What to track first

When a business decides to “take data seriously,” the first mistake is usually trying to track everything at once: every click, every scroll, every button. The result is crowded reports, events nobody understands, and a long wait before anything useful appears.

The smarter start is to track less, but what matters more. This article sets out a practical order for what to track first, so you get numbers that help decisions within weeks, not months.

The rule: start from the result, then work backwards

The first thing to know precisely is the result your business lives on: a purchase, a service request or a booking. Then step back: where did the people who achieved it come from? Then step back again: where did the people who didn’t achieve it stop? This order ensures everything you build is tied to what matters.

Level one: core results

This is the absolute priority; no analysis means anything before it:

• For online stores: a completed purchase, with its value and order ID, so it is recorded once and can be matched with the store system.
• For service businesses: a successfully submitted request form, and the type of service requested.
• For bookings: a completed booking.
• For direct contact: a click to call or WhatsApp, bearing in mind that a click isn’t necessarily a real conversation.

Test each of these events yourself: complete a test action and confirm it appeared in your analytics tool only once, with the right value, on both mobile and desktop.

Level two: sources

Once you know the results, you need to know where they come from:

• Tracking parameters in campaign links (UTM) with one naming rule, as we explain in How to build a measurement framework.
• Linking your analytics tool to the ad platforms you use, and sending result events to them so their systems can learn.
• The question “How did you hear about us?” in the form or at first contact. Simple, but it reveals sources the tools don’t capture, such as recommendations and conversations.

Level three: the main journey steps

Now add the steps that come just before the result, so you know where people stop:

With these steps you can draw a numerical path that shows the biggest gap — the starting point for any work on conversion optimization.

Level four: after the conversion

The result recorded on the website isn’t the end of the story. An order may be cancelled, a lead may not be serious, and a deal may close weeks later. So you need to know what happens after the conversion:

• How many orders were actually completed and not cancelled or returned?
• How many service requests became contracts, and which source did they come from?
• How many customers came back to buy a second time?

This information usually comes from the store system or a customer relationship management (CRM) system, not the analytics tool. Linking it to the source is what turns your reports from “which channel brings requests?” into “which channel brings profitable customers?”

What tools do you need at the start?

You don’t need a large toolset to begin. The practical minimum for most businesses:

• A website analytics tool, such as GA4, with core result events set up and tested.
• A tag manager, such as Google Tag Manager (GTM), which makes adding and changing events easier without editing the website each time.
• The measurement tools in the ad platforms you actually use, connected to the same result events.
• One place where orders or customers are recorded, whether the store system, a CRM, or even an organized spreadsheet at first.
• A way to manage consent when you use tracking tools that require it, in line with the regulations in force.

Add other tools when a question arises that these can’t answer, not before.

What can wait?

Some things commonly tracked at the start add little before the earlier levels are in place:

• Scroll depth on every page.
• Clicks on every link and image.
• Time on page as a success indicator in itself.
• Social engagement indicators without a link to visits and results.

These aren’t worthless, but they help with specific questions later. Add them when a decision needs them, not because they are available.

Signs your tracking needs a review

• Order numbers in your analytics tool are far from those in the actual system.
• A large share of visits or requests has no known source.
• Sudden jumps in the numbers without a clear business reason.
• Nobody knows when a particular event was added, or why.

Any one of these means decisions based on the current numbers need caution until it is fixed.

A suggested timeline

1. Week one: core result events, tested and confirmed.
2. Week two: the campaign naming rule, linking analytics to the platforms, and the “How did you hear about us?” question.
3. Weeks three and four: the main journey steps, and a first report showing the path and its gaps.
4. After that: linking actual results from the internal system to sources, then adding what specific decisions need.

The timeline is approximate and depends on the size and technology of the business, but the sequence is what matters: result, source, journey, then after the conversion.

In summary

Don’t start by tracking everything; start with what answers the most important business question. Make sure the core result is recorded accurately, then learn its sources, then map the steps before it, then link it to what happens afterwards. Four levels, built in order, are enough to improve most decisions. For the broader framework behind all this, read Measurement that drives decisions.