Marketing Attribution Window
A marketing attribution window is the period after a click or impression during which a resulting conversion is credited back to that marketing touchpoint.
Key takeaways
- The window is a lookback rule evaluated when the conversion fires, not at click time.
- Click and view windows are configured separately and can credit the same conversion.
- Match window length to the observed time from first touch to closed deal.
- Different conversion events, such as quiz start, lead and deal, need different lookbacks.
- Platforms running unequal windows cannot be compared until the settings are normalised.
In depth
An attribution window is a lookback rule applied when a conversion is recorded, not when the ad is served. The moment a conversion fires, the platform searches its log of that user's prior interactions and credits the conversion only if a qualifying touchpoint falls inside the configured period. Clicks and views carry separate windows, so a single conversion can be evaluated against two different lookback rules. Whichever touchpoint the window admits then feeds the attribution model that decides how credit is split.
Window length is the main dial. Lengthening it raises reported conversions per channel because more historical touchpoints become eligible, which flatters upper-funnel and awareness placements. Shortening it concentrates credit on interactions close to the purchase and favours retargeting and branded search. Sales-cycle length, purchase frequency and consideration depth should set the number: a window shorter than the typical time from first ad to signed deal will systematically under-report. The trade-off is noise, since long windows sweep in coincidental impressions that had no influence.
In practice, teams pick one window per conversion event and hold it constant across platforms so channel comparisons mean something. Pick the event carefully too: a quiz completion, a captured lead and a closed deal each sit at a different distance from the click, so each deserves its own lookback. A scorecard funnel makes this tractable because the quiz timestamp gives a reliable mid-funnel marker, letting you set a short window on quiz starts and a longer one on qualified leads.
Windows only describe what a tracker could observe. Cookie expiry, cross-device journeys, browser restrictions and offline conversations all truncate the record, so a generous setting cannot recover data that was never stored. Windows also say nothing about causation: a touchpoint inside the period may have been irrelevant, and one outside it may have started everything. Where several platforms each claim the same conversion under their own window, the sum of reported conversions exceeds the real total and needs reconciling against a single source.
Example in practice
How to measure it
Start by measuring the thing the window is supposed to fit: the distribution of days between first recorded touch and conversion. Plot it and read the percentile where the curve flattens; a window covering roughly ninety percent of that distribution is a reasonable rule of thumb. If most conversions cluster in the first two days but a long tail runs for weeks, the tail is where budget decisions go wrong.
Then check for over-claiming. Divide the total conversions reported by all platforms by the count in your own database for the same period; a ratio well above one means windows are overlapping and the same lead is being claimed twice. Re-run key reports at two or three window lengths and watch which channels change rank. Channels whose ranking is stable are safe to fund; those that flip depend on the setting.
Common mistakes
The most common error is leaving every platform on its default and then building a budget spreadsheet from those numbers. Ad networks ship with different click and view defaults, so the resulting table compares measurements that were never on the same scale. Set one window per conversion event, apply it everywhere the platform allows, and where a setting cannot be changed, flag the mismatch in the report instead of quietly summing the columns.
The second is changing the window mid-quarter and reading the jump as performance. Widening a window retroactively re-credits older touchpoints, so conversions appear to rise on a day when nothing about the campaigns changed. Treat any window change as a break in the series: annotate the date, keep the old setting running in a parallel view if the platform allows it, and compare like with like for at least one full sales cycle afterwards.
Frequently asked questions
What is a typical attribution window length?
Common defaults include a 7-day click and 1-day view window on paid social, and up to 30 or 90 days for click-based attribution on search and analytics tools. The right length depends on how long your sales cycle takes to convert.
Does a longer attribution window inflate conversions?
A longer window can credit more touchpoints and may appear to increase conversions, especially if channels overlap. The key is to keep windows consistent across channels so you compare them fairly rather than double-counting.
How should I set the window for a quiz funnel?
Match the window to your real time-to-conversion from quiz completion to qualified lead. If your sales cycle is several weeks, a one-day window will undercredit the top-of-funnel campaigns that actually source your best-fit leads.
What is a good default attribution window?
There is no universal default; the window should cover the typical gap between first touch and conversion for the specific event you track. A short lookback of a day or two suits impulse purchases, while considered B2B deals often need weeks. Measure your own time-to-conversion distribution first, then pick a length that captures the bulk of it.
What is the difference between a click window and a view window?
A click window credits a conversion when the person clicked the ad within the lookback period; a view window credits it when they merely saw the ad without clicking. View-through credit is far weaker evidence of influence, which is why view windows are usually set to a day or less while click windows run much longer.
Does changing the attribution window change past data?
Yes, in most platforms the change is retroactive: historical conversions are recalculated under the new lookback, so yesterday's report may not match today's for the same date range. Export the old figures before you change anything, annotate the switch date in your dashboards, and avoid comparing periods that straddle the change without noting it.
Why do ad platforms report more conversions than my CRM?
Because each platform applies its own window to its own touchpoints, and several can claim the same conversion independently. Your CRM counts each lead once. Treat platform numbers as channel-level directional signals and your own database as the total, then allocate that total across channels using one consistent window rather than adding the platform figures together.
Should the window differ for lead capture and closed revenue?
Yes. A lead capture sits close to the click, so a short lookback is usually enough to explain it. A closed deal can arrive months later, so the same short window would credit almost nothing to the campaign that started the journey. Configure a separate window per conversion event and report them as distinct measurements.
How does the attribution window interact with the attribution model?
The window decides which touchpoints are eligible; the model decides how credit is divided among them. A last-click model with a thirty-day window still gives everything to one interaction, just from a wider pool. Changing the window without changing the model shifts which single touchpoint wins, so read the two settings together when a channel's numbers move.