Yesterday's spend has arrived. Some of the sales it bought may not have. Cut Google Ads bids because that unfinished day looks expensive, and you may be reacting to a reporting gap rather than a worse campaign.

The sensible response is not another automatic waiting period. Give the recent numbers a fair comparison while checking whether the business has actually stopped receiving orders or leads. Those are separate jobs, and neither should wait for the other.

Compare the age of the evidence

Google's conversion-lag guidance explains why delayed conversions can make CPA look higher and ROAS lower before reporting catches up. Its forecasts estimate how those measures may change; they do not promise recovery.

That makes a comparison between yesterday and a settled older period suspect before any bidding argument begins. Both date ranges may contain the same number of days while having very different opportunities to collect conversions.

Start with the conversion action that matters to the campaign. Inspect its usual time to conversion and how earlier periods changed as they aged. Compare recent results with earlier snapshots taken at a similar age, where those snapshots exist. If you only have today's settled history, say so; it cannot tell you exactly what last month's dashboard showed the morning after the click.

This is more useful than borrowing a three-day or two-week rule from another account. A form submission and a later closed sale do not finish on the same schedule.

Keep two clocks on the report

Google distinguishes interaction-date reporting from conversion-time reporting . The standard conversion columns credit a later sale back to the ad interaction; Conversions (by conv. time) shows when the conversion happened. The former aligns with the spend's date. The latter helps investigate activity alongside external records.

Keep both visible, but give each a job. Use the interaction-date view to assess the traffic bought in that period. Use the conversion-time view when reconciling events with orders or CRM records. Align the date range, time zone and conversion action before interpreting a mismatch.

Do not divide today's spend by today's conversion-time total and assume you have repaired campaign CPA. That total can include conversions from earlier interactions. Changing the clock changes the question; it does not finish yesterday's cohort.

Estimate what is missing without booking it as revenue

Historical maturation can show whether an early result is unusually weak or merely incomplete. Preserve successive snapshots of the same interaction period, then look at how its conversion count and value develop. Use that account's pattern as a planning estimate, not a credit note for future sales.

An estimated adjustment is worth showing separately from observed conversions. Otherwise a helpful forecast becomes another number the team treats as earned. No historical pattern establishes that the current gap will close, particularly if the purchase path has changed or stopped working.

A broken checkout does not need a mature cohort

The strongest objection to waiting is simple: money is leaving now. If the form no longer submits or orders are disappearing from the store itself, a discussion about reporting maturity offers no protection.

Test the actual purchase or lead path. Compare the relevant business records with what the platform records. Inspect whether spend and clicks changed as well. A normal order flow with a reporting gap calls for a different investigation from a real collapse in incoming orders.

Then make a bounded decision:

  • If the path is broken, repair it and decide whether continued spend is defensible.
  • If the path works but recent reporting is immature, set a review point based on that account's lag and a spend limit the business can tolerate.
  • If similarly mature results still deteriorate, investigate the campaign rather than extending the waiting period by habit.

This is narrower than a general argument for patience during learning. The question is whether the conversion evidence is comparable yet. Let the right cohort finish counting, but do not give a real failure extra budget merely because reporting can be late.