Google made both of the conversion lookback windows in Google Analytics fully configurable, a change documented in the "What's new in Google Analytics" Help Center release notes on 11 August 2026 with no blog post and no press release attached to it. Engaged-view conversions, which were locked at a fixed 3 days for every property, can now be set to any whole number from 1 to 30 days, and click-through conversions, previously limited to six presets of 1, 7, 14, 30, 60 and 90 days, can now be set to any whole number from 1 to 90 days.
PPC Land reported the change on 14 August in its write-up of the release note, placing it one day after the 10 August launch of the campaign data import validation report and six days before the 17 August Google Ads bidding target enforcement change. Google's stated reason in the release notes is that the update "allows Google Analytics to align conversion attribution lookback windows more closely to your unique business cycles."
What changed in Google Analytics on 11 August 2026
Two separate settings moved at the same time. The first is the engaged-view conversion window. An engaged-view conversion is credited when someone watches a video ad without clicking it and then converts within the window. Until 11 August that window was a fixed 3 days in every property, with no control exposed to the account owner. It is now an open range: any integer from 1 to 30 days.
The second is the click-through conversion window, which decides how long after a click a conversion can still be attributed to that click. Google Analytics previously offered a dropdown of six values: 1, 7, 14, 30, 60 or 90 days. That dropdown is gone as a constraint. Any integer from 1 to 90 days is now valid, so 45 days, 21 days or 5 days are all selectable where before you had to round to the nearest preset.
Nothing in the release notes says the defaults changed. The change is about the range of values you are allowed to pick, not about Google picking a different value on your behalf. If you never open the setting, the source gives no indication that anything moves.
What a conversion lookback window actually does to your reported numbers
A lookback window is not a reporting filter that hides rows. It decides whether an interaction gets credit at all. When a conversion happens, the attribution model looks backwards from the conversion for eligible interactions. Anything older than the window is not a candidate. A click that happened 40 days before the purchase is invisible to a 30 day window and visible to a 60 day window, and the conversion either lands on that campaign or lands somewhere else, usually on a later touchpoint or on direct traffic.
Two consequences follow. The first is that conversion counts attributed to paid channels move when the window moves, without a single change in real customer behaviour. Widen the window and slow-converting campaigns look better because purchases that were previously orphaned now find an earlier ad interaction to attach to. Narrow it and the same campaigns look worse. Nothing about the business changed. The measurement rule did.
The second consequence is the one that costs money. Smart Bidding strategies are fed by conversion data. Target CPA and Target ROAS learn from the conversions the system can see and the interactions those conversions are attached to. Change the lookback window and you change the training signal, both its volume and its distribution across campaigns and keywords. That is why a lookback window edit is not an analytics preference. It is an input to your bidding, and it belongs in the same mental category as changing a conversion action's inclusion setting or its counting method. Anyone running Google Ads campaigns against Analytics conversions should treat it that way.
Why the fixed 3 day engaged-view floor was a mismatch for long consideration
The old 3 day engaged-view window assumed that if a video view was going to influence a purchase, the purchase would happen inside 72 hours. For an impulse product that is a defensible assumption. For a category where people research for weeks, it quietly wrote off most of the effect. A homebuyer, a B2B software buyer, an insurance shopper or someone comparing a large appliance rarely converts within three days of a video view, so the view registered as having produced nothing.
The practical effect was that video ad spend looked worse than click-driven spend in properties with long consideration cycles, and it looked worse structurally rather than because it performed worse. That is the mismatch Google's release note language points at when it talks about business cycles. Retail and ecommerce operations with short repurchase cycles were closer to the fixed 3 day assumption than considered-purchase categories ever were.
The honest caveat here is that a wider window is not automatically a truer window. Extending the engaged-view window to 30 days does not prove that a video view 29 days before a purchase caused the purchase. It only means the model is now allowed to credit it. Correlation gets easier to collect as the window grows. Causation does not.
The two windows side by side, before and after 11 August 2026
The table below sets out exactly what was available before the release note and what is available now, based on the values Google published.
| Setting | Before 11 August 2026 | After 11 August 2026 |
|---|---|---|
| Engaged-view conversion window | Fixed at 3 days for every property, not adjustable | Any whole number from 1 to 30 days |
| Click-through conversion window | Six preset choices: 1, 7, 14, 30, 60 or 90 days | Any whole number from 1 to 90 days |
| How values are chosen | A fixed value and a short preset list | A free integer inside the allowed range |
Where to find and change the conversion lookback window
Google's release note gives the path: Advertising, then Conversion management, then the more options icon, then Settings. The same controls are also reachable from the linked Google Ads conversion management interface, so accounts with an Analytics and Ads link can arrive at the setting from either side. That second entry point matters more than it sounds, because it means a change made by whoever owns the Ads account shows up in the Analytics numbers the SEO or analytics team is reading, and the two teams may not be talking to each other that week.
Before you touch it, write down the current value. The release notes do not promise a change log you can consult later, and a setting that was previously fixed at 3 days has no history for anyone to remember.
The retroactivity question the release notes do not answer
This is the most important practical caveat in the whole story, and it is an open question rather than a finding. The release notes do not state whether adjusting a window applies retroactively to conversions that were already recorded, or only to conversions recorded from that point forward. Both behaviours exist in analytics products, and the difference is not cosmetic.
If the change is prospective only, your history stays as it was and you get a visible seam in the data at the date of the change: numbers before the seam were produced under one rule, numbers after it under another. If the change is retroactive, previously reported figures can move under you. A month you already reported to a client or a board could recalculate. Nobody sends you a note when that happens, and a widened window would generally shift reported conversions upward across the affected period, which is the direction least likely to prompt anyone to investigate.
