View-Through Attribution

View-through attribution (VTA) credits a conversion to an ad IMPRESSION the user SAW but did NOT click — recognizing that ads influence people even when they don't click (the display banner, the video ad, the social impression that registered, shaped perception, or prompted a later search, without a click). It contrasts with CLICK-THROUGH attribution (which credits only clicked ads), and it's measured within a 'view-through window' (the period after the impression during which a conversion is still credited to the view — often a day, sometimes longer). The premise is real: viewed-but-unclicked ads can genuinely influence behavior. The danger is equally real: view-through credit is dangerously easy to over-count.

Why the signal is real but easily over-credited, and the window: view-through attribution's premise is sound — ads do influence people without clicks (the impression that built awareness, shaped consideration, or prompted a later branded search or direct visit — the influence that click-only measurement misses entirely, especially for upper-funnel, display, video, and social ads whose value is often in the view, not the click). But VTA has a structural over-crediting problem that makes it the most abused metric in attribution: an impression is a low bar (the ad was merely served and viewable — the person may not have noticed it at all), so VTA can claim credit for conversions that would have happened anyway (the person who was going to buy regardless, who happened to be served an impression along the way — the correlation-not-causation core of the INCREMENTALITY-VS-ATTRIBUTION problem, in its most inflated form), and the longer the view-through window, the more conversions get swept into the credit (a long window credits the impression for conversions days later that it plausibly had nothing to do with). So VTA systematically over-credits — it's the metric that makes display and programmatic campaigns look far more effective than they incrementally are, by claiming a conversion the person would have made anyway. The disciplines that keep VTA honest: a SHORT view-through window (limiting credit to conversions close to the impression, where influence is more plausible — a day rather than weeks), VIEWABILITY standards (only counting impressions that were actually viewable — on-screen long enough to be seen — not served-but-unseen, the VIEWABILITY-and-vCPM connection), de-duplication and sensible weighting against click and other touches (not double-counting, not over-weighting the view), and crucially VALIDATION WITH INCREMENTALITY (the holdout test that reveals how many view-through conversions were actually incremental versus would-have-happened-anyway — almost always far fewer than VTA claims). The honest framing: view-through attribution captures a real influence (viewed-but-unclicked ads do affect behavior, which click-only measurement misses), but it's structurally prone to severe over-crediting (an impression is a low bar, it claims would-have-happened-anyway conversions, and a long window inflates it further); the discipline is using VTA with a short window, viewability standards, and sensible weighting to capture the real view influence — while never trusting the raw view-through number as incremental, and validating it with the holdout testing that reveals how little of it actually was.

View-through attribution matters most for upper-funnel and impression-led media (display, video, social, programmatic, CTV) — where much of the genuine value is in the view, not the click, and click-only measurement undercounts the real influence. It matters as a real signal worth capturing (viewed ads do influence behavior) AND as the most over-credited metric in attribution (the structural over-counting that makes impression-based media look far more effective than it incrementally is). The discipline is using VTA with a short view-through window, viewability standards (count only genuinely viewable impressions), sensible de-duplication and weighting against clicks and other touches, and — the decisive control — validating view-through credit with incrementality holdout testing rather than trusting the raw number, which almost always claims far more than was actually caused. Treat VTA as a directional signal of view influence to be earned through incrementality, not as incremental conversions to be banked.

View-through attribution arose with display and programmatic advertising to capture the influence of viewed-but-unclicked ads that click-only measurement missed; its structural over-crediting - an impression is a low bar, and a long window inflates credit - made it the most abused attribution metric, disciplined by short windows, viewability standards, and the incrementality testing that reveals how little view-through credit is actually causal.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where view-through attribution is a core concern: