Unlocking Attribution: Click and View Attribution in Programmatic Advertising

Two numbers describe the same campaign and they do not agree. One counts the people who clicked. The other counts the people who were exposed, did not click, and converted anyway. Most arguments between an advertiser and a media partner are a disagreement about which of those two numbers is the real one — and very few of them get settled, because nobody agreed in advance which number governs.

This is what click attribution and view attribution actually measure, what each one can and cannot prove, and why the gap between them is usually the larger part of the result rather than a rounding error.

What click attribution measures

Click attribution credits the ad a person clicked immediately before converting. It is precise, it is easy to audit, and it is the model almost every analytics tool defaults to. If someone clicks a display ad and deposits ten minutes later, click attribution captures that cleanly.

Its limit is not accuracy. It is scope. Click attribution can only see people who clicked, so it is a complete record of one behaviour and a blank space everywhere else.

What view attribution measures

View attribution credits exposure — a person who was served an impression, did not click, and converted later within an agreed window. On The Trade Desk this is tracked through conversion pixels and identity signals, and reported against a window you set before launch.

View attribution is the harder number to defend, which is exactly why the terms around it matter more: the window, who can change it, and whether the platform figure is ever reconciled against your own backend. A view-through number with no agreed window is not a measurement, it is an assertion.

The gap is not noise. It is most of the result.

Across six independent measurements — five operators, two continents, six years — the post-view share of conversions has landed between 89% and 96.2%. Every one of those was customer-ID-level or platform-verified. None was modelled. An operator reading clicks alone is seeing roughly one conversion in ten.

The shape underneath it is a ratio: 27.8 impressions per converting journey against 0.12 clicks. There are two clocks running. The execution clock — the click and the session immediately around it — is short, and a session-scoped analytics tool measures it well. The influence clock is long, and that same tool cannot see it by construction.

A third clock runs in the same account, and it is the one budget calendars collide with: how long an acquisition engine needs before it can hold a peak. The Seasonal Engine Lens puts that one in your hands — twenty-one months of measured first-deposit data from a single operator, with the cost curve and the build date as controls rather than assertions.

None of this is an incrementality claim. It says nothing about what would have happened with no media at all. It is a statement about measurement completeness only: how much of the delivered outcome a click-only report can physically observe. Those are different questions and they need different tests.

The full evidence, the archive it came from and the six accounts it cost to learn are set out in The 2% Problem — a fully anonymised case study, operators A to E, no client named.

Why Google Analytics shows you the smaller number

Google Analytics is a site analytics tool. It records what happens on your property and attributes it to the last click that brought someone there. That is a reasonable default and it is not a defect — it is a scope decision built into the product.

The consequence is structural rather than accidental. A person served an impression on connected TV on Sunday, who searches your brand on Wednesday and deposits on Thursday, appears in Google Analytics as organic or direct. The exposure that started the sequence is not in the data at all, because it never touched your site.

So an operator comparing a DSP report against Google Analytics is not finding an error in one of them. They are comparing two systems counting two different populations, and then treating the disagreement as a credibility problem for whichever number is larger. That is the moment most accounts go wrong.

What to fix, and when

The fix is procedural and it is cheap, but only if it happens before the money moves. Attribution education delivered inside a performance report — after the spend, in the appendix — does not change behaviour. Reconciliation before spend does.

• Agree which number governs — platform-attributed or backend-confirmed — and write it down before launch.

• Fix the window per objective, in the contract, and make changing it require your signature. Window-shifting is the oldest trick in performance reporting.

• Reconcile monthly, in writing, against your own backend: event definitions, time zones, currencies, deduplication, naming and UTM logic.

• Label anything unreconciled as platform-attributed, every time, until the backend is connected. Numbers that have never been checked should not be presented as facts.

• Ask for the raw path-to-conversion export. A desk that cannot produce full journey data is not analysing it either. Screenshots are not data.

If you want to test a partner against all of this rather than take it on trust, the 20-question partner scorecard runs the published due-diligence standard against any desk — including ours — and scores it against a bar of 24 out of 30. It takes a few minutes, asks for no email, and nothing leaves your browser.

Frequently asked questions

Which is more effective, click attribution or view attribution?

Neither, on its own. They answer different questions, so they will always disagree, and the useful information is in the size and direction of the gap. Last-touch flatters retargeting and lower-funnel display; multi-touch reveals what video, CTV and prospecting contributed upstream. Run both, side by side, and require the disagreement to be explained rather than hidden.

Can Google Analytics track post-view conversions?

No. It attributes to the last click and does not observe impressions that never produced a site visit. That is by design. Post-view measurement requires conversion tracking in the DSP, reconciled against your own backend — not a different setting in Analytics.

What attribution window should we use?

It depends on the objective, and it should be agreed before launch and written into the contract: tighter for lower-funnel registration and first-deposit activity, longer where CTV or digital out-of-home influence is genuinely part of the journey. The number matters less than the fact that it is fixed, documented, and cannot be moved after results are visible.

How do we know the post-view figure is real and not modelled?

Ask how it was derived. A defensible post-view number is customer-ID-level or platform-verified against known conversions, and the method is stated. A modelled estimate is a legitimate tool but it is a different class of evidence, and any partner should tell you unprompted which one you are looking at.

The longer version

This is one chapter of a larger argument. The iGaming Operator's Guide to Programmatic covers first-deposit economics, cohort logic, the ten red flags that show up in media reports, and a 90-day partner evaluation plan — 30 pages, free. ThumbAd is an independent programmatic trading desk in Oslo, trading on The Trade Desk across Europe, Africa, North America and Latin America since 2012, with betting and gaming at the core.