Paid media
Correct platform over-reporting into a true blended ROAS you can bank on.
- Category
- Paid media
- Signup
- None
- Your data
- Stays in your browser
- Also in
- Claude, via MCP
How do you correct platform-reported ROAS?
Compare what the ad platforms claim with what your backend actually recorded. True blended ROAS = actual revenue ÷ ad spend, and the correction factor = true ROAS ÷ platform ROAS. If platforms claim ₹42 lakh but your store shows ₹30 lakh on ₹8 lakh spend, platforms over-report by 40% and every platform ROAS should be multiplied by 0.71.
Your platforms claim 5.25× but true blended ROAS is 3.75×, about 40.0% of over-reporting from overlapping attribution. Apply a ~29% haircut to platform numbers when you scale.
Every platform takes credit for the same conversion, so summed platform revenue routinely exceeds real revenue. Blended, backend-anchored ROAS is the honest number.
About this calculator
Since Apple's App Tracking Transparency rollout, ad platforms lean harder on modeled and probabilistic attribution, and the practical effect is that every platform tends to claim credit for the same conversions your other channels also claim. This calculator compares the sum of what your platforms report against your real backend revenue, quantifying exactly how much of that claimed revenue is double-counted overlap rather than genuinely incremental spend.
How to use it
- Enter platform-claimed revenue, the sum of revenue every ad platform reports across all your channels.
- Enter actual revenue from your backend, store, or finance system for the same period.
- Enter total ad spend across all channels.
- Read the true blended ROAS versus platform-reported ROAS, the over-reporting percentage, and the double-counted revenue figure.
Methodology
Platform-reported ROAS is platform-claimed revenue ÷ ad spend, and true blended ROAS is actual backend revenue ÷ ad spend, the same denominator, two different revenue numerators.
Over-reporting is how much platform-claimed revenue exceeds actual revenue, expressed as a percentage, and double-counted revenue is that same comparison in currency, claimed minus actual.
A correction factor is derived as true blended ROAS ÷ platform ROAS, giving you a single multiplier to discount platform-reported numbers toward reality when comparing across periods or making budget decisions.
Anything above roughly 15% over-reporting is flagged as inflated. This threshold is a practical planning heuristic, not a precise statistical cutoff, some overlap between platforms is normal and expected even with clean tracking.
Platform ROAS, true blended ROAS and the correction factor
Platform ROAS = Platform-claimed revenue ÷ Ad spend True blended ROAS = Actual backend revenue ÷ Ad spend Over-reporting = Platform-claimed revenue ÷ Actual revenue − 1 Correction factor = True blended ROAS ÷ Platform ROAS
- Platform-claimed revenue
- Revenue each ad platform attributes to itself, added up across Meta, Google and others for the same period.
- Actual backend revenue
- Revenue your store, billing system or CRM recorded from customers in the period.
- Ad spend
- Total spend across the same platforms.
Over-reporting comes from two sources: several platforms claiming the same sale (each counts a view or click in its own window), and modelled conversions that estimate sales the platform cannot observe directly. Actual revenue here usually includes sales that ads did not cause at all, so true blended ROAS is not incremental ROAS either. It is a more honest ceiling, not the final answer.
Worked example: a D2C brand on Meta and Google
In one month, Meta and Google together claim ₹42,00,000 of revenue from ₹8,00,000 of spend. The brand's Shopify store recorded ₹30,00,000 of revenue in total over the same month.
- Platform ROAS = ₹42,00,000 ÷ ₹8,00,000 = 5.25×.
- True blended ROAS = ₹30,00,000 ÷ ₹8,00,000 = 3.75×.
- Over-reporting = ₹42,00,000 ÷ ₹30,00,000 − 1 = 40%.
- Double-counted revenue = ₹42,00,000 − ₹30,00,000 = ₹12,00,000.
- Correction factor = 3.75 ÷ 5.25 = 0.71. A campaign showing 6.0× in the platform is closer to 4.3× on this basis.
Use 0.71 to discount platform figures when comparing campaigns or setting targets, and recalculate it every month because it moves with spend mix and tracking changes. For the whole-business view, use the ROAS to MER converter. To estimate what ads truly add, run a holdout or geo test; Meta's open-source GeoLift is one method.
How much over-reporting is normal?
Some overlap is normal whenever more than one platform runs, because each counts the sales it touched. The tool flags over-reporting above 15% as inflated; that is a planning threshold, not a published standard.
Use three checks:
- Trend: a stable gap is easier to correct for than one that jumps after a tracking change.
- By platform: compare each platform's claimed revenue with orders tagged to it by UTM in your backend.
- Against tests: a holdout test tells you what the platforms really add, which is often less than either number.
Attribution gaps in India vs the US
India is mostly an Android market, so Apple's App Tracking Transparency affects a smaller share of traffic than in the US. The bigger gaps for Indian D2C brands are usually cash-on-delivery orders that never complete and orders placed through WhatsApp or phone that platforms cannot see. Correct for delivered orders first with the RTO-adjusted ROAS calculator.
iPhone users make up a large share of US traffic, and Apple's App Tracking Transparency requires apps to ask permission before tracking users across other companies' apps and websites. Platforms responded with modelled conversions, which fill gaps with estimates. Server-side events through Meta's Conversions API improve the data the models start from.
Common attribution correction mistakes
- Adding up revenue claimed by every platform and treating the total as real.
- Using total store revenue, including organic and repeat sales, as if all of it came from ads.
- Comparing periods with different attribution windows or settings.
- Applying one correction factor to every platform when the gap differs between them.
- Cutting the platform with the biggest gap without testing what it actually adds.
Sources
- User privacy and data use, Apple Developer. Accessed 2026-10-19.
- Conversions API, Meta for Developers. Accessed 2026-10-19.
- GeoLift introduction, Meta Open Source. Accessed 2026-10-19.
FAQ
They compute similar math but for a specific reason: this tool is framed around the post-iOS14 tracking problem specifically, comparing what platforms claim in aggregate against verified backend revenue to isolate the double-counting effect that modeled, probabilistic attribution introduced. Use it when you specifically suspect platform over-reporting from privacy-driven attribution changes, rather than as a general MER check.
With less device-level tracking available, platforms shifted toward modeled conversions and broader attribution windows to fill the gap, which increases the odds that a single sale gets attributed, and counted as revenue, by more than one platform simultaneously.
Apply it as a rough discount when a platform reports a ROAS or revenue figure for planning purposes, treating the raw platform number as the reliable one leads to over-scaling budget on inflated numbers. Recompute the correction factor periodically since the gap tends to drift as channel mix and platform attribution methods change.
Usually, yes, if the gap is large and consistent across many periods. A modest, stable gap can simply reflect normal multi-touch overlap. A gap that keeps widening, or is unusually large versus your historical baseline, is the stronger signal that attribution has genuinely degraded and platform numbers need heavier discounting.
Each platform credits itself if a buyer saw or clicked its ad within its attribution window. A buyer who clicked a Google ad and saw a Meta ad can be counted by both.
Blended ROAS is total revenue divided by total ad spend across all channels, measured from your own backend rather than from platform reports.
Incrementality is the revenue that would not have happened without the ads. It is measured with experiments such as holdout groups or geo tests, and is usually lower than attributed revenue.
Numbers looking off?
If the model says one thing and the pipeline says another, that gap is usually where the revenue is. A 30-minute call finds it.
It starts with a 30-minute call. Pick a time below.