And what DTC brands should use instead.
Platform-reported ROAS from Meta, Google, and TikTok overstates true channel performance by 30–200% due to cross-platform double-counting, view-through attribution inflation, and probabilistic modeling that consistently overcounts conversions after iOS privacy changes. DTC brands should use Marketing Efficiency Ratio (MER) as their north star metric and Marketing Mix Modeling for channel-level budget allocation instead of platform ROAS.
If you add up the ROAS your ad platforms report, you’d need 10x your actual revenue to make the numbers work. That’s not a rounding error. It’s a structural problem with how digital advertising attribution works — and it’s costing DTC brands millions in misallocated spend.
Meta says it drove a sale. Google says it drove the same sale. TikTok saw the user too, so it also takes credit. Each platform attributes the full conversion to itself because each platform only sees its own touchpoint.
The result: if you sum platform-reported ROAS across all channels, the total implies you’re generating far more revenue than your bank account shows. This isn’t fraud — it’s each platform doing exactly what it’s designed to do: maximize the appearance of its own impact.
Meta’s default attribution window includes view-through conversions: someone sees your ad, doesn’t click, then buys within 1–7 days. Meta counts that as a Meta-driven sale — even if the person was already going to buy, or was driven by a Google search, an email, or a friend’s recommendation.
For brands running broad awareness campaigns, view-through attribution can inflate reported ROAS by 2–5x above true incremental impact.
Apple’s App Tracking Transparency (ATT) update, launched in 2021, allowed iOS users to opt out of cross-app tracking. According to mobile analytics firm Flurry, approximately 75–85% of iOS users opted out. This removed direct conversion visibility for a large percentage of mobile users on Meta, TikTok, Snap, and other platforms that relied on device-level tracking for attribution.
The ad platforms responded with probabilistic modeling — statistical estimates of conversions they can no longer directly observe. Meta’s Aggregated Event Measurement (AEM) and TikTok’s Advanced Matching use modeled data to fill the gap. The modeling is directionally useful, but it is inherently noisy and consistently biased toward overcounting rather than undercounting. The platforms have a financial incentive to report higher ROAS, and their modeling reflects that incentive.
Google was less affected because Search ads rely on click-based attribution (user clicks ad, lands on site, converts), which doesn’t depend on device-level tracking. However, Google’s Performance Max campaigns, which include YouTube and Display inventory, use modeled conversions similar to Meta’s approach — with the same overcounting risk.
The degree of ROAS overstatement varies by channel and campaign type, but research and incrementality testing consistently show that platform-reported ROAS exceeds true incremental ROAS by 30% to 200% or more. Upper-funnel awareness campaigns (Meta prospecting, TikTok reach) show the largest gaps because they rely heavily on view-through attribution. Lower-funnel campaigns (Google branded search, retargeting) have smaller gaps but still overcount by capturing conversions that would have occurred organically.
Here’s what incrementality testing typically reveals:
| Channel / Campaign Type | Platform ROAS | True Incremental ROAS | Overstatement |
|---|---|---|---|
| Meta broad prospecting | 4.0–8.0x | 1.5–3.0x | 60–170% |
| Meta retargeting | 8.0–15.0x | 2.0–5.0x | 100–200% |
| Google branded search | 10.0–20.0x | 3.0–8.0x | 100–150% |
| Google Shopping (non-brand) | 3.0–6.0x | 2.0–4.5x | 30–50% |
| TikTok awareness | 2.0–5.0x | 0.8–2.0x | 100–200% |
These ranges are based on common patterns observed in geo-holdout tests and MMM analysis across DTC brands. Your specific numbers will differ — which is exactly why measuring true incrementality matters.
Platform ROAS measures channel-specific return as reported by each ad platform (total attributed revenue divided by spend on that platform). Blended MER (Marketing Efficiency Ratio) measures total business revenue divided by total ad spend across all channels. MER uses real revenue from your bank account and real spend from your ad accounts — no attribution model in between. It is the single most honest metric for evaluating overall marketing efficiency.
The key difference: platform ROAS can be gamed by attribution windows, double-counting, and modeled conversions. MER cannot. If your Meta Ads Manager shows 5.0x ROAS and your Google Ads shows 4.0x ROAS, but your MER is only 3.0x, something doesn’t add up — and MER is telling you the truth.
Calculate your MER now: Use our free MER calculator to see where you stand against industry benchmarks.
DTC brands can improve attribution accuracy by adopting a three-layer measurement framework: MER as the north star metric for overall efficiency, Marketing Mix Modeling (MMM) for channel-level budget allocation, and incrementality testing for validating high-stakes channel decisions. This approach uses real business outcomes rather than platform-reported metrics that conflict with each other.
Three approaches give you more honest measurement: