A Google update is rolling and your graph moved. The next 48 hours decide nothing about your rankings, and everything about the quality of your response: sites that follow an ordered checklist respond to patterns, and sites that panic respond to noise.
Here is the checklist, including its two deliberate waits.
Step 0: confirm it is actually an update
Before anything: is this Google's weather or your own plumbing? Check Google's Search Status dashboard for a confirmed rollout and a volatility index like the Semrush Sensor for storm confirmation.
No confirmed update and calm sensors while your graph dives means the problem is local, and you want the silent-change forensics, not this checklist. The overlap case is real too: sites break themselves mid-update and blame the weather.
Step 1: the first wait
Updates roll out over days to weeks, and positions swing and reverse mid-flight. Damage assessed on day two routinely un-happens by day ten.
During the wait: publish normally, maintain normally, change nothing structural. Collect the data; spend none of it.
Step 2: scope the impact
Rollout complete, now measure. In Search Console, compare the settled week against the last clean pre-update week: queries, pages, and both directions, because your gains are as diagnostic as your losses.
Watch the mix-shift traps while comparing: an average that "collapsed" because long-tail impressions vanished is a different event than money keywords sliding, and only the filtered, per-page view tells them apart.
Step 3: classify the pattern
| Pattern | Suggests | Your response lane |
|---|---|---|
| Sitewide, most query types | Quality or trust reassessment | The long game: quality sweep |
| One topic cluster | Intent shift or topical authority moved | Re-read those SERPs, realign |
| One page type or format | Format preference changed | Study the winning format |
| Scattered, no pattern | Ordinary re-shuffling amplified | Mostly wait; fix nothing big |
Step 4: read the SERPs that changed
For your five most important affected queries, study the new page one like an analyst: who entered, who left, and what the new winners share. Fresher dates? Different format? More firsthand experience? Bigger brands? A new feature box absorbing the clicks?
The shared trait of the winners is the update's message to your niche, and it is more reliable than any commentary thread about what the update "targeted".
Step 5: respond to the pattern, not the pain
Cluster and format patterns get targeted responses: realign the affected pages to the intent and format that now wins, which is ordinary refresh work with a sharper brief.
Sitewide patterns get the honest audit: your weakest content, thin pages, stale sections, everything you already knew was mediocre, addressed as a program over months, because quality reassessments respond to bodies of work rather than tweaks.
And the not-to-do list, which is most of the value: no mass deletions in week one, no redesigns, no disavow files, no rewriting your best pages to chase a theory. Post-update flailing converts recoverable dips into self-inflicted declines.
Step 6: the second wait, documented
Write the incident down: the update, your impact scope, the pattern class, the responses shipped. Then expect the full verdict at the next comparable update, because broad recoveries frequently land in cycles, and your alert thresholds will tell you the moment it starts.
The document is what turns updates from recurring traumas into a dataset: three incidents in, you know your site's exposure patterns better than any pundit does.
The one-line takeaway: confirm it was an update, wait out the rollout, scope both directions, classify the pattern, read the new winners, and respond in the pattern's lane with structural panic banned. Two waits, one document, zero flailing.