It is not a penalty
Google says core updates are broad changes that do not target specific sites or pages. A site can move down because other pages became better matches, not because it did something wrong.
When traffic falls after a core update, the instinct is to change something immediately. Google’s documentation says the opposite: wait, measure properly, assess the site honestly, and avoid quick fixes. This guide covers what a core update actually is, how to confirm that one caused your drop (and not a migration, a penalty or a seasonal dip), how to judge your pages against Google’s published questions, what to fix first, and what recovery timing Google does and does not promise.
// note: Built from Google Search Central documentation and industry reporting on update dates. Nobody can promise a recovery, and I have not recovered a site for this article. Sources are linked and listed at the end. For hands-on help, see my Google penalty recovery service.
| step | what to do |
|---|---|
| 1 | Confirm rollout finished on the status dashboard |
| 2 | Wait a full week after completion |
| 3 | Compare that week with the week before it began |
| 4 | Review top pages and queries |
| 5 | Split by Web, Images, Video, News |
// source: Google Search Central, core updates.
Google says core updates are broad changes that do not target specific sites or pages. A site can move down because other pages became better matches, not because it did something wrong.
Google says some improvements can show within days, but its systems may need several months to confirm a site consistently produces helpful content. No effect is guaranteed.
Google advises against removing page elements because someone said they hurt SEO. Focus on changes that serve users and last.
A core update is a broad, significant change to Google’s search systems, released several times a year, meant to keep results helpful and reliable as the web changes.
Three details from Google’s documentation shape everything that follows.
It is not aimed at you. Core updates do not target specific sites. Google says sites that drop are not necessarily doing anything wrong, because other sites may have become better matches for the same searches.
Google announces them. They are listed on the Search Status Dashboard with start and end dates. The documentation does not state how long a rollout lasts, so the dashboard is the source to check. Recent rollouts have taken about twelve days according to reporting by Search Engine Roundtable and Search Engine Journal.
Smaller changes happen between them. Google says it also makes smaller, unannounced core updates, so improvements can appear without a big named update.
// per Search Engine Roundtable and Search Engine Journal reporting; confirm on Google’s dashboard.
Different problems need different responses. Misdiagnosis wastes months.
| Core update | Spam update | Manual action | Technical issue | |
|---|---|---|---|---|
| Cause | Broad re-evaluation of quality and relevance | Detection of spam practices | A reviewer finds a policy violation | Crawling, indexing or speed fault |
| Notice | Dashboard entry, no site notice | Dashboard entry | Message in Search Console | Reports and logs |
| Review request | None | None | Yes | Not applicable |
| Fix | Improve overall quality over time | Remove the practice | Fix, document, request review | Fix the fault |
Check the Manual Actions report first. If there is a message, you have a manual action and my penalty recovery service page explains the process. If your problem is spammy links, read the Penguin update guide. If the drop began on the day you launched a new site, read the website migration checklist first, because redirects and noindex mistakes mimic a core update.
Google gives a specific procedure for checking Search Console after a core update. It is worth following exactly, because it prevents the most common error: analysing during the rollout, when results are still moving.
Google also gives scale guidance. A small slip, for example from position 2 to 4, generally calls for no drastic action, and well-performing content should not be changed needlessly. A large fall, for example from position 4 to 29, deserves a deeper assessment.
Dates that overlap an update are not proof. Run these checks first.
| Question | How to check | If yes |
|---|---|---|
| Did you change the site? | Compare release logs, plugin updates, theme changes and redirects to the drop date | Fix the change first |
| Is anything blocked or noindexed? | URL Inspection on lost pages, robots.txt, meta robots | Remove the block |
| Are pages not indexed? | Page indexing report; see crawled, currently not indexed | Treat as an indexing problem |
| Is it seasonal? | Compare with the same weeks last year | Do not react |
| Did a spam update land too? | Check the dashboard for overlapping dates | Review spam policy exposure |
| Is there a manual action? | Manual Actions report | Follow the manual action process |
| Are AI features changing clicks? | Impressions stable but clicks fall for question queries | See how to rank in AI Overviews |
| Did a competitor improve? | Look at who now ranks for your lost queries | Compare their pages to yours |
A tracking tool’s volatility chart can tell you an update is rolling out, but it cannot tell you why your site moved. Your own Search Console data is the evidence that matters.
Google tells you to look at the pages that lost the most and ask whether other pages on the web serve searchers better. A useful way to do this is to group the losses and look for what they share.
Then open the search results for three or four of the lost queries and read what now ranks. Do not ask “what keywords do they use”. Ask what they give the reader that your page does not: original data, first-hand evidence, clearer structure, a more complete answer or a more trusted publisher.
| group | pages | clicks change |
|---|---|---|
| Listicle posts | 34 | -48% |
| Service pages | 9 | -3% |
| How-to guides | 21 | -12% |
| Location pages | 40 | -35% |
// illustrative numbers, not a real site.
