How to Make Data-Driven Decision

Making a decision implies several steps, from analysis of the situation to formulating an action plan. Let’s see how a data-driven decision is different from just making things happen. 

7 Steps of Data-Driven Decision

Key topics of the article:

Why Do We Need Data-Driven Decisions?

A one word answer is complexity. Any decision that we make is based on data. You decide to cross a street and, intuitively, you gather a lot of data: traffic, weather conditions, data about the behaviour of other pedestrians.

In the business context, we can also make decisions by intuitively following the instinct, but their complexity is high, so this natural approach won’t take us far. At a certain stage, we need to switch to data-driven decision-making.

Key Features of Data-Driven Decisions

Data-Driven Decision-making (also known under DDDM abbreviation) is a practice of gathering and analyzing relevant data to support the decisions.

There is no agreement about a specific process to follow. While many authors approach DDDM from the data side, I’d like to show how data-driven decision-making matches the ideas of strategic planning.

Here are the main features of a disciplined, data-driven approach to decision-making:

  • Tracking KPIs. Use Key Performance Indicators (KPIs) aligned with a strategy; beware of vanity metrics and simple metrics.
  • Keeping rationale on record. Writing down what led you to this decision, the reasons for the decision.
  • Learning from mistakes. Analyzing the outcomes of bad and good decisions, creating learning and improving loops.

Below, we discuss what these features mean in practice.

Alternatives to Data-Driven Decisions

Let’s discuss some alternatives to data-driven decision-making.

KPI-Driven Management

A term “KPI-driven management” typically refers to the practice of building a hierarchy of KPIs and making business decisions according to the KPI trends.

  • Basically, we are talking about a KPIs scorecard. In good hands, it will be aligned with a business context and will guide the organization to the right targets.

Sometimes, the term is used ironically to describe the bias of the management team that focuses on the performance measurement rather than on creating real business value.

Big Data

When discussing data-driven decisions, we assume that data is already available. In contrast, with big data initiatives, we are focusing on extracting information from large volumes of complex data.

When properly collected, analyzed and reported, big data will be a source for a data-driven decision.

Data-Driven Decisions vs. No-Data Decision

I’ve compared the steps of a typical no-data decision-making process (let’s call it “making things happen”) with data-driven decision-making.

1. Understand Business Context

“Making things happen” Data-driven decision-making
We have a great idea – let’s do it! It’s not about just one decision, it’s about understanding context, prioritization, and finding the cause-and-effect connections. Let’s look at a strategy map first.

Ideally, decision-making should be a tactical part of an organized strategic planning process.

Good decisions are coherent with a shared vision of an organization that is presented on a strategy map.

2. Define KPIs

“Making things happen” Data-driven decision-making
The idea looks interesting! Let’s find some KPIs on the Internet for it. We want to make a decision more tangible and more specific with KPIs. We design tailor-made KPIs. Here are the questions we ask:

  • How will we control the execution’s progress (leading indicators)?
  • How will we validate the achieved results?
  • When do we plan to achieve these results? (set target values)

KPIs are the pain point of any performance management system. Some people prefer to work without KPIs, while some prefer to use KPIs from the long list of indicators found on the Internet. In my opinion, it makes sense to invest time in finding performance indicators specifically for your business challenges.

3. Visualize

“Making things happen” Data-driven decision-making
Let’s do some dashboards! It’s easier to catch the trends and anomalies when we have our data on a dashboard. Let’s put the performance data for leading and lagging indicators on the same chart.

It’s not a problem to design an impressive dashboard with any software tool. The problem is to make this information contextual enough to make the difference for decision makers.

  • I’m sure you know the cases when fancy dashboards are used for quarterly reports, but the real decisions are made using tons of custom-made spreadsheets.

In this sense, a strategy execution software (here we talked about the difference between a dashboard and Balanced Scorecard) moves your data one step closer to the strategic challenges.

4. Action Plan

“Making things happen” Data-driven decision-making
We have a plan, and we have a budget! We formulate an action plan based on our current understanding of the situation. We describe:

  • The decision and specific activities it implies
  • The rationale
  • The time and budget needed

The budget is an important part of any action plan, but it is even more important to formulate all the details that stand behind the decision. Such an approach makes it easier to onboard new members of the team, cascade decisions across the organization, and analyze the results (see step 7).

5. Prioritize Decisions

“Making things happen” Data-driven decision-making
Stop doing what you were doing and switch to a new idea. We use priority scorecard to compare new decisions with competing ideas. The one with a higher score normally comes first.

Strategy is about choosing the priorities, deciding what to do first, and what to ignore. Sometimes, it’s enough to have a quick look at the idea to approve or reject it (see step 1), while in other cases, you create your own prioritization framework that takes into account factors important for your organization.

6. Execute

“Making things happen” Data-driven decision-making
We have a plan, someone will execute it. A person who was involved in the discussion now will follow the approved plan. We use leading and lagging metrics as control points. We note down any unexpected findings.

It’s nice to have a picture of how things are doing in real-time, but be careful with KPIs. In most cases, KPIs used for direct control will fail. Instead, use performance measurement as a base for discussion and improvement.

7. Analyze Results

“Making things happen” Data-driven decision-making
It looks like the idea worked/didn’t work for us… For each decision, we plan to analyze the outcomes. We use gap analysis or the OKR framework to review the results formally.

That’s where a rationale described in detail (see step 4) will help. The final performance data is not as important as the job your team did along the way. Don’t just do the “evaluations” – analyze deep reasons for failure/success and suggest strategic improvements.

8. Learning Loop

“Making things happen” Data-driven decision-making
Bad decisions are inevitable… We review the decision-making process itself:

  • What principles were helpful?
  • What approaches need to be improved?

We are improving our decision-making culture: we look for returning problems, remove unnecessary complexities, update templates and standards.

It’s your chance to talk with yourself in the past. Use this step as a retrospective view on the principles your team followed. Improve communications, improve infrastructure, align the internal mechanics better with the value creation for the end users.

If you plan data-driven decisions seriously, look at the KPIs for big data that will help to quantify data sourcing, analysis, and reporting efforts.

Conclusions

Making data-driven decisions is not just about looking at nice BI dashboards. It’s more about a disciplined approach to formulating the problem, quantifying control points, and then following up with the progress and results.

Properly implemented business frameworks, such as Balanced Scorecard or OKR, support  data-driven decision-making “by design.”

What practices does your team use for data-driven decision-making? Feel free to share in the comments.

Cite as: Alexis Savkín, "How to Make Data-Driven Decision," BSC Designer, July 28, 2020, https://bscdesigner.com/data-driven-decision.htm.