Home → Blog → GA4 and Google Ads conversions

Why don't GA4 and Google Ads conversions match?

Google Ads says 42 conversions last month and GA4 says 27, or the other way round. Neither is lying. Most of the time the two aren't counting the same action at all, and when they are, they count it on different clocks, credit it by different rules and fill in what a cookie banner hides in different ways. A gap of some size is normal. Its direction, and whether the causes below can explain its size, tells you whether yours is.

This article takes the causes in the order worth checking them, says which way each one pushes the numbers, and ends with which of the two to use for bidding, for reporting and for setting the budget. The platform facts are from Google's own help pages, checked 15 November 2026.

The usual cause: they aren't counting the same action

I've seen this hundreds of times, and the cause is usually the simplest one: the two platforms aren't counting the same thing. Most often it's naming. The "conversion" in GA4 is a soft conversion action, like a click on a call-to-action button, while the one in Google Ads is a high-intent conversion action, a submitted lead form. No wonder they don't line up.

How far apart can those two be? The nearest published figure with a stated method is Zuko's benchmark of the forms it tracks: of visitors who see a contact form, 14.5% submit it, and of those who start filling one in, 47.4% do. It covers 739 forms, leans towards financial services and gives no date range, so treat it as a sense of scale rather than a norm. A visitor who clicks a button to reach the form is at or before the first of those steps, so a soft count several times the size of the hard one is what you'd expect, not a fault in either platform.

GA4's own vocabulary doesn't help. In March 2024 it renamed its conversions "key events", and a "conversion" in GA4 now means a key event shared with Google Ads for measuring and bidding. Google's stated reason was that the old GA conversions "were measured differently from how Google Ads conversions are measured, leading to discrepancies".

1. Clicks your ad Google Ads records a click 2. Clicks "Get a quote" GA4's conversion: a soft action 3. Submits the form Google Ads' conversion: high-intent
The most common mismatch: two actions with the same name, counting different steps of one visitor's path. Many visitors who reach step 2 never get to step 3, so GA4's count runs higher.

So the first check is not a setting. It's a list: open Goals → Conversions → Summary in Google Ads and the list of key events in GA4, and list what each named action actually fires on. The same problem inside Google Ads alone, soft actions sitting in the Primary column, is covered in the four conversion tracking mistakes.

The causes, which way each pushes and how to check

Once the two are counting the same action, a gap is still normal. Google's own help page on the differences says it "is common and can occur for several reasons, even with a correct setup". These are the main ones, with the direction each one pushes:

CausePushesWhat to do
Fix Setup faults: put these right
Different actions under one nameEither wayList what each action fires on, then count the same action in both
The same form counted twice in Google Ads, by a tag and an imported GA4 key eventAds higher, often about doubleSubmit one test enquiry and watch it arrive; keep one primary action per outcome
Count setting: "One" per click in Google Ads, every event in GA4GA4 higher on "One"Read the counting setting on each action and choose deliberately
Paid visits lost on the way in: a redirect strips the click ID, or auto-tagging is offGA4 lower for paid trafficLoad your landing page with ?gclid=test and see if it survives; turn auto-tagging on
Expected How the platforms work: allow for these
Click date against conversion dateEither wayCompare using the "Conversions (by conv. time)" column in Google Ads
Credit: Google Ads only counts conversions after its own clicks, GA4 shares credit across every channelAds higher than GA4's Google Ads figure GA4's all-channel total higherView GA4 under "Google paid channels last click"
Consent Mode modellingAds higherCheck whether your Consent Mode is basic or advanced, and your ad clicks a week
Conversion windows: 30 days after a click in Google Ads by default, a 90-day lookback in GA4Either wayRead both settings, and align them if you want closer numbers
Engaged-view conversions from video adsAds higherCheck for campaigns running skippable, in-feed or Shorts video ads
Cross-device conversions, which Google Ads includes by defaultAds higherNothing to check
Calls straight from the adAds higherLook for call conversions in the Google Ads list
Time zones, set separately in eachEither way, at the edges of a dayCompare the two settings
Imported GA4 conversions arriving lateAds lower for the last day or twoCompare periods that ended at least 48 hours ago

A few of those need a sentence more. Google Ads reports a conversion "from the date and time of the click that led to the successful action", while GA4 files it under the day it happened, so a click on 31 October and a form on 1 November land in different months. On attribution, both now default to a data-driven model for most actions, but Google Ads only ever counts conversions that followed one of its own clicks, so the two models are splitting different pies. View-through conversions, from ads someone saw without clicking, don't affect the main column at all: Google keeps them out of "Conversions" and shows them only in the "View-through conversions" and "All conversions" columns. Engaged-view conversions are different, and do count: in Google Ads, someone who watches 10 seconds of a skippable video ad, or 5 of an in-feed or Shorts ad, and converts within three days, by default. GA4 has its own, longer "engaged view" for its attribution model, so the two terms don't mean the same thing.

GA4 joins someone's phone and laptop through User-ID, which needs a sign-in most lead-generation sites don't have, or by modelling under the Blended reporting identity. Google Ads models those journeys from people signed in to Google. And a call from a call asset on the ad is a Google Ads conversion that never touched the website, so GA4 can't see it at all; the call tracking guide covers how to count calls properly.

The time zone row has a catch: Google says time zone changes "are no longer supported" in standard Google Ads accounts, so if the two differ, it's GA4's you'd change, and only for data going forward.

On a UK site, anything that feeds Google Ads needs consent under PECR (the Consent Mode v2 article covers what the law asks for), and some visitors never give it. What each platform does about the people who decline is another source of the gap, and several of the pages ranking for this question get it backwards.

