CLP DECODES · DATA FOR EVERYDAY DECISIONS

Data Literacy: Turn Data into Better Decisions

Data literacy is not just knowing how to use a dashboard. It is the ability to ask a useful question, understand where the numbers came from, interpret patterns carefully, and explain what the evidence supports—and what it does not.

← Back to CLP Learn Library

Start with a decision, not a chart

Imagine a team notices that customer complaints increased this month. A chart can show the trend, but the decision requires better questions: Which complaints increased? Did the number of customers or orders also change? Was there a product, process, or measurement change? What action would the answer inform?

Begin with a decision or question, then identify the evidence that would help answer it. This reduces the temptation to search a dashboard for a story after the fact.

Five checks before trusting a number

1. Definition: what exactly is being counted? Make sure terms such as “active user,” “conversion,” or “delay” have a consistent definition.
2. Source and collection: who recorded the data, when, and for what purpose? Understand missing records and possible measurement changes.
3. Context and comparison: compare equivalent periods, groups, and units. A larger total may reflect a larger population rather than a worse rate.
4. Variation and outliers: distinguish a sustained pattern from a one-off event or a small sample. Check whether the average hides important differences.
5. Alternative explanations: ask what else could produce the pattern. A relationship between two measures does not, by itself, prove causation.

Example: did a new process improve performance?

Suppose a team reports that average handling time fell after a new process was introduced. Before declaring success, check whether the mix of cases changed, whether quality or rework also changed, and whether the comparison covers enough time to account for normal variation. Consider both speed and the outcome the process is meant to improve.

A responsible conclusion might be: “Average handling time fell in the observed period. We need to check case mix and quality before attributing the change to the new process.” That is more useful than a confident claim unsupported by the available evidence.

Turn analysis into a clear explanation

Use a simple four-part structure:

• Question: what decision are we trying to make?
• Evidence: what did we observe, and what is the source?
• Interpretation: what is a plausible explanation, and what remains uncertain?
• Next step: what should we test, monitor, or decide now?

AI tools can help summarise data or suggest questions, but they can also misread tables, overlook context, or produce unsupported explanations. Verify calculations against the source and keep sensitive data within approved systems.

Try it with a real, low-risk question

1. Write one question that could inform a decision.
2. Choose a small dataset you are allowed to use.
3. Define each measure and check for missing values or inconsistent time periods.
4. Write one observation and at least one alternative explanation.
5. State what you can conclude, what you cannot conclude, and what evidence you need next.

Key takeaway

Data becomes valuable when it improves a decision—not merely when it produces a chart. Ask a clear question, inspect the context, challenge the first explanation, and communicate uncertainty honestly.