How to decide what to track

Google Analytics can record more than six hundred events out of the box. Your team could create thousands more. I have walked into accounts where every click and scroll produced a new metric, yet nobody could explain which numbers mattered at Monday’s growth meeting.

After fifteen years fixing those bloated setups I have learnt one rule: track only what guides budget or roadmap decisions. Everything else clutters reports and hides real insight. This chapter walks you through a simple framework that turns lofty business goals into a lean, consistent event plan that everyone signs off before a single tag ships.

We will start by mapping the conversions that truly drive revenue so the rest of the spec builds on solid ground.

Open a blank document and write the four funnel stages down the page: awareness, consideration, conversion, and expansion. Under each heading list actions that prove a prospect has progressed. Awareness needs only session_start and a pair of scroll_depth checkpoints on cornerstone pages. Consideration might keep pricing_page_view and newsletter_signup because both reveal genuine interest in value.

Conversion events deserve extra care. Choose meeting_booked for pipeline creation or purchase for checkout success. Each of these fires once, carries clear monetary weight, and links directly to revenue modelling. Expansion completes the loop with upgrade_click and renewal_paid. These two events let you measure lifetime value growth without a second database.

Now score every candidate on a simple scale: does this event change budget, improve targeting, or alter product priority? If the answer is no, delete it. The list should shrink fast, leaving perhaps eight to ten events that steer real decisions.

The next task is to name those events so future reports stay clean, which we cover in the following section.

Consistent naming prevents duplicate metrics and endless debate. Use snake_case for every label, keep letters lower case, and avoid spaces or dashes. GA4 treats signup and sign_up as different events, and that typo will split your numbers in two.

Add only the parameters that slice data in meaningful ways. newsletter_signup might carry plan_type and cta_location. meeting_booked could include calendar_name and booking_country. Stay well below the twenty-five parameter limit so you never hit cardinality walls in GA4.

Pin these rules at the top of your tracking specification so nobody improvises under deadline pressure. When a new idea arrives, the first step is to check whether its proposed name follows the convention. If not, it waits until the next revision cycle.

With naming locked you are ready to draft the full specification, the subject of the next section.

Create a table in Notion or Sheets with five columns: event name, trigger, parameters, owner, and date added. Work through your trimmed list row by row. For meeting_booked set the trigger to Calendly success, list the parameters, assign a lifecycle manager as owner, and record today’s date.

Audit existing tags as you go. If an old event overlaps with a new one decide whether to rename, merge, or delete. This step prevents duplicate data flows and keeps historical reports interpretable.

Apply a strict rule: no spec row means no implementation ticket this sprint. The boundary keeps scope stable and debugging manageable. New ideas join a backlog that you can review in the next planning round.

Once the table is complete you need team approval before development begins, which we handle in the final section.

Share the specification link with Growth, Revenue Operations, Data, and Product. Ask one clear question: will these events answer your key performance indicator questions? Require an explicit reaction from each function. A single thumbs-up emoji in Slack counts, but silence does not.

When all stakeholders agree, lock editing permissions until the next sprint. This freeze guards against late additions like hover_teaser_video that creep in during crunch time and break reports a week later.

Store the approved document in the same workspace where you track projects. Future hires will see the context, understand naming logic, and avoid inventing near-duplicate events. A living, visible record is the cheapest form of data governance.

With decisions documented and frozen you can move on to wiring events in Tag Manager, confident that every hit links back to revenue impact.

You now have four assets: a pared-down list of meaningful events, a strict naming convention, a detailed specification, and written cross-team approval. Together they form a single source of tracking truth that developers can implement once and analysts can trust for quarters.

This groundwork prevents the common nightmare where a dashboard shows twenty conversions but finance books only ten deals. By agreeing on events upfront you remove guesswork, speed up troubleshooting, and turn analytics into a decision engine rather than a debate club.

The next chapter covers how to fire these events through Tag Manager and verify them in GA4, so your new specification moves from plan to live data without surprises.