Short answer
Wait until the rollout completes before measuring or acting, because core updates run twelve to twenty-four days and rankings move in both directions throughout. Sites regularly recover part of a loss before an update finishes with no intervention at all. Acting mid-rollout makes the result impossible to attribute.
Waiting is the hardest instruction in this work and the cheapest advice I give.
It costs nothing, it protects everything you do next, and almost nobody follows it, because doing nothing looks like negligence when a chart is falling.
How long is the wait?
It depends entirely on what you have. These are very different clocks.
| Situation | Wait before acting | Why |
|---|---|---|
| Technical fault suspected | Do not wait at all | Fix it now. Nothing is being reassessed |
| Manual action present | Do not wait | You have an explicit problem statement |
| Core update rolling out | Until it completes, 12 to 24 days | Rankings move both ways throughout |
| Rollout finished | 28 days for a clean reading | Effects continue settling after the end date |
| No confirmed update matches | 2 weeks | Rules out normal variance |
The first two rows are the exception people miss. Waiting applies to quality reassessments, not to broken tags or a penalty message.
Why rankings move during a rollout
A rollout is not a switch being flipped. Changes propagate across the index over days or weeks.
During that period a site can fall on day four, partly recover by day twelve, and settle somewhere different by day twenty. None of those intermediate positions is the verdict.
Google has said this explicitly. The June 2021 core update was announced as incomplete, with a second part following in July. Sites that reacted in June were reacting to a state Google had already described as interim.
What acting early actually costs
Three things, and the third is the expensive one.
- You may fix nothing. The drop you reacted to might have partly reversed on its own.
- You may cause harm. Panic deletions and disavow uploads are both slow to reverse.
- You lose attribution forever. If the site recovers, you will never know whether your change did it or the rollout finished.
That third cost compounds. Without attribution you cannot repeat what worked or avoid what did not, on this site or the next one.
What to do while waiting
Waiting is not idleness. The diagnostic work is exactly what should happen now.
- Export sixteen months of Search Console data before retention removes your baseline.
- Check Manual Actions and the Pages report. Both are fast and both change everything.
- Find the exact break date from daily data with the drop analyzer.
- Establish scope: site-wide, one template, or a handful of pages.
- Write the hypothesis down, with what evidence would disprove it.
That is a week of real work with nothing changed on the site. It is also something to show people who are asking what you are doing.
After the rollout: the second wait
Most people do not know about this one.
Once your fixes ship, there is another wait before they can be judged, and its length depends on the system that demoted you.
- Technical fixes take effect on recrawl. One to four weeks.
- Spam policy cleanups are reassessed continuously. Two to eight weeks.
- Core quality work waits for the next core update. Historically 7 to 147 days apart.
That range is the part to plan around. Work finished in January can sit unassessed until May, and that is normal rather than a failure. The core update countdown shows the current gap and the full historical spread.
When waiting is the wrong answer
Do not wait if any of these apply.
- Manual Actions is not empty. You have an explicit problem and a defined process.
- Pages are dropping out of the index. That is technical and it gets worse while you wait.
- You just shipped a change. A redesign, migration or plugin update near the break date is the lead to chase now.
- The site is broken. Server errors, blocked crawling, a staging noindex. Fix immediately.
The distinction is straightforward: wait for judgements, act on faults.
Freeze the site once the work is in
The second wait has its own discipline.
Continuing to change structural things during the wait destroys attribution again. Ordinary publishing is fine. Templates, URLs, navigation and indexing settings are not.
Rebuilding twice during a long gap is worse than waiting once, because you end up with no idea which version was assessed. Verification is covered in how to know if the recovery actually worked, and the reactions that cause most self-inflicted damage are in what not to do in the first week.
Should you wait, or is yours a fault?
Waiting protects a quality diagnosis and costs you weeks if the real cause is technical. Send your Search Console access and within 48 hours I will tell you which you have, and what the honest timeline looks like.
Get free recovery audit