View-through conversions on Meta inflate reported ROAS by thirty to seventy percent in most accounts because the model credits a conversion to the ad even when the user saw the ad once, scrolled past it, and converted via another channel entirely. In accounts where VTC accounts for less than fifteen percent of total conversions, the inflation is small and the number is roughly usable. Above twenty-five percent, the inflation starts driving budget decisions toward audiences that may not be incremental at all. The specific VTC share that reflects real lift is almost never visible in platform reporting — it is only visible in a geo holdout or lift test, and most accounts have not run one.
VTC inflates reported ROAS by 30-70 percent in most accounts. Only 10-20 percent of view-through conversions reflect real incrementality. Below 15 percent VTC share, trust the number. Between 15-25 percent, discount it. Above 25 percent, exclude it from budget decisions until a lift test runs.
What exactly does a view-through conversion count?
A view-through conversion counts a user who saw an ad (an impression served, often for as little as one second with 50 percent of pixels in view) and converted within one day without clicking the ad. The user did not interact with the ad in any traceable way. Meta's model credits the ad because the impression was served before the conversion happened, and the correlation is tagged as attribution. The window is one day by default and can extend based on the account's attribution settings. VTC is distinct from click-through conversions, which require a click, and from engagement-through conversions, which require a documented interaction like a comment or share.
Why do view-through conversions overstate incremental lift so dramatically?
VTC overstates incremental lift because most of the conversions it credits would have happened anyway through other channels. A customer who already receives the brand's email, follows the brand on Instagram organically, and types the brand name into Google directly will convert through those channels regardless of whether Meta served them an impression yesterday. Meta's model sees the impression plus the conversion and credits itself. The customer did not need the impression to convert. The incrementality coefficient — the share of VTC that would not have happened without the ad — is typically 10-20 percent in Advize's experience across lift-tested DTC accounts. The other 80-90 percent is correlation, which inflates reported ROAS without reflecting business value. This is one of the primary reasons blended ROAS and platform-reported ROAS almost never match.
When is VTC data actually credible enough to use?
VTC data is credible when three conditions hold. First, VTC accounts for less than 15 percent of total reported conversions on the campaign — at that share, even if 90 percent of VTC is non-incremental, the inflation on total ROAS is manageable. Second, the campaign is prospecting-only to cold audiences with no prior brand exposure — these users have fewer competing attribution paths, so the view-through signal is more likely to be causal. Third, the account has run a lift test in the last 180 days establishing the incrementality coefficient, so the VTC share can be discounted appropriately for budget decisions. All three conditions together are rare, which is why most VTC data is either over-trusted at face value or dismissed entirely.
What is the specific threshold where VTC starts corrupting budget decisions?
VTC starts corrupting budget decisions when it exceeds 25 percent of total reported conversions on a campaign. Below that share, the correlation-inflation can be reasoned around and the budget allocation still reflects underlying truth. Above 25 percent, the inflation can push a campaign's reported ROAS above the account average when its incremental ROAS is actually below average, which routes additional budget toward spend that produces no additional revenue. This is particularly common in retargeting campaigns, where VTC often reaches 40-60 percent of reported conversions because the warm audience would have converted through email or direct anyway. Retargeting campaigns running at 5x reported ROAS routinely show 1.5-2x incremental ROAS when a holdout runs.
How do you measure the real VTC incrementality coefficient?
1) Run a geo holdout: identify two comparable metros, pause Meta spend in one for 30 days, measure revenue delta against the active metro and historical baseline. 2) Calculate the lift coefficient: incremental revenue divided by Meta-reported revenue. 3) Separate the VTC-attributed portion of Meta-reported revenue from the click-attributed portion. 4) Apply the lift coefficient to VTC specifically to isolate the view-through incrementality. 5) Record the account-specific VTC coefficient. Use it to discount VTC in budget decisions. 6) Rerun quarterly because the coefficient drifts as audience saturation and channel mix change. Confidence in multi-touch attribution has dropped below 50 percent across the market and lift tests have become the baseline for budget-weighting decisions in sophisticated accounts.
What should you do if you cannot run a lift test right now?
If a lift test is not feasible in the current quarter, use a conservative discount: assume VTC is 15 percent incremental and 85 percent correlation. Multiply VTC-attributed revenue by 0.15 and recalculate blended ROAS. The resulting number is almost always closer to truth than the platform-reported figure. Flag any campaign where this adjustment pushes ROAS below the account's break-even — those campaigns are candidates for budget reduction regardless of what Meta's dashboard shows. The 0.15 figure is based on the average incrementality coefficient Advize sees across lift-tested DTC accounts and is a reasonable default until a holdout provides an account-specific number. For B2B accounts, use 0.25 — the incrementality tends to be higher because competing attribution paths are fewer.
Conclusion
View-through conversions are credible in a narrow band and noisy outside it. Below 15 percent of total conversions, the number is usable at face value. Between 15 and 25 percent, it needs discounting. Above 25 percent, it should be excluded from the budget decision entirely until a lift test establishes the incrementality coefficient for the account. The practical next step for any team currently anchoring budget decisions to Meta's default 7-day click plus 1-day view window is to measure what fraction of reported conversions are VTC and recalibrate accordingly.