Back to the Journal

Analytics without questions is a screensaver

A dashboard nobody interrogates is decoration. The chart was never the point. The question you brought to it is.

There’s a TV mounted near the ceiling in a lot of offices we’ve walked through, cycling through charts. Sessions. Users. Engagement, whatever that meant this quarter. A line going up and to the right, which is the direction lines are supposed to go.

Nobody looks at it. Or everybody looks at it the way you look at a fish tank: pleasant, vaguely reassuring, completely non-actionable. If the line dipped tomorrow, not one person in that office could tell you what they’d do differently on Thursday.

That’s not analytics. That’s a screensaver with better typography.

Watching versus using

Watching numbers feels like work. You open the dashboard, you nod at it, traffic is up or it’s down, you close the tab feeling informed, and somewhere in there you never actually decided anything, which is the tell. The dashboard performed its little show. You performed your little audience. The site is exactly what it was an hour ago.

Using numbers looks different, and the difference shows up before the chart ever loads. It starts with a question: a real one, with a decision attached. Should we kill the second contact form or fix it? Is anyone reading the case studies we spend three weeks apiece producing? Did the redesign change what people do, or just how we feel about the site?

A chart is an answer. An answer without a question is trivia.

This is also why “check the analytics” never survives as a habit. Tasks with no finish line don’t get finished; they get postponed. People will faithfully open a dashboard for two weeks after it launches and then never again, not because they’re undisciplined but because staring at numbers with nothing riding on them is one of the most boring activities a person can perform at a desk.

Curiosity needs a stake.

Questions that earn their charts

Most sites don’t need more dashboards. They need a short list of things somebody genuinely wants to know:

  • Where do people give up? Every form, checkout, and signup flow has a step where humans quietly leave. Find the step.
  • What do visitors search for and not find?
  • Which pages get read, and which just get landed on? Scroll depth and time tell you whether the words are working or merely present.
  • When we changed something last month, did the number we changed it for actually move?

Notice what each of those has that “how’s traffic” doesn’t: an action waiting at the end. If people abandon the form at the phone number field, you know what Thursday looks like. Good questions come pre-loaded with next steps. Metrics alone just come with more metrics.

What a question does for you

Three things, quietly.

It tells you what to measure, which is mostly a list of what to ignore. The tools will happily report forty numbers about your site. The question shrinks that to two or three, and everything else becomes noise you’re allowed to stop feeling guilty about.

Then it tells you when you’re done. Open-ended dashboard sessions end when you get bored. A question ends when it’s answered, and answered questions close tabs.

And it gives the number a job. “Bounce rate is 62 percent” is a fact. “Bounce rate on the pricing page is double the rest of the site, and we’re deciding whether to rewrite it” is a project. Same number. Entirely different building.

The uncomfortable version

Here’s the part clients don’t always love hearing. If nobody in your company has asked the dashboard a question in six months, the honest move isn’t a better dashboard. It’s admitting the data was serving a decorative purpose: proof that you’re the kind of organization that has data, displayed where visitors to the office can see it.

There’s no shame in that exactly. But decoration has a cost. Someone maintains those charts, someone pays for the tools behind them, and worst of all, the glow of the screen creates a comfortable feeling that measurement is happening when what’s happening is ambiance.

The fix costs nothing. Before you open the analytics tab tomorrow, write down one question with a decision hanging on the answer. Just one. If you can’t think of one, don’t open the tab, and take that as its own finding.

The fish will be fine without you.