Meta Ads Reported vs Real Orders: What Two Stores Found
We make Real ROAS. The numbers on this page come from Real ROAS run on two stores' real ad and order data.
Published · Updated · Facts checked on · 11 min read
On Store 1, Meta Ads Manager reported 1.89 times the real orders across three sales campaigns over 30 days. On Google, the gap ran in both directions: YouTube and Demand Gen campaigns reported more than the store counted, while Search and Shopping campaigns reported fewer. This page shows the campaign-level data, the method and its limits. Two stores, 30 days.
How far apart are reported and real orders?
The gap runs in both directions, depending on the campaign type. On Store 1, Meta's total was 1.89 times the real count. On Google, six of eight campaigns under-reported while two YouTube campaigns over-reported. Five findings:
- Store 1's three Meta sales campaigns reported 1,372 purchases against 727.3 real orders on ₹1,119,497 of spend (data set 1).
- On Store 1's eight Google campaigns, six reported fewer orders than the store counted; the two YouTube campaigns reported more (data set 2).
- Store 2's Google Demand Gen campaign reported 9.8 purchases against 1.0 real order; its brand search campaign reported 3.9 against 9.0 real (data set 3).
- Of 1,755 sales on Store 1, 556 had more than one tracked visit; 191 of those buyers first visited 0 to 7 days before the sale, and 146 first visited 31 to 90 days before (data set 5).
- Across Store 2's 2026 results to 1 November, Meta claimed 4.27 times the revenue GA4 put down to social, and Google claimed 1.90 times GA4's Google paid figure (data set 6).
How we measured it
Two D2C stores in India, 14 October 2026 to 12 November 2026, measured with Real ROAS. Store 1 sells cosmetics; Store 2 sells women's fashion. Two stores is a small sample.
The method in plain words:
- "Reported" means what Meta Ads Manager or the Google Ads API showed for each campaign at the account's own attribution setting, for those dates.
- "Real" means purchases recorded in GA4, joined to the ad that brought the visitor by its click ID or UTM names, with each order credited once across all channels. The model used is position-based: 40% to the first touch, 40% to the last touch, 20% shared among the middle touches. Credit can split across channels, so some order counts end in a decimal.
- Store 1 had no store order export loaded for these dates. Its "Real" counts are GA4 counts raised by the store-to-GA4 ratio measured on Store 1 from January to June 2026 (data set 4). The tool marks these as estimates.
- Store 2's counts are GA4-matched counts with no raise applied.
- Two cross-checks ran on Store 1's Meta campaigns: GA4's own attribution and a Markov model.
Limits: two stores, one 30-day window; matching rests on GA4 purchase events and ad click IDs and UTMs; a full store order record would make it tighter. No other test was run on these accounts for this period.
How much does Meta over-report purchases?
Meta Ads Manager reported 1.89 times the real orders on Store 1's three sales campaigns over 30 days, on ₹1,119,497 of spend.
| Campaign type | Spend | Meta reported purchases | Meta ROAS | Real orders | Real ROAS |
|---|---|---|---|---|---|
| Largest sales campaign (CBO) | ₹1,003,938 | 1,199 | 2.00 | 663.2 | 1.14 |
| Testing campaign (CBO) | ₹83,846 | 122 | 2.25 | 51.3 | 1.02 |
| Smaller sales campaign (CBO) | ₹31,713 | 51 | 1.87 | 12.8 | 0.49 |
| Together | ₹1,119,497 | 1,372 | — | 727.3 | — |
Two cross-checks ran on the same three campaigns: GA4's own attribution counted 540.5 orders and a Markov model counted 799.4. Both sit well below Meta's 1,372. The store took 1,733 orders in those 30 days from all channels. When your store-level revenue total and your campaign ROAS tell different stories, how MER and ROAS answer different questions explains why the two figures measure different things.
Does Google Ads over-report too?
On Store 1, the two YouTube campaigns over-reported while Search, Shopping and Performance Max campaigns reported fewer orders than real. On Store 2, the Demand Gen campaign carried the over-count.
Store 1: eight Google Ads campaigns
| Campaign type | Spend | Google reported purchases | Google ROAS | Real orders | Real ROAS |
|---|---|---|---|---|---|
| Brand search | ₹36,031 | 199.7 | 8.96 | 219.1 | 10.18 |
| Performance Max, new audiences (brand targeting) | ₹23,438 | 21.2 | 1.58 | 35.3 | 2.65 |
| YouTube, acquisition and retargeting | ₹19,657 | 24.8 | 2.02 | 11.9 | 0.92 |
| Performance Max, competitor search terms (1) | ₹17,609 | 14.1 | 0.98 | 18.9 | 1.38 |
| Performance Max, competitor search terms (2) | ₹16,393 | 20.5 | 2.03 | 27.9 | 2.99 |
| Performance Max, new audiences (acquisition) | ₹16,367 | 23.0 | 2.79 | 36.8 | 4.44 |
| Demand Gen on YouTube, warm audience | ₹16,175 | 32.0 | 2.78 | 5.2 | 0.44 |
| Shopping, own brand searches | ₹14,219 | 33.6 | 3.81 | 38.2 | 4.39 |
Together, these eight campaigns spent ₹159,889; Google reported 368.9 purchases against 393.3 real orders, with six of the eight campaigns reporting fewer orders than the store counted.
