Running ads · Updated October 20, 2026
Why do Google Ads, Meta and Microsoft Ads report different numbers for the same leads?
Updated October 20, 2026
Because each platform credits its own ads and, by default, never checks the others. Google Ads, Meta and Microsoft Ads each count a lead when one of their ads reached that person inside their own window, so one lead can count three times. Run each platform on its own number and split the budget on your own lead records.
If the platforms add up to more leads than you got, that is expected
If Google Ads, Meta and Microsoft Ads together report more leads than your inbox and call log hold, nothing is necessarily broken. Each platform counts a lead when one of its own ads reached that person inside its own window. By default, none of them asks whether another platform's ad reached the same person, so a lead that touched three platforms is counted by all three.
Google Ads gives credit only to Google ads. Its data-driven model, the default for most conversion actions, shares credit among interactions on Search, YouTube, Display and Demand Gen. Google says its attribution reports reflect "only the keywords and ads in your Google Ads account." A Meta or Microsoft click in the same path is not part of the calculation, unless the conversion is imported from Analytics with Paid and organic channels selected.
Meta counts views as well as clicks. Its default setting counts a conversion within 7 days of a link click, 1 day of seeing the ad, or 1 day of another interaction with it, such as a share or a video watched for 5 seconds. Someone who scrolled past your ad this morning and clicked a Google ad this afternoon counts for Meta.
Microsoft Ads credits the last Microsoft ad clicked. Last click is its default attribution model and 30 days its default conversion window, so a click three weeks ago still earns the lead when a Google ad came after it.
For example
In June a business's own records show 40 leads from its website and ads. Google Ads reports 30, Meta 22 and Microsoft Ads 6: 58 in total. Take one of the 40. On Thursday afternoon she sees the business's ad in her Meta feed and keeps scrolling. That evening she searches, clicks the Microsoft ad and leaves. On Friday morning she searches again, clicks the Google ad and sends the form. Meta counts her, because she sent it within a day of seeing its ad. Microsoft counts her, because its last click was inside 30 days. Google Ads counts her, because its click was inside its 30 day window, and it would have counted her even if the Microsoft click had come later. Analytics, which sees only the two clicks, credits her once in all. Her email arrives once.
Run each platform on its own number, and split the budget on yours
Use three numbers for three jobs, and never add the platforms' numbers together. The sum counts every lead that touched two platforms twice, and the more platforms you run, the further it drifts from the leads you received.
- Each platform's own count
- The Conversions column in Google Ads and Microsoft Ads, Results in Meta. It is what each platform's automated bidding works toward.
- Judge campaigns against each other on it. Comparing two Google campaigns on Google's count is fair. Comparing Google's count with Meta's is not, because they claim the same people by different rules.
- Analytics
- One rule for every channel that sends visitors to your website, so a lead from two ads adds up to one lead, however the credit is split.
- It sees clicks, and engaged views of YouTube ads, but no Meta or Microsoft ad views. A lead Meta claims through a 1 day view has no Meta visit for Analytics to credit, and a call from the button in a Google ad never reaches the website at all.
- Your own lead records
- Every form fill, call and instant form lead counted once, with where it came from saved on it.
- It is the only count that removes the overlap. Work out cost per lead as total ad spend divided by these leads, and per platform as that platform's spend divided by the leads your records credit to it.
Use it inside that platform
Use it as the referee for clicks
Use it for the budget
This split is a judgment, not any platform's advice. Each vendor documents its own numbers and none of them says how to weigh one against another. Your records are the one count that has no reason to favor any platform, which is why the budget question belongs to them.
They still credit one source per lead, usually the last click. A platform that mostly starts journeys, as feed ads often do, looks weaker in your records than its share of the work. Before you cut it on that basis, cut it for a few weeks and watch total leads, not its own count. If total leads hold, the spend was not earning them.
One platform claiming more leads than your records hold in total is not overlap. Overlap can push the sum of the platforms past your total, never a single platform. That is a counting error inside that platform, and the counting step below is where to start.
