A lookback window is the maximum stretch of time before a conversion during which an earlier ad click or visit is still allowed to take credit for it. When someone converts today, the attribution system walks backwards through their touchpoints — but only as far as the window reaches. A click 20 days ago gets credit under a 30-day window and gets nothing under a 7-day one. The same concept also appears under the name attribution window, and in Google Ads as the conversion window. This entry covers the two lookback settings GA4 actually exposes, how the window interacts with your attribution model, and the mismatches it creates between GA4 and Google Ads numbers.
What a Lookback Window Is
Attribution answers “which touchpoint earns credit for this conversion?” The lookback window answers a quieter question first: “which touchpoints are even in the running?” It is a hard cutoff. Whatever your model — data-driven, last click, first touch — the model only sees touchpoints inside the window. Everything earlier is invisible, no matter how influential it was.
That makes the window a bigger lever than most people expect. Two teams with identical traffic and the same model but different windows will report different campaign performance, because they are not judging the same set of touchpoints.
The Two Lookback Settings in GA4
GA4 keeps the setting under Admin → Data display → Attribution settings, and splits it by the type of key event:
| Key event type | Options | Default |
|---|---|---|
Acquisition key events (first_open, first_visit) |
7 or 30 days | 30 days |
| All other key events | 30, 60 or 90 days | 90 days |
The split exists because the two event families measure different things. Acquisition events fire once, at the very start of a relationship, where only recent touchpoints plausibly matter. Purchase-style events sit at the end of consideration cycles that can run for months — hence the far longer default.
How the Window Changes Your Numbers
Shortening a window pushes credit toward the touchpoints closest to conversion; the early ads that opened the journey fall outside the cutoff and their measured contribution drops to zero. In practice:
- Long consideration products (B2B software, high-ticket purchases) lose upper-funnel credit fastest under short windows — the discovery click routinely happens 60+ days before the deal.
- Impulse products barely notice the setting; nearly all journeys fit inside even 30 days.
- Channel mix shifts. Channels that start journeys (paid social, display) look weaker under short windows; channels that end them (brand search, direct) look stronger. The traffic didn’t change — the accounting did.
Lookback Window vs Attribution Model
The two settings are frequently confused because both live in attribution. The distinction: the window decides eligibility, the model divides the prize. A 90-day window with last-click attribution still gives all credit to one touchpoint — it just could be a touchpoint up to 90 days old. A 30-day window with data-driven attribution spreads the credit — but only across touchpoints from the past 30 days. Change either one and reported channel performance moves; change both at once and you cannot tell which caused what.
GA4 vs Google Ads: Different Windows, Different Numbers
Google Ads maintains its own conversion windows, configured per conversion action, independently of GA4’s attribution settings. When the two systems disagree — a common finding in analytics audits — window mismatch is one of the standard causes, alongside attribution-model differences and the GCLID-based counting Ads uses. If a conversion happens 45 days after the click, a 90-day GA4 window counts it while a 30-day Ads window does not. Before declaring a tracking bug, compare the windows on both sides.
Changing the Setting: Forward Only
Lookback window changes in GA4 apply going forward — Google’s documentation states that changes “apply going forward and are reflected in all reports” from that point. Historical rows are not recomputed. Two habits keep this manageable:
- Annotate the change date wherever your team keeps analytics changelog notes; window changes are invisible in the data itself and mysterious a quarter later.
- Avoid comparing attribution-sensitive metrics across the change date, for the same reason methodology changes always poison comparisons.
One related nuance from Google’s attribution documentation: a conversion can be reattributed for up to 7 days after it happens, as late-arriving data settles. Yesterday’s campaign numbers are softer than last week’s.
A Worked Example: One Journey, Three Windows
A visitor’s path to a purchase on day 0, walking backwards:
- Day −75: clicks a display ad (first contact)
- Day −40: returns from an organic search
- Day −12: clicks a paid search ad
- Day −1: arrives direct and buys
Under a 90-day window, all four touchpoints are eligible; data-driven attribution can spread credit across display, organic and paid search. Under a 30-day window, the display click and the organic visit vanish — only paid search (day −12) and direct remain, and paid search suddenly looks like the journey starter. Under a 7-day window, everything except the direct visit is gone, and the conversion reports as close to unattributed as it gets. Same person, same journey, three different stories — the window decided which one you read.
Multiply that by thousands of journeys and the pattern from the channel-mix section becomes visible in your own reports: shrink the window, and the channels that open journeys quietly bleed credit to the channels that close them.
Where the Window Shows Up in Reports
The lookback window applies wherever GA4 attributes key events to sources:
- Advertising workspace — the attribution reports (model comparison, conversion paths) are computed inside the active windows. The conversion-paths view also shows path length in days, which is the fastest way to check how long your real journeys run before deciding a window is “too long.”
- Traffic acquisition dimensions — session- and user-scoped source dimensions follow their own rules, but event-scoped attribution for key events respects the window.
- Exports to Google Ads — what GA4 hands to Ads reflects GA4’s attribution; Ads then applies its own conversion windows to its own click-based counting, which is exactly why the two never match to the digit.
What the Lookback Window Does Not Control
Three neighbouring concepts are frequently mistaken for it:
- Audience membership duration. How long a user stays in a GA4 audience is its own setting on the audience, unrelated to attribution windows.
- Data retention. The 2/14-month retention setting controls how long event-level data survives for Explorations; the lookback window controls attribution eligibility. A 90-day window works fine with 14-month retention.
- Session timeout. When a session ends after inactivity is a collection setting; the lookback window operates across sessions, not inside them.
Choosing a Window Deliberately
The defaults (30 acquisition / 90 everything else) are sensible for most properties, and there is a real cost to shortening them casually: credit silently exits your reports. Shorten the window when your sales cycle is demonstrably short and you want reports to react faster to campaign changes. Keep it long when journeys are long — and check your actual journey length in GA4’s path and time-lag style explorations rather than guessing. The window should describe how your customers behave, not how fast you want reports to move.
Frequently Asked Questions
Is a lookback window the same as an attribution window?
Yes — “attribution window” is the generic industry term, “lookback window” is what GA4’s interface calls it, and Google Ads calls its equivalent the “conversion window.” All three name the same idea: how far back from a conversion the system searches for creditable touchpoints.
Does changing the lookback window rewrite my historical data?
No. GA4 applies the new window from the change onward; previously recorded attribution stands. This is different from switching the attribution model, which GA4 does apply to historical data in reports.
What lookback window should an online store use?
For typical e-commerce with days-to-weeks consideration, the defaults work: most purchase journeys fit comfortably inside 90 days, and the 30-day acquisition window captures how people first found you. Shorten only if your own time-to-purchase data says journeys are fast.
Does the lookback window affect audiences or data retention?
No to both. Audience membership duration is set on each audience, and the 2/14-month data-retention cap is a separate property setting. The lookback window governs one thing: how far back attribution may search for creditable touchpoints.
Related Terms
- Attribution — the discipline the window is part of
- Attribution Model — divides credit among eligible touchpoints
- First Touch / Last Touch — the model endpoints the window constrains
- Conversion — the event that triggers the backwards search
- GCLID — how Google Ads ties conversions to clicks on its side
- Campaign — what ultimately receives the credited conversions