledge of marketing dynamics and explores a range of possible outcomes, making it especially robust even when data might be sparse or imperfect.

Additionally, MMT is committed to innovation, constantly developing its suite to ensure clients have the most advanced analytical tools at their disposal. In essence, MMT's diverse suite, ranging from quick assessments to detailed analyses, equips clients with the insights needed to optimize marketing strategies, enhance returns, and foster business growth.

Wrapping Up

In a competitive environment where maximizing every marketing dollar matters, the advantages of Bayesian MMM are worth exploring. By accounting for uncertainty, utilizing prior knowledge, providing detailed insights, and performing well even with limited data, Bayesian MMM is an invaluable tool in a marketer's arsenal. It's not just about understanding the 'what,' but also the 'why,' helping marketers make well-informed, data-driven decisions.

-

The most frequently asked questions about marketing mix modeling

I’m assessing the performance of my current campaign and I want to know how much each channel contributed to the revenue. Is it feasible with MMM?

Of course - this is one of the key purposes of Marketing Mix Modeling. The results of modeling include weights assigned to every relevant channel or publisher. Moreover, it’s possible to see how that contribution changes over time, especially when certain channels or publishers are inactive. Revenue is not the only target KPI that can be used in modeling - it’s possible to assess the contribution to any target KPI affected by your advertising activity and important to your business. By pairing together the contribution and the amount of advertising spend, MMM also helps assess the effect of each channel on ROI.

Is it possible to include both digital and classical marketing channels into one Marketing Mix Model?

The answer is yes - but there are certain preconditions. To obtain meaningful results, you need the spend and reach (or contacts/impressions - depending on the channel) as well as your target KPI data over the same period of time and with the same granularity. For example, if you have daily sales, your advertising budget should also be available on a daily basis - and the same applies to contacts. With digital channels, this should be quite easy; however, with classical media, there are some challenges to overcome: for example, if you place your ads in a magazine and you model on a daily level, you don’t know exactly how many people see your ad every day, so you need some convention how to distribute the numbers over time based on the total print run.

I’m running an online store. Our target KPI is revenue and we advertise only in digital channels. Will it be enough for MMM if I provide the revenues, advertising budget, and impressions?

You will also need to take into account external factors. It’s hard to imagine a business that’s completely free from external impacts, such as competitor activity or the weather.

Incorporating such factors in the MMM model and thus capturing additional sources of influence on your target KPI helps improve the model’s accuracy and gain a deeper understanding of how these external factors interact with advertising efforts to drive results.

Based on your experience, you should be able to tell which external factors are affecting your sales, revenues, or other target KPIs. Make a list of those factors and gather the data over the same time period as the target KPI and advertising data. This will enhance the model's ability to accurately attribute the contributions of different channels and help provide a more holistic view of the marketing landscape, leading to more informed decision-making and optimized marketing strategies.

I’m already assessing the effectiveness of my digital campaigns based on last-click attribution. Does it make sense for me to try MMM - I will get the same results anyway?

Last-click attribution is a widely used model but it only takes into account the final step of the funnel that leads to the conversion - and only if that step belongs to a digital channel. Remember how many times you converted online after seeing an ad for the first time ever? Probably not so many, if ever at all. Decision-making is a complicated process that normally involves multiple touchpoints and some time in between. Quite often we cannot even remember how many times we saw the relevant ads before we made the decision, or where and when we saw those ads. As a result, the impact of the last step in the funnel tends to be overrated, while the effect of supporting channels gets underestimated or completely neglected. Moreover, assessing the impact of offline media is impossible with last-click attribution.

The essence of MMM is that it compares the dynamics of the target KPI - that can be sales or app installs or registrations or any other conversions - to the dynamics of the advertising activity and budget and derives the correlations, taking into account all the channels involved as well as external factors that could affect the target KPI. MMM results show the whole picture and assess the indirect effects and interrelations, which is impossible with last-click models.

And what happens when third-party cookies are disabled? Will marketing mix modeling still be possible?

