View Through Conversions: A Sceptical Read Before You Pay for Them

October 30, 2026

View through credit pays a platform for purchases by people who may never have noticed the ad. How to test whether those conversions would have happened anyway.

A view through conversion is a purchase credited to an ad that someone was served, did not click, and may not have consciously seen. Platforms report them alongside click conversions, blended, which flatters the channel in exactly the way a media buyer should expect a seller of media to flatter it. None of this means view through value is zero. It means the burden of proof is on the claim, and the default settings will never carry that proof for you.

Start with the window. A 30 day view through window on a frequently purchased product will hoover up repeat orders from loyal customers who were always going to buy, and present them as advertising outcomes.

Separate, shorten, compare

Report view through separately from click conversions, always, in every internal document. The moment they are blended, the number stops being a measurement and becomes a sales aid. Then look at window sensitivity: pull the same channel at 1 day, 7 day, and 30 day view windows. Real influence decays. If reported conversions keep climbing steeply with window length, you are mostly measuring time passing near your existing demand.

Compare the audience honestly. Retargeting pools are built from your own visitors, the people with the highest baseline purchase rate you have. Crediting view through conversions in a retargeting pool without a holdout is close to paying for your own traffic twice: once to earn the visit, once to stand near the purchase it produced.

The holdout is the argument

The clean test is a holdout: a random slice of the eligible audience that is not served the ads, compared on purchase rate over the same period. If the exposed group does not buy more than the holdout, the view through conversions in the dashboard are attribution, not incrementality, and budget should move accordingly.

Platforms offer lift studies that approximate this, and they are worth running, with the caveat that the platform designs, runs, and reports the exam it sits. Where spend justifies it, run your own holdout at the audience or geography level and keep the method boring and consistent so results are comparable quarter to quarter.

What to pay for in the meantime

Until a holdout says otherwise, manage the channel on click and post purchase survey and incrementality evidence, and treat view through as a diagnostic: useful for spotting frequency build up and reach into new audiences, not as revenue. Set frequency caps with that in mind. The most expensive view through conversions are usually the highest frequency ones, served to people who had already decided.

Write the rule down before the results arrive. Teams that decide what evidence would change the budget after seeing a flattering dashboard rarely change the budget.

Frequently asked questions

What is a view through conversion?

A purchase credited to an ad impression where the person did not click. It assumes the impression influenced the purchase, which is exactly the assumption that needs testing.

Why do long view windows inflate results?

A long window captures more purchases that would have happened anyway, especially repeat orders from existing customers. If conversions climb steeply as the window lengthens, you are measuring elapsed time near existing demand.

How do I test whether view through value is real?

With a holdout: withhold ads from a random slice of the eligible audience and compare purchase rates. No lift over the holdout means the reported conversions are not incremental.

Should view through conversions be reported at all?

Yes, but separately from click conversions, with the window stated. Blending them hides the assumption inside a revenue number.