How to really transition to data-driven decision-making?

We have developed a guide for you to facilitate the planning of this high-potential transformation for your business. It addresses the most pressing questions that managers will have to cover to set out on this exciting journey:

• What are data-driven decision-making solutions?

• Why invest in data-driven decision-making solutions?

• When to invest in data-driven decision-making solutions?

• How to invest in data-driven decision-making solutions?

• Which steps are needed to get started with the transition?

• How to estimate the costs?

What are data-driven decision-making solutions?

Technical solutions which enable the handling and leveraging of data, aiming to improve business outcomes through transparency, cost-saving, and internal efficiency. Datasets could originate from various sources:

• Internal: Supply Chain Management, Customer Service, Advertising, and so on
• External: market data, competitors, legal environment, and other forces which can influence the company's business.

The resulting outcomes can include production costs, sales, loyalty, retention, margin, average basket price, ROI… These solutions are named Data Science, Business Intelligence, Data Analytics, and so forth. Their common point is that they are objective-based and drive tangible improvements enabled through data.

Why invest in data-driven decision-making solutions?

Because no business is capable of analyzing its data manually anymore. It is not only about collecting, organizing, and presenting data but also about making sense of it and leveraging the analysis to reach the company’s goals.

Some companies may have implemented super powerful excel macros to convert part of their data into insights. But not to the fullest. Why?

• Because The World Economic Forum estimates that by 2025, 463 exabytes of data will be created each day globally – that’s over 210 million DVDs per day! That means a LOT of data!
• Because more data is being created while number crunching
• Because number crunching is time-consuming, so decision-makers have to wait
• Because manual solutions cannot encompass the full spectrum of data
• Because companies cannot resort to a truncated view of their businesses

The pandemic exposed companies’ weaknesses in terms of anticipation and adaptability. The ones that have digitized show better ability to navigate and act in times of crisis. Do I really need to detail this further?

When to invest in data-driven decision-making solutions?

If the above “why’s” are not relevant, Companies may start a transition to the data-driven model for the following purposes:

• Efficiency purposes: data-driven solutions provide analytical support for continuous improvement to drive a business towards its goals
• Transparency and quality control: automation and centralized reporting help deter fraud and reduce human error which could generate great savings
• Agility: a single source of truth at any given time reinforces the company’s capacity to decide quickly based on a clear common vision.

If you want to save on resources and energy as well as avoid unnecessary stress - then it's about time to invest in data-driven solutions.

How to invest in data-driven decision-making solutions?

Transitioning to data-driven business decisions needs some preparation to guarantee success as well as quick internal take-up of the solution after the team onboarding (just like in any well-managed project, as my expert project management team member would confirm).

• “I need a dashboard to keep an eye on everything”,
• “Can you show me the performance results of the top 20 markets first thing tomorrow?”
• “Can you send me this client’s sales volumes and commitments at the regional level over the last five years?”

Typical requests that we hear daily at work, requiring hours of manual work, which involves multiple contacts within the company, even on a global scale. In order to keep pace, the company needs a robust communication network.

Acquiring a BI SaaS license, purchasing a visualization suite, or running a data analytics project will give you a fish, but it makes more sense to learn how to fish yourself: only a long-term solution will bring sustainable results.

5 Pillars to get started with transitioning - change management

The transition to data-driven decision-making requires the following: Involving all management levels will help build the right momentum to make sure the project comes to fruition. To encourage the evolution of the decision-making process, the actual decision-makers need to feel comfortable with what is being brought to life (sounds like common sense, doesn’t it?), better write a change management plan. 

1. Long-term objectives:

For this, it would have to start with a clear definition of the objectives to be achieved. It often requires an assessment of each department's needs and future developments to visualize holistically the potential economy of scale and data synergies to be leveraged. 

While the list of objectives can be endless, the implementation timing and budget are always limited, but we will come to that.

2. Leaders of the data-driven transition:

How many people are going to be affected in their daily job? Is data-driven model implementation an IT project? Or a Finance one? Who is leading the transformation of the company?

This comes as a hurdle for many businesses: limiting the process to one department creates silos while involving everyone hinders decision-making. This is where external partners prove useful to handle multidisciplinary projects involving all stakeholders while keeping the focus on implementation timings and overall goals.

