View-Through vs Click Attribution: What SaaS Loses

Ad platforms spent 2026 quietly deleting a category of conversion that SaaS teams had been counting for years. In January, Meta removed its longer view-through attribution windows; by March it had narrowed what even counts as a click. Google had already made click-based credit its default. The headline effect was a sudden drop in reported conversions for many advertisers — but the drop was definitional, not real. No customers disappeared. What disappeared was credit for conversions that the platform used to attach to an ad someone saw but never clicked.

For a SaaS growth or product lead, this is less a paid-media story than a measurement story. If your dashboards read platform-reported conversions as the scoreboard, a change in how a platform defines a conversion looks exactly like a change in performance — and you will spend a week chasing a demand collapse that is really an arithmetic edit. This post explains the difference between view-through and click credit, why the removal matters for a reporting baseline more than for the business, why SaaS reads it differently than ecommerce, and how to re-base your paid contribution on data a platform cannot revise.

What actually changed: platforms are deleting view-through windows

The concrete events are specific and dated. On 12 January 2026 Meta removed the 7-day-view and 28-day-view attribution windows from its Ads Insights API, leaving 1-day view and 7-day click as the defaults — reporting that advertisers can no longer pull the longer view-through windows from the API. Then on 3 March 2026 Meta redefined click-through to mean link clicks only, pushing likes, comments, shares and short video views into a new engage-through bucket, as documented in detail by practitioners tracking the change. Two edits, months apart, both narrowing what counts as a credited conversion.

The reported impact was large and immediate. Some advertisers saw 30-40% of reported conversions vanish overnight on the change dates. Read naively, that is a catastrophe. Read correctly, it is a platform counting fewer things: the conversions that used to be credited to a passive impression simply stopped being credited. The business did not lose a third of its customers on 12 January. It lost a third of a particular, generous way of counting them. That gap between what changed and what it looks like is the entire problem, and it is the same class of trap we describe when two analytics tools report different numbers for the same reality.

View-through vs click-through: the credit model, not the window length

The distinction that matters here is not how long the window is but what the window credits. Click-through attribution gives an ad credit only when the user actively clicked it and then converted. View-through attribution gives an ad credit when the user merely saw it — no click — and converted within the window afterward. A click is an active signal of intent; an impression is passive exposure. As one explainer of the two models puts it, view-through is concerned with impressions, while click-through focuses on actions. That is a different axis entirely from the lookback-length question we cover in attribution lookback and conversion windows: you can shorten a window and still credit views, or keep a long window and credit clicks only.

Because a view is a weaker signal than a click, any honest model weights view-through conversions below click-through ones — and many SaaS teams should arguably not count platform view-through at all, because it systematically flatters the channels that buy the most impressions. Removing view-through windows therefore does something subtle: it does not just cut the count, it changes the ranking of channels, because the channels that leaned hardest on cheap impression volume lose the most credit. A paid-social line that looked efficient on view-through credit can drop below an organic or referral line once only clicks count — which is why the removal should trigger a fresh read of your multi-touch attribution model, not just a note that the totals fell.

Why SaaS reads this differently than ecommerce

Most of the loud commentary on the 2026 window changes came from ecommerce and performance-marketing teams, where the purchase is fast and the attribution window is short. SaaS buys behave differently, and that difference cuts both ways. A B2B SaaS purchase has a long consideration cycle: a buyer can see an ad, not click, disappear for three weeks, and return through a branded search or a direct visit to start a trial. Under a generous view-through window that return gets credited to the original impression; under click-only credit it is invisible to the ad platform and reappears as organic or direct. The sales cycle routinely outlasts the attribution window, so the credit model decides a lot.

That means the view-through removal genuinely destroys some real signal for SaaS — the impression-assisted, delayed, branded-return path is a true part of how enterprise software gets bought, and click-only credit throws it away. But view-through also overcounted badly for SaaS, because a passive impression in a long cycle is a weak claim on a conversion that may have been driven by a dozen other touches. The honest position is that neither the old view-through regime nor the new click-only regime told the truth about SaaS contribution; they had opposite biases. The change is a reason to stop treating any single platform's credited conversions as the answer, not a reason to mourn view-through or celebrate its removal.

Re-baselining paid contribution without view-through credit

The first operational task is to treat the change dates as a measurement break, not a trend. A view-through window removal splits your history into two regimes that cannot be compared directly: conversions counted one way before 12 January, another way after. Plot them on the same axis without annotation and every trend line, every month-over-month, every year-over-year lies to you. Annotate the break in the dashboard, freeze the pre-change series as historical-only, and rebuild your trend from the new definition forward. This is the same discipline required whenever a platform changes its default attribution logic underneath a reporting series you have been trending.

The second task is to re-rank channels on the new basis before you move any budget. The channels that lose the most credit under click-only are precisely the impression-heavy ones, so a mechanical read of the post-change numbers will tell you to cut them — possibly correctly, possibly not, because some of that lost view-through credit was real assist and some was noise. Resist the reflex to reallocate on the first post-change week. Hold budget steady through at least one full cycle under the new definition, watch whether total pipeline actually moved or only the attributed share did, and only then decide. The platform changed what it counts; your job is to find out whether the business changed, which is a different and slower question.

Own the source of truth: first-party event-based attribution

The deeper lesson of 2026 is that platform-reported conversions are not a measurement you control. They are computed by the platform, in the platform's commercial interest, under definitions the platform can and does change without warning. Building your growth model on that number is building on rented ground. The alternative is to own a first-party, event-based record of the funnel: stamp the acquisition source on the user at signup, record the activation and the paid conversion against your own identity and timestamps, and keep those events in a store you control. When a platform redefines a window, your owned ledger does not move, because it never depended on the platform's definition in the first place.

This is the same cookieless, owned-instrumentation posture that survives consent decline and tracking-prevention, which we argue for in full-funnel cookieless attribution and cookieless attribution accuracy benchmarks. Platform-reported conversions still have a job — they are a fast, useful input for in-flight bidding and campaign optimization, where the platform's own model is what the algorithm is optimizing against. But they should be one biased input into your contribution picture, not the scoreboard. Cross-check them against a low-tech, model-free signal like self-reported attribution on your signup form, which is immune to every window change because it asks the customer directly.

What to run this sprint

Start with a reconciliation, not a reallocation. Pull your paid-channel conversions for the three months straddling each 2026 change date and put them next to your own first-party signups and activations for the same period. If the platform number fell but your owned signups held, you have proof the drop was definitional and your budget decisions should ignore it. If both fell together, you have a real demand question worth chasing. Either way you have replaced a scary platform chart with a decision you can defend, and you have found out whether your owned instrumentation is complete enough to answer the question at all.

Then make the posture permanent. Annotate every platform-definition change on the relevant dashboards so no future break gets read as performance, set the rule that strategic budget moves wait for owned-data confirmation rather than platform-reported swings, and close any gap the reconciliation exposed in your first-party event coverage. The ad networks will change their attribution rules again — they did it twice in three months in 2026 alone. The teams that are not whipsawed by it are the ones whose source of truth is their own event data, with the platform's credited conversions kept in their proper place as a useful, revisable, second opinion.