Attribution time basis is the reporting setting that decides which date a conversion is filed under: the date of the ad impression or click, or the date the conversion happened. It is not the attribution window, which decides whether a conversion counts at all; attribution time basis only changes the date a counted conversion lands on.
Choose the wrong basis and your spend and conversions sit on mismatched dates, so cost-per-conversion for a recent week looks worse than it will once delayed conversions finish reporting. Get it right and you can compare periods honestly, reconcile ChatGPT Ads numbers against your analytics, and avoid cutting a campaign that is working.
In conversion reporting, attribution time basis is the rule that assigns each attributed conversion to a calendar date, either the moment of the ad interaction or the moment the person completed the action. The conversion count does not change. Only the date it appears on moves, shifting how a week reads.
Two bases are common. Impression or click time files the conversion on the date of the ad event that won it, so a purchase made days later is backdated to the click. Conversion time files it on the date the purchase occurred. Delayed conversions behave differently under each: impression-time totals for a recent period keep rising as late conversions attribute backward, while conversion-time totals stay put once the day closes.
Attribution time basis sits alongside the attribution model and the attribution window, and the three are easy to confuse. The model decides which touchpoint earns credit, the window decides how long after an ad a conversion still counts, and the basis decides the date. In ChatGPT Ads, picking a basis and holding it steady is what makes your reporting comparable with Google Ads, Meta, and your own analytics.
Resources: See how to structure ChatGPT Ads campaigns and read their reporting correctly.
Every conversion your campaign earns carries two timestamps: when the ad was seen or clicked, and when the person acted. The basis decides which of those two dates the conversion is reported on, and the platform applies the rule in order.
Record the interaction. When someone sees or clicks your ChatGPT Ad, the platform logs that event with its date and time.
Capture the conversion. When that person later completes the tracked action, the conversion pixel or API fires and records its own date and time.
Attribute the conversion. The attribution model and window link the conversion back to the qualifying ad interaction, confirming it counts.
Apply the basis. The reporting setting files the counted conversion against either the interaction date or the conversion date.
Populate the report. Your dashboard places the conversion in the chosen date's row, so totals for recent days keep updating as late conversions resolve.
The report tells you how many attributed conversions landed on each date under the basis you picked. It does not tell you your true number of sales, and it does not change how the campaign bids.
The basis you report on drives the budget decisions you make every week. If spend and conversions sit on different date logic, every efficiency number you calculate for a recent period is distorted, and you may move money toward or away from campaigns for the wrong reason. For ChatGPT Ads, where the channel is new and under scrutiny, a distorted early read can get a working campaign killed before it matures.
Honest period comparisons: A fixed basis lets you compare this month with last month without the date logic silently changing underneath you.
Clean reconciliation: Matching your basis to how your analytics and finance teams count conversions keeps your team debating strategy instead of debating whose number is right.
Protection against premature cuts: The failure mode is judging a recent week on impression-time data that is still filling in, seeing an inflated cost-per-conversion, and pausing a campaign that would have looked fine a week later.
SEO managers use attribution time basis to line up ChatGPT Ads conversion dates with the organic conversion dates in their analytics before building a blended channel report.
Growth marketers use attribution time basis to decide whether a recent week's cost-per-conversion is final or still filling in before they shift budget.
Demand gen leads use attribution time basis to match ad-reported conversions to the dates their CRM and finance team recognize revenue.
Impression time files each conversion on the date of the ad impression or click that earned it, which means totals for any recent period keep climbing for days as delayed conversions are attributed backward to their original interaction date.
Conversion time files each conversion on the date the action actually happened, so a day's total stops changing once that day closes, but your spend is recorded on the interaction date and never shares the same date basis as these conversions.
Conversion lag is the elapsed time between the ad interaction and the conversion it produced, and it is the reason the two bases diverge, with longer lags and lower-funnel events such as purchases showing the biggest reporting shift when you change the basis.
Align ChatGPT Ads conversion dates with Google Analytics 4 and your CRM for a single, defensible timeline.
Read recent-week performance correctly instead of reacting to impression-time totals that are still filling in.
Compare month-over-month efficiency on stable date logic.
Reconcile ad-reported revenue with what finance books.
Explain reporting gaps between ChatGPT Ads and Meta without guessing.
Defend budget decisions in reviews by showing spend and conversions on matched dates.
Pick one basis and document it so every report and every teammate uses the same date logic.
Match your basis to your primary system of record, usually conversion time if finance recognizes revenue on the event date, so numbers reconcile.
Let recent periods settle before judging them, because impression-time totals climb for days as delayed conversions attribute backward.
Note the basis directly on every dashboard and export, so no one compares two reports built on different logic.
Re-pull historical data after a basis change instead of splicing old and new reports together.
Keep attribution model, window, and basis labeled separately, because changing one explains a shift the others do not.
Avoid the common mistake of comparing a fresh impression-time week against a fully settled one and reading the difference as a performance drop. Competent marketers do this constantly, then cut spend on a campaign whose recent numbers simply have not finished reporting.