Yes! And Marketing Mix Modeling is actually one of the first-choice alternatives to performance assessment methods based on third-party cookies. The imminent death of third-party cookies is one of the reasons why MMM is becoming so popular. If you still have no experience with MMM and you don’t want to be left behind, it’s about time to jump on the bandwagon!

Our top managers want to cut down on TV ads because they focus on last-click data and think that only digital channels are working, but I have a strong feeling that TV ads are also contributing to sales. Is there a way to find out who is right?

Indeed, Marketing Mix Modeling should be the key to solving your dilemma. The modeling helps find out how much each channel contributes to the target KPI. To have more accurate results, it’s important that your data cover periods when there was only digital advertising and no TV ads, as well as periods with TV ads and no or very little digital activity. Optimally, you should also include periods without any advertising to serve as a baseline for the model.

By analyzing the relative performance of TV ads in comparison to other channels and accounting for various contextual factors, MMM enables a robust estimation of their true contribution to your target KPI. This empowers you to make informed decisions about the allocation of your advertising budget, optimizing your marketing strategy based on data-driven insights.

I want to run a new flight of my campaign similar to the previous one but I aim to get more effect from the same advertising budget. Can MMM help me with that?

Absolutely, this is a perfect use case for Marketing Mix Modeling. Based on the results of the previous flight, you can reallocate your budget more effectively. Another use case is when you need to build the optimal allocation plan with a higher or lower budget - and this is also possible with MMM. Ultimately, modeling can help determine the optimal share of each channel involved in the budget allocation. This way, you can make data-driven decisions that ensure your budget is allocated in the most efficient and effective manner, driving higher returns on your investment and achieving better outcomes for your marketing campaigns.

We are planning to launch an out-of-home campaign next month. We don’t have any experience with OOH advertising yet. Can we use MMM to estimate the optimal budget and expected reach?

No, Marketing Mix Modeling is designed to assess the impact of advertising media that have already been part of the company’s advertising campaigns. If there is no past data, the channel cannot be included in the modeling.

Our campaign started a month ago. Can we already run MMM to assess the results, if we want to redistribute the budget before it’s too late?

Even if you take a daily breakdown, 30 days are not enough to apply MMM and get meaningful outputs. Either you need to rely on the results of previous campaigns or wait until a sufficient amount of data is accumulated. For example, with a weekly breakdown, best practice suggests using at least 2 years of data - which means you need to learn from your past experience.

Is there a chance that MMM results show that all the channels are inefficient and advertising is not generating any additional revenue?

It is not very likely in real life that an advertising campaign is completely inefficient, but this scenario cannot be completely ruled out - so yes, it’s possible. But when this happens, there must be some really obvious signs and reasons: for example, you will already see that your sales are not growing or growing too slowly. It’s also possible that some external factors outweigh the effect of advertising. For example, think about the poor airlines that were advertising heavily as the COVID pandemic broke out.

We are new in the market and we want to plan our advertising budget but we don’t have any ideas, about which channels to use and how much money to spend. Can MMM help us?

Not at this point. Marketing Mix Modeling relies on past data and results, and it doesn’t work from scratch. You will need to start with some trial and error. One possible option would be to learn from competitors’ experience - if you can gain access to their spends and conversions data, maybe in some anonymized form. Depending on what industry your business belongs to, you might be able to find a partner who already runs MMM in this sector and knows which channels work better or which don’t work at all.

We are not sure which target KPI to choose in order to optimize our advertising efforts. Can MMM provide us with relevant recommendations?

No, the purpose of Marketing Mix Modeling is to build optimized budget allocation depending on what KPI is selected as the target. The allocation can vary quite drastically if you model for different KPIs - even if the budget and reach data remain the same. The choice of the appropriate target KPI should be based on your priorities, business experience, and understanding of your market.

Marketing Mix Modeling is a valuable tool but in order to really benefit from it, it’s essential to understand its capabilities and limitations. After all, every tool is only good for its purpose and only when applied properly.

The MMM process is based on complex machine-learning algorithms and therefore it might be daunting for business professionals without a sound mathematical background to start out on this journey alone - and even more daunting to make actual decisions based on the outputs of MMM. That’s why it’s important to have an experienced navigator to guide you in this way and help you build your own expertise in this field.

