Google Ads conversion delay: why recent days look worse

Google Ads conversion delay is the gap between an ad interaction and the conversion that follows it. Because the standard conversion columns in Google Ads credit each conversion to the date of the click, not the date of the sale or lead, the most recent days in any report are always short of conversions that have not happened yet, so they look worse than they will once the data settles.

That single reporting rule is behind a lot of unnecessary panic: a budget cut on Tuesday because Monday's cost per conversion doubled, a target CPA lowered after three bad-looking days, an experiment ended early because the trial arm seemed to be losing. This guide explains how the delay works, which columns show it, how Smart Bidding treats it, how long to wait before judging a change, and where to find your own account's delay.

What conversion delay (conversion lag) is

Google uses both names, conversion delay and conversion lag, for the same thing. If someone clicks your ad on Monday and buys on Thursday, that conversion has a delay of three days. The delay belongs to your business, not to Google Ads: a food delivery order may follow the click within minutes, while a condo enquiry, a B2B software demo or a university application can take weeks.

Two settings decide how much of that delay Google Ads will ever see.

• The conversion window. Each conversion action has a click-through conversion window. The default is 30 days, and depending on the conversion source it can be set anywhere from 1 day up to 30, 60 or 90 days. A conversion that happens after the window closes is not counted at all. Engaged-view conversions have their own window (3 days by default) and view-through conversions another (1 day by default).
• Changes only apply forward. If you lengthen a window today, conversions from older clicks that already fell outside the old window are not counted retroactively. Lengthening the window does not repair past data.

On top of the behavioural delay there is a small processing delay: conversion data can take 24 to 48 hours to appear in reports. That matters for the last day or two; the behavioural delay matters for everything else.

Why recent days look worse in Google Ads reporting

Cost and conversions arrive on different clocks. The cost of a click is recorded on the day of the click, in full, and cost data is usually refreshed within about an hour. The conversions credited to that same click date keep arriving over the following days and weeks, up to the end of the conversion window. So for any recent date you are looking at complete cost and incomplete conversions.

A worked example with illustrative numbers: suppose Monday's clicks cost ฿10,000 and will eventually produce 20 conversions, a cost per conversion of ฿500. If only 8 of those conversions have been recorded by Tuesday morning, Monday's cost per conversion reads ฿1,250. Nothing went wrong on Monday. The other 12 conversions simply have not happened yet, and when they do, Google Ads adds them back to Monday.

That back-filling has three side effects worth knowing:

• Past numbers change. Last week's report will show more conversions next week than it does today. A spreadsheet exported on Friday will not match the same date range viewed a month later.
• Conversion rate and ROAS dip at the end of every chart. The downward hook on the right side of a 30-day trend line is usually delay, not decline.
• Short date ranges suffer most. "Last 7 days" in an account with a two-week delay is mostly unfinished data. "Last 90 days" is mostly finished, with only the tail affected.

The cost of misreading this is real. Pausing keywords, cutting budgets or tightening a target CPA on the strength of unfinished days removes traffic that was performing normally, and Smart Bidding then has to relearn after the change.

"By conversion time" vs "by time of interaction" columns

Google Ads gives you two ways to date the same conversions. The standard columns (Conversions, Conv. value, Cost / conv. and so on) are dated by the time of the ad interaction. A set of six "by conv. time" columns dates them by when the conversion itself happened:

• Conversions (by conv. time)
• Conv. value (by conv. time)
• Value / conv. (by conv. time)
• All conv. (by conv. time)
• All conv. value (by conv. time)
• Value / all conv. (by conv. time)

Add them from Columns, then Modify columns, in the Conversions group. Their data goes back to March 2019, and store visit and store sales conversions are not included in them.

Which one to read depends on the question you are asking:

One trap with the "by conv. time" view: if you divide cost by Conversions (by conv. time), you are dividing today's spend by conversions that partly came from older clicks. Over a month that is usually fine. Over a few days, or in a week when spend jumped or dropped sharply, the ratio can mislead in either direction.

How Smart Bidding accounts for delayed conversions

Smart Bidding learns from both historical and new conversion data. Where data is available, the bid strategy report shows how many conversions Google Ads predicts are still to be reported, based on your typical delay, and the performance chart marks the date range where more conversions may still arrive. Hovering over Conversions, Cost / conv. or Conv. rate in that report shows the average length of the delay.

A few practical consequences follow:

• Set targets from mature data. When you choose a target CPA or target ROAS, base it on a period that leaves out the most recent days still affected by delay. A target set from the last week of unfinished data will be stricter than your real performance and will choke volume.
• Expect a learning period. After a significant change, a "Learning" status can show for up to seven days, and the strategy can take one to two conversion cycles to settle on a new goal. With a long delay, a cycle is long, so settling takes longer too.
• Late data teaches late. Bidding can only learn from conversions it has received. If conversions reach Google Ads days after they happen, for example through a weekly offline upload, the system is always working from an older picture of what works.

How long to wait before judging a change or an experiment

Google's guidance for bid strategy changes is to evaluate over two conversion cycles or a two-week window, whichever is longer, while accounting for conversion delay. A conversion cycle is the typical time from click to conversion for that conversion action.

