GA4, Google Ads, Meta and the CRM report different conversion numbers because each one counts different things: different attribution models, different dates, different windows, different counting rules, and different amounts of modeled data. Some gap is normal and stable. A gap that is large, sudden or growing usually points to a tracking fault, and the way to tell the difference is to reconcile each platform against actual CRM records, not against each other.
Every agency gets the client email: "Google Ads says 48 leads, Analytics says 31, Meta says 40 and we got 35 in the CRM. Which is right?" This guide explains where each difference comes from, what range is reasonable, and a reconciliation method you can run in an afternoon.
The short answer for the client
Before the detail, a framing that works in client conversations: each ad platform reports how many conversions it believes it influenced. Those claims overlap, because one person might click a Google ad on Monday and a Meta ad on Wednesday before submitting a form. GA4 tries to share credit across channels. The CRM records what actually arrived. So ad platform totals added together will usually exceed real leads, and that is expected. The CRM is the count of record; the platforms are the signals each one uses to optimize.
Reason 1: Attribution models and who gets credit
- Google Ads only credits Google Ads interactions. If a conversion action uses data-driven attribution, credit is shared across Google Ads clicks (and eligible engagements) in the path, but never given to Meta, email or organic search.
- GA4 shares credit across all channels using the property's reporting attribution model, usually data-driven. A lead that Google Ads claims in full may appear in GA4 as partly paid search and partly organic or email.
- Meta credits conversions to Meta ads within its attribution setting, which by default includes clicks and views.
- The CRM usually records first or last touch from UTMs or a hidden field, or a "how did you hear about us" answer, depending on how it was built.
Fractional credit is why GA4 can show 31.6 conversions for a campaign. That is not a bug.
Reason 2: Conversion date vs click date
Google Ads reports conversions against the date of the ad click (the interaction), not the date the conversion happened. If someone clicks on the 28th and submits a form on the 3rd of next month, Google Ads adds that conversion to the 28th. GA4 reports the key event on the 3rd.
This has two consequences. Recent days in Google Ads keep growing for a while as late conversions arrive, so last week always looks weaker than it will in a month. And month-end totals in Google Ads and GA4 will never line up exactly. Google Ads offers "by conversion time" columns that report on the conversion date; use them when comparing with GA4 or the CRM.
Meta Ads Manager also ties conversions back to the ad interaction date by default, so the same effect applies there.
Reason 3: Counting settings
Each Google Ads conversion action is set to count "every" or "one" conversion per click. For leads, "one" is normally right; for purchases, "every". If a lead action is set to "every", one person who submits two forms counts twice in Ads.
GA4 key events can count once per event or once per session. A mismatch between the GA4 setting and the Google Ads setting creates a gap on its own. Meta counts events, deduplicated only if browser and server events share an event ID. The CRM may merge duplicate contacts, which lowers its count further.
Reason 4: Conversion windows and lookback windows
Google Ads conversion windows are set per action (commonly 30 days for clicks, configurable up to 90). GA4 has its own lookback windows in attribution settings. Meta's attribution setting is chosen per ad set, with 7-day click and 1-day view as the usual default. A conversion 45 days after a click can be counted in one system and dropped in another.
Reason 5: Consent and modeled conversions
When users decline consent, tags either do not fire (basic consent mode) or send cookieless pings (advanced consent mode). Google Ads can then model conversions it could not observe, and GA4 can apply behavioral modeling if the property meets the thresholds. Modeled conversions appear in Google Ads but not in any CRM, and the GA4 equivalent only appears when the reporting identity is blended and the thresholds are met.
For sites with significant EEA or UK traffic, consent alone can create a meaningful gap. The setup is covered in our Consent Mode v2 implementation guide.
Reason 6: Cross-device and enhanced conversions
Google Ads includes cross-device conversions, where the click happened on one device and the conversion on another, matched through signed-in Google users. Enhanced conversions add matches from hashed first-party data. GA4 can see some cross-device journeys through Google signals and user ID, but it does not see everything Google Ads does. Expect Google Ads to report somewhat more than GA4 for that reason alone.
Reason 7: Imported vs native Google Ads conversions
A Google Ads account can count conversions from its own tag, from GA4 key event imports, or both. They behave differently:
| Native Google Ads tag | Imported GA4 key event |
| Speed | Usually appears within hours | Often takes longer to appear |
| Source of the data | Google Ads tag on the site | GA4 events, subject to GA4 filters and consent |
| Enhanced conversions | Supported on the tag | Depends on the GA4 setup |
| Risk | Duplicates if both are primary | Duplicates if both are primary |
The single most common fault we find is both: a native Ads lead action and the GA4 generate_lead import, both set as primary, so every lead counts twice in the Conversions column and in bidding. Pick one as primary and leave the other as secondary for observation. Our Google Ads conversion tracking setup guide covers how to choose.
