An attribution window is the time limit a marketing platform uses to decide whether a conversion counts as credit to a past ad interaction. It is a measurement setting, not a fact about your customers. Each platform picks its own default window, which is why reported conversions rarely match between Meta, Google Ads, GA4, and your CRM.
Key Takeaways
- An attribution window is a configurable timer, not a truth. It decides which past ad touch a platform will credit for a conversion.
- Different platforms ship different defaults, so the same campaign can report wildly different conversion counts in the same period.
- Click-through and view-through windows measure different behaviors, and combining them can double-count soft impressions as real demand.
- Set your window from your measured sales cycle, not from a platform's convenience default.
- Reconcile platform numbers against warehouse or CRM data on a fixed, deduplicated key to see what actually closed.
What Is an Attribution Window?
An attribution window is the lookback period a platform applies after an ad interaction before it stops counting that interaction as a possible cause of a conversion. If a user clicks your ad and buys nine days later, whether that purchase shows up as a conversion depends entirely on whether the platform's window still covers day nine.
The window is a setting baked into how the platform's reporting model assigns credit. It does not describe how humans actually decide to buy. A long B2B sales cycle does not change because Meta's default window is seven days. The window simply stops counting touches that fall outside it, which quietly drops real conversions from your reported totals or, when set too long, pulls in conversions that had nothing to do with the ad.
Because every platform owns its own reporting database, every platform also owns its own default window. That is the root cause of the familiar frustration where Meta says 400 conversions, Google Ads says 300, GA4 says 220, and the CRM shows 180 actual closed deals. None of them are lying. They are measuring with different timers.
How Do Click-Through and View-Through Windows Differ?
Click-through attribution counts a conversion if the user clicked an ad and then converted within the click window. View-through attribution counts a conversion if the user only saw the ad (no click) and then converted within the view window. The two answer different questions: click-through captures active intent, while view-through attempts to capture assisted memory or brand lift.
The trap is that view-through conversions are easy to inflate. A user who was already going to buy sees a retargeting banner, does not click, and converts anyway. Under a generous view-through window, the impression gets full or partial credit for a sale it did not cause. This is why many performance teams either disable view-through entirely or keep it on a tight window like one day while letting click windows run longer.
When you compare platforms, always separate the two numbers. A platform reporting "7-day click, 1-day view" is telling you it credits clicks for seven days and passive views for one day. The "7-day click 1-day view" label is one of the most common configurations you will see, and the asymmetry is intentional: clicks are stronger signals, so they earn a longer window.
What Are the Default Attribution Windows on Major Ad Platforms?
Defaults vary, and they change over time, so always confirm in the platform UI before you rely on them. The table below describes the general shape of each platform's window model. Where a specific default number is uncertain, the description is left generic rather than guessed.
| Platform | Window type | Typical default behavior | Configurable? |
|---|---|---|---|
| Google Ads | Click primarily; view-through separate | Conversion window settable per action, often a longer click window by default with a separate view-through window | Yes, per conversion action |
| Meta | Click and view combined | Commonly ships a "7-day click, 1-day view" style default that can be widened or narrowed | Yes, within preset options |
| Click and view combined | Supports configurable click and view windows; defaults tend to be shorter on view | Yes, within preset options | |
| TikTok | Click and view combined | Offers selectable click and view windows; view windows are typically shorter than click | Yes, within preset options |
| GA4 | Lookback window for attribution | Uses a lookback window that determines how far back acquisition credit extends; adjustable in reporting | Yes, via lookback setting |
The practical point is not the exact numbers but the mechanism: each platform lets you set a window, ships a default that suits its own reporting incentives, and reports only what falls inside its window. The GA4 lookback window in particular is easy to forget because it sits in a different reporting layer than your ad accounts, yet it reshapes how credit flows to channels.
How Do You Choose the Right Attribution Window for Your Business?
Stop copying platform defaults. Choose a window from your actual measured time-to-conversion. Here is a step-by-step process you can run in your warehouse or even a spreadsheet.
