A client asked me recently, “Mal, why has our conversion rate dropped when our sales are actually up?” Brilliant question. And honestly, it’s one I hear more and more these days. So I thought I’d write it all up properly, because if one person’s scratching their head over it, plenty more probably are too.
Here’s the short answer: there isn’t just one way to measure conversion anymore. Google Analytics now gives you two different lenses to look through. And depending on which one you’re using, your performance can look like a roaring success or a worrying dip – even when nothing about your actual sales has changed.
Let me walk you through it.
What conversion rate actually tells you
At its heart, conversion rate answers a simple question: out of everyone who visits your site, how many actually do the thing you want them to do – usually buy something? It’s one of the most important numbers in your business, because it tells you how hard your website is working for you.
The trouble is, the maths behind that number can be done in more than one way. And the different methods can tell completely different stories.
One quick note on language first. Google now calls these “key events” rather than “conversions” in GA4. A key event is simply any action you’ve flagged as important – a purchase, a sign-up, a lead. For most ecommerce businesses, the key event that matters is the purchase, so that’s what I’ll focus on here.
The traditional way of thinking
For decades, the classic ecommerce conversion rate has been about as simple as it gets:
Converting sessions ÷ total sessions × 100.
So if you had 100 visits and 5 of those ended in a sale, your conversion rate is 5%. Clean, honest and easy to understand.
The beauty of this approach is that everything is in there. Every single visit to your site counts, including people who pop back two or three times. It paints a wide angle picture of how the whole site is performing, which makes it brilliant for tracking sales volume and comparing year on year.
If you want a true read on money in the till, this kind of session based thinking is your friend.
GA4 gives you two lenses, not one
Here’s the bit that catches a lot of people out. GA4 doesn’t replace the old model with a single new one. It actually offers you two ways to measure your key event rate – and they answer different questions.
1. Session key event rate
This is the one closest to the traditional model:
Sessions with a key event ÷ total sessions × 100.
It tells you what percentage of visits included a purchase. Great for judging the immediate punch of a landing page, an ad campaign or a single visit.
One thing worth knowing: this is based on sessions with a key event, not the raw number of purchases. So if someone buys twice in one visit, that still counts as a single converting session, not two. It also means every extra non converting visit a customer makes can nudge this number down.
2. User key event rate
This is the newer lens, and it’s the default you’ll often see in GA4:
Users with a key event ÷ total users × 100.
Spot the difference? This one focuses on people, not visits. If someone visits five times but only buys once, they still count as a converted user. The four non-buying visits don’t drag the number down. Because of that, this rate almost always looks higher than the session based one – which feels great at first glance!
It’s a genuinely useful upgrade, especially for longer buying journeys where people research over days or weeks before they commit. It gives you a much fairer read on how good your site really is at winning a person over time.
Why the two numbers look so different
This is where heads start to hurt. Neither lens is wrong. They’re simply answering different questions.
The session lens asks: how effective is each visit at driving a sale?
The user lens asks: what share of my audience has bought at least once?
One is about visits. The other is about people. Both are useful – but only if you know which one you’re looking at.
The year on year trap
Here’s where leaders can get caught out, and it’s exactly what sparked my client’s question.
Once a user has bought, the user key event rate locks them in as “converted”. Every return visit after that doesn’t change their status. That’s lovely for measuring people – but it can play tricks when you compare periods.
Picture this. You run a clever repeat-purchase campaign this year – something that gets customers coming back again and again – but you didn’t run it last year. All those extra visits flood in. Under a session-based view, every one of those returning, non buying visits dilutes your converting session percentage, so your rate can look softer year on year, even while your tills are busier than ever.
Flip to a different lens or a different definition and the story changes again. Compare the wrong number across two periods and you could end up worrying about a problem that doesn’t actually exist.
So what should you do?
My honest advice is this: use GA4’s user key event rate for everyday reporting and future-proofing, because it’s the modern standard and it’s not going anywhere. But when you’re doing proper analysis or comparing this year to last, lean on a clean, session based view for an honest, like for like read on sales performance.
The real win isn’t picking the one “right” number. It’s understanding what each lens is telling you -and making sure you’re comparing like with like – so you can make smart, confident decisions instead of chasing a figure that was never the full story.
If your conversion data is leaving you puzzled, that’s exactly the sort of thing we love untangling. Give us a shout – we’ll help you see what’s really going on.
by Malachi Simons – Founder of Hot Dog Solutions
