Every company collects more data than it uses, and every team has a dashboard somebody built two years ago that nobody looks at anymore. This is not a failure of tooling. The tooling is excellent. It’s a failure of clarity about what analytics is for.
This guide is about treating performance analytics as a decision-making discipline, not a reporting one. Numbers are only valuable when they change behavior.
What analytics is supposed to do
Useful analytics answers two kinds of questions: what’s happening right now, and what should we do differently. A dashboard that only answers the first one is a status report. A dashboard that helps you do the second one is a tool. Most companies have a lot of the first and very little of the second.
The shift from status report to tool happens when each metric has an owner, a normal range, and an action that follows from being outside that range. Without those three things, a metric is decoration.
The two questions every metric must answer
Before you spend a single engineer-hour building a dashboard, hold every candidate metric to two questions. First: what decision would change if this number doubled or halved? Second: who would make that decision, and how often? If you can’t answer both clearly, the metric isn’t load-bearing yet, and the dashboard work is premature.
This filter alone will eliminate most of what teams ask for. Reports that “the leadership team would find interesting” almost always fail it. The point isn’t to make leadership feel informed — it’s to make a specific decision easier to make correctly.
How to spot a vanity metric in five seconds
A vanity metric is one that goes up over time and feels good but doesn’t predict anything you care about. The five-second test: if this number went up 50% next month, would you change anything you’re doing? If the answer is “I’d celebrate, but no,” it’s vanity.
Common offenders: total registered users, cumulative downloads, social media followers, page views without context. None of these are useless, but treating them as the headline metric pulls focus away from the numbers that actually predict business outcomes — retention, revenue per cohort, conversion at each step of a funnel.
Building dashboards people actually use
Used dashboards have a few things in common. They’re short — five to seven metrics, not fifty. They’re updated automatically — nobody is hand-pasting numbers from Stripe. They show direction and magnitude at a glance — you can tell in three seconds whether anything is worth investigating. And they live where people already look — usually a Slack channel or a wiki page, not a separate analytics app that requires login.
The dashboards nobody uses tend to be encyclopedic. They try to be comprehensive, which means they’re impossible to scan, which means they get ignored. Comprehensive belongs in the raw data layer; the dashboard layer is curation.
Where to go from here
The spoke articles in this guide pick up specific subskills: choosing metrics for the stage your business is in, designing dashboards that drive action, running an analytics review meeting that doesn’t waste an hour, and using cohort analysis to understand retention.
If your current dashboard situation makes you defensive when someone asks “what does this number mean?”, that’s the symptom worth solving first.