If you didn’t find an answer to your question in our article or if you are still not sure if Marketing Mix Modeling could help you cope with your challenges, feel free to contact us directly!

Marketing Mix Modeling is a statistical analysis method that determines the effectiveness of marketing activities and helps to optimize the allocation of marketing resources across channels, tactics, and/or campaigns. It quantifies the Return on Investment (ROI) for each marketing activity and recommends how to increase it.

Self-Service Marketing Mix Modeling takes this process one step further by empowering marketers to perform analyses themselves, using intuitive and easy-to-use Software-as-a-Service (SaaS) solutions. It enables marketers to evaluate the effectiveness of each marketing activity and compare them whenever needed.

In this article, we will answer the following questions:

• Why is Self-Service MMM an important trend within marketing departments?

• Which steps are typically involved in a Self-Service MMM project?

• What are the key benefits of Self-Service MMM?

Why is Self-Service Marketing Mix Modeling an important trend within marketing departments?

The marketing landscape is becoming increasingly complex: With the rise of digital marketing, marketers have access to tons of data and a whole bunch of analytics tools, which can help them to better understand their customers and improve their marketing performance. However, this overwhelming availability of data and tools can be difficult for marketers to master and use effectively. This is where Self-Service Marketing Mix Modeling comes in - it empowers marketers to focus on the relevant data, train highly accurate statistical models that find out the important patterns, and derive outcomes like Contribution, Return on Investment, Ad Stock Effects, and Saturation Curves. 

Demand for more accountability and transparency: In today's data-driven world, stakeholders expect marketers to be able to measure and report on the impact of their marketing efforts with a high degree of accuracy and transparency. Self-Service Marketing Mix Modeling enables marketers to validate their own data and measure the impact of their marketing activities more effectively, which helps to build trust and credibility within an organization and for external stakeholders.

Need for fast insights and flexibility: Traditional Full-Service Marketing Mix Modeling can be time-consuming and expensive, and it may not provide the flexibility and agility that marketers need to act in fast-changing environments. Self-Service Marketing Mix Modeling empowers marketers to run MMM on demand when needed with high efficiency because everything needed can be bundled in the marketing department without external dependencies.

What are the key benefits of Self-Service Marketing Mix Modeling?

Self-service MMM platforms are designed to be user-friendly and intuitive, even for non-technical users, with drag-and-drop interfaces, easy-to-understand data visualization as well as built-in guidance and support.

It empowers marketers to become more data-driven and agile, by enabling them to run their own marketing mix modelings without relying on externals and to build a deep understanding of the root causes of their marketing performance themselves. This helps marketers to make faster, more informed decisions about their marketing activities and to optimize their marketing performance more effectively.

Which steps are typically involved in a Self-Service Marketing Mix Modelling project?

Platform Setup & Configuration: Although vendors provide their solutions as Software-as-Service nowadays to speed up processes, some work still has to be done in advance, like spinning up a client-specific environment with resources attached, defining roles and authorizations as well as providing initial access for each user.

Onboarding & Support: If users work with the solution for the first time good vendors provide an onboarding procedure & ongoing support afterward to get users up to speed. Depending on the vendor, onboarding procedures can have completely different formats ranging from classical onboarding sessions with an instructor to complete self-learning based on onboarding guides and/or videos.

Creation and provisioning of modeling dataset: Either the vendors or the marketers themselves create a modeling dataset, which combines the relevant data for the target KPI and the explanatory factors. The modeling dataset will be made available within the Self-Service Marketing Mix Modeling Software, so everything is ready for model training.

Functional Tests and Feature Engineering: After a modeling dataset is made available in the software a few test runs will be performed, to make sure that the model training works and that all factors as well as target KPIs can be used for the model training. On top factors can be transformed (e.g. normalized, inverted, indexed) so that they are more readable for MMM algorithms.

Model training & Result Interpretation: Via the software, marketers are enabled to train different models themselves by e.g. changing the set of explanatory factors that are used for the model training or even changing the target KPI. Via these iterations of model training and interpreting the results, marketers develop a good understanding of the importance of each factor and also the media variables themselves.

