Analytics insights

Analytics tells you what people do and where. Research tells you why. Use one to aim the other.

Steps

1. Start from a question: "Where do people drop out of the quote flow?" not "What do the numbers say?"
2. Check the data first: date range, tracking gaps, bots and internal traffic excluded, sample size per segment. Note anything that makes it unreliable.
3. Find the pattern: funnel steps with the biggest drop, pages where people leave, segments that behave differently (device, new or returning, source).
4. Look closer: heatmaps for clicks on things that aren't links and content nobody reaches; session recordings for rage clicks, back and forth, and hesitation. Watch at least ten sessions of the problem step.
5. Write hypotheses, not conclusions: "Because a quarter of mobile clicks land on things that aren't links, we believe people expect the cards to be tappable." Each one says how to test it (a usability test, an A/B test, an interview question).
6. Triangulate with research and support tickets where possible.

Output

The question, data caveats, the top three patterns with numbers and screenshots, and a hypothesis plus test for each.

Don't

• Don't explain the why from numbers alone. Don't compare periods without checking for seasonality or tracking changes. Don't include anything that identifies a person.