Google points site owners to its helpful content guidance. The questions group into four areas, summarised here in my words.
// source: Google Search Central, helpful, reliable, people-first content. Grouping and wording are mine.
Google also suggests asking people unaffiliated with your site to run the same assessment. That is good advice, because owners tend to rate their own pages generously.
For each affected page, choose one action. Google encourages improving content by rewriting or restructuring. It treats deleting as a last resort, though it says removing whole sections built for search engines rather than people may help.
Whatever you choose, keep a dated log. If you remove or merge pages, follow the redirect rules in the migration checklist, because the same mistakes (chains, homepage redirects) apply.
| page state | action |
|---|---|
| Right topic, weak depth | Improve |
| Duplicates of each other | Merge |
| Written only for search | Remove |
| Small dip, still strong | Leave |
These are the changes that match Google’s criteria. None is a trick.
Original data, screenshots of your own work, test results, case notes or quotes from people who did the thing. Google asks whether content shows first-hand experience.
Add a real byline linked to a real author page. Google says fabricated authors or credentials are deceptive.
If you compare products, say how you tested or that you did not. Disclose AI use where readers would ask.
Correct errors, refresh prices and dates, and update when something truly changes. Do not fake freshness.
Add the follow-up questions a reader has. Remove padding that exists only to reach a word count.
Fix intrusive ads, slow templates and confusing layouts. Google points to page experience as part of the picture.
My own reviews follow the same approach: vendor-stated prices, a clear note when something was not tested, no affiliate links and a statement of what I could not verify. That kind of honesty is a quality signal readers can see.
Google tells you to review your whole site, and core updates evaluate sites broadly. If large parts are weak, polishing the best pages will not offset them.
Google gives a candid answer. Some improvements can show results within a few days. But its systems may need several months to confirm that a site consistently produces helpful content. If nothing changes after a few months, Google says to keep improving and expect the next core update to be the next point where changes may be reflected. It also stresses that rankings are dynamic and can rise or fall for reasons unrelated to your site, and that there is no guarantee of a noticeable effect.
The practical consequence: plan in quarters, not weeks. A sensible rhythm is diagnose in the first two weeks, make the main improvements across the next two months, then wait and measure.
Recovery also does not mean returning to the old position. If other pages overtook you because they were genuinely better, the goal is to become competitive, not to restore a past rank.
| when | focus |
|---|---|
| Weeks 1 to 2 | Confirm drop, rule out other causes, group losses |
| Weeks 3 to 6 | Improve, merge or remove the worst pages |
| Weeks 7 to 12 | Site-level fixes, new useful content |
| Months 4 to 6 | Measure; keep improving; await later updates |
// the pacing is my suggestion; Google gives no fixed schedule.
Fewer pages with real depth beat many near-duplicates. Merge before you add.
Build topical depth across linked pages rather than chasing every trend.
Track clicks and impressions by page group in Search Console, and annotate site changes.
Re-run the helpful content questions on your top pages and fix gaps.
Avoid link buying and automated mass content. Both create risks beyond core updates.
Email, direct and referral audiences soften the effect of any single algorithm change.
I read Google Search Central’s core updates page and its helpful content guidance. Update dates come from reporting by Search Engine Roundtable and Search Engine Journal, and they should be confirmed on Google’s Search Status Dashboard, which I did not open directly. The grouping of Google’s questions, the decision guide and the pacing are my own suggestions. I did not recover a site for this article, and there are no affiliate links here.
A broad, significant change to Google’s search systems, made several times a year to keep results helpful and reliable. It does not target specific sites.
Wait until the rollout is complete on the Search Status Dashboard, wait a further week, then compare that week with the week before the update began in Search Console. Rule out technical issues and manual actions first.
No. Google says core updates do not target specific sites, and that sites which drop are not necessarily doing anything wrong.
Google says some improvements may show within days, but its systems may need several months to confirm consistent helpfulness. It guarantees no effect.
Google says that if you see no effect after a few months of improving, waiting for the next core update may be needed. It also says smaller unannounced core updates occur between named ones.
Google calls deletion a last resort. Improving through rewriting or restructuring is encouraged. Removing whole sections made for search engines rather than people may help.
No. Review requests exist only for manual actions. A core update drop has no request form.
Not as a direct reaction. Results move during a rollout, so analyse once it ends. Fix genuine technical faults at any time.
Google’s guidance targets unhelpful content made mainly for search, and says using automation to manipulate rankings violates its spam policies. It also asks you to disclose AI use where readers would expect it.
No. Treat guarantees as a red flag. See my recovery service for what I do and do not promise.
Read the Penguin update guide, the negative SEO guide, the technical SEO guide and the migration checklist.