Module 1: Analytics Mindset

Making Decisions with Data

Every manager makes decisions under uncertainty: which product to stock, which branch needs support, whether a promotion worked, whom to hire. Experience and judgement matter, but they can also mislead, especially when a business grows beyond what one person can see. Data does not replace judgement; it sharpens it. This lesson explains what data-informed decision making really means and how to start practising it in your own unit.

Data-informed, not data-obsessed

There are three common ways managers make decisions:

The aim of this course is the third approach. Data tells you what is happening, sometimes why; your judgement adds context, values and strategy.

Why gut feel is not enough

Human judgement is affected by well-documented biases. A few that commonly affect business decisions:

• Confirmation bias: noticing evidence that supports what we already believe and ignoring the rest.
• Recency bias: giving too much weight to what happened last week.
• Availability bias: judging how common something is by how easily examples come to mind, such as one loud customer complaint.
• Survivorship bias: studying only successes (for example, branches that are still open) and missing lessons from failures.
• Sunk cost thinking: continuing a project because of what has already been spent rather than its future value.

Data helps by putting these instincts to a test. It is not immune to bias either, which is why later lessons focus on asking good questions and reading analysis critically.

A simple decision framework

For any important decision, work through five steps:

1. Define the decision. What exactly are we choosing between, and by when?
2. State what you believe now. Write down your current view and how confident you are.
3. Identify the evidence that would change your mind. What would you need to see to choose differently?
4. Gather and examine that evidence. Use existing data first; collect new data only if needed and worth the cost.
5. Decide, record and review. Make the call, record the reasoning and what you expect to happen, and check later whether you were right.

Step 3 is the most powerful and the most often skipped. It forces you to be honest about what matters before you see the numbers, which protects against confirmation bias.

A worked example

Grace manages five pharmacies in Nairobi. The Westlands branch manager asks for an extra pharmacist because "we are always busy". Grace's gut says yes; Westlands feels like the busiest branch when she visits.

Using the framework, she defines the decision: hire an additional pharmacist for Westlands this quarter, or not. She writes down her belief: about 70 percent sure it is needed. She decides that the evidence that would change her mind is the number of prescriptions per pharmacist hour by branch and by time of day, and customer waiting times if available. The point-of-sale data shows Westlands is busy only from 5 p.m. to 8 p.m., while the morning is quiet; another branch is consistently busier per pharmacist all day. Grace decides instead to shift one pharmacist's hours at Westlands to cover the evening peak, and to review the other branch's staffing. She records the decision and agrees to check waiting times in two months.

Notice that the data did not make the decision. It reshaped the question from "more staff or not?" to "staff at the right times in the right places".

Common mistakes to avoid

• Seeking data only to justify a decision already made.
• Waiting for perfect data. Many decisions need to be made with the best available evidence; be clear about the uncertainty instead.
• Ignoring what the data cannot show, such as staff morale or customer relationships.
• Not recording decisions, so nobody learns whether the reasoning was sound.
• Treating every decision the same. Small, reversible decisions do not need a full analysis; large or irreversible ones do.