Model Selection, Insights Creation & Optimization: With a good understanding of each factor and the media variables, marketers can choose their favorite model and use it for optimized budget allocations in order to increase efficiency for future marketing investments (e.g. next campaign, next years allocation)

Process management: The Self-Service MMM project has to be managed like every other project in order to make sure that expectations are fulfilled, deliverables are on time, and project goals are set and reached in the end.

Decision-Making & Operationalization: Good solutions also provide possibilities to use the created insights for briefings and/or setting tactical as well as strategic goals.

How far can you go with Self-Service? Which steps can be performed internally?

Marketers that have all crucial capabilities e.g. for the dataset creation or have them in-house available can perform nearly all steps themselves. Only the first two steps “Platform Setup & Configuration” plus “Onboarding & Support” are always done by the vendors. Here is an overview of which ways of working make sense depending on the capabilities and resources available in-house.

Brands are realizing the need for new tools and for expertise to accurately assess their marketing effectiveness. This presents a common dilemma: whether to build a customized Marketing Mix Modeling (MMM) solution in-house or seek out third-party assistance. To facilitate this decision-making process for you, we've compiled a guide outlining the benefits and drawbacks of each approach, including the involvement of specialized companies or the application of ready-to-use software solutions.

Questions addressed in this article

• What resources are needed to build a Marketing Mix Model in-house?
• How long will it take to build MMM in-house?
• What is the cost of building Marketing Mix Modeling in-house?
• How much will it cost to buy Marketing Mix Modeling software?
• What to look out for when buying Marketing Mix Modeling Software?

How to Build a Marketing Mix Modeling In-House?

Rule of thumb: You can't build Marketing Mix Modeling in-house without a team of good data scientists, data analysts, and data engineers with the following skills:

• Good knowledge of the market (or at least experience in analyzing various markets or branches) in order to identify possible influencing factors
• Basic knowledge of how media channels work
• Good overview of possible statistical methods as well as their advantages and disadvantages, and experience with their application
• Ability to use analytical software or scripting languages
• Capability to connect available external data sources and merge the data with the internal media data
• Experience in data preparation (merging, transforming, and assessing data quality) and ability to recognize correlations

Building an internal data intelligence team is often not easy, because professionals in this field are in high demand and well paid, as listed further down in the article. This should be taken into account.

Building your own Marketing Mix Modeling solution can be a challenging task due to the complexity of your company's marketing activity. We would like to highlight a few specific difficulties:

• The first challenge is to specify the model correctly. That means, all the relevant influencing factors have to be included, but overfitting has to be avoided. Moreover, sufficient data points for individual influencing factors must be available.
• The presence of multicollinearity, or correlation between marketing channels, can make it difficult to create accurate models and has to be taken into account. This happens when marketers simultaneously book different media channels (which of course often happens) or when several external events (e.g. holidays, Black Friday, etc.) occur at the same time.
• Non-linear correlations, in particular time-delayed media effects like the adstock effect and saturation, have to be mapped, and the same applies to interaction effects.
• It is important to take into account long-term brand development as well as the relationship between media budget and media performance, which makes budget allocation in particular more complex.
• Budget allocation is not an easy task, as it requires knowledge of optimization algorithms and involves definition of allocation procedures.
• Marketing Mix Modeling becomes more complex if it covers multiple regions and/or multiple brands, which requires a hierarchical approach.
• As data quality and availability of all the relevant data in general is the key to successful Marketing Mix Modeling, it's crucial to ensure that data pipelines are automated.

Furthermore, the model needs to comply with a long checklist of necessary features, charts, tables, and recommendations for decision-making. At MMT, we believe that to provide comprehensive insights into your marketing efforts, a successful Marketing Mix Model should include the following features:

• Attribution modeling: Quantifiying the contribution of media      channels to the target metric and their efficiency (ROI analysis).
• Recommendation: Optimizing the media mix
• Predictive analytics: Creating forecasts for campaign scenarios
• Hierarchical approach: Generating individualized models for different regions or brands, if required.

