Almost every new client asks me some version of this in the first week. Google Ads says one number, GA4 says a smaller one, the CRM says something else, and someone wants to know which is right.
Usually all three are. They're just counting different things.
The numbers were never meant to match exactly, so I don't spend much time trying to make them. What I want to know is whether the gap makes sense, or whether it's hiding a tracking problem. Those are very different situations.
They're answering different questions
Google Ads only looks at people who interacted with your ads, and it's grading its own homework. GA4 looks at everything that happened on your site, from every channel, and tries to share the credit out.
Once you keep that in mind, a lot of the differences stop looking strange.
Different days
Google Ads puts a conversion on the day of the click. GA4 puts it on the day the conversion actually happened.
Say someone clicks your ad on the 28th and pays on the 3rd. Google Ads counts that sale in last month. GA4 counts it in this one. If your customers take a week or two to decide, this alone can open up a big gap, especially if you're comparing short date ranges.
It also explains why last week's numbers in Google Ads keep creeping up for a while after you first look at them. Don't judge the last few days too early.
Different credit
Most people touch more than one channel before they buy. They see an ad, come back through Google search a few days later, then click a link in an email.
Google Ads will usually give itself credit for that sale, because one of its ads was in the path. GA4 might give it to organic search or to the email. Neither is lying. They're using different rules.
Before comparing anything, check which attribution model each tool is using. In GA4 the advertising reports show how credit gets split across paid and organic, and I often find that view more useful than either headline figure.
Different counting
This is the one I find most often on accounts I take over.
In Google Ads, every conversion action has a counting setting: count every conversion, or just one per click. For a purchase, "every" makes sense. For a lead form it usually doesn't. If somebody fills in your form twice, you've got one lead, not two. I regularly find lead forms set to "every", which quietly inflates the numbers and gives the bidding system a slightly wrong idea of what's working.
The conversion window matters too. That's how many days after a click Google Ads will still count a conversion. And Google Ads can count people who watched or saw an ad without clicking, which GA4 doesn't treat the same way.
It takes ten minutes to open each conversion action and write down how it's set up. It's worth doing.
Google Ads fills in gaps
Some people decline cookies. Some click on their phone and buy on a laptop. Some use browsers that block tracking. Those conversions still happen, you just can't see all of them directly.
Google Ads tries to fill that gap with modelling, consent mode and enhanced conversions. If those are switched on in Google Ads but GA4 isn't getting the same signals, Google Ads will normally show more conversions. That's expected, and for bidding it's usually helpful.
What I check is that these things are actually working. A setting that's switched on but receiving no data is surprisingly common.
Different events
This is where real problems live.
Sometimes the two tools simply aren't tracking the same thing. Google Ads might count a thank-you page view while GA4 counts the form submission. A tag might fire twice when someone refreshes the page. A page in the checkout might be missing the tag altogether.
The worst version is double counting. The account has its own Google Ads conversion tag, someone has also imported the GA4 version, and both are set as primary. Google Ads now counts every lead twice and bids as if you're getting twice the results.
My rule is one primary conversion action for each real business outcome. Then I go through the journey myself, fill in the form, and watch what fires.
Which one should you believe?
It depends on what you're deciding.
For day-to-day campaign decisions, I go with Google Ads, because that's the data the bidding system learns from. It needs to be right more than anything else.
For working out how to split budget between channels, I'd look at GA4, since it's the only one that sees all of them.
And for whether the business is actually making money, I trust neither. I look at the CRM, the order system or the bank account. That's where you find out if those conversions turned into revenue.
A quick check you can do today
Set both tools to last month. In Google Ads, look only at your primary conversion actions. In GA4, look at the same key event, filtered to traffic from Google Ads.
If the gap is modest and roughly the same from month to month, you're probably fine. If it's big, or it suddenly changed, look at counting settings and double tags first. In my experience that's where most of the trouble is.