How a view-through conversion actually gets recorded
An ad platform fires a tracking pixel the moment an ad is served to a person. Not just when it gets clicked. That impression pixel logs a timestamp against whatever identity the platform has for that browser, device, or logged-in account. If the same identity later completes a conversion event, whether on the same device or one the platform believes belongs to the same person, the platform counts it as a view-through conversion inside a lookback window it defines, commonly one day by default and adjustable up to seven or more depending on the platform and campaign objective.
The mechanism is the important part: there was no click, no landing page visit, and often no click ID to tie the two events together in the way a normal referrer would. The match runs entirely on the ad platform's own identity graph, built from logins, device IDs, and hashed contact data it already holds, none of which your analytics tool ever sees.
Impression attribution is the same idea under a different name
You will also see this called impression attribution, or sometimes post-impression attribution, in ad platform documentation and reporting tools. All three terms describe the same mechanism: crediting a conversion to something a person saw rather than something they clicked. The naming varies by platform and by whoever is writing the help doc, but the underlying pixel-on-serve, match-on-identity, count-within-window logic is identical. If a report or a vendor uses one term and your team uses another, confirm out loud that everyone means the same mechanism before comparing numbers.
View-through vs click-through attribution
| View-through attribution | Click-through attribution | |
|---|---|---|
| Trigger event | Ad impression served, no click required | Ad clicked, then landing page reached |
| Matching method | Platform's own device or account identity graph | Click ID or UTM parameter carried through to the landing page |
| Typical lookback window | 1 day by default, sometimes configurable to 7 | Usually longer, often 28 days on major ad platforms |
| Where it is reported | Only inside the ad platform that served the impression | Both the ad platform and any independent analytics tool with matching UTMs |
| Independently verifiable | No, the platform is the only party that saw the impression | Partially, if UTM parameters and referrers survive the click intact |
The lookback window is a dial the advertiser can turn
Unlike a click-through window, which is largely fixed by how long a cookie or click ID survives, a view-through lookback window is a setting inside the ad account (often buried a few menus deep, which is part of why so few advertisers ever check it). Widening it from one day to seven does not create a single additional buyer. It just expands how many existing conversions the platform gets to claim credit for. Two campaigns with identical real-world performance can show very different view-through totals purely because one account has a wider window configured, which is worth checking before comparing view-through rates across campaigns or across accounts run by different people.
Why view-through inflates ad platform reporting
View-through conversions get added to click-through conversions inside most ad platform dashboards, often without a clear label distinguishing the two in the default view. That combined total is what gets compared to spend to produce a return-on-ad-spend figure, and because view-through counts anyone who was already going to buy and happened to have seen the ad at some point in the lookback window, it tends to skew that number upward. A shopper who searched your brand name directly, converted, and had also seen a retargeting ad in the prior 24 hours gets counted as a view-through win for that ad, whether or not the ad had anything to do with the decision.
Why you cannot reconcile it against your own analytics
Most explanations gloss over this part. Your website analytics tool, and any click-tracking or link-in-bio tool sitting in front of it, only ever records a request that actually reached its own server. An ad impression that happened inside Instagram's feed or a video player never touches your infrastructure at all, so there is no event for your analytics tool to log even in principle. And no amount of better tagging changes that. This is not a mistake anyone made, it is a structural limit: the impression and the eventual site visit are two separate systems, and only one of them, the ad platform, was present for both. A tool like Raydar can tell you with certainty which click landed on a bio link and where it came from, because that click hit its redirect. It has no way to see, and makes no claim to see, an ad impression that never produced a click in the first place, which is exactly why view-through numbers live and die inside the platform that generated them.
When view-through attribution is still worth looking at
None of that makes it useless. It just means the number needs a different kind of trust than a click-through conversion gets. For awareness and video campaigns, where the goal is exposure rather than an immediate click, a rising view-through conversion rate over several weeks can be a real, if imprecise, signal that the campaign is building intent that later shows up elsewhere, in direct traffic or branded search. The mistake is treating it as equivalent to a click-through conversion in a blended ROAS number, when the two are measuring fundamentally different behaviors with fundamentally different confidence levels.
The honest way to use it
Report view-through and click-through conversions separately rather than blending them, and treat a big gap between an ad platform's total conversions and what your multi-touch attribution model or your site analytics shows as expected, not suspicious. It is one of the more mechanical reasons behind why your analytics numbers never match across tools, and no amount of better tagging closes that particular gap, because the two systems were never both present at the same moment in the customer's journey.