Meta Ads Data Integrity: How Bad Data Costs You Money

Picture an account where Ads Manager shows a ROAS of 4. The owner doubles the budget. A month later the bank balance has dropped.

Nothing about the ads changed. The numbers were wrong, and every decision built on them was wrong with them.

This post covers how bad data misleads you on Meta and how to audit your account before it costs you money. For how Meta uses the events you send, read Meta Ads Algorithm Explained.

What is data integrity in Meta ads?

Data integrity means the numbers in your dashboard match what happened in your business. Each event is the right one, counted once, carrying the right value, read in a window you understand and tied to cash that reached your account.

When one of those breaks, the dashboard still shows a number. It just points you in the wrong direction.

Five ways bad data fools you

1. The same sale gets counted twice

Most stores send purchases through two paths, the browser Pixel and the server-side Conversions API. Meta can merge them into one purchase only when both events carry the same event name and the same event ID. Meta's documentation says it discards the later copy when it finds a match within 48 hours.

It does not deduplicate when you use only one source. If the IDs differ, Meta counts two purchases for one order.

Meta then optimizes against inflated volume, and every ROAS and CPA you read is off with it.

How to check: Compare Meta purchases with store orders for the same dates. Then open Events Manager and confirm both sources send matching event IDs.

2. Meta changes how it counts

On January 12, 2026, Meta removed the 7-day view and 28-day view attribution windows from the Ads Insights API. Meta announced it in October 2025. Click windows and the 1-day view window stayed.

Agency analyses put the drop in reported conversions between 15 and 40 percent for affected accounts. One agency, Williams Randall, notes that Meta changed reporting visibility and left ad delivery alone.

A second change came in March 2026, according to agency reports. Click-through attribution narrowed to real link clicks. Likes, shares and saves moved to a new engage-through bucket with a one-day window.

If you compare this quarter with last quarter without knowing this, you may cut a campaign that never got worse.

How to check: Write down the date of every attribution change and set a new baseline from that date. Compare Meta against store revenue.

3. Meta takes credit for sales it did not cause

Platform ROAS counts conversions credited to the ads. It does not count conversions the ads caused. Some buyers would have bought anyway.

The evidence conflicts, and that is the lesson. Haus reports an average Meta lift of about 19 percent across 640 experiments, and says retargeting often runs 40 to 70 percent below its reported ROAS. Stella reviewed 46 holdout studies from DTC brands that chose to test. On average Meta drove about 21 percent more than its dashboard claimed, and the reported number did not predict the result. The sample was small and self-selected.

So the platform number can mislead in either direction. Optimizing blindly to it can push you to cut cold prospecting and feed retargeting, the spend that causes the least new demand.

How to check: Run a holdout test or a Meta Conversion Lift study when your volume allows. Track MER every month. MER is total revenue divided by total ad spend, and it does not care which platform claims the sale.

4. Orders are not cash

ROAS counts an order the moment it is placed. Returns, cancellations and refused deliveries come later and never appear in Ads Manager.

This hits hardest where cash on delivery is common. A 2024 paper in the International Journal of Public Policies in Egypt reports that about 62 percent of customers prefer cash on delivery. eGrow, an Egyptian COD platform, lists high return-to-origin rates and weak order confirmation among the main threats to profitability.

An order that bounces back cost you the ad spend, the shipping both ways and the stock tied up in transit.

How to check: Report ROAS on delivered orders, net of returns and refusals. Keep the order-placed ROAS next to it so you see the gap.

5. The math underneath is wrong

Breakeven ROAS equals one divided by gross margin. If the margin you plug in ignores shipping, payment fees, tax or returns, the breakeven is wrong and every campaign is judged against the wrong line.

Here is an illustration with made-up numbers. Meta reports a ROAS of 4. Half of those purchases are duplicates, so real ROAS is 2. Your true margin after costs is 40 percent, which puts breakeven at 2.5. The dashboard says you sit 60 percent above breakeven. You are below it and losing money on every sale.

How to check: Rebuild your margin from the bottom up and recalculate breakeven. Make sure the purchase value Meta receives follows the same tax and shipping rules as your margin.

How wrong data collapses a business

The damage runs in a chain:

1. A wrong number looks healthy.

2. You scale the campaign that looks best.

3. Spend grows faster than real revenue.

4. Cash drops while the dashboard still looks fine.

You find out when cash runs short, and by then the budget is spent.

The seven-step Meta ads data audit

Do these steps in order:

1. Match Meta purchases to store orders for the last 30 days. Use the same time zone and the same window.

2. Check deduplication in Events Manager. Confirm matching event IDs and event names across Pixel and Conversions API.

3. Write down your attribution settings and the date of every change. Set a new baseline after each one.

4. Report results on delivered orders, net of returns and refusals.

5. Rebuild breakeven ROAS from your real margin after shipping, fees and returns.

6. Track MER next to Meta ROAS every month. Watch the gap.

7. Run a holdout test when you have the volume, and keep a change log of every tracking edit.

The result is a set of numbers you can bet budget on.

Meta ads data integrity FAQ

Why does Meta show more purchases than Shopify?
Common causes are duplicate Pixel and Conversions API events, overlapping attribution and orders later canceled or returned. Start by checking deduplication.

Does Meta double count purchases?
It can. Meta merges Pixel and Conversions API events only when the event name and event ID match within 48 hours.

What changed with Meta attribution in 2026?
Meta removed the 7-day view and 28-day view windows on January 12, 2026. Agencies report a second change in March that narrowed click-through to link clicks.

Is Meta ROAS accurate?
Treat it as a credited number, not a measured one. Incrementality studies show it can overstate or understate real impact.

What is MER?
MER is total revenue divided by total ad spend across all channels. It shows what the business earned for what it spent.

Where the evidence ends

Meta's documentation confirms the deduplication rules and the January 2026 window removal. The March 2026 click change and the incrementality figures come from agencies and vendors, and several of them sell measurement tools. The ranges disagree with each other, so use them as direction. The audit steps are my method.

I run media buying and creative strategy together for e-commerce brands on Meta. If you want a Meta ads freelancer to audit your tracking before you scale, get in touch through the form on this site.

Sources

Meta documentation

• Handling Duplicate Pixel and Conversions API Events: https://developers.facebook.com/docs/marketing-api/conversions-api/deduplicate-pixel-and-server-events

• Server Event Parameters: https://developers.facebook.com/docs/marketing-api/conversions-api/parameters/server-event

• Conversions API End-to-End Implementation: https://developers.facebook.com/documentation/ads-commerce/conversions-api/guides/end-to-end-implementation

• Ads Insights API Metric Availability Updates, Oct 16, 2025: https://developers.facebook.com/blog/post/2025/10/16/ads-insights-api-metric-availability-updates/