Store 2: seven Google Ads campaigns
| Campaign type | Spend | Google reported purchases | Google ROAS | Real orders | Real ROAS |
|---|---|---|---|---|---|
| Demand Gen, mixed | ₹27,889 | 9.8 | 1.83 | 1.0 | 0.10 |
| Shopping, non-brand | ₹14,119 | 0.8 | 0.25 | 1.0 | 0.79 |
| Shopping, brand | ₹13,518 | 12.5 | 4.69 | 11.0 | 4.26 |
| Performance Max, one product line, non-brand | ₹12,177 | 1.3 | 0.41 | 2.0 | 0.46 |
| Performance Max, one product line, top cities, non-brand | ₹11,412 | 3.0 | 1.08 | 1.0 | 0.54 |
| Search, brand | ₹6,896 | 3.9 | 2.39 | 9.0 | 6.37 |
| Search, non-brand | ₹2,460 | 0.0 | 0.00 | 1.0 | 4.43 |
Together, these seven campaigns spent ₹88,471; Google reported 31.3 purchases against 26.0 real orders, with the Demand Gen campaign carrying most of the over-count (9.8 reported against 1.0 real).
Cutting a campaign on its platform ROAS alone can cut the wrong one. Brand search on Store 2 showed a platform ROAS of 2.39 and a real ROAS of 6.37. The store's most effective campaign looked average in the platform view.
Why do platform numbers drift from your store?
Each platform credits itself for any sale inside its own attribution window, with no view of what the other platform counted. The sum of what each platform claims can exceed your store's total.
Four reasons the numbers split:
- Each platform keeps its own score: a sale inside its window is its sale, with no view of what other channels counted.
- Meta's default setting counts a purchase within 7 days of a link click, 1 day of a like or a 5-second video view, or 1 day of a view with no click.
- Platforms add modelled conversions to fill gaps where tracking breaks.
- Search and Shopping campaigns can under-report when a buyer's click is lost to cookie consent or browser privacy rules, while the store still records the order.
Google click IDs can carry a date mismatch. On Store 1, some click IDs found in checkout cookies were 12 to 35 days old; the full window logic is in how Meta's attribution window counts a purchase. Real ROAS counts a Google click ID only within 7 days of the click.
Buyers also take time. Of 1,755 sales on Store 1, 556 had more than one tracked visit before the purchase. For 191 of those, the first tracked visit came 0 to 7 days before the sale. For 146, it came 31 to 90 days before. For 47 each, it came 91 to 182 days and over 182 days before. Any attribution window that closes before the buyer returns will miss the sale.
Copy this sheet to measure your own gap
Pull your own numbers from your store and your ad accounts and fill in each blank.
Date range: ____ | Attribution setting in Ads Manager: ____
Campaign name | Spend | Platform reported | Store orders (UTM/click ID) | Gap
Campaign 1: ____ | ____ | ____ | ____ | ____
Campaign 2: ____ | ____ | ____ | ____ | ____
Campaign 3: ____ | ____ | ____ | ____ | ____
Store total orders for the dates: ____
Sum of orders every platform claims: ____
Overlap = sum claimed / store total: ____
GA4 orders by channel: ____
Store orders / GA4 orders: ____
Decision: keep / cut / test
Campaign 1: ____
Campaign 2: ____
Campaign 3: ____
- Export your store's orders for the dates with UTM source, medium and campaign columns attached to each row.
- For each campaign, count the orders where the UTM campaign name matches, or where the click ID in the order links back to that campaign. That is your matched count.
- Divide the platform-reported number by the matched count. A result above 1 means the platform claimed more than you can confirm; a result below 1 means the platform may have under-counted.
How do you see real orders on every ad row?
Each store order is credited once to one channel and one ad path, then placed on the same row as the platform's figure in the ads table.
Real ROAS shows store orders next to Meta's number inside the Ads Manager table. It puts real orders, revenue and ROAS on each Meta and Google row, alongside the platform's own figure. Each row carries a match grade from A to D: A for an order matched by ad ID, B by UTM names, C by GA4 and ad ID, D by GA4 and names. The tool reads only: nothing in your ad accounts or store is created, paused, edited or deleted. It costs $5 a month. See it inside the Ads Manager table.
Frequently asked questions
These four questions cover what buyers and media teams ask when comparing their ad platform numbers to their store counts.
Is a gap between Meta and your store always a tracking error?
A tracking error is one cause, but platform attribution windows, modelled conversions and double claims between platforms explain a large part of the gap. Measure it per campaign: a campaign with a large over-count and a low real ROAS is a different problem from one with a small over-count and a high real ROAS. The two stores here show both patterns in the same 30-day window.
Can Meta under-report orders?
Yes. When a buyer opts out of tracking and Meta loses the signal, it misses that sale. Data set 4 also shows that GA4 under-counts store orders: on Store 1, the store-to-GA4 ratio ran from 1.15 to 1.42 for Meta traffic in every month from January to June 2026, so a GA4-only check gives Meta a lower score than its actual contribution.
How often should you check the gap?
Check it each month, and again after any change to your tracking setup, checkout flow or attribution setting. A gap that grew after a specific change points to that change as the likely cause.
Do these numbers apply to my store?
These are two stores from one 30-day window. Your gap will differ based on your campaign mix, attribution setting and the share of buyers who touch more than one channel. Use the sheet above to measure your own gap this month.
Start measuring your real orders
Real ROAS costs $5 a month, one plan. Cancel any time on your account page.
Start Real ROAS for $5 a month
You can also run a free Meta Ads account audit on your account.
Related reading