Marketing Mix Modeling is a complex process that requires time and resources. According to Facebook (Meta), it takes 12-22 weeks to manually build the initial model, not taking into account continuous development and maintenance work required for automation. Smaller companies can achieve a "good enough" model in just a few weeks, but for larger organizations, the typical timeline ranges from several months to more than half a year.

Assuming a team of two data scientists working full-time, it could take approximately six months to build and automate an initial model. The average salary of a data scientist in Germany in 2023 is around €65,000 according to job platforms, with potentially higher salaries in major cities like Berlin, Hamburg or Munich. Once the initial model is finalized, ongoing maintenance will require at least one day every two weeks as well as a four-week update cycle each year.
The costs of data procurement, connection, and automation must also be taken into account, and you might even need to hire a dedicated data integration engineer for this purpose.

What to look for when buying Marketing Mix Modeling Software?

Marketing Mix Modeling (MMM) has become a popular tool in the marketing field, with various vendors offering their own approaches. It’s important to understand the scientific basis behind individual approaches in order to select the one that best suits your needs. One major differentiation within MMM is between frequentist and Bayesian methods. Traditional MMM relied on ordinary least squares (OLS) regression, whereas Facebook's Robyn algorithm utilizes a more modern ridge regression approach. Alternatively, Google has adopted Bayesian methods for their modeling techniques, which allows for greater flexibility in incorporating previous knowledge into the model and producing more realistic outcomes.

The biggest difference in cost and maintenance will depend on your decision to either use the standard automated software or customize your MMM solution. Most companies offering Marketing Mix Modeling can build something custom for your business, if needed. A model tailored to the company's own needs can be capable of answering exactly the right questions instead of providing general guidelines. In addition, you can receive advice regarding the implementation of the entire analysis cycle and the interpretation of the results. However, the manual approach requires a lot of trust because MMM is susceptible to human bias. Automated tools can use modern data pipelines to update the model regularly, so decisions can be made more often without having to repeat the analysis. Automated models are also able to automatically generate budget allocation recommendations, which makes them far more actionable.

For us at MMT, it is crucial that advertisers could turn their MMM insights into action. Therefore our Marketing Mix Modeling platform provides different media mix scenarios that could be directly adapted to media plans within the same platform. Our solution is a self-service tool, giving advertisers the possibility to model on their own to really understand how the MMM works and thus not rely on trust. This way you have full control over the modeling process and build up the expertise in your own company.

When it comes to selecting a vendor, the following questions might help you find the right approach for your company:

• Do they offer an off-the-shelf solution or will they build a custom MMM tool?
• Have they already worked with similar companies and thus gained industry knowledge?
• How much media expertise do they have?
• What methods do they use? Linear regression, ridge regression or Bayesian methods?
• What’s the cost of the initial model building?
• What are the ongoing costs associated with maintenance / running the model regularly?
• Do they offer after-sales services?
• Do they have the key features mentioned above?
• Do they allow different levels of automation?
• Does the level of involvement required from your company match your own skills/available resources?
• Can they explain the applied methods and possible outputs in a comprehensible way?
• Will they give useful, competent, and transparent advice on how to answer your specific questions based on the modeling insights?

Once you've made that decision whether you want to use scalable software or have the provider build something custom for you, it's essential to consider in detail what the model will do. This can be challenging, given that there is no one-size-fits-all solution in this emerging field. Some things to keep in mind include the law of diminishing returns (which states that channels may become saturated at higher spends) and the adstock effect (the delayed impact of advertising on performance). Finally, it’s important to incorporate all relevant key features into your model to reflect diverse influence factors and deal with trends and seasonality. Choosing more advanced approaches that account for changes in channel performance over time or for the variance in correlations for different products or regions can significantly increase the accuracy of your model's outputs.

