Here’s the thing nobody tells you when you switch to Google Analytics 4: there is no single GA4 conversion rate. There are two of them, they measure different things, and they can disagree by a factor of two on the same data set. If you paste “the” conversion rate into a client report without knowing which one you pulled, you are quoting a number you don’t understand. This guide fixes that — the mechanics behind each rate, a worked example that shows the split, and a plain decision rule for which one belongs in your report.
GA4 doesn’t have “conversions” anymore — it has key events
Start with the vocabulary, because Google changed it in 2024 and half the confusion traces back here. What Universal Analytics called a “conversion,” GA4 now calls a key event: any event you’ve flagged as mattering to the business — a purchase, a form submit, a demo booking, a qualified scroll-plus-signup. “Conversions” as a word still appears in GA4, but it now refers specifically to Google Ads conversions, which is a narrower thing. The underlying counting didn’t change; the label did. So when you compute a GA4 conversion rate today, you are really dividing something by your count of key events, and the first job is to make sure you’ve flagged the right three-to-five events as key and nothing junk.
The two rates: session conversion rate vs user conversion rate
This is the split that trips everyone up. GA4 exposes two distinct conversion-rate metrics, and they differ only in the denominator:
- Session key event rate = converting sessions ÷ total sessions. It answers “what share of visits ended in a key event?”
- User key event rate = converting users ÷ total users. It answers “what share of people eventually converted?”
On a site where people buy on their first visit — impulse ecommerce, a cheap app — the two numbers sit close together. On a site with a long research cycle, they diverge hard. A B2B buyer might visit six times over three weeks before filling in the form. That’s six sessions and one user, one of which converted. The session rate sees 1-in-6. The user rate sees 1-in-1. Same human, same outcome, two rates that are 6× apart. Neither is wrong. They answer different questions.
A worked example, so the split is concrete
Say organic search sends you 1,000 users in a month, and those users generate 1,600 sessions because many come back. Of them, 40 people complete a key event, but because some browsed once before converting on a later visit, only 40 of the 1,600 sessions contain the key event. Your session key event rate is 40 ÷ 1,600 = 2.5%. Your user key event rate is 40 ÷ 1,000 = 4.0%. If your paid-search channel, by contrast, converts mostly on the first click, its two rates might both sit near 3.5%. Report session rate across the board and organic looks weaker than paid; report user rate and organic looks stronger. The channel didn’t change between those two sentences — only the denominator did. That is the entire trap.
Where to actually find the GA4 conversion rate
Neither rate shows up on the default reports out of the box, which is the second-most-common source of frustration. To get them:
- Customize a standard report: open a report like Traffic acquisition, hit the pencil/customize icon, and add “Session key event rate” or “User key event rate” as a metric column.
- Build an Exploration: in the Explore section, drag in a dimension (Session default channel group, Landing page, Device) and pull either rate as the value. This is where you’ll do real analysis, because you can segment freely.
- Filter to a specific key event using the “Key events” dropdown so the rate reflects one outcome (say, purchases) rather than all key events blended together — an all-events rate is usually meaningless because it mixes newsletter signups with sales.
If the metric reads 0% or blank, the usual cause is that no event is marked as a key event, or you’re looking at a date range before the event existed.
Ecommerce conversion rate adds another layer
For stores, GA4 also derives an ecommerce-specific rate from the purchase event, and this is where tracking bugs quietly wreck the number. Three recurring culprits inflate or deflate it: payment-gateway self-referrals (Stripe or PayPal starts a new session on return, so the purchase gets credited to “referral / stripe.com” instead of the source that actually earned it); duplicate purchase events firing on a thank-you page that customers refresh, doubling your conversion count; and malformed items arrays where the purchase fires but the ecommerce parameters are incomplete, so revenue reports break even though the count looks fine. Configure your referral-exclusion list, dedupe purchases on transaction_id, and validate the event payload in DebugView before you trust any ecommerce conversion rate.