With Consent Mode in place, Google Ads models conversions for visitors who declined once the account reaches 700 ad clicks over seven days, per country and domain, and those modelled conversions sit in the main Conversions column. GA4 models too, but in two separate ways. It models key events for visitors who declined, and Google states that "any Google Ads conversions created based on Google Analytics key events will include modeling". So the common claim that GA4 imports carry only observed conversions is wrong by Google's own documentation.

What most small sites won't get is GA4's other kind, behavioural modelling, which fills in sessions and users. It needs advanced Consent Mode, at least 1,000 events a day with analytics consent denied for seven days, and at least 1,000 consenting users a day on seven of the previous 28, and it shows only under the Blended reporting identity. An account buying 100 ad clicks a day crosses Google Ads' threshold, while the site behind it may never come near GA4's. That is one reason GA4's sessions from paid search can look thin beside Google Ads' clicks while its key events look nearer.

Which mode your banner runs in decides how much of this applies. The Consent Mode v2 article explains basic and advanced and what each sends, and the Tag Manager consent guide shows how to set it up.

When the gap means something is broken

Google's help says "discrepancies, often up to 20%, are expected". It gives no data behind the figure, and the 10–30% that vendor pages call normal has none either. A percentage is the wrong test anyway, because a healthy account can sit outside it for the reasons in the table. Better tests are a direction or a size that those reasons can't explain:

  • Google Ads is almost exactly double GA4, or double your enquiries. Two primary actions are counting one form. Google sets conversions created from GA4 to secondary "to prevent counting the same event twice for bidding", so a doubling usually means someone has made a second copy primary.
  • GA4 shows almost no paid visits. Few or no Google Ads sessions in GA4 while Google Ads reports clicks: the click ID is being stripped or auto-tagging is off.
  • GA4 is a multiple of Google Ads. GA4 is counting a soft action, the case above.
  • The gap changed overnight. Something changed that day: a tag, a banner, a site rebuild, a form plugin. Google Ads' change history shows changes to the account; a tag, banner or form change has to be traced on the site.

And if neither number matches the enquiries or sales the business actually recorded, the problem isn't the gap between them. That one is covered in the conversion tracking mistakes article.

Which number to use for bidding, reporting and budget

When an owner tells me Google is inflating its numbers to sell more ads, I tell them it isn't. The answer is almost always more straightforward than that: conversion actions with the same name that are counting different things on the website.

For bidding, I'd almost always use the high-intent action. There are rare cases where testing a softer conversion action is worth it. For reporting, use the lower-funnel actions, because they're closer to real business outcomes. The exception is when someone is actively working on the site's conversion rate (CRO): they should be reporting on the soft actions too.

The rare case is an account where real leads are too few for Google's bidding algorithm to learn from. Google says Target ROAS on Search needs at least 15 conversions in the past 30 days, and recommends judging Target CPA over at least 30. In Google's 2026 bidding video, a Google Ads product manager recommends bidding on "the lowest conversion action in the customer journey that will have volume and is a high-quality indicator of value" (as reported by PPC Land). The cost is the one Navah Hopkins names in Search Engine Journal: "Google Ads will report an artificially high conversion rate." Hopkins's advice is to keep it as primary only while the account doesn't yet have enough of the real conversions.

Once you know which conversion action is which, set the right one as the primary action in Google Ads and check that its numbers match what the business is actually seeing. Then set the budget on that performance and on what the total marketing budget allows.

A checklist, in the order to work through it

  1. List what every conversion action in Google Ads and every key event in GA4 actually fires on.
  2. In Google Ads, read the Primary column: one high-intent action per real outcome, and any imported GA4 copy of a tagged action left as secondary.
  3. Submit one test enquiry through your own site. It should appear once in each platform.
  4. Read the counting setting on each Google Ads action, so you know whether two forms from one click count as one or two.
  5. Load your landing page with ?gclid=test on the end and check it survives, and check auto-tagging is on.
  6. Compare like with like: Google Ads by conversion time, the same date range, and a period that ended at least 48 hours ago.
  7. View GA4 under "Google paid channels last click" for the closest match to how Google Ads credits.
  8. Check which Consent Mode your banner runs, and whether the account is over Google's 700 ad clicks in seven days, per country and domain.
  9. Only then put what's left down to windows, video ads, devices, calls and time zones.

Quick answers

Which number to trust

For bidding, almost always the high-intent action, set as primary in Google Ads and checked against what the business actually records, though there are rare cases where testing a softer action is worth it. For reporting, the lower-funnel actions, because they're closer to real business outcomes, with soft actions reported as well where someone is working on the site's conversion rate.

How big a gap is normal?

Google's help says discrepancies "often up to 20%" are expected, but gives no data behind the figure. A better test is whether the known causes explain the direction and size of your gap. Almost exactly double, or GA4 several times Google Ads, points to a setup fault.

Why is Google Ads higher?

Most often because the two are counting different actions under the same name. Once they count the same one, Google Ads can still run higher because it models conversions for visitors who declined cookies, includes cross-device and engaged-view conversions, and counts calls from ads, none of which GA4 sees the same way. If it is close to exactly double, check for one form counted by two primary actions.

Are imported GA4 conversions modelled?

Yes. Google states that Google Ads conversions created from GA4 key events include modelling, and that reports based on imported conversions may include modelled conversions as estimates.

If the two still won't line up

Untangling which action is which, and what each platform should count, is what a conversion tracking audit and repair covers. A free 30-minute intro call is enough to look at your two numbers together, say which of the causes above is most likely, and what it would take to confirm it.