GA4 conversion rate is two metrics, not one, and GA4 will happily show you either without making the difference obvious. Session conversion rate divides converting sessions by total sessions. User conversion rate divides converting users by total users. On a site where people research over several visits, those two numbers can differ by a factor of two or more.
Add the 2024 rename of conversions to “key events” and it is easy to see why this metric generates so many arguments in reporting meetings. Here is how to find it, how to read it, and what quietly distorts it.
Key events replaced conversions — mostly
In 2024 Google renamed GA4 conversions to key events. The mechanic is unchanged: you mark an event as important, and GA4 starts counting it separately. What changed is the vocabulary and the boundary with Google Ads.
Key events live in GA4 and drive GA4 reporting. Conversions are what Google Ads calls the key events it imports for bidding. So a screenshot from 2023 says conversions, your GA4 admin panel says key events, and your Ads account says conversions again. All three refer to the same underlying thing in most setups. Nothing broke; the label moved.
The conversion rate metrics in GA4 still carry the conversion name — session conversion rate and user conversion rate — which is a small mercy for anyone writing documentation.
How to find conversion rate in GA4
Neither conversion rate metric appears in standard reports by default. Two routes get you there.
Customise a standard report
Open a report such as Acquisition → Traffic acquisition, click the pencil icon at top right, choose Metrics, and add Session key event rate and User key event rate (the naming has shifted with the rename, so look for both wordings depending on your property’s interface version). Apply, then save either to the existing report or as a new one. This is the fastest permanent fix and it survives across sessions.
Build an exploration
In Explore, start a free-form exploration, drag in dimensions such as Session default channel group or Landing page, and add the conversion rate metrics. Explorations let you apply segments and compare date ranges, which is where the metric earns its place. Standard reports monitor; explorations investigate.
Either way, you can scope the rate to a specific key event using the dropdown above the metric, rather than lumping every marked event together. A blended rate that mixes newsletter signups and purchases is not a rate anyone can act on.
Session conversion rate vs user conversion rate
Choose deliberately. The rule of thumb that holds up:
- Session conversion rate for channel and campaign performance, and for anything you are optimising at the visit level — landing pages, ad creative, checkout flows. It answers “of the visits we bought or earned, how many converted?”
- User conversion rate for business reporting and for long consideration cycles. It answers “of the people who found us, how many eventually acted?”
For SEO specifically, user conversion rate usually tells the truer story. Organic search is often the first touch in a journey that finishes through direct or branded search weeks later. Session rate systematically undersells it. Report both if you can, label them precisely, and never let a deck silently switch between them month to month.
Ecommerce conversion rate in GA4
For ecommerce, the rate you want is scoped to the purchase event, not to all key events. GA4’s Monetisation reports expose ecommerce purchases, and once purchase is marked as a key event you can select it in the conversion rate dropdown to get a clean purchase rate.
Three things routinely corrupt ecommerce conversion rate in GA4:
Self-referrals from the payment gateway. If checkout hands off to a third-party processor and back, the return trip can start a new session attributed to the processor’s domain. That inflates session counts, splits credit, and drags the rate down. The fix is cross-domain measurement plus an unwanted referrals list.
Duplicate purchase events. Thank-you pages that fire on refresh, or a server-side and client-side tag both sending purchase, produce inflated revenue and conversion counts. Deduplicate on transaction_id and verify in DebugView.
Missing items array. A purchase event without a properly structured items array will still count as a conversion but will leave your product-level reports empty, which people then interpret as a tracking outage.
Why your GA4 conversion rate looks wrong
Before you conclude the site got worse, work through the mechanical explanations.
Too many key events. Marking eleven events as key events makes the blended rate meaningless and mechanically inflates it. Keep the list to genuine business outcomes and scope your reporting to one at a time.
Enhanced measurement double-tracking. Enhanced measurement already collects scrolls, outbound clicks, site search, video engagement, file downloads and form interactions. Teams add manual GTM tags for the same interactions, then mark both as key events, and the count doubles.
Attribution is retroactive. GA4 defaults to data-driven attribution, and changing the reporting attribution model recalculates historical figures rather than only affecting future data. Last-click is a setting, not the default. If channel-level conversion rates shifted overnight and nothing was deployed, check whether someone changed the model.
Processing delay and thresholding. Standard reports take 24 to 48 hours to settle, and data thresholding withholds rows when Google Signals is on and the audience is small. Both routinely masquerade as a tracking failure.
Benchmarks, and why most of them are useless
Published conversion rate benchmarks are the weakest input in this whole exercise. The number depends entirely on what you defined as a key event, and no two properties define it the same way. A site counting newsletter signups reports 8%; the identical site counting only purchases reports 1.4%.
Use internal comparisons instead. Compare organic against paid for the same landing pages. Compare this quarter against last for the same key event with the same attribution model. Compare your worst-converting page that still gets meaningful organic traffic against your best. Those comparisons are apples to apples and they point at work you can actually do.
When you do need an external reference point, benchmark against the weakest competitor still holding a page-one position for your target term, not against an industry median. Beating the weakest page-one result is achievable this quarter. Beating a median calculated across a thousand unrelated sites is not a plan.
Connecting conversion data back to search
Conversion rate tells you what happened after arrival. It says nothing about which queries produced the arrivals, what position you held, or which impressions you never converted into clicks. A conversion rate that improves while revenue falls usually means you lost volume on a broad term and kept a smaller, better-qualified slice — a story GA4 cannot tell on its own.
SEO Rocket connects Google Search Console and GA4 and treats them as ground truth for your own site’s performance, showing real clicks and sessions alongside third-party index estimates with each source clearly labelled. That lets you read a conversion change against a query and position change instead of guessing. It does not replace GA4, build custom GA4 reports, or implement tracking code for you — your property, your tag, your measurement plan.
A workable setup
Define three to five key events that map to revenue or qualified pipeline, and no more. Scope every conversion rate you report to a single named key event. Decide once whether the organisation reports session or user conversion rate and write it into the template. Verify in DebugView that each key event fires exactly once per real outcome. Then leave the attribution model alone unless you can explain to everyone reading the report why historical numbers just moved.
