A business can have more reports, more dashboards and more research than ever, yet still struggle to decide what to do. Collecting another dataset will not necessarily close that gap.
The useful question is what the evidence helps us understand. A number describes something. Insight explains something that matters to a decision.
This is the first of four practical guides on using insight. The examples are illustrative, not findings from a client project.
Distinguish three different things
Data is the material you start with: transactions, interview answers, observations, measurements. It may need cleaning or checking before you can use it.
An observation describes a pattern in that material. For example, purchases from a particular product range are falling faster in smaller stores than larger ones.
Insight adds a supported explanation: what need, belief, behaviour or circumstance could account for the pattern, and why that changes your view of the problem.
An observation can be commercially important before you understand its cause. The distinction helps you avoid presenting an explanation as established when the evidence only supports the pattern.
Work through an example
Imagine a retailer finds that its premium ready-meal range is underperforming.
The sales report gives you the size and location of the decline. It does not tell you whether customers dislike the products, cannot find them, consider them poor value or are buying a different kind of meal.
A store visit reveals customers examining the packs and leaving without selecting one. That adds an observation, not yet an explanation.
Discussions and further investigation suggest that some shoppers see the meals as too small for the occasion they have in mind. The problem may be the fit between portion size, price and the meal people are trying to provide.
That explanation points towards different options from a general awareness campaign. You might test a different pack, change how portions are communicated or focus on a more suitable occasion. You would still need to establish how widespread the issue is and whether the response makes commercial sense.
Ask what would make the explanation wrong
A plausible story is not enough. Could stock availability, promotions, distribution or a change in who shops the store explain the same pattern?
Look for evidence that challenges your preferred account. Combine sources where they contribute something different: transactions show purchasing; observation shows behaviour in context; conversation helps explore the reasons people attach to it.
Agreement between sources strengthens an explanation. Disagreement is useful too: it may reveal a missing part of the picture.
Use a practical insight test
Before putting an insight in a presentation, ask:
- What exactly did we observe, and how reliable is that evidence?
- What is our explanation, and what remains uncertain?
- What does it reveal about people or their circumstances?
- Which business decision would change if it were true?
- What could we test or do differently as a result?
You do not need every insight to be surprising. A clear explanation of an expensive, persistent problem can be more valuable than an arresting line with little behind it.
Try it with your team
Take one claim from a recent report. Write it in three parts: the evidence, the explanation and the decision it affects.
If the explanation is missing, decide what would help establish it. If no decision follows, ask why the finding belongs in the report.
Next: Ask questions people can actually answer.
Have a perspective or a related question?
Discuss this piece