3. Enablement of the data-driven transition:

Do we contract an external partner or build an in-house infrastructure? Do we buy the servers or host them in the cloud? What are the requirements and restrictions of both alternatives?

It entails building a vision for long-term usage because it influences the scalability and agility that the system will require overtime. This is an extremely important part of the process as it determines the financial scope of the project.

We will dedicate an article to the topic of structure and - we will link it to this section.

4. Timing:

There are few limits to what programming and machine learning can do - except being humans: these are rooted in urgency as we explained earlier in “Why focus on code quality in media and marketing software development?” article.

It's essential to find the balance between complexity, scalability, and speed because needs will become more sophisticated as the data maturity of the organization grows.

There is no right or wrong answer to the debate whether to buy off-the-shelf or get a development team to build your proprietary solutions internally, it all depends on the objectives and ambitions of the business with respect to data. But you know: if you want to master driving a Ferrari at Le Mans, you may start training with a Skoda in a parking lot, but it is no use in the long run.

5. Roll-out plan:

Launching a new process across the entire organization and making everyone adapt overnight is truly a challenge - but it is not impossible. It requires good coordination and strong will as well as organized prep time: internal communication, training, test sessions, and planned release to get the teams up to the challenge.

Launching sequentially (per region, per department, etc.) would generally take the same time, but organized into a series of smaller launch phases with shorter cycles. 

"Designing a roll-out plan including training, assessment, and iteration phase at each step of the process will secure higher efficiency and company take-up."

Aiko Müller
Director Implementation & Support, MMT

How to estimate the cost?

As a starting point: the basic calculation deducts the resource savings (man-hours saved on reporting/data gathering or number crunching) from the solution costs (which would usually be a SaaS license).

The costs of designing, building, and implementing a more sophisticated data-driven solution have to be visualized in the long run as benefits and needs will evolve along with data maturity.

The benefits of transitioning to data-driven processes are ranging from faster reporting, more transparent decision making, and information sharing, to better cost control and resource allocation. It can boost the agility of future business decision-making but it will not happen overnight.

This could be enough to start exploring investments in many basic data-driven solutions but it would be shortsighted to limit the calculation to this, as the cost of data-driven decision-making is 4-fold - all offering the choice between internal and external resources:

• Cost of data: the gathering (data connexion and provision) and storing it (servers or cloud infrastructure and its maintenance)
• Cost of resources: IT (for servers), data engineering (for cloud computing and provisioning), data analysts (for modeling and visualization), business intelligence (for insights)
• Cost of implementation: setting up the team, training and follow-ups, support and knowledge base management => without forgetting the time that teams need to invest into the training and usage learning
• Cost of change management support: to guarantee successful implementation, it is important to monitor the changes and push innovations forward. Such investments can be an absolute waste of money when no assistance is provided to the people who have to make their practices evolve. Change management is what would actually achieve data maturity across the organization

Then the nature of savings:

• Resources savings: FTE spend on collecting, preparing, and visualizing data that now can be automated; reporting preparation (internally or across suppliers)
• Company savings: that real-time decision making has allowed
• Efficiencies: in the company operations and supply-chain that data maturity offers.

Conclusion

As we love to call it, the disruptive aspect of data in our world is so mind-blowing that companies have difficulty comprehending it, sometimes even reluctant to accept it. The reason often originates from the risk-averse leaders who foresee disruption as a potential negative impact on the business in the short run and would rather let the next senior team take the plunge, and keep their own legacy protected.

Sometimes disruption is just too much, and transition offers a “safer” - more acceptable - perspective to those risk-averse managers. It creates the potential for learning and growing with the new approach given the necessary time for the team to reach mastery and become more comfortable with the new practices. That is why “change” is to be managed and not enforced, it brings stronger business benefits over time.

Do not hesitate to get in touch!

Sources:

• Olga’s amazing help to review this piece of content.
• https://www.weforum.org/agenda/2019/04/how-much-data-is-generated-each-day-cf4bddf29f/
• https://www.linkedin.com/pulse/please-another-dashboard-sergio-maldonado/
• https://medium.com/privacycloud/mydata-business-models-e3c26c14367f
• https://www.mercurymediatechnology.com/en/blog/code-quality-in-media-marketing/