Line them up, in this order
Compare a finished month, in one time zone, on the right dates
Pick a month that ended more than 30 days ago and set the same dates in every platform. A month that ended a week ago works for a rough check, but Google Ads and Microsoft Ads will still add to it. Then line up two things that move leads between days: the date each platform files a conversion under, and its clock.
Google Ads and Microsoft Ads file a lead under the day of the click. Meta files a website conversion under the day it happened. Meta's Marketing API documentation, which it says mirrors Ads Manager, says off-Meta actions, such as a purchase on your website, "leverage conversion-based reporting time," while on-Meta actions such as link clicks use the impression date. A click on the 30th and a form on the 2nd is last month in Google Ads and Microsoft Ads and this month in Meta.
In Google Ads, add Conversions (by conv. time). Google suggests these columns for comparing its data with other tools. Microsoft has no conversion-time column, so it stays on click dates in your comparison.
Then check the clocks, if you compare single days. Over a whole month the clocks only move leads near midnight on the first and last day. Each platform reports in its account's time zone. Google Ads no longer lets a standard account change its time zone. Meta's only way to change one closes the ad account and opens a new one, and new Meta accounts start on Pacific time. Microsoft Ads sets the time zone for the whole account and lets you pick a different one for a single report, so match Microsoft's report to the other two.
Why last month's numbers keep rising in Google Ads and Microsoft Ads
Both file conversions under the click date and keep adding them until the action's window closes, 30 days after the click by default. Microsoft says this "can temporarily inflate your cost per acquisition." A month compared the week it ended looks low against your records, and on the default 30 day windows, a month that ended more than 30 days ago has stopped moving.
Take out the leads only one platform can see
Before you look for overlap, set aside the leads that only one platform could ever count. They make that platform look stronger in a side-by-side comparison without any double counting.
- Calls from the button in a Google ad. The caller goes straight to your phone without loading your website, so no other platform and no website tag can see them. Compare them with the calls in your call log that came from that number.
- Meta instant form leads. The form is filled in on Meta, not on your website. Meta's Results column leaves out organic leads and your own test leads, though both appear when you download the leads, so a download can hold more than the report. Meta keeps leads available for 90 days, so download each month's before they go.
- Leads a platform imported from you. A booked job you upload to Google Ads with its click ID counts in Google Ads only, as it should. It is still one lead in your records.
Make each platform count one conversion per lead
Set every lead action to count once per click. Two of the three platforms default to counting every conversion, which suits sales and inflates leads.
- Google Ads: website actions default to Every. Set a lead action to One under Goals, then Conversions, then Summary: open the action, then Edit settings, then Count.
- Microsoft Ads: goals default to All. Microsoft calls Unique "Common choice for leads." It also says it "does not currently include any functionality to detect or eliminate test, fraudulent or duplicate conversions," and that the same UET tag installed twice double counts.
- Meta: a lead sent by both the Meta Pixel and the Conversions API is counted once when Meta can match the pair. Its recommended method is the same event name and event ID reaching the same pixel within 48 hours; matching browser or external IDs is a fallback. With neither, Meta can count the lead twice.
If Google Ads alone counts more conversions than you received, the cause is inside Google Ads, and "Why does Google Ads show more conversions than I actually got?" walks through finding the action behind it.
Read view-through conversions on their own
Separate conversions from people who saw an ad from those who clicked one before you compare platforms. A view is the weakest claim a platform makes on a lead, and each platform treats it differently by default.
Meta includes them. Its default counts a conversion within 1 day of someone seeing the ad, and since a change Meta began rolling out in March 2026, interactions other than link clicks, such as shares and saves, count in a separate 1 day engage-through window.
Meta's default view window
1 day
A conversion within a day of someone seeing your ad, without clicking, counts in Meta's Results.
