Synonyms: analytics audit, data review, quantitative review, behavioral data audit, metrics review
Do not index
Definition
An analytics review is a focused investigation that uses existing analytics data to decide where deeper research should happen next. Rather than collecting new data, the team reviews funnels, drop-off points, search logs, and segment differences to identify the highest-value questions to explore through usability testing, interviews, or session replay.
Use cases
Skip the review and you'll spend research time chasing the wrong problem. The review narrows a large product down to the few places most worth investigating.
Fixing the wrong page: A team rebuilds the homepage because it "feels off," while the analytics review would have shown the real leak sitting on step 3 of checkout.
Aiming a small research budget: With money for only a few usability sessions, a review points them at the exact flow that's bleeding users instead of a guess.
Killing a stakeholder myth: "Users hate the new nav." A review checks it against actual behavior in an hour, before the team spends two weeks acting on a feeling.
How it's used in practice
Start with the biggest anomaly: Find the biggest drop-off first. The largest leak is usually where the review pays for itself.
Read on-site search: Search queries are free voice-of-customer data showing what people can't find or name.
Split by segment: Check mobile vs. desktop and new vs. returning. Problems often hide inside an average that looks fine.
Hand off questions, not verdicts: Convert each finding into a question for qualitative research ("why do mobile users abandon step 3?"), since the review shows where, never why.
🪄
Pro-tip: An analytics review's job is to generate good questions, not final answers. The deliverable should not be: "users drop off at step 3," but should be: "why do mobile users drop off there twice as often as desktop?" handed to a usability session.
Treat it as the cheap scoping step that aims your expensive research, and you'll stop wasting studies on the wrong screen. Exit pages and search terms are the fastest signals of where things break.
Challenges & limitations
It inherits bad data: If events are mistagged or missing, the review produces confident, wrong conclusions from broken numbers.
No question, no focus: Walk into the dashboards without a hypothesis and you'll drown in charts and leave with nothing you can act on.
Free resources
VWO: Funnel Analysis — a practical guide to finding and reading drop-off in a funnel.