Reason 8: View-through and engaged-view conversions
Google Ads can count view-through conversions for display and video, and engaged-view conversions for YouTube. Meta includes view-through by default in 1-day view. These are reported in separate columns or included depending on settings. GA4 does not record impressions from these platforms, so it cannot see these conversions at all. If Meta's number is far above GA4's for the Meta channel, check how much of it is view-through.
Reason 9: Time zones and currencies
- The Google Ads account time zone is set when the account is created and cannot be changed. If GA4 or the CRM uses a different time zone, conversions near midnight fall on different days.
- Values in different currencies are converted using each platform's own rates. Small value differences between platforms are normal when currencies differ.
Reason 10: Things that are actually faults
All of the above produce gaps you can explain. These produce gaps you should fix:
- Duplicate tags or duplicate imports inflating one platform.
- A key event on a thank-you page view that counts reloads.
- Self-referrals from payment or booking domains taking credit away from paid traffic in GA4. See GA4 cross-domain tracking and self-referrals.
- Auto-tagging off or GCLIDs stripped by a redirect, so GA4 cannot attribute paid clicks.
- A form changed and the event stopped firing in one platform but not another.
- Consent banner blocking tags for everyone, or for no one.
- Meta browser and server events without matching event IDs, so leads count twice. See Meta Conversions API deduplication.
What gap is normal and what is a fault
There is no universal percentage that defines a healthy gap. It depends on the share of EEA traffic, the sales cycle, the channel mix and the attribution settings. What we look for instead:
| Probably normal | Probably a fault |
| The gap is stable week to week | The gap jumped on a specific date |
| Ad platforms sum to more than the CRM | One platform reports more leads than the CRM received in total |
| Google Ads is somewhat higher than GA4 for paid search | GA4 shows almost no paid conversions while Ads shows plenty |
| Recent days in Ads keep increasing | Conversions in Ads are exactly double GA4 |
| Differences explained by consent share and modeling | Zero conversions in one platform since a site update |
A single platform reporting more leads than actually arrived is the clearest signal. An exact 2:1 ratio almost always means double counting.
A step-by-step reconciliation method
Reconcile each platform to the CRM, one at a time, for a period that has closed (for example a full month that ended at least two weeks ago, so late conversions have arrived).
- Pull the CRM leads for the period. Exclude spam, duplicates and existing customers. Note how many have a GCLID, an fbclid or UTMs captured. If the CRM does not capture these, that is the first fix; see CRM lead source attribution.
- Pull GA4 key events for the same dates using the event date, in the same time zone. Compare the total with the CRM total. GA4 should normally be equal or a little lower (consent, blockers). If it is higher, look for duplicates.
- Pull Google Ads conversions by conversion time for the primary lead action only. Compare with CRM leads that have a GCLID or a Google paid UTM. Ads may be somewhat higher because of modeling and cross-device.
- Pull Meta results for the lead event with the attribution setting noted. Compare click-through only with CRM leads that have an fbclid or Meta UTM, then look at what view-through adds.
- Match individual records where possible. Take ten CRM leads from paid search and check that each has a GCLID and that a conversion exists for that date. Missing GCLIDs point to redirects, forms that drop URL parameters or auto-tagging problems.
- Write down each explained gap and each unexplained one. Explained: consent share, view-through, date basis. Unexplained: send to fixing.
- Repeat monthly. The first reconciliation is work. After that it is a quick check, and a changing gap is the earliest warning that something broke.
How to present the gap in client reporting
Once you understand the gap, stop it from coming up in every monthly call. A few reporting habits help:
- Lead with the CRM number. Put real leads or orders at the top of the report, with cost per lead calculated from them. Platform numbers come underneath, labeled as each platform's own attribution.
- Label every platform number with its basis. "Google Ads, data-driven, 30-day click, by click date" tells the reader why it differs from GA4 without a separate explanation.
- Show the trend of the gap, not just the gap. A simple line of platform conversions divided by CRM leads per month makes a sudden change obvious to everyone, including the client.
- Note tracking changes on the report. Consent mode going live, a form rebuild or a new primary conversion action all move the numbers. A one-line annotation saves a long email thread.
- Never sum platform conversions. Adding Google Ads and Meta conversions together double counts overlapping journeys and produces a total higher than reality.
If the client wants one blended view, build it from CRM records with captured click IDs and UTMs rather than from platform totals. That is the only number that does not depend on a platform's view of its own contribution.
What we check in an audit
- Which conversion actions are primary in Google Ads, and whether any lead is counted by two of them.
- Counting, windows and attribution model on each action, compared with GA4 settings.
- GCLID capture through redirects and into the CRM.
- Consent mode setup and the share of traffic that declines.
- Meta deduplication and attribution settings.
- Time zones and currencies across GA4, Ads and the CRM.
- A sample of CRM leads traced back to platform conversions.
Getting the numbers to make sense
If a client's numbers have drifted apart and nobody can explain why, our GA4 audit reconciles GA4, Google Ads and the CRM and separates normal gaps from faults, starting with a free tracking check.