A simple routine that works for budget, bid, ad and landing page changes:

1. Measure your delay with the Days to conversion segment (steps below) and note the day by which most conversions have arrived. Use the main conversion action, not all of them together.
2. Wait for the evaluation period to pass: two cycles or two weeks, whichever is longer.
3. End your comparison date range that many days ago. If most conversions arrive within 10 days (an illustrative figure), the "after" period should end at least 10 days before today.
4. Compare equal-length periods with the same lag excluded from both, so the "before" period is not fully matured while the "after" period is still filling in.
5. Change one thing at a time. If budget, target and ads all change in the same week, a delayed result cannot be traced to any of them.

Experiments follow the same logic. An experiment that looks decided on its last day still has unfinished conversions in its final stretch. Keep reading the results for at least one delay length after the end date before applying or discarding the trial arm.

How to see your own account's delay

Google Ads has several places that show delay directly. You do not need to estimate it.

The Days to conversion segment

1. Open Campaigns, Ad groups or Search keywords from the left menu.
2. Set a date range that ended at least 30 days ago, or longer if your conversion window is longer, so the data is complete.
3. Click the Segment icon and choose Conversions, then Days to conversion.

The conversion columns split into up to 19 rows by days to conversion. Add up the rows to see the share of conversions that arrive on the same day, within a week, within two weeks and later. That cumulative share is your working definition of "mature".

Hover cards and the bid strategy report

Hover over a campaign in the Conversions column, or over the scoreboard figures, and Google Ads shows conversion lag information where it applies. Choose a recent date range and annotations appear when the selected data is affected. For Smart Bidding campaigns, click the Bid strategy type column (add it if it is hidden) to open the bid strategy report with its predicted conversions.

Path metrics in attribution reports

The Path metrics report, under attribution reports in Goals, shows how long it takes and how many interactions happen before a conversion. The "Measure from" menu switches between the first and the last ad interaction, and "Measure in" switches between days and hours. Measuring from the first interaction usually shows a longer delay than the last, which is useful when deciding how long a conversion window should be.

Offline imports and delay

Offline conversion imports add a second delay on top of the customer's: the time between the sale being recorded in your CRM and the upload reaching Google Ads.

• The 90-day limit. Conversions uploaded with a GCLID more than 90 days after the associated last click are not imported and do not appear in conversion statistics. The conversion action's own window still applies within that limit.
• Uploads back-fill old dates. Each imported conversion lands on its original click date, so a large upload can lift numbers for days or weeks that you had already reported.
• Batch size shapes the chart. If uploads happen once a week, the most recent week always looks empty for that conversion action, and Smart Bidding learns from it a week late. A short, regular upload schedule keeps both reporting and bidding closer to reality.
• Long sales cycles need an earlier milestone. If deals close after the window or the 90-day limit, import an earlier stage that does fall inside it, such as a qualified lead, and keep the closed deal for your own reporting.

Common mistakes and what each one costs

• Judging yesterday's cost per conversion. It is always inflated by delay. Acting on it cuts traffic that was performing normally.
• Lowering a target CPA from unfinished data. The target ends up stricter than real performance, and the strategy bids down and loses volume.
• Expecting a longer window to fix past data. Window changes only apply going forward.
• Calling a click-date vs sale-date gap a tracking fault. Google Ads standard columns and most CRMs date conversions differently. Compare with "by conv. time" columns before opening a tracking investigation.
• Weekly offline uploads. Reporting and bidding both run a week behind.

Conversion delay in the Thai market

Several common purchase paths in Thailand can stretch the measured delay beyond the moment a customer decides to buy:

• Chat-led journeys. When an ad sends people to a LINE Official Account or another chat channel, the sale is often agreed in chat and recorded later by a sales team. If that sale reaches Google Ads through an offline import, its delay includes the human follow-up time and the upload schedule.
• Bank transfer and PromptPay checkouts. If the purchase conversion fires only when a transfer or QR payment is confirmed rather than when the order is placed, the recorded delay includes the confirmation step. Know which event you are counting.
• Cash on delivery. A conversion counted on successful delivery arrives days after the order, by design.
• Long holidays. Around Songkran or the New Year break, both buying behaviour and back-office confirmation can slow down. Compare those weeks with the same period in an earlier year, not with the week before.

If your team shares daily numbers in a LINE group or a morning report, mark the last few days (your measured delay) as provisional. That one label prevents most knee-jerk changes.

FAQ

What is conversion delay in Google Ads?

Conversion delay is the time between someone interacting with your ad and completing a conversion. Because standard columns credit conversions to the click date, recent days show fewer conversions than they will once those delayed conversions arrive.

Why does yesterday's cost per conversion look so high?

Yesterday's cost is complete but its conversions are not. Conversions from those clicks will keep arriving until the conversion window closes, and each one is added back to yesterday's date, which brings the cost per conversion down.

Does changing the conversion window fix conversion delay?

No, a longer window only lets Google Ads count conversions that happen later, and it applies only to conversions going forward. It does not shorten the delay or add back conversions that fell outside the old window.

How long should I wait before judging a Smart Bidding change?

Wait at least two conversion cycles or two weeks, whichever is longer, and leave out the most recent days still affected by delay when you compare. Use the Days to conversion segment to find your cycle length.

Do offline conversion imports change past data?

Yes, imported conversions are credited to the original click date, so an upload can raise conversion numbers for days you have already reported. GCLID uploads more than 90 days after the click are not imported.

If your reports swing every time you look at them, the fix usually starts with measurement rather than bids. The Google Ads management team at Relevant Audience sets up conversion tracking with GA4, enhanced conversions and offline conversion imports so Smart Bidding works from real revenue data, and the GA4 setup service covers chat-led journeys and bank-transfer checkouts. To talk through your own account, start with our paid search service.