View-through vs. click-through conversions: the credit category most worth distrusting

View-through vs. click-through conversions: the credit category most worth distrusting

Display and video reports separate clicks from views. A view-through conversion credits an impression a user saw but never clicked, then converted within a window. When is that credit real?

The asymmetry
A click is a revealed action — weak evidence of intent, but evidence. A view-through is an exposure with no behavioral confirmation the ad registered at all. Yet view-through windows (often 1-30 days) sweep in vast numbers of conversions, because in a 30-day window almost everyone who buys saw some impression.

Why this is the easiest credit to fake
View-through is where attribution and selection bias fuse most dangerously. Retargeting view-throughs are the extreme case: you show impressions to people who already visited your site (high intent), they convert anyway, and the view-through claims it. The longer the window, the more "free" conversions the impression harvests. This is correlation dressed as causation by design.

What experiments find
Ghost-ad and PSA holdout studies repeatedly show view-through-attributed conversions have far lower true incrementality than click-through ones — often a small fraction of the claimed volume.

Bottom line for practitioners: Treat click-through and view-through as different evidentiary tiers, never a single "conversions" total. Shorten view-through windows aggressively (a day, not a month), exclude view-through entirely from retargeting evaluation, and require any channel leaning on view-through credit to pass a holdout test before it keeps its budget. View-through is the single line item most likely to be measuring conversions you already owned.