Conclusion

Quite often the answer to the question of whether a company should set up marketing mix modeling on its own or buy already existing software starts with "it depends". The key factor is the resources available. Building a model requires expertise and takes time. There is a lot of room for error in the calculations, which can lead to high losses if the marketing budget is used in a suboptimal way. And for marketers to really take advantage of the valuable insights, they need smart functionalities, as well as domain and media expertise. If it's your first experience with MMM or you have doubts about certain aspects, it might be worthwhile to involve a consultant/MMM specialist to guide you on your way and help you avoid unnecessary efforts and wrong marketing decisions.

Sources

• https://www.stepstone.de/gehalt/Data-Scientist.html
• https://en.wikipedia.org/wiki/Marketing_mix_modeling
• https://www.glassdoor.de/Geh%C3%A4lter/data-scientist-gehalt-SRCH_KO0,14.htm

Are you still feeling confused? No worries! In this article, we will point out the differences between these two concepts, look at their similarities, and help you decide, which one of the two approaches - or maybe both - could be the right one for you.

In this article we will answer the following questions:

• What are the similarities of, and differences between, Marketing Mix Modeling and Media Mix Modeling?

• How to set up a model?

• What are the use cases for Media Mix Modeling and Marketing Mix Modeling?

• What questions could I answer with the models?

• Could I combine Media and Marketing Mix Modeling together?

What is the difference between Marketing and Media Mix Modeling?

Marketing and Media Mix Models are types of statistical analysis involving multiple influencing factors, developed by data scientists to estimate the impact of numerous drivers on the target KPI. Both models are important tools for marketers that help measure the impact of their activities, reallocate budgets, and predict future performance.

The two approaches differ mainly in terms of the scope of influencing factors taken into account and the addressed target KPIs, i.e. the purpose of modeling. The time horizon and the level of detail also vary.

Media Mix Modeling uses data-driven advertising insights to measure the impact of individual media channels such as display, TV, radio, social media, OoH, and so on. The model shows the relation between the media performance such as the number of contacts, gross rating points, impressions or budget, and the brand development, e.g. ad or brand awareness, to generate clustered insights about what's working in a media campaign and what’s not. Media Mix Models analyze historical campaign data to predict future results. It allows marketers, especially media planners, to better understand how their media mix affects their marketing KPIs. This approach helps optimize the media mix, reallocate budgets between the different media channels, and predict the outcome of advertising campaigns.

Questions that could be answered with a Media Mix Model:

• How can I achieve the highest possible impact with an advertising campaign?
• What is the most effective media mix?
• How should my marketing budget be distributed across different channels over a planned period of time?

Marketing Mix Modeling, on the other hand, has the purpose to figure out the effectiveness of each element, or ‘ingredient’, in a company’s marketing strategy. It evaluates the performance of all available marketing levers, not only the media - as media mix modeling does - but also all kinds of promotions, product launches, price changes, and other drivers. It's a more holistic approach that takes into account various external factors like weather, seasonality, or competitors' activity. By understanding how different elements interact with each other, marketers can allocate resources in a way that produces the maximum impact. Often the key target in marketing mix modeling is sales or ROI. You could learn more about Marketing Mix Modeling here.

Questions that could be answered with Marketing Mix Modeling:

• What marketing activities are driving my target KPI, e.g. sales revenue or profit?
• What external factors have the greatest impact on my target KPI?
• How should I best allocate the marketing resources?
• In which marketing activities should I invest and in which not?
• What is the optimal level of marketing spending?
  
  

When comparing the approaches, it becomes clear that Media Mix Modeling has a more specific scope - i.e. optimizing the overall media mix and advertising campaigns in particular, whereas Marketing Mix Modeling takes a broader range of influencing factors into account. Media Mix Models could deliver recommendations for immediate adjustments in ongoing advertising campaigns, whereas Marketing Mix Models can help optimize the overall marketing strategy. However, as Marketing Mix Models include media channels as influencing factors, it could be argued that marketing mix modeling is an expanded version of Media Mix Modeling. In the end, it depends on the target KPI that the marketers want to maximize.

How to set up a Model?

To set up either a Media or Marketing Mix Model, historical data is needed. The accuracy of the results heavily depends on the data quality and completeness of the data inputs. If the data is inaccurate or incomplete, the output of the analysis might be distorted. The length of the data inputs is also important for the accuracy of the model. Depending on the granularity of the model and whether the data is daily, weekly, or monthly, an average of one to three years of data may be sufficient.

