Haus Just Proved Meta Under-Reports Its Own Performance by 15%
Founder, BTB Audits. $150M+ in ad spend managed across Meta and Google
The Haus findings sit alongside Meta's own disclosure that 31 percent of incremental Meta conversions get credited to other channels in the typical attribution dashboard. Two independent data points, both pointing in the same direction: the typical D2C operator dashboard substantially under-represents Meta's actual contribution. The Haus paper is the more interesting of the two because it includes the off-DTC finding, which most operators have never measured at all.
What happened
What most operators will get wrong
The popular D2C operator complaint about Meta has been, for years: "Meta over-reports my conversions. The platform sees touchpoints it should not be claiming credit for." The complaint is so common it has become an article of faith in operator group chats.
The Haus data contradicts that complaint directly.
Meta is the conservative number, not the inflated one. The 15 percent under-reporting on click attribution means that the conversion count in Ads Manager is roughly 85 percent of the conversions Meta actually drove. The remaining 15 percent is real, measurable, incremental conversions that Meta is not getting credit for in its own dashboard. They show up in your back-end as direct visits, email recoveries, or organic search, but Haus's holdout testing proves Meta caused them.
The bigger surprise is the off-DTC finding. Roughly one-third of Meta's total incremental impact for a typical advertiser does not land on the advertiser's own DTC site at all. It lands in Amazon orders (someone saw the ad, then searched the brand on Amazon and bought there), in retail visits (someone saw the ad, then bought in a physical store), or in marketplace listings. Most operators do not measure these channels at all in the context of Meta attribution. They treat them as separate revenue streams driven by separate marketing.
The operators most affected by this finding are brands with meaningful Amazon or retail presence. If 25 to 35 percent of your total brand revenue runs through channels other than your own DTC site, then your Meta dashboard is showing you a fraction of Meta's real contribution. Cutting Meta budget based on the DTC dashboard alone is structurally wrong for these brands. The DTC ROAS may look weak. The full picture (DTC plus Amazon plus retail) tells a different story.
The misread that costs the most is concluding "Meta works for DTC brands but not for omni-channel brands." The Haus data says the opposite. Meta works better for omni-channel brands than the dashboard shows; the operators are just measuring two-thirds of the work.
What you should actually do
Run this 3-step check on your account this week. Most of the work is in your back-end and retail reporting, not in Ads Manager.
This is the calculation most CFOs have not seen and most operators have not done. The full diagnostic for setting up off-DTC measurement lives at Stage 8 of the Meta ad audit method. The 31 percent misattribution post covers the DTC-side under-counting in more detail.
How this changes the audit method
Stage 8 of the Meta audit method already covers cross-channel measurement reconciliation. The Haus data sharpens two checks inside Stage 8.
First, the DTC under-counting is now a quantified expected gap. Auditors comparing what Meta reports to what the brand's full revenue dashboard reports should expect Meta's number to be roughly 15 percent lower than reality, not equal to reality. The audit reconciles the gap rather than treating it as a Meta-reporting error.
Second, the off-DTC measurement is now a required line item for any brand with retail or marketplace distribution. Before the Haus data, the audit either skipped this entirely or treated it as a strategic conversation rather than a measurement step. Now there is a specific number (roughly one-third of total Meta impact) to anchor the analysis against. The auditor's job is to pull the brand-search lift from Amazon, the velocity data from retail, and reconcile against the Meta spend pattern.
These are the only changes to the Meta ad audit method. Stage 8 still comes eighth. What changes is the auditor has two more concrete benchmarks (15 percent DTC under-count, 33 percent off-DTC) to anchor the conversation against, rather than arguing from first principles every time.
