View-Through vs Click-Through Attribution: What Each Number Is Really Claiming
Two windows, two different claims about cause. How to use view-through credit without letting it swallow every order in the store.
Every ad platform offers two kinds of credit: click through, for people who clicked and later bought, and view through, for people who saw an ad, never clicked, and bought anyway. View through is where attribution debates go to die, because the claim is unfalsifiable in the platform's own report: some of those buyers were genuinely influenced, and some would have bought regardless, and the report cannot tell them apart.
Used carelessly, view through makes every campaign look essential and total attributed revenue exceeds actual revenue by a wide margin. Used well, it measures something real about how display and video actually work. This guide separates the claim from the credit.
What each window is actually claiming
A click is evidence of intent: the shopper saw, chose, and arrived. Click through attribution claims the click caused the visit that caused the order, which is at least a connected story. A view is evidence only of an impression served, often for a second, possibly below the fold of attention entirely. View through claims the order would not have happened without an exposure the buyer may never have registered.
That does not make view through fiction. People are influenced by things they do not click, especially video and repeated display, and click only measurement systematically undervalues channels that work by familiarity. The honest position is that view through measures a real effect with an unknown error rate, while click through measures a smaller effect with a clearer mechanism. Reports should carry that distinction instead of blending both into one 'attributed revenue' line.
The window length is the argument
Most view through inflation is window inflation. A one day view window makes a modest claim. A thirty day view window claims that an impression a month ago caused today's order, across every other exposure, search, email, and need that intervened. Platforms default to long windows because long windows credit more, and credit justifies spend. Shortening the view window is the single fastest way to make view through numbers defensible.
Match the window to the buying cycle honestly. A considered purchase with a two week research phase supports a longer window than an impulse buy, where a seven day view claim is already generous. When someone presents view through results, the first question is always the window, and the second is whether it changed recently, because a quietly lengthened window manufactures growth that no campaign created.
Keep the two numbers in separate columns
The reporting discipline that keeps everyone honest: click through and view through are reported side by side, never summed, never blended into a single attributed figure. Decisions scale with the click number and get context from the view number. The moment the two are added, every channel's claim grows and the total detaches from the store's actual revenue, which is the anchor that keeps attribution tethered to reality.
Apply the same separation retroactively with care. Historical reports built on blended numbers will make any honest new report look like a decline. Flag the methodology change in the report itself, restate the old period in the new format where possible, and let the team argue about the change once, openly, rather than rediscovering it every quarter.
Test influence instead of assuming it
The way to learn what view through is worth in your store is a holdout: a region, an audience slice, or a time period where the ads do not run, compared on total orders rather than attributed ones. Holdouts are unglamorous and occasionally unwelcome, because they sometimes show the influence is smaller than the report said. They are also the only evidence that distinguishes influence from coincidence at scale.
Run them periodically, not once. Influence decays as creative ages and audiences saturate, so last year's holdout does not govern this year's budget. A program that re tests its view through assumptions a couple of times a year can defend display and video spend with a straight face, in numbers leadership can check against the bank account.
Related reading
If this topic is useful, these earlier guides go deeper on connected questions: Attribution for Influencer and Creator Seeding Programs, How to Attribute Sales From In-Store and Pop-Up Events, and Attribution When the First Visit Is Research and the Purchase Comes Weeks Later.
Frequently asked questions
What is the difference between view through and click through attribution?
Click through credits orders from people who clicked the ad; the mechanism is visible. View through credits orders from people who were served an impression and bought without clicking; the influence is plausible but the platform cannot separate genuine influence from buyers who would have purchased anyway. They are different strength claims and should never be blended.
What view through window is reasonable?
As short as the buying cycle honestly allows: a day to a few days for most commerce, longer only for genuinely considered purchases. Long default windows credit stale impressions for orders with many intervening causes. The window is the main lever on how inflated view through numbers get, so set it deliberately.
Should view through revenue ever be added to click through?
No. Report them in separate columns, scale decisions on the click number, and use the view number as context about familiarity effects. Summing them double counts shared journeys and inflates totals beyond the store's real revenue, which destroys the anchor that makes attribution useful.
How do we find out what view through is really worth?
Holdout tests: pause the ads for a region, audience slice, or period and compare total orders, not attributed ones. That measures actual influence. Repeat periodically, because creative fatigue and audience saturation change the answer over time.