How to Recover From a Google Algorithm Update
There is no procedure, no appeal and nothing to undo. Recovery from a core update is genuine improvement followed by waiting, which is slower and less satisfying than fixing a fault. This guide sets out what that actually involves and what it may not deliver.
There Is Nothing To Undo
No action was taken against your site. There is no notice, no case, no queue and nobody to appeal to. Most people arrive on a page like this looking for the thing to reverse, so it is worth saying at the top that the thing does not exist.
What a reassessment is. A change to the judgement.
A core update alters how quality is weighed across everything Google ranks. Your pages were reassessed alongside billions of others. Nothing was aimed at you, which is why there is no record of anything happening to you.
What a manual action is instead. The other case entirely.
Something applied to one site for a specific breach of published policy, notified in Search Console and resolvable through a process. If you have no notice, this is not your situation. Our material on risky practices owns that route and it is worth reading before you assume anything.
Why people look for the appeal anyway. Every other system has one.
A fine has an appeal, a suspension has a review and a dispute has a hearing. Search does not work that way for algorithmic movement. Hunting for the form to fill in burns the first fortnight.
Why we open with the discouraging part. It saves a year.
A business that believes it was punished spends months removing things. A business that understands it was reassessed spends those months improving. Same effort. Opposite direction. Only one of them works.
Which Means Recovery Is Improvement
If nothing was applied, the only route back is being a better answer than you were. That is the whole of it. Google's published guidance says as much, that there is no specific fix tied to a core update and that a drop does not necessarily mean anything is wrong with a page.
Why that guidance frustrates everybody. It offers no lever.
Businesses want a fault, because a fault can be fixed on Thursday. An instruction to be genuinely more useful than the sites now above you is harder, slower and correct.
What improvement means in practice. Substance, not polish.
Answering the question more completely, bringing knowledge nobody else has, showing who stands behind the page and removing the parts that exist for search rather than for a reader. Our guide to what Google means by helpful content is the fullest description of the standard being applied.
What it does not mean. A technical exercise.
Speed, headings, image compression and internal linking are all worth doing and none of them is what moved. A faster page that says the same thin thing is a faster thin page.
Why this page has no checklist. A checklist would be a lie.
Ten boxes to tick implies that ticking them produces the outcome. Nobody can promise that, so what follows is a method of investigation and a set of things not to do instead.
What To Actually Look At, In Order
The work after a drop is comparative. You are not looking for faults on your own site, you are working out what the reassessment rewarded and where you sit against it. Four things, in this order.
Which pages and sections fell. Start with the distribution.
Compare the settled period after the rollout against a settled period before, split by page and by search term. A decline spread evenly across everything means something different from one confined to a single section.
What now ranks in their place. The part everybody skips.
List who occupies the results you lost. This is the evidence about what changed and it exists nowhere else, which is why the next block is entirely about it.
What those pages do that yours do not. Read them properly.
Depth, first hand knowledge, a named author with a reason to be trusted, evidence or simply a page built for that exact question rather than a general page hoping to cover it.
Whether the fall is concentrated or general. It changes the response.
Concentrated points at a category of page or subject you handle weakly. General points at how your site is assessed as a whole, which is a bigger and slower piece of work. Google's own core updates documentation publishes a set of self-assessment questions worth working through here.
Look At What Won, Not At Your Own Site
A site audit tells you what is technically imperfect. The update was not about that. Examining what replaced you is the only reliable guide to what actually changed. It is the step most businesses never take.
Why an audit feels productive. It generates a list.
Audits produce dozens of findings, which feels like progress and can be invoiced. Most of those findings were true before the drop as well, when the same pages were ranking perfectly happily.
What comparison produces instead. One uncomfortable sentence.
Frequently something like: every page above us is written by somebody who does this work. Ours was written to cover a keyword. That sentence is worth more than forty audit findings.
How to run it. Ten searches, five results each.
Take the searches you lost most on, record what ranks now and note what the winners share. A pattern across all of them is the change. One competitor beating you is chance.
