I Audited My Own GA4 Key Events and Un-Marked 2 of the 7

The short version: I marked 7 GA4 events as key events on this site. Two of them, linkedin_click and view_search_results, could never be more than Secondary conversions in Google Ads, so I un-marked both. Five real key events left, with room to add one later.

GA4 has more than 40 possible event types: 3 automatically collected events that fire no matter what (session_start, first_visit, user_engagement), roughly 10 enhanced measurement events you can toggle on with zero code (page views, scroll, outbound clicks, site search, video engagement, form starts and submits, file downloads), 35+ recommended events with predefined names for things like sign-ups and e-commerce, and unlimited custom events you define yourself.

I’ve marked 7 of those as key events, GA4’s current name for conversions, on andyforsberg.com:

Key event	What fires	Filtered on
newsletter_signup	form_submit	form_id = newsletter-form
contact_form_submit	form_submit	form_id = contact-form
linkedin_click	outbound click	URL contains linkedin.com
toolkit_resource_click	click	link_classes contains tk-card
chrome_extension_click	outbound click	URL is the Chrome Web Store, for my extension WPSNIFFER
wp_plugin_click	outbound click	URL is wordpress.org/plugins, for my plugin Sexy Author Bio
view_search_results	built-in on-site search event	none, used as-is

Six of those seven are enhanced measurement events (form_submit, click) narrowed down with a filter, not true custom or recommended events. That’s a normal way to build key events, but the underlying event still fires for every form and every click on the site. My key event is a slice of that.

If I ever change a form’s ID, a link’s URL, or the tk-card class, the conversion breaks silently. No error, no warning. The number quietly drops in the reports, and for toolkit_resource_click, it drops for all 204 tools at once, not one at a time.

GA4 also has recommended events built for exactly two of my use cases: sign_up for the newsletter form and generate_lead for the contact form. Using those instead of a filtered form_submit would survive a URL or form-ID change, since the event fires on its own name, not on a value that can drift.

Why this decision matters

A key event decides two things: what GA4’s reports treat as “success,” and, if I ever run Google Ads or paid social, what the algorithm optimizes my spend toward.

Mark the wrong action as a conversion, and you train a bidding algorithm to chase it.

GA4 caps you at 30 key events per property, but the cap isn’t the target. Best practice:

1. Log the event first.
2. Confirm it fires correctly in DebugView.
3. Only then mark it as key.

Never mark speculatively. Keep the list short enough that “conversions” still means something. Most small sites do best with 3 to 7 business-critical key events, not 20-plus diluted ones.

Grading my own 7

Mine split into two tiers:

Tier	Key events	Why
Strong	newsletter_signup, contact_form_submit, chrome_extension_click, wp_plugin_click, toolkit_resource_click	Each maps to real downstream value: a lead, a subscriber, product intent, or which of 204 tools someone cared enough to click
Soft	linkedin_click, view_search_results	Shows curiosity, not commitment

toolkit_resource_click used to be toolkit_page_view, which only proved someone landed on the page, not that they used it. Switching the trigger to clicks on the tk-card class, shared by all 204 tool cards on my digital marketing toolkit, fixed that with one filter instead of 204 separate ones.

Primary vs. secondary, if I ever run paid

When you import GA4 key events into Google Ads, each one gets marked Primary or Secondary:

	Drives Smart Bidding	Shows in reports
Primary	Yes, the algorithm spends to get more of these	Yes
Secondary	No influence on bidding	Yes, for analysis only

If I ran a campaign today and left all 7 imported as Primary by default, I’d be telling Google’s algorithm that a LinkedIn click and a search-results view are worth exactly as much as a newsletter signup or a plugin install. That’s a fast way to burn budget optimizing for the wrong signal, on top of the campaign URL mistakes that can already muddy where that budget is coming from.

Sorting them that way raised a better question: if linkedin_click and view_search_results can never be more than Secondary, should they be key events at all?

Marking an event as a key event doesn’t change whether GA4 collects it or whether I can analyze it. Both still show up in Explore, funnels, and segments either way. What key event status controls is narrower: whether the event lands in the Key Events report, and whether it’s eligible to import into Google Ads as a conversion action. Neither linkedin_click nor view_search_results earns that. So I un-marked both.

GA4 still records every click and every search. I stopped treating curiosity as a conversion.

That leaves 5 real key events: newsletter_signup, contact_form_submit, chrome_extension_click, wp_plugin_click, and toolkit_resource_click. All 5 as Primary if I ever run paid, no Secondary tier needed, because nothing weak enough to require one is still marked. Five also sits inside the 3-to-7 range from earlier, with room to add a real one later.

Next: switching the newsletter and contact forms to fire sign_up and generate_lead directly, so they survive a form-ID change the same way toolkit_resource_click now survives everything except a tk-card rename.