GA4 shows fewer conversions than Google Ads or Meta Ads Manager because each tool measures a different thing: GA4 splits credit across all channels and dates conversions to when they happen, while ad platforms credit their own ads, date to the click and add modeled conversions. A stable GA4 conversion discrepancy is normal; a sudden one is a tracking bug.
A GA4 conversion discrepancy is the difference between the conversions GA4 reports for an action and the conversions an ad platform reports for the same action over the same period. This guide, part of the analytics, tracking and privacy hub, lists the documented causes, gives a rule to tell a normal gap from a broken one, and a reconciliation order you can run in an afternoon.
What causes a GA4 conversion discrepancy?
A GA4 conversion discrepancy comes from design choices, not from one tool being wrong. GA4, Google Ads and Meta Ads Manager each decide differently which touchpoint gets credit, which date a conversion belongs to, how long after an interaction it still counts, how often one user can convert, and whether unobserved conversions are estimated.
Google's conversions vs. key events page says the differences "are rooted in the distinct purposes of these metrics" and names attribution model, lookback window, counting method, behavioral modeling and reporting scope. An attribution model is the rule that decides which touchpoints receive credit for a conversion. A key event is a GA4 event you mark as important to your business. A conversion is created from a GA4 event and used to measure and bid on Google Ads campaigns.
The causes fall into two groups:
- Measurement differences that are always present: attribution model, conversion date, lookback window, counting method, modeled and cross-device conversions (a modeled conversion is a conversion Google Ads estimates for users it cannot observe directly), view-based credit (a view-through conversion is a conversion after someone saw an ad without clicking it).
- Tracking problems that appear and disappear: a tag missing on some pages, a tag firing twice, duplicate browser and server events, a consent banner that blocks one tool and not the other.
How do GA4, Google Ads and Meta Ads Manager count the same conversion?
GA4, Google Ads and Meta Ads Manager count the same purchase or lead under different rules for credit, date, window and estimation. Comparing them on those criteria side by side shows which part of the gap is structural before you start debugging tags, and which settings you can align to narrow it.
| Criterion | GA4 (key events) | Google Ads (conversions) | Meta Ads Manager |
|---|---|---|---|
| Who gets credit | All channels, by the reporting attribution model | Google Ads interactions only, data-driven by default | Meta ads only |
| Date of the conversion | When the conversion happened | Date of the ad click (standard "Conversions" column) | Not stated in Meta's developer docs |
| Default window | 90 days, 30 for acquisition key events | 30 days click-through, adjustable 1 to 90 | Set per ad set by the attribution setting |
| Credit after an ad view | No | Yes: view-through and engaged-view | Yes, when the attribution setting includes views |
| Estimated conversions | Behavioral modeling in reports | Modeled and cross-device conversions in "Conversions" | Not stated in Meta's developer docs |
| Counting rule | Once per event or once per session | "Every" or "One" per conversion action | Per event, deduplicated by event ID |
The GA4 windows come from Google's attribution settings page, the Google Ads window defaults from About conversion windows. Meta's view options come from the ad set reference, its dedup rule from Meta's developer documentation.
Why do attribution models and conversion dates split the numbers?
Attribution models and conversion dates split the numbers because GA4 shares one conversion across every channel the user touched, while Google Ads gives all of it to Google Ads interactions, and because Google Ads files that conversion under the click date. The same sale can appear in different amounts and in different weeks.
In Google Ads, "the 'Data-driven' model is the default attribution model for most conversion actions," according to About attribution models. GA4 credits organic search, email, direct and other paid channels too, so a Google Ads campaign receives a smaller share of the same purchase. Since November 2023, GA4 no longer offers first click, linear, time decay or position-based models, per Google's attribution settings page. The attribution models guide covers what is left.
Dates are the second split. Google Ads "has always reported conversions by the date your ad was clicked," says the conversions by time page, while GA4 records the date the conversion happened. A click on October 28, 2026 that converts on November 3, 2026 is an October conversion in the standard Google Ads column and a November key event in GA4. The "Conversions (by conv. time)" column puts Google Ads on the same date basis as GA4. Recent days always look worse in Google Ads until late conversions arrive.
How do consent, modeling and view-based credit widen the gap?
Consent choices, conversion modeling and view-based credit widen the gap because they add conversions to the ad platform that GA4 never observes. A user who declines cookies, switches device or only watched a video ad can still count in Google Ads or Meta, while GA4 records nothing or credits a different channel.
Consent Mode is Google's setting that adjusts how Google tags collect data based on each visitor's cookie consent choice. Since June 15, 2026, Consent Mode is the single control for Google Ads cookies and identifiers collected through Google Analytics, per the key events page. Your cookie consent banner affects both numbers.
On modeling, Google Ads estimates conversions for unconsented users it cannot observe and reports them in the "Conversions" column once an account reaches 700 ad clicks over 7 days per country and domain grouping, per About consent mode modeling as of November 2026. Google Analytics "only exports observed conversions to Google Ads," says the legacy comparison page.