Why your GA4 conversion rate “looks wrong” compared to the old number
People migrating from Universal Analytics almost always see their rate drop, and assume something broke. Usually nothing did. GA4 counts differently: it’s event-based rather than session-based at the core, its default attribution is data-driven rather than last-click, and its session definition resets at midnight and on campaign change, which fragments sessions more than UA did. A “20% lower conversion rate in GA4” headline is frequently just the platforms measuring two different quantities. The fix isn’t to reconcile them to the decimal — you can’t — it’s to pick GA4 as your ground truth going forward and stop comparing to a system Google retired.
Benchmarks are mostly noise — here’s the honest version
You’ll find “the average conversion rate is 2–4%” repeated everywhere. Treat those numbers as close to useless for decisions. A rate blends intent (branded vs cold traffic), offer (a $9 ebook vs a $40k contract), industry, device mix, and how loosely you defined “key event.” A site that flags newsletter signups as a key event can post a 15% rate that means nothing next to a store’s 2% purchase rate. The only benchmark that earns its keep is your own trend and your own channels against each other, measured the same way. Is organic’s user rate rising quarter over quarter? Is your top landing page converting better than the site average? Those comparisons hold; cross-site averages don’t.
The decision rule: which rate to report
Cut through it with one rule. Use session rate to judge and optimize a channel or campaign; use user rate to report the business outcome, especially over long consideration cycles. Session rate is the right lens for a landing-page A/B test or a paid campaign, where the unit of decision is the visit. User rate is the right lens for a board deck or an SEO retainer report, where the unit of value is the person who eventually bought — and where organic search, which often plants the first touch weeks before the purchase, would be unfairly penalised by session-only counting. Whichever you choose, label it explicitly and keep it consistent. The unforgivable move is switching silently between the two.
Connecting conversion rate back to the keywords that earn it
A conversion rate on its own is a scoreboard, not a strategy. The leverage is upstream: which queries bring visitors who convert, and which bring bounces. GA4 tells you a landing page converts at 4%; it won’t tell you that the page ranks for three commercial keywords worth chasing harder and a dozen informational ones that inflate traffic but never buy. That’s where keyword-level SEO data closes the loop. In SEO Rocket, AI keyword research runs on real Ahrefs data so you can see the volume, difficulty, and intent behind the terms feeding a page, then map high-converting pages to the keywords actually worth defending. Pair that with rank tracking and you can watch a keyword climb, traffic follow, and — in GA4 — the user conversion rate for that landing page move in step. It’s the same weakest-competitor, evidence-first playbook we’ve run across 1,000,000+ ranking pages: measure the outcome in GA4, but fix the input at the query.
A workable setup you can ship this week
Keep it lean. Flag three to five genuine key events and delete the vanity ones. Standardise on one attribution model so month-over-month comparisons mean something. Add both session and user key event rates as columns to your acquisition report, and build one Exploration that breaks user rate down by landing page. Exclude your payment processors from referrals, and dedupe purchases. Then report user rate to stakeholders, session rate to your optimisation team, and never quote a blended all-events rate. That setup gives you a GA4 conversion rate you can actually defend in a meeting — which is more than most analytics dashboards manage.
Frequently asked questions
What is a good GA4 conversion rate?
There’s no universal “good” number, because it depends entirely on what you flagged as a key event and how commercial your traffic is. A lead-gen site might see 3–5% user rate; a cold-traffic ecommerce store might sit at 1–2%. The useful benchmark is your own trend over time and your channels compared to each other using the same rate.
Why is my GA4 conversion rate lower than Universal Analytics was?
Because the two tools count differently — GA4 is event-based, uses data-driven attribution by default, and fragments sessions more aggressively. The gap is a measurement difference, not a real drop in performance. Adopt GA4 as your baseline and stop reconciling to the old figure.
Should I use session or user conversion rate for SEO reporting?
User conversion rate, in most cases. Organic search frequently drives the first of several visits before a purchase, so session-based counting understates its true contribution. User rate credits the person who eventually converted, which is fairer to long research cycles.
Why does GA4 show a 0% conversion rate?
Most often no event is marked as a key event, so there’s nothing to divide by. Check Admin → Events, toggle the relevant event to “key event,” and confirm your date range falls after the event started firing.