AirOps: tracks how your brand shows up across AI answer engines so you can connect ChatGPT Ads reporting decisions to the organic AI visibility those same prompts drive.
Google Analytics 4: records conversions on the event date in the user's time zone, giving you a conversion-time reference to reconcile your ad platform basis against.
Google Ads: exposes the same choice directly as its "Conversions" versus "Conversions by conversion time" columns, a useful model for how the two bases behave.
Find the setting. This week, open your ChatGPT Ads reporting controls and locate the attribution time basis option so you know which basis your current numbers use.
Pick your system of record. Decide which platform your team treats as the truth for conversions, usually your analytics or CRM, since that determines the basis you should match.
Set a single basis. Choose impression time or conversion time to align with that system of record, and apply it as the default across every report. Switching later is possible, but it forces another re-baseline, so decide deliberately now.
Document and label it. Write the chosen basis into your reporting template and onto each dashboard, so anyone reading a report knows the date logic behind it.
Re-baseline your history. Re-pull past performance under the new basis before you compare periods, so your trend lines sit on one consistent date logic.
Attribution time basis sets which date a conversion is reported on: the ad interaction date or the conversion date.
You apply it by choosing one basis in your reporting settings and holding it steady across every report.
The main constraint is conversion lag, which makes recent impression-time totals keep rising until delayed conversions finish attributing.
The main risk is misreading an unsettled recent period and cutting a campaign that is actually performing.
The leverage is matching your basis to your system of record so ad, analytics, and finance numbers reconcile.
They answer two different questions, and only one changes whether a conversion counts. An attribution window is the eligibility rule: it sets how long after someone sees or clicks your ad a conversion can still be credited, so a purchase outside that window is dropped entirely. Attribution time basis assumes the conversion already qualified and only decides which calendar date to file it under, the ad interaction date or the date the person acted. You can change the basis and your total conversion count stays identical, because no conversion is added or removed; the numbers simply move between dates. Change the window instead and the count itself rises or falls, because you are expanding or shrinking the pool of eligible conversions. In practice you set the window to capture the conversions you believe your ads influenced, then set the basis to control how those conversions line up against spend and against your other reporting systems.
Rarely, and ideally never mid-flight. Set your attribution time basis once, before or at launch, and keep it fixed for the life of the campaign so your trend data stays comparable from week to week. The honest answer to how often you should touch it is almost never: each change forces you to re-pull historical reports under the new basis, and any period you fail to re-baseline becomes an apples-to-oranges comparison. There are a few legitimate reasons to switch. Your finance or analytics team adopts a different system of record, you discover your current basis does not match how revenue is recognized, or you are standardizing reporting across several ad platforms. When one of those applies, make the change at a clean boundary such as the start of a month or quarter, re-pull the affected history immediately, and announce the switch to everyone who reads the reports so no one compares old and new periods by mistake.
Your totals for any given date range move because conversions carry two timestamps, and the basis decides which one the report obeys. When someone clicks today but buys three days from now, impression-time reporting credits that sale back to today, while conversion-time reporting records it on the day of purchase. Over a settled, far-past period the two bases converge, because every delayed conversion has long since landed. Over a recent window they diverge sharply: impression-time totals for the last several days keep climbing as late conversions attribute backward, and conversion-time totals for those days look lighter until the lagging sales arrive. The size of the gap tracks your conversion lag, so campaigns driving quick, low-value actions barely differ between bases, while those driving considered purchases can look very different depending on which basis you read. None of this reflects a change in real performance. It is the same conversions, filed against different dates.
Yes, the basis is yours to set. It is a reporting toggle you control inside the platform, and no campaign performance is required to turn it on. Within the ad platform's reporting controls you select whether conversions report on impression or interaction time or on conversion time, and you can usually apply that choice to dashboards and exports. What you cannot influence is the underlying conversion lag that makes the two bases differ, because that is driven by how long real people take to act after seeing your ad. You also cannot change the timestamps themselves; the platform records when the interaction and the conversion happened, and the basis only picks which of those stamps the report displays. So treat the basis as a lever you control deliberately and the lag as a fact you plan around. Set the basis to match your system of record, then let the natural delay between clicks and conversions inform how long you wait before trusting a recent period.
A good setup is a consistent one: a single basis, chosen on purpose, documented, and matched to the system your business trusts for conversions. There is no universally correct basis, so judge quality by three practical signs instead of by which option you picked. First, everyone reading your reports knows which basis is in use, because it is labeled on the dashboard and in the export. Second, your basis agrees with how your analytics and finance teams date conversions, so the same sale is not counted on two different days across systems. Third, you give recent periods enough time to settle before you draw conclusions, which tells you your team understands the conversion lag behind the numbers. If a new analyst can open any report, see the basis, and reconcile it against your source of truth without a side conversation, the setup is working. If reconciling takes an argument, the basis is either undocumented or mismatched.