On views, Google Ads counts view-through conversions (1 day default window) and engaged-view conversions (3 day default), where an engaged view is at least 10 seconds of a skippable in-stream ad, per About Engaged-view conversions as of November 2026. Meta ad sets can credit view-through and engaged video view conversions too, per Meta's ad set reference as of November 2026, and Meta adds on-device conversion measurement.
When is a gap a tracking problem and not a measurement difference?
A gap is a tracking problem when the ratio between platform conversions and GA4 conversions moves outside its own normal range, usually right after a site release, a tag change, a consent banner change or a new server-side setup. A gap inside its usual range is a measurement difference to document, not a bug.
The Gap Baseline Check turns that rule into a routine. For one conversion action, divide weekly platform conversions by weekly GA4 key events, both by conversion time. Take eight weeks as the baseline and note the lowest and highest ratio. Investigate any new week outside that range.
Hypothetical example, invented numbers: the Google Ads to GA4 purchase ratio ranged from 1.15 to 1.30 for eight weeks. After a checkout redesign it reads 260 to 130, a ratio of 2.0, so suspect a GA4 tag missing on the new confirmation page.
| Symptom in the Gap Baseline Check | Likely cause | First check |
|---|---|---|
| Ratio jumps up, GA4 falls | GA4 tag missing on new pages or blocked by consent | Tag Assistant on the confirmation page |
| Ratio jumps up, platform rises | Platform tag firing twice, or browser and server events not deduplicated | Event count per order ID, event_id |
| Ratio drops, GA4 rises | GA4 event firing twice, or "Once per event" on a reloadable page | DebugView on a page reload |
How do you reconcile GA4 with your ad platform?
To reconcile GA4 with an ad platform, compare one conversion action at a time on the same date basis, time zone and counting rule, then test the tags on the pages that fire it. Align the settings first and debug tags second, because most of the gap disappears or becomes explainable once both tools describe the same thing.
- Pick one conversion action. Match it to the exact GA4 key event, not the sum of all key events.
- Use the same date range and time zone. A different property or account time zone moves conversions between days.
- Switch Google Ads to conversion time. Add the "Conversions (by conv. time)" column.
- Compare counting settings. In GA4, check "Once per event" or "Once per session" under Key events, per Google's counting method page. In Google Ads, check "Every" or "One" under conversion counting options.
- Check consent behavior. Decline cookies on your own site and see which tags still fire.
- Test tag coverage. Run Tag Assistant and GA4 DebugView through the funnel, then reload the confirmation page to catch double firing, especially after a GTM optimized tagging experience upgrade.
- Check deduplication. The Pixel eventID must match the Conversions API event_id, and Meta deduplicates only within 48 hours of the first event, as of November 2026.
Which number should you use for bidding and which for reporting?
Use the ad platform's own conversions to steer that platform's bidding, because Smart Bidding and Meta delivery optimize on their own data. Use GA4, or one agreed source of truth such as a CRM, to compare channels and report results. Never sum platform conversions: each platform claims its full share.
Google positions conversions for ad measurement and bidding and key events for analyzing acquisition and site journeys, per the key events page. That split is the decision rule:
- Bidding and in-platform tests: platform conversions, ideally from the native tag. Native Google Ads tracking typically shows within 3 hours and imported data can take up to 24 hours, per Google's legacy comparison page, still online as of November 2026.
- Channel mix and budget across platforms: GA4 key events or CRM revenue, with one attribution model for every channel.
- Proof that ads caused sales: neither; run a holdout test such as Google Ads conversion lift.
Label the source of every ROAS figure and show the Gap Baseline Check ratio beside it.
Common mistakes when comparing GA4 and ad platform conversions
The most common mistakes when comparing GA4 and ad platform conversions come from comparing totals that were never meant to match: different conversion actions, different date bases or different counting rules. Each mistake makes a normal gap look like broken tracking, or hides real broken tracking inside a gap everyone has learned to ignore.
- Comparing all key events with one conversion action. GA4 totals often include sign ups, leads and purchases together. Compare the single key event that matches the platform action.
- Comparing click-date and conversion-date numbers. The standard Google Ads column dates conversions to the click. Use "Conversions (by conv. time)" before you judge recent weeks.
- Changing the GA4 lookback window to fix history. Lookback changes in GA4 only apply going forward, per Google's attribution settings, so past reports stay the same.
- Running the Meta Pixel and Conversions API without a shared event ID. Without a matching eventID and event_id, Meta cannot deduplicate browser and server events, and the Meta count climbs.
- Adding up platform conversions for total results. Google Ads and Meta can both claim the same purchase. Report the business total from GA4 or the CRM.
- Ignoring a stable gap until it moves. Without a recorded baseline, nobody can tell whether a new gap is normal. Start the Gap Baseline Check before you need it.
