Business AnalyticsInstitute
Back to blog

When dashboards lie — and how to know

Most teams don't have a data problem. They have a 'looks fine in the chart, breaks in production' problem. Three checks every analyst should run.

BAI Editorial·April 16, 2026·1 min read

When dashboards lie

Most analytics teams don't have a data problem. They have a "looks fine in the chart, breaks in production" problem.

In this piece we walk through three checks every analyst should run before publishing a dashboard or briefing.

1. Is the denominator stable?

Half of all dashboard arguments end up being denominator arguments. Pin yours, document it, and re-check it every quarter.

2. Are you summing rates?

Sums of rates are almost always wrong. Re-derive the rate from the underlying counts.

3. Is the join doing what you think?

LEFT JOINs to event tables are a common source of silent inflation. Audit the row counts before and after.

A good check is cheap, fast, and runs every time. A great check is automated.

That's it. Three checks. Run them weekly.

Stay in the loop

Get next week's edition in your inbox

Field notes on AI, analytics, and shipping. Weekly. Free.

Subscribe