View-through attribution is a lie your ad platform tells you
Crediting a sale to an impression nobody engaged with is not measurement, it is a correlation with a marketing budget attached. Here is how to test it on your own account this week.
view-through credit ∝ impressions served, not intent demonstrated
Turning view-through off typically drops reported conversions by 15–30% while actual sales stay flat. If a metric can be halved by a settings change and your revenue does not move, that metric was not measuring your revenue.
There is a legitimate brand-awareness argument for view-through. It is not an argument for putting it in the same column as clicks.
A view-through conversion means: this person was served your ad, did not click it, and later bought something.
Stated plainly, it sounds like what it is. Stated as “1-day view attribution window,” it sounds like a technical setting, and it has been sitting on by default in your account since you opened it.
The mechanical problem
Consider two people:
Person A scrolled past your ad on Tuesday without registering it, then bought on Thursday after a friend recommended you.
Person B scrolled past nothing at all, then bought on Thursday after a friend recommended you.
Your platform credits Person A to the ad campaign. Person B goes to direct. The only difference between them is that A happened to have your ad served in a feed they were scrolling past.
Scale that up. The more impressions you buy, the more purchases will coincidentally have an impression in the preceding 24 hours — regardless of whether the impression did anything. View-through credit scales with impressions served, not with intent demonstrated.
That is not a measurement. That is a correlation with a budget attached to it.
The test
Turn view-through off in your reporting settings — 7-day click, no view — and compare the same fortnight before and after.
| What you observe | What it means |
|---|---|
| Reported conversions drop 15–30%, revenue flat | Normal. That share was never click-driven. |
| Reported conversions drop 40%+, revenue flat | Your channel is heavily impression-credited. Treat every CPA from it as fiction. |
| Reported conversions drop and revenue drops with it | You changed something else at the same time. Try again cleanly. |
| Barely changes | Congratulations, you are already reporting honestly. |
The near-universal result is the first or second row. Revenue does not move, because turning off a reporting setting cannot affect what customers do.
The stronger test
The settings test tells you how much of your reported number is view-through. It does not tell you whether the ads work.
For that you need incrementality: turn a campaign off completely in one comparable geography for two to four weeks, and compare total revenue between regions — not platform-reported conversions, which are precisely the thing under suspicion.
It is uncomfortable, it costs some revenue if the campaign is genuinely working, and it is the only honest answer. Every mature performance team runs holdouts, and most of them were surprised the first time.
The steelman
There is a real case for view-through, and it deserves stating properly.
Impressions do influence behaviour. Brand advertising has worked for a century without a click. A customer who saw your ad three times and then searched your brand name was influenced by those impressions, and last-click will credit branded search for work the display campaign did.
That is all true. It is an argument for measuring brand impact with holdouts and brand-lift studies. It is not an argument for adding impression-credited conversions to click-credited conversions and dividing spend by the total.
The error is not “counting impressions.” The error is mixing two different units of evidence in one number and then doing arithmetic on it.
What to do
- Set click-only windows for every campaign whose CPA drives a budget decision.
- Keep view-through in a separate report labelled “influenced”. Nobody should ever compute a CPA from it.
- Run a geo holdout on your largest campaign this quarter. Once a year, minimum.
- Reconcile against your own ledger. A first-party record of what actually happened is what makes any of these comparisons possible.
If you want the size of the problem before doing any of the work, the ghost-conversions checker prices the total gap across every platform in about two minutes.
Frequently asked questions
Is view-through attribution completely worthless?
No. Impressions genuinely influence behaviour, and for brand campaigns, view-through is a directional signal worth watching. The problem is reporting it in the same column as click conversions and then computing CPA from the total.
How do I test whether mine is real?
Run a geo holdout. Turn a campaign off entirely in one comparable region for two to four weeks and compare total revenue, not platform-reported conversions. Incrementality is the only honest test.
What window should I use instead?
Click-only, 7-day, for anything that drives budget decisions. If you want a view-through view, keep it in a separate report labelled 'influenced' so nobody computes a CPA from it.
Why do platforms enable it by default?
Because it makes the platform look more effective, and because there is a defensible theoretical case for it. Both things are true at once. Defaults are a business decision, not a measurement one.
Does anyone still buy on view-through?
Programmatic display and some CTV buying essentially cannot work without it. Which is a good reason to hold those channels to a higher incrementality standard, not a reason to accept the metric at face value.
See what your platforms are claiming in total
Add up every platform's reported conversions and compare to what you actually banked. View-through is usually the single largest contributor to the gap.
Stop guessing which ad made the sale.
Flowsk Signals stitches the anonymous click to the email to the purchase — first-party, server-side, de-duplicated. One snippet, $29/mo, and every conversion comes with a receipt you can inspect.