To be able to allocate budgets most efficiently and maximize the impact of advertising campaigns, the following data is needed for Media Mix Modeling:

• performance data (e.g. impressions, reach, clicks, leads, GRP)
• spends per channel or publisher

To be able to identify trends and create insights about business performance, the following data is needed for Marketing Mix Modeling:

• sales figures (volumes sold, revenues)
• data on competitors’ activity
• media variables (both performance, e.g. reach or impressions, and spend) per channel
• promotions
• external factors like weather, inflation rate, or holidays

As more factors are included in Marketing Mix Modeling, more data sources have to be processed and analyzed, which makes the modeling more complex and the development more time-consuming. But in the long run, Marketing Mix Modeling could be used for forecasting the impact on different KPIs.

If you would like to learn how to set up Marketing Mix Modeling in more detail, you could read our article about it here.

What are Use Cases for Media Mix Modeling and Marketing Mix Modeling?

A typical use case for Media Mix Modeling is a media agency that wants to know how to optimize its media plans towards a certain KPI. Let's assume we are a company selling ice cream. We have a brand new product “vegan ice cream” and brief our media agency to plan a campaign to increase brand awareness for this new product. As we already have experience with media campaigns for other ice cream products in the last 4 years, the media agency could use this data to predict the impact of the new campaign. It must be mentioned that accurate results can only be provided for the channels that have already been used by the brand in the past. For example, we have done Social Media, Display, TV, and Out of Home campaigns but no Radio. Therefore we are unable to forecast the uplift in brand awareness that could be achieved with radio spots. After processing the data and combining it with information from some external sources like weather data and competitor sales of vegan ice cream, we can proceed with media mix modeling to predict the optimal media mix and budget split between Social Media, Display, TV, and Out of Home. During the campaign, the media agency gathers intermediate results and can use the model to optimize the media mix and use the allocated budget most effectively. In our example, we see that we should invest mostly in the summer months and shift the budget toward social media. The insights of Media Mix Modeling could help media planners develop different scenarios taking into account the ad stock effect, competitors' advertising activities, or external factors like the weather.

Marketing Mix Models are often used by marketing departments within companies to optimize their overall marketing strategy and mix. If we were a company selling electric bikes, Marketing Mix Modeling could help us decide which marketing channels have the greatest impact on our sales revenue, taking into account that, for example, in summer, and for Christmas people are buying more bikes than during other seasons, or considering our new promotion with special prices or the actual inflation rate. Similar to Media Mix Modeling, we have to collect media data from our campaigns over at least the last 3 years as well as sales data, information about promotions, etc. Assuming we did a great job here, we could process the data and enrich it with external information. As mentioned in this example, seasonality, inflation, and competitors’ activities are the factors that may have a significant impact on sales. We could now use the marketing mix model to allocate our marketing budget toward the most promising marketing channels. In the long run, it will be possible to analyze which influencing factors have a high impact on sales. In our example, we get higher sales if we do a promotion in combination with a TV campaign right before Christmas.

How to Combine Media and Marketing Mix Modeling

As advertising usually comprises a tangible share of marketing activities, Media Mix Modeling could be an important part of a holistic Marketing Mix Modeling process aimed at optimized budget allocation and media mix. As Media Mix Modeling is less complex, companies could start with it as the initial step and in the long run include more drivers into the model for a more comprehensive view of marketing activities. 

Conclusion

Both Media and Marketing Mix Modeling are powerful tools for measuring and optimizing the performance of marketing activities. With efficient Media Mix Modeling, marketers can gain valuable insights about their advertising campaigns and make informed decisions based on the available data. With Marketing Mix Modeling, marketers go one step further, taking multiple marketing and sales activities as well as external factors into account. The potential impact of both modeling approaches is huge, and they are must-have tools for any marketers who want to maximize the value of their marketing efforts. The decision on which approach to choose depends on many factors, including the company’s goals, the data available, the current expertise as well as the existing resources.