Conversion delay
reporting lag / conversions still coming in / why does last week look bad / time lag
Conversion delay is the gap between the click and the sale. Microsoft states conversions are reported on the click date rather than the conversion date, so recent days are always incomplete, which temporarily inflates cost per acquisition and deflates return on ad spend.
Microsoft defines it as the time between someone clicking your ad and completing a conversion, which is the obvious half. The half that changes how you read a report comes next: conversions are reported based on the click date rather than the conversion date, so there may be a delay before all conversions are fully reflected. A sale that happens today is added to the day the click happened, which could be three weeks ago. Days you have already looked at keep filling in behind you.
The error therefore runs one way. Costs are complete the moment they are spent; the conversions they produced are not. Microsoft says plainly that this can temporarily inflate your cost per acquisition or deflate your return on ad spend. Every fresh report is pessimistic, and the newest days are the most wrong.
Microsoft also ships the measurement, which is the useful part. Its conversion delay column shows how many days it typically takes to collect 90 percent of conversions for each campaign in the range you selected, and the grid will point at the date on which performance can be judged inclusive of delayed conversions. That turns a vague warning into a number you can wait for.
Google approaches it from the other end, through the time lag report in attribution reports, which shows how the time between ad interactions and conversions varies. It attaches a limit worth knowing before quoting the figure: that report reflects time lag only for ad interactions on google.com, and does not show it for conversions attributed to the Display Network or other search partners.
In practice
This settles the most common argument about a bad week, because it says when to look rather than what to conclude. If it takes twelve days to collect 90 percent of conversions, then any window inside twelve days is a partial count, and last Friday is the least finished number on the page. Judging it on Monday is judging the incomplete half of it.
Not the same as
- Conversion window
- That is the setting deciding whether a late conversion is ever counted. Delay is how long they take to arrive inside it.
- Modeled conversions
- Those estimate events the platform cannot observe. These are real events that simply have not landed yet.
Why it matters to you
Decisions get made on the freshest data because it feels the most current, and it is the data most certain to be wrong, in a direction that makes good campaigns look bad. Knowing your own collection time converts that from a feeling into a date. It also settles the reverse case, where a comparison sets a finished month against a half-reported one and calls the difference a trend.
What to ask or check
- 01How many days does it take to collect 90 percent of conversions in this account?
- 02Does this report end far enough back to be complete, or does it run to yesterday?
- 03Are the two periods being compared both finished, or is one still filling in?
What people get wrong
That a report is a finished record of what happened. Microsoft states conversions are reported on the click date rather than the conversion date, so the most recent days keep changing after you have read them.
Red flags
- A performance review run on data ending yesterday.
- A campaign paused on a figure from inside the collection window.
- A comparison between a completed month and one still filling in.
Who owns it
The platforms define the reporting rule, and Microsoft publishes a column that tells you your own collection time. Whoever schedules the reporting decides whether that number is respected.
Where you will see it
In every weekly and monthly ads report, and most sharply in the last few days of one.