Can you recover from a Google core update?

Short answer

Often, yes, but not by undoing something, because there is usually nothing to undo. A core update re-weighs what was already true about your site. Recovery means changing what is true, then waiting, sometimes until the next core update. Quick fixes made in the first fortnight usually make the diagnosis harder.

The question I hear most is phrased as “how do I reverse the update”. It is an understandable way to put it, and it points people in the wrong direction. You cannot reverse a core update. You can change your site so that the next assessment comes out differently.

What a core update is

Google describes core updates as significant, broad changes to its ranking systems, made several times a year. Its documentation is direct that they don’t target specific sites or individual web pages, and it makes a point of saying that pages which move down are not necessarily bad. Other pages may simply have been assessed as more useful.

That is the most important thing to understand about a core update drop. It is not a penalty and there is no list of violations. The systems looked at the whole web again with different weights, and your pages came out lower relative to the competition.

Why “trigger” and “cause” are different

When a client tells me they lost traffic to a core update, I treat the update as the trigger. The cause is whatever the update re-weighted. It might be content that answers the query less directly than competitors. It might be pages that exist to rank rather than to help. It might be a section of thin pages pulling the rest of the site down with it. It might be that the results page changed shape and now favours forums, or video, or big brands.

Skipping straight to fixes means guessing which of those it was. In my experience that is where most failed recoveries come from.

What recovery actually involves

Confirm the timing. Line the drop up against the update’s rollout dates. If your traffic fell a week before the update started, it was something else.
Find what lost. Which pages, which templates, which queries. A core update rarely hits everything evenly.
Look at what won. For the queries you lost, study the pages now ranking. What do they do that yours do not? Be honest.
Change the substance. Improve the pages that lost, merge or remove the ones that cannot be made useful, and fix whatever made the site look like it was written for search engines rather than people.
Wait, and measure the right thing. Track the pages and queries that lost, not the sitewide total.

Google’s own people-first content self-assessment is genuinely useful here. The Who, How and Why questions in it are exactly the ones I would ask: who wrote this, how was it made, and why does it exist. If the honest answer to “why” is “to rank”, that is usually the problem.

How long it takes

Longer than anyone wants. Google says it could take several months for its systems to learn and confirm that a site has improved, and that in some cases the effect will not show until the next core update. Smaller unannounced updates happen in between, so some sites recover gradually, but many see the change land on the next big update date.

That has a practical consequence. If you make changes in March and the next core update is in June, you may see nothing for three months even though the work was right. Decide in advance what you will measure and how long you will wait before concluding it did not work.

What I would not do

Do not make quick changes in the first two weeks. Google advises against it too. During the rollout, rankings swing and the data is noisy. Changes made then are made blind.

Do not mass-delete content as a first move. Google describes removing content as a last resort, and pruning removes the wrong pages more often than you would think. I have seen a pruning exercise make a drop worse.

Do not rewrite everything in a new style because someone said the update favoured a particular format. Updates reward usefulness, not a template.

When recovery is not realistic

Some losses are permanent. If the results page for your main queries has been taken over by forums, marketplaces or AI answers, there may be less room for any site like yours, however good it is. If a competitor has genuinely outgrown you, fixing your own pages narrows the gap but may not close it.

That is worth knowing before you spend six months on a recovery plan. Part of a proper diagnosis is an honest read on the ceiling.

A sign you are on the right track

Before the next update lands, watch the pages you improved for small movements in impressions on long-tail queries. They often appear first, weeks before the head terms move. It is not proof, but a flat line on everything you changed is a reason to revisit the diagnosis rather than keep waiting.

If a core update took your traffic and you want to know what it actually re-weighted, my core update recovery work starts with the dates and the segments, not the fixes, and tells you up front if the loss looks permanent.