- Pull closed conversions from your CRM with a timestamp for the conversion and a timestamp for the first and last paid touch, joined on a stable user or account key.
- Compute the elapsed time between the paid touch and the conversion for every deal that you are confident was influenced by paid media.
- Plot the distribution. Find the percentile (often the 80th or 90th) where most genuine conversions have already occurred.
- Set your click window near that percentile. If 90 percent of real conversions happen within 14 days, a 7-day window is silently discarding the back third of your demand.
- Keep the view-through window short, typically one day, unless you have a measured brand-lift reason to extend it.
- Document the chosen window per platform and review it quarterly as your sales cycle changes.
For longer B2B cycles, this process often reveals that a 7-day window is reporting only a fraction of true influenced pipeline. That is not a performance problem to fix with more bids; it is a measurement problem to fix with a longer window and with modeling that reaches beyond the ad platform, which is exactly the gap covered in our marketing attribution models guide.
Why Do Platform Conversion Numbers Never Match Your CRM?
There are several independent reasons, and they stack. First, each platform uses its own window, so the same conversion may be in-date for one and out-of-date for another. Second, platforms count on their own identity graph: Meta knows Meta users, Google knows Google cookies, and neither sees the anonymous research session on a different device that later became a CRM deal.
Third, deduplication differs. A single user who clicks three platforms before buying may be counted once in your CRM but three times across ad platforms if you sum them naively. Fourth, post-conversion events like returns, refunds, or disqualified leads rarely flow back into ad platforms automatically, so the platform keeps the conversion while your CRM later marks it invalid.
The reconciliation fix is structural, not a spreadsheet hack. Send server-side conversions with a stable order or lead ID, and build a warehouse table that maps each platform-reported conversion to the CRM record on that ID. Then you can report "platform-assisted" versus "truly closed" without guessing. Our offline conversion tracking guide walks through the server-side setup that makes this mapping possible.
How Should You Audit and Change an Attribution Window Safely?
Changing a window rewrites your historical reporting, so treat it like a config change with a before-and-after. Start by exporting current reported conversions per platform for a fixed date range under the existing window. Note the totals. Then change the window and export the same range again.
Expect the numbers to move, and resist the urge to interpret a longer window as "more performance." A longer window usually reveals conversions that were always there but previously fell outside the timer. Communicate this to stakeholders before you switch, or a reporting jump will look like a campaign win or loss that did not happen.
For bidding, be careful: platforms like Google Ads that optimize toward a conversion action will retrain their models when you change the window. A longer window feeds them more signal but also more lag, which can shift pacing. Make one change at a time, watch cost-per-acquisition and volume for two to three weeks, and only then judge the new setting. If your goal is efficient spend, pair window changes with a clear target ROAS definition so the optimizer has a consistent objective.
See also: cross-device attribution shows why aligned lookback windows matter when the same user converts on a second device.
Frequently Asked Questions
What Is an Attribution Window in Plain Terms?
An attribution window is the time limit a platform uses to decide if a past ad click or view still counts as the cause of a conversion. If the conversion happens after the window closes, the platform does not credit the ad, even if the ad genuinely helped.
Should I Use Click-Through or View-Through Attribution?
Use click-through as your primary signal because it reflects active intent. Keep view-through on a short window like one day, or turn it off, to avoid crediting ads for sales that would have happened anyway. The right mix depends on whether you are measuring direct response or brand assist.
Why Does My Google Ads Conversion Window Differ from Meta?
Each platform sets its own default window and lets you configure it independently, so the same conversion can be in-window for one and out-of-window for the other. Google Ads and Meta also model identity differently, which adds a second layer of mismatch on top of the window difference.
How Long Should My Attribution Window Be for a Long Sales Cycle?
Measure your actual time-to-conversion from CRM data and set the click window near the percentile where most real deals close, often 14 to 30 days or longer for B2B. A 7-day default will systematically undercount demand in any cycle that runs longer than a week.