Google Ads mostly keeps them apart. View-through conversions sit in their own column and in All conv., not in Conversions, except on App, Demand Gen and Performance Max for store goals. Engaged-view conversions, from someone who watched a video ad long enough (10 seconds of a skippable in-stream ad, 5 seconds of an in-feed or Shorts ad), are in Conversions. The default window is 3 days for web conversions, or 1 day on Demand Gen and Performance Max.
Microsoft Ads has its own column too. Its view-through window defaults to 1 day and applies to audience ads and search ads extended to audiences. Two Microsoft pages disagree on whether they also count in Conversions by default, so read both columns in your own account.
On a small account, give a view-through lead the least weight when you decide where money goes. That is a judgment: a person who saw an ad and converted the same day may have been on their way anyway, and none of the platforms can show you which.
For the Google Ads side, "Keep view-through conversions out of the Google Ads numbers you judge" walks through the settings.
Expect estimated conversions in the platforms and none in your records
Accept that part of each platform's count is an estimate. Your inbox counts people; the platforms also count conversions their models believe happened but could not observe.
- Google Ads says it reports "both modeled and observed conversions" in the Conversions column, including many conversions that started on one device and finished on another, and conversions from visitors who declined cookies where consent mode is set up.
- Meta says that where data is partial or missing, statistical modeling may be used for some results, and that it may distribute conversions among campaigns, ad sets or ads.
- Microsoft Ads can add modeled conversions to its Conversions column, as fractions, for accounts that use consent mode, get at least 700 ad clicks a week in the targeted market and pass a confidence threshold. It is generally available in the European Economic Area, Switzerland and Great Britain, and in a pilot in the US.
A local account with a few hundred clicks a week is below Microsoft's 700 click threshold. Google and Meta publish no minimum, so modeled conversions can appear in their counts at any size. They are estimates, so they will not match your records lead for lead, even when the total is close.
Why older guides describe Meta windows that no longer exist
Meta stopped returning data for its 7 day view and 28 day view windows through its reporting API on January 12, 2026. On March 3, 2026 it announced that click-through attribution for website conversions would count link clicks only, moving shares, saves and other clicks into engage-through, and that its engaged video view would drop from 10 seconds to 5. The rollout is gradual, and Meta's own help page says some accounts may still use the earlier versions.
So a guide or a screenshot from before 2026 can show settings your account no longer offers, and a year-on-year comparison in Meta can move because the rule changed, not the results.
Get one count you can trust
Save where each lead came from on the lead itself
Record the source on every lead as it arrives, and the overlap question answers itself. Tag every ad link with UTM tags, save them and the click ID in hidden form fields, give each platform its own tracking number for calls, and note instant form leads as Meta's.
Then count by source in your records. Leads with a Google click ID are Google's, and so on. A lead that arrived through two ads is still one row, credited to whichever source your form saved, which is usually the last click.
Use Analytics as the check on your website leads. Its reports credit each key event across paid and organic channels, by default, under one rule for every channel. It cannot see Meta or Microsoft ad views or calls from ad buttons, so it will credit Meta less than Meta does, and it agrees with neither platform by design. "Why is my conversion rate different in Google Ads and in Analytics?" covers lining Analytics up with Google Ads.
"How do I save the Google Ads click ID with each form lead?" sets up the hidden field. The free Conversion tracking check under Free tools on this site sends a test form on your website and shows which platforms heard it, which is the quickest way to see whether a platform is missing a lead or claiming one it could not have seen.
If your own count is right, leave the platform settings alone
Do not change a platform's window or attribution model to make its number agree with another platform's. Each platform's bidding works toward its own count, so a setting changed for a tidier report changes what the campaign chases.
Fix counting errors, not definitions. An action set to Every, a tag installed twice or a Pixel and Conversions API pair with no deduplication set up are errors. A 1 day view window or a 30 day click window is a definition, and it is why the numbers differ, not a fault.
Some changes reach further than they look. Analytics says changing which channels can receive credit applies going forward to every linked Google Ads account, and may change the conversions Google Ads bids on and reports.