GA4 EventsNaming what matters, marking what counts

Why event design is a business decision#

Teams often treat GA4 events as a developer task: add a tag, check it fires, move on. That misses the point. The events you choose and the ones you promote to key events become the organisation's working definition of success. If they are wrong, every report and every bidding algorithm downstream optimises for the wrong thing.

An event plan should therefore start in a meeting with marketing, sales and whoever owns revenue, not in the tag manager. The question is not 'what can we track?' but 'which actions tell us someone is moving towards becoming a valuable customer?'

Track generously, mark sparingly.

Step 1: map the journey to a short list of actions#

List the steps a visitor takes from first arrival to becoming a customer, then the actions that signal progress at each step. For a B2B service site that might be viewing a service page, downloading a guide, starting a contact form and submitting it. For a D2C store it is viewing a product, adding to cart, beginning checkout and purchasing.

Every action on the list should earn its place by answering a question someone will actually ask. If no report or decision would use it, leave it out. You can add events later; cleaning up years of noisy event data is much harder.

Include the actions that happen off the website too. In many Indian businesses the real conversation starts on WhatsApp or a phone call, not a form. A click on a WhatsApp button or a tap-to-call link is worth tracking as an intent event, and the eventual qualified lead should come back from the CRM so that the journey is complete.

Step 2: choose names using recommended events first#

Google publishes a list of recommended events with set names and parameters for common actions such as login, sign_up, generate_lead, add_to_cart, begin_checkout and purchase. Use these wherever the action fits, because GA4's reports, ecommerce features and integrations recognise them. Check the current list in Google's documentation before building.

For actions unique to your business, create custom events with a strict convention. We recommend lowercase snake_case, an object_action pattern (for example quote_request or brochure_download), and no spaces, dates or personal data in names. Event names are case-sensitive, so Form_Submit and form_submit become two separate events.

Step 3: decide the parameters#

Parameters turn a count into an insight. A generate_lead event with no parameters tells you how many leads arrived. The same event with service_line and form_id tells you which offer and which form produced them. Add a value parameter wherever you can estimate one, even roughly, so that reports and bidding can tell a high-value enquiry from a low-value one.

Custom parameters will not appear in standard reports until you register them as custom dimensions or metrics in the GA4 admin. Register only those you will report on, since properties have limits. Never send personally identifiable information such as names, emails or phone numbers in parameters; it breaches Google's policies and, in many jurisdictions, privacy law.

Step 4: mark key events, and choose what feeds bidding#

Mark as key events only the actions that represent real commercial progress. Three to five is plenty for most sites, and each should be something a sales head or finance lead would recognise as progress without explanation. Then make a separate, more conservative decision about which of those are imported into Google Ads (or sent to other platforms) as conversions for bidding.

This second decision matters more than most teams realise. Automated bidding optimises towards whatever you tell it is valuable. Feed it form starts and it will find people who start forms. Feed it qualified leads, ideally via offline conversion tracking from your CRM, and it learns to find customers.

Step 5: test before you trust#

Use Google Tag Manager preview mode and GA4's DebugView to confirm that each event fires once, at the right moment, with the right parameters. Test with consent granted and denied. Then, after a week of live data, compare key event counts with your CRM or order system.

Reading key event rates sensibly#

Once events are clean, the simplest useful number is the rate at which sessions produce a key event. Use the calculator below to compare landing pages or channels, but read results alongside volume: a high rate on a handful of sessions proves little.

Two refinements make the rate more honest. First, segment it: a blended rate across all channels hides the fact that branded search and cold social traffic behave very differently. Second, weight it by value. If generate_lead carries an estimated value per service line, a page producing fewer but larger enquiries can be seen to outperform one producing many small ones.

Step 6: govern the event dictionary#

Event schemes decay. A campaign team adds promo_click_diwali, a developer adds formSubmit2, an agency adds its own lead event for a landing page builder. Within a year the property has three events meaning the same thing and reports that double count. The defence is an event dictionary: one shared document listing every event, its trigger, its parameters, its owner and the date it was added or changed.

Make the dictionary the gate. No event goes live until it is in the document and reviewed by its owner. Review the list each quarter, retire events nobody uses, and note retirements with dates so that historical reports remain interpretable. This is unglamorous work, and it is the difference between a property people trust and one they quietly stop opening.

For a fuller picture of drop-off between steps, move to funnel analysis. For the wider setup, see our GA4 guide.

Published by Fabulous.Media, a network of specialist marketing agencies. Updated 9 October 2026. Platform features change often; check current official documentation before acting on platform-specific detail.