View-through conversions and Click-through conversions at a glance
Every ad platform that tracks conversions has to decide which ad interactions may claim a later conversion and for how long. The Lookback WindowThe time after an ad interaction in which a conversion can still be credited to it. sets the time limit, and most systems keep two: a click-through window for conversions that follow a click, and a view-through window for conversions by people who saw an ad without clicking it. Meta's help documentation on attribution settings describes clicks taking precedence over views when both occurred, so view credit applies only where no qualifying click exists; Google Ads documents conversion tracking per action, with view-through conversions reported for display and video separately from the main column.
A click-through conversion has the stronger claim. The person took a deliberate action on the ad, arrived on the site through it, and converted within the window. That does not prove the ad caused the purchase, since people click ads for products they were already going to buy, but the evidence of influence is direct. A view-through conversion has a weaker claim: the person was served an ad, which the platform records as an impression, and converted later without clicking. Many such people would have converted anyway, which is why the view-through conversion glossary entry treats the count as a sequence that happened rather than an effect that was caused.
How each platform sets the windows, and what changing them does to delivery as well as reporting, is covered in the Meta attribution settings guide. This page is about the weight each kind of conversion deserves once it has been counted.
Compare table
View-through conversions vs Click-through conversions
| Row label | View-through conversions | Click-through conversions |
|---|---|---|
| What triggers credit | An impression without a click, followed by a conversion inside the view window | A click on the ad, followed by a conversion inside the click window |
| Typical window | Short, because the link between an unclicked impression and a sale is weak | Longer, because a click is a deliberate step towards the conversion |
| Which formats depend on it | Video, display and other formats that are watched or seen far more often than clicked | Search, shopping and any format built to send people to a site |
| Visibility in analytics | None: an unclicked impression leaves no session for analytics to record | Visible: the click starts a session that analytics can attribute |
| Risk of over-counting | High: it credits people already close to converting, such as remarketing audiences | Lower, though still present for warm audiences and branded searches |
| Deduplication across platforms | Poor: several platforms can each claim the same buyer by impression | Better, since a click usually leaves one trail, though platforms still overlap |
| Use in bidding | Optional and risky: including it teaches delivery to value unclicked exposure | The standard input for conversion-based bid strategies |
| How to validate | A holdout or lift test comparing exposed and unexposed groups | A holdout for causality; analytics for a cross-check on the count |
The rows that matter most are visibility in analytics and validation. Click-through conversions can be cross-checked against a second system; view-through conversions can only be checked by an experiment.
When to choose View-through conversions
Count view-through conversions when the campaign's job is to be seen rather than clicked. Video, display, connected television and much of social advertising build memory and intent without asking for an immediate response, and a channel judged on clicks alone will be undervalued, sometimes to the point of being cut when it was working. In those cases the view-through figure is the platform's attempt to account for influence that leaves no click, and ignoring it entirely is as misleading as trusting it entirely.
The condition is that view-through conversions are reported separately and priced with evidence. A lift study or a geo test that withholds the ads from a comparable group shows how much of the view-through count is genuine, and the ratio between the test result and the platform count can then be used to discount the figure between tests. IncrementalityConversions caused by marketing. testing is the only method that does this; the comparison of marketing mix modelling and multi-touch attribution covers how a model can extend a handful of tests across a whole budget.
Without that evidence, view-through conversions are a hypothesis, not a result. They belong in a report as context for an awareness campaign, clearly labelled, never added to click-through conversions to produce a single total.
When to choose Click-through conversions
Use click-through conversions as the default for reporting and bidding. They rest on a deliberate action, they are visible to analytics as well as to the platform, and they are less exposed to the problem of crediting people who were going to convert regardless. For search, shopping and most campaigns built to drive visits, they are the conversions that matter and the ones bid strategies should optimise towards.
Click-through conversions are also the fairer basis for comparing channels. Two platforms with different view windows will report very different totals for the same spend, but on a click-only basis with matching windows the comparison is much closer to like with like. The comparison of platform and analytics attribution explains what each kind of system observes, and why platform totals still run ahead of analytics even on clicks.
Click-through is not the same as causal. Remarketing and branded search attract clicks from people who were already buying, so a click-only figure still overstates what those campaigns caused. It is the better default, not the final answer.
A bicycle retailer splits its conversion lines
Pellingham Cycles, an online bicycle retailer, runs a video campaign and a search campaign. Its weekly report adds the two platforms' conversions together and shows 480 purchases, against 300 orders in the shop's own records. Splitting the platform figures shows that the video campaign reported 190 conversions, of which 170 were view-through and 20 click-through, and search reported 290, all click-through. The team rebuilds the report with click-through conversions as the main line and view-through on a separate, labelled line. It then runs a geo test, withholding the video ads from a matched set of regions for six weeks, and finds the video campaign added about 40 orders over the period rather than the 1,020 view-through conversions the platform reported for the same weeks. It keeps the video budget, now judged at that rate, and removes view-through conversions from the column used for bidding.
Decision table
| Situation | Use | Why |
|---|---|---|
| Search or shopping campaign | Click-through conversions | The format exists to earn clicks and the click is the evidence |
| Video or display campaign built for awareness | View-through on a separate line, validated by a test | Few people click, so clicks alone undervalue the channel |
| Remarketing campaign | Click-through, with a holdout before trusting either | The audience is already close to buying, so view credit flatters most |
| Choosing the conversion column for bidding | Click-through, unless a lift result supports views | Including views teaches delivery to pay for unclicked exposure |
| Reconciling platform totals with analytics | Remove view-through first | Analytics cannot see unclicked impressions |
| Comparing two platforms | Click-through on matching windows | Different view windows make totals incomparable |
Setting windows and reporting rules
Write the rules down once and apply them to every platform. Decide the click window and the view window for each conversion action, matched to how long real customers take to decide; decide whether view-through conversions may enter the column used for bidding; and decide how the report will show them. The usual answer is a main line of click-through conversions on matched windows, a separate line of view-through conversions labelled with its window, and a note beside each channel that has been tested giving the ratio between the test result and the platform count. In Google Ads, where windows and the attribution model are set per conversion action, the guide to conversion windows and attribution settings sets each window from the account's own conversion lag.
Keep the windows stable. A change in either window changes the count without a single customer behaving differently, so any change belongs in a change log and breaks the trend line at that date. When the total of platform conversions runs ahead of the orders the business recorded, the diagnosis of channel conversions exceeding sales works through the causes, of which view-through credit is usually the first to check. For LinkedIn specifically, the diagnosis of Campaign Manager conversions running higher than the CRM starts by splitting the column into click and view credit.
Common mistakes
- Adding view-through and click-through conversions together into one total, as if they were the same evidence.
- Leaving view windows at platform defaults without deciding whether they suit the product's buying cycle.
- Including view-through conversions in the bidding column by default, which steers delivery towards people already likely to convert.
- Comparing platforms with different view windows and concluding one performs better.
- Dropping view-through entirely for awareness formats, which undervalues channels whose job is to be seen.
- Treating click-through conversions as causal, when remarketing and branded search clicks come largely from people already buying.
- Expecting analytics to confirm view-through counts, when it cannot see an impression that was never clicked.