Measurement mindset: before and after
| Aspect | Before | After |
|---|---|---|
| Default operator complaint about Meta | "Meta over-reports my conversions. The platform takes credit for sales it did not drive." | "Meta under-reports my conversions by 15 percent, and another 33 percent of impact lands off my site. My dashboard shows two-thirds of the real picture." |
| Where to look for Meta's contribution | Only on the DTC site. Amazon and retail are treated as separate channels driven by separate marketing. | DTC site plus Amazon brand search plus retail SKU velocity. Meta drives the awareness that lifts all three. |
| What "ROAS" actually means | Revenue from the DTC site divided by Meta spend. Easy to calculate, structurally incomplete. | Revenue from the full attribution picture (DTC + Amazon + retail + 15% Haus uplift) divided by Meta spend. Harder to calculate, structurally closer to truth. |
| What the audit checks at Stage 8 | Reconcile Meta, attribution platform, and back-end DTC revenue. | Same plus: pull Amazon brand search lift and retail SKU velocity for any brand with multi-channel distribution. Quantify the off-DTC contribution. |
| Risk of misreading the data | Lower. Operators knew DTC ROAS was an imperfect measure but had no specific industry benchmark. | Higher in one direction: an operator who hears the 15 percent number and treats it as license to keep growing Meta spend without doing the off-DTC math is using the data to justify a decision they were going to make anyway. |
Frequently asked questions
Common questions
About the study
Who is Haus and why does this study matter?
Haus is an independent incrementality measurement vendor used by major DTC and e-commerce brands. They run holdout-based experiments rather than attribution-model-based reporting, which means their numbers are not influenced by attribution windows or model choices. The 640-experiment dataset cited at the May 2026 Performance Marketing Summit reflects a large sample of brands across categories, which makes the findings more generalizable than a single-brand case study.
Is the 15% under-reporting consistent across all brands?
No. 15 percent is the average across the 640 experiments. Individual brand-level under-reporting varies based on creative strategy, audience targeting, customer journey length, and the strength of the brand's Conversions API setup. Brands with strong CAPI deduplication and broad targeting tend to see smaller gaps. Brands with weak signal layer and narrow targeting tend to see larger gaps. Run your own lift test to find your specific number.
Does the off-DTC finding apply to DTC-only brands?
Less directly. The one-third off-DTC number is heavily influenced by brands with Amazon and retail distribution. A pure DTC brand selling only through its own site will see most of Meta's impact land on the DTC site, with smaller leakage to organic search and direct. For these brands, the more important number is the 15 percent under-reporting on the DTC dashboard itself.
What to do next
How do I measure Amazon attribution from Meta ads?
The cleanest approach is brand-search lift analysis. Pull Amazon Brand Analytics data for your branded keyword search volume over the last 12 months. Overlay your Meta spend pattern (or specific Meta campaign launches). If brand search lifts during Meta-heavy periods and falls when Meta spend pauses, the lift is most likely from Meta-driven awareness. Quantify the incremental brand-search-driven Amazon revenue using your average conversion rate and AOV on Amazon. This is a directional measurement, not exact, but it captures impact your DTC dashboard misses entirely.
Should I shift budget from DTC measurement tools to incrementality tools?
Not shift. Add. DTC attribution platforms (Triple Whale, Northbeam, GA4) are still the right tool for daily and weekly decisions. Incrementality testing (Meta Conversion Lift, Haus, INCRMNTAL) is the right tool for periodic calibration of those daily decisions. Most healthy D2C operations run incrementality calibration twice a year and use the results to tune the daily attribution platform's weighting.
What if my CFO does not trust third-party measurement?
Run a Conversion Lift test inside Meta Ads Manager. Meta's in-platform Conversion Lift uses a true holdout group and is not affected by the same attribution-model bias as Triple Whale or Northbeam. The output is auditable, comes from Meta directly (so harder to dismiss as third-party noise), and produces account-specific numbers rather than industry averages. Use the Meta in-platform lift test as the bridge to having the Haus-style measurement conversation.
The Haus data is the strongest third-party case yet that operators are systematically under-measuring Meta. The 15 percent DTC under-count is a calibration adjustment. The 33 percent off-DTC impact is a whole new measurement category. Run the 3-step check above on your account this week. The 30 minutes to pull the numbers may unlock 25 to 50 percent more Meta-attributable revenue your dashboard never showed you.
If you don't have four to six hours, or you want a second pair of eyes that's managed $150M+ across Meta and Google, the Free Quick Scan is what I built for that. I'll record a private 5 to 7 minute Loom walking through the leaks I find on your account using public data only. You'll have it in 48 hours.
Get Your Free Quick Scan →Keep reading on Meta measurement
Aditya Chaturvedi is the founder of BTB Audits. He has managed $150M+ in ad spend across Meta and Google for DTC, SaaS, and lead-gen brands ranging from $10K per month to $500K per month. Industry Updates from BTB Audits cover platform changes and what they actually mean for operators, not what the headlines say they mean. Read more on the BTB Audits blog.