Your analytics dashboard is not telling you what to fix

You open your website analytics, see a line that went up or down, feel briefly worried or pleased, and close the tab. Nothing changes. That is the normal experience for most small-business owners, and it is not because the numbers are wrong or too shallow.

The problem is that the dashboard was assembled from whatever your tools happened to measure, not from a question you actually needed answered. A screen full of accurate charts can still leave you with no idea what to do on Monday.

Start with the question, not the chart

“How is my website doing?” cannot be answered. It has no metric, no group of people, and no action attached to it. “Which pages do people land on before they call us?” can be answered, and the answer tells you where to spend your next hour of effort.

Write the question down before you open anything. Good ones for a small site tend to sound like: where do people give up on the contact form, which service page brings in the most inquiries, or how many visitors on phones reach the booking step. Each one implies a change you could make.

This also protects you from the opposite trap. If you start from the report your builder or analytics tool shows by default, you will end up interpreting whatever it chose to display. Bounce rate, session duration, and page views are real measurements, but none of them tell you why someone left without calling.

Numbers hide shapes

The statistician Francis Anscombe made a point in the 1970s that is still the best argument for looking at data rather than reading summaries of it. He built four sets of numbers with identical averages and identical statistical relationships. Charted, they look nothing alike.

The small-business version of this happens constantly. An average page load time of three seconds sounds fine until you split it by device and find that phone visitors, who are most of your traffic, are waiting far longer. A steady monthly inquiry count can hide the fact that one page collapsed while another grew. The average is true and useless at the same time.

So before you accept a single headline number, break it apart by the split that matters to your business: phone versus desktop, new versus returning, one service versus another. If a breakdown changes what you would do, it belongs in your regular view.

Decide what would make you act

Ask yourself, in advance, what result would cause you to change something. If the answer is nothing, you are looking at a chart for reassurance, not for a decision, and you can stop tracking it.

This is also how you avoid expensive false alarms. Say inquiries drop noticeably in a month. A dashboard that only shows the total sends you off redesigning a contact form that works perfectly well. A dashboard that shows inquiries next to where your traffic came from might reveal that a promotion or a social post brought in a wave of people who were never going to buy, which diluted the rate without anything on the site breaking.

The cost of getting this wrong is not abstract. It is paying someone to fix a page that was fine.

What to ask your website provider for

You do not need a custom reporting tool. You need a small, boring view you will actually open. If someone else manages your site, that is a reasonable thing to request in plain terms.

Ask for a single page showing the handful of numbers tied to your questions, split by mobile and desktop, with a comparison to the same period last year rather than last week. Ask that form submissions and phone taps be counted as events, so you are measuring inquiries rather than guessing from traffic. Ask for the report to be emailed to you on a schedule, because a dashboard you have to remember to visit is a dashboard you will not visit.

Keep the set small. Four or five measures you understand will change more decisions than thirty you skim.

The same rule applies to charts on your own site

If you publish figures for customers, on a pricing page, a results page, or a case study, the same discipline holds. Edward Tufte’s long-standing advice was that every mark on a chart should carry information rather than decorate it, and that is a good default. But stripping a chart bare can also remove the context a reader needs to understand what they are seeing.

A customer reading your numbers is not an analyst. Label the axis in words they use, keep the color meaning identical everywhere it appears, and put the takeaway in a sentence above the graphic. If the chart needs a paragraph of explanation to make sense, the sentence is doing the work and the chart is decoration.

Background reading: Rethinking Data Visualisation: A UX Approach To Dashboards That Actually Drives Decisions