Because the source does not resolve this, treat it as unknown and protect yourself procedurally rather than guessing:
- Annotate the exact date and time you change either window, along with the old value and the new value, in whatever annotation or change log your team actually reads.
- Before changing anything, export a stable comparison period at the current setting: conversions by channel, by campaign, and total, for a period long enough to be more than noise.
- After the change, re-export the same historical period. If the figures for those past dates have moved, the change behaved retroactively in your property, and you now know something the release notes did not tell you.
- Keep both exports. They are the only evidence you will have if a reported number is questioned later.
This is a five minute procedure that turns an unanswered question into an answered one for your own property, without waiting for documentation.
Do not move a lookback window during an active Smart Bidding learning period
If a campaign is in a learning period after a bid strategy change, a target change or a significant budget change, the system is actively recalibrating against the conversion signal it can see. Changing the lookback window in the middle of that alters the signal itself, and you lose the ability to say what caused what. Performance moves, and there are now two candidate explanations sitting on top of each other.
The safer sequence is to wait for the learning period to settle, record a baseline, then change the window, then leave the bid strategy alone long enough to read the result. If you must change both, change them far enough apart that each has its own readable period, and write down both dates. The cost of skipping this is not just an unclear report, it is a target that gets adjusted for the wrong reason.
What to check in your own property this week
- Open Advertising, Conversion management, the more options icon, then Settings, and note the current click-through window value. Do not change it yet.
- Note the engaged-view window value. If your property has always run at the fixed 3 days, that is the value you are comparing everything against.
- Ask whether your actual purchase cycle is shorter or longer than the current window. Use your own data on time from first interaction to purchase, not an industry rule of thumb.
- Check who else has access through the linked Google Ads conversion management interface, so a change does not arrive from a direction you were not watching.
- Check whether any campaign is currently in a learning period. If so, postpone.
- If you do change it, annotate the date and keep the before and after exports described above.
What the source did not say
Several things are genuinely unresolved, and it is worth being precise about them rather than filling the gap with assumption. The release notes do not say whether the change is retroactive. They do not give a recommended value or any guidance on how to choose one. They do not describe a rollout schedule or say that availability differs by region or account type. There was no blog post or press release, so there is no accompanying Google commentary beyond the single release note sentence about business cycles. PPC Land's report is dated 14 August and covers the release note itself.
Anything beyond that, including the specific effect on any particular account, is analysis rather than reporting, and this article has labelled it as such.
What this means for Thai marketers
The source is a product release note and a trade report on it, neither of which says anything specific about Thailand. What follows is reasoning about the implication, not a claim from the source.
Two features of the Thai market make the engaged-view change worth attention. The first is how heavily video sits in the upper funnel here, and how often the measurable outcome arrives days or weeks later through a completely different surface. Under a fixed 3 day rule, a large part of that delayed effect simply was not counted. The window can now be set to match the real gap.
The second is the split-channel reality of Thai purchase journeys. A viewer sees a video, then researches, then converts through a chat conversation, a marketplace or an offline visit. Not all of that lands in Google Analytics 4 in the first place, and a longer lookback window does not fix an untracked step. Widening a window when the conversion itself is not being recorded correctly widens nothing. The measurement problem worth solving first is usually the missing conversion, not the length of the window that would have caught it.
The third point is procedural and applies to every agency and in-house team. Reports go out weekly and monthly in this market. If a window change can move reported history and nobody wrote down the date, the conversation with a client about why last month moved becomes unwinnable. Annotate first, change second.
Frequently asked questions
Is this live in Thailand?
The source does not say anything about regional availability. Google documented the change in the "What's new in Google Analytics" release notes on 11 August 2026 without describing a staged rollout or any regional restriction, and PPC Land's 14 August report does not mention one either. The only reliable check is to open Advertising, Conversion management, the more options icon, then Settings in your own property and see whether the fields accept a free value.
Do I have to do anything?
No. This is a widening of the values you are allowed to choose, and nothing in the release notes says defaults change on their own. Doing nothing leaves the property where it was. The reason to look is to find out whether the current window matches your actual purchase cycle, which is a question worth answering whether or not the setting had changed.
Will changing the window rewrite conversions I have already reported?
The release notes do not state whether adjusting a window applies retroactively to already-recorded conversions or only to future ones, so this is unknown. Treat it as possible: export a historical period before you change anything, re-export the same period afterwards, and compare. That tells you the answer for your property even though the documentation does not.
Should I just set both windows to the maximum?
No, and the maximum is not a safe default. A longer window credits older interactions, which inflates the apparent contribution of upper-funnel activity and feeds a different signal to any bid strategy reading those conversions. Pick a window from your own measured time to purchase, and change one window at a time so you can read the effect.
Does this affect Google Ads reporting as well as Google Analytics?
The settings are reachable from the linked Google Ads conversion management interface as well as from Analytics, and conversions fed into bidding come from the same attribution rules, so a change here is not contained inside Analytics reporting. The release notes do not spell out every downstream effect, which is another reason to record the date of any change.
Before you touch the setting
The lazy version of this whole story is one sentence: you now have a dial that used to be a fixed number, and turning it changes the numbers everyone reports on. If you want a second pair of eyes on whether your current window matches your real purchase cycle, or on whether your conversions are being recorded properly in the first place, that is a conversation worth having before the dial moves rather than after.
