View-Through vs. Click-Through Attribution
Crediting an ad the user never clicked
Click-through attribution credits a conversion to an ad only if the user clicked it. View-through attribution credits a conversion to an ad the user merely saw — an impression served and rendered, with no click — provided the conversion happens within a (typically much shorter) view-through window after that impression. The justification is real: display and video ads are known to influence purchase decisions even when never clicked, the same way a billboard or TV spot can. The problem is that view-through attribution is also uniquely easy to inflate, because an impression is a much lower bar to clear than a click, and most users are shown far more impressions than they click.
Why this structurally favors high-impression, low-engagement channels
Display remarketing and programmatic video campaigns can serve the same user dozens of impressions across a single browsing session through standard ad exchange frequency. If any one of those impressions falls inside the view-through window of an eventual conversion — commonly 1 day, sometimes up to 7 — the channel gets credit, regardless of whether the user ever consciously registered seeing the ad. A user who was already going to buy, saw a retargeting banner they don't remember, and converted the next day will show up as a view-through conversion for that banner even if the banner's actual causal influence was zero.
User journey (24 hours):
09:00 Organic search for product (genuine interest signal)
09:15 Served 14 programmatic display impressions
across news sites while browsing (never clicked any)
09:20 Direct navigation back to site
09:45 Purchase
Click-through attribution:
-> Display channel: 0 credit (no click occurred)
View-through attribution (1-day window):
-> Display channel: full credit for the purchase
-> Reported ROAS for display campaign jumps accordingly
The display ads may have had zero actual influence —
the user was already searching with purchase intent.View-through inflation is the single biggest reason display ROAS looks better in-platform than it is
Click-through vs. view-through attribution
| Dimension | Click-through | View-through |
|---|---|---|
| Trigger event | User clicked the ad | Ad was served/rendered, no interaction required |
| Causal confidence | Higher — explicit user action | Lower — exposure alone, easy to over-credit |
| Typical window length | 7-30 days | 1-7 days, usually shorter |
| Most affected channel types | Search, email, affiliate | Display, programmatic video, some social |
| Inflation risk | Lower, but still subject to window-length effects | High — frequency/reach alone can trigger credit |
| Best validated by | Platform click data cross-checked with GA4 | Incrementality testing (geo-lift, PSA holdout) |
When view-through attribution is legitimate to include
View-through isn't inherently invalid — brand and upper-funnel video genuinely can move someone toward a purchase without a click, and excluding view-through entirely would systematically undercount upper-funnel channels the same way an overly short attribution window does. The difference between legitimate use and inflation is usually window length and channel type: a 1-day view-through window on a video ad with a clear brand message is a defensible measurement of genuine recency effect; a 7-day view-through window on a programmatic display network running dozens of impressions per user is a setup that will attribute conversions to ads nobody consciously saw.
GA4 doesn't model view-through at all
Validating view-through credit with holdout testing
The only way to know whether a channel's view-through conversions represent real incremental lift or just exposure-correlation is to test it directly — run a holdout where a comparable audience segment or geography sees no impressions at all, and compare conversion rates. If the holdout group converts at a similar rate to the exposed group, the view-through credit was measuring correlation, not causation. This is exactly the logic behind geo-lift incrementality testing, which extends the same holdout principle to a full campaign rather than a single view-through setting.
What's next
Holdout-based validation is the throughline connecting this lesson to incrementality testing more broadly — geo-lift testing applies the same exposed-vs-holdout logic at the regional level to measure a channel's true incremental impact without relying on pixels or cookies at all.
Next: Geo-Lift Incrementality Testing →
I build these systems professionally.
Whether it's a RAG pipeline, analytics migration, or AI workflow — let's talk.