2026-08-05

Analytics in 5 minutes a week: the only numbers you need

You opened the analytics tab three weeks ago. You closed it two minutes later.

Not because you did not care. Because the screen asked you a question before it answered one.

Which report? Which date comparison? Users or active users? And what is an engaged session anyway?

So you closed the tab and went back to work. Now you run your site on a feeling. Some weeks feel busy, some feel quiet.

You cannot tell whether the article you spent Saturday on did anything at all.

Here is what I want you to take from this article. The problem is not that you are bad with numbers. The problem is that you were handed a tool built for a different job and then made to feel slow for not enjoying it.

You need four questions and five minutes. That is the whole system.

You are not bad at this

Let us deal with the feeling first, because it is what actually stops people.

Analytics anxiety is a design problem wearing a costume. A tool that needs a course before it answers a basic question has failed at its job, not at yours.

"Don't make me think!"

Steve Krug, Don't Make Me Think, his first law of usability

Krug wrote that about web pages. It applies exactly to the tool you use to measure them.

If you have to remember which report holds the answer, the tool made you think. Every one of those small hesitations is a reason to close the tab and they add up until you stop opening it.

So the fix is not to try harder. The fix is to have a routine so short that skipping it takes more effort than doing it.

The five minute routine

Do this once a week. Same day, same time. Friday afternoon works well, because the week is finished and the decision is for next week.

Open your dashboard and answer four questions in order. Do not go looking for anything else.

1. Did more people come than last week?

One number, compared with the previous week. Up, down or flat.

That is it. Do not open a segment. Do not check the year on year.

If it is roughly flat, you are done with question one in ten seconds. If it moved a lot, question two will usually explain it.

2. Where did the new people come from?

Look at your traffic sources, ranked. You are looking for a change in the order, not for exact percentages.

Did search grow? Did something bring a burst from a social platform? Did a partner link start sending people?

A new name near the top of that list is the single most useful thing you will see all week. Somebody linked to you and you did not know.

3. Which page did the work?

Your top pages, ranked. Again, look for movement rather than absolute numbers.

If a page you wrote three months ago suddenly climbed, that is a signal about what your audience wants. It is also the cheapest content decision you will ever make, because the topic is proven and you already know how to write it.

4. Did anyone do the thing you wanted?

Signups, orders, contact forms, whatever your site is for.

If you only ever track one thing on purpose, track this. A week with more visitors and fewer of these is a worse week, not a better one.

Four questions. If your tool shows all four on one screen, this takes about ninety seconds. The five minutes is for thinking, not for clicking.

Still with me? Good, because the next part is the one people skip and then wonder why nothing changes.

What to do when a number moves

Looking is not the point. Deciding is. Here is the whole decision table.

• Traffic up, source unchanged. Something you published is working. Write the next one on the same topic.
• Traffic up, new source in the list. Go and look at that source. Someone mentioned you. Thank them or at least learn what they said about you.
• Traffic down, search down. Check whether the pages that lost traffic still load and still exist. Most sudden search drops on small sites are technical, not editorial.
• Traffic flat, conversions down. The problem is on the page, not in the marketing. Open the page and try to complete the action yourself on a phone.
• Traffic down, conversions flat. You lost low intent visitors. This is often fine. Do not panic and do not change anything.

Notice that three of the five lead to doing nothing or doing something small. That is correct. Most weeks should not produce a change of plan.

The fifteen minutes once a month

Once a month, do one extra thing. Not a report. One question.

Pick the page that gets the most traffic and check what people do next. Do they go deeper into your site or do they leave?

If they leave, that page needs one clear link to the next step. This is the highest value fifteen minutes in the entire routine and most small sites never spend it.

What to ignore on purpose

Permission to skip things is part of the system. These are the ones that eat time and rarely change a decision on a small site.

• Bounce rate as a quality score. A visitor who read your article completely and left satisfied is a bounce. On a blog, a high bounce rate is often the sound of a job well done.
• Session versus user counts. Interesting to an analyst. Not to your Friday afternoon.
• Average time on page. Heavily distorted by how it is measured. Use it for rough comparisons at most.
• Anything you cannot act on. If you cannot name the decision a metric would change, you do not need it this week.

The point of a routine is not to know everything. It is to notice the few things that should change what you do.

Where StatJolt fits in

I built the dashboard around those four questions, in that order.

You open it and the answers are already on the screen. Visitors against the previous period, ranked sources, top pages, conversions. Nothing to configure, no report to build, no saved view to set up first.

This is the whole screen:

The numbers are real time so a Friday check covers the week that just ended, not the week that ended two days ago.

There are no cookies and no consent banner so you are also not making Friday decisions on a sample that quietly dropped everyone who declined.

If you only read one box

Analytics anxiety is a tool design problem, not a competence problem and you can run a small site well on four questions a week. Did more people come, where from, which page worked and did anyone do the thing you wanted. Most weeks the honest answer is that nothing changed and knowing that in ninety seconds is the point.

Back