What to do with the finding. Rebuild the answer, not the site.
Improve the specific pages that lost against the specific standard you have just identified. That is narrower, cheaper and considerably more likely to work than a redesign.
Why Some Sites Recover And Others Do Not
Recovery is not evenly distributed and the pattern is fairly consistent. Sites that make substantial genuine change tend to regain a great deal. Sites that make cosmetic change generally do not. A third group never returns at all.
What the recovering group has in common. They changed what the site is.
New material written by people who know the subject. Weak pages consolidated or improved rather than shuffled. A clear reason for the site to exist beyond ranking. That is months of work rather than weeks.
What the group that stays down has in common. They changed how it looks.
Headings rewritten, publish dates refreshed, paragraphs reordered and a new template applied. Nothing about what the page is worth to somebody reading it changed, so nothing about the assessment changed either.
The third group. The standard moved past them.
Some sites ranked because they were the best available answer in a weak set of results. Once genuinely better answers exist, no amount of work on the old page recreates the conditions that gave it the position.
Why we say this rather than sell a package. You can check it.
Any firm can promise recovery. What nobody can do is promise the standard has not moved past the thing you are asking them to restore, which is a question worth putting to anybody you speak to.
You May Not Get Back To Where You Were
A position built on being the best available answer at the time is not owed back to you when better answers exist. That is uncomfortable and it is the thing that makes everything else on this page credible.
Why the position was never yours. It was a judgement.
Ranking is a comparison rerun constantly against everything else available. You held a position because of how you compared. If the field improved, the comparison changed, whatever you did or did not do.
What that means for your target. Aim at being the best answer now.
Not at the position you had. The position follows from the first thing and cannot be pursued directly, which is why plans built around getting back to page one tend to fail on their own terms.
What a firm promising the old position is promising. Something nobody controls.
No agency decides where you rank. Any guarantee is either a misunderstanding or a sales tactic. It is the clearest single signal that the firm in front of you should be declined.
What often happens instead. Something better than restoration.
Businesses that take this seriously frequently end up stronger across a wider set of searches than the ones they were fixated on. That is a better outcome than the old position and it is not the outcome they asked for.
Recovery Follows The Next Assessment
Improvement is not recognised the moment you publish it. Google's guidance states that recovery comes over time and that the more significant change tends to arrive with a later core update. We do not attach a period to that, because nobody credible can.
What happens between the work and the result. Very little, visibly.
Pages get recrawled and reassessed against the current standard. Smaller unannounced changes happen continually, so partial movement is possible. The larger reassessment arrives when it arrives.
Why this gap is where businesses give up. Nothing confirms the work.
Weeks of genuine improvement with a flat graph feels identical to weeks of wasted effort. That is precisely the point at which correct work gets abandoned and replaced with something worse.
How to survive the wait. Judge the work, not the graph.
Ask whether the pages are now genuinely better for somebody reading them. If they are, the work stands regardless of the timing. If you cannot answer that, the problem is the work rather than the wait.
What we will not tell you. A number of weeks.
Anybody quoting a recovery timeline is quoting an average of situations that have nothing to do with yours. It is the commonest thing said on this subject and there is no basis for it.
Do Not Churn The Site
Repeatedly changing things while waiting is the commonest self-inflicted damage after an update. It destroys the only thing you had, which was the ability to tell what your improvements actually did.
What churn looks like. A change every few days.
Titles rewritten on Monday, pages deleted on Thursday, a template swapped the following week and links disavowed the week after. Every alteration overlapping the last one still being assessed.
Why it defeats measurement. Too many variables.
With six changes in a month, movement afterwards cannot be attributed to any of them. You learn nothing, so the next round of decisions is guesswork built on the last round of guesswork.
What to do instead. Deliberate rounds.
Decide what you are improving, do it properly across a defined set of pages, record the date and then leave it alone long enough to be assessed. Then look, then decide again.
Why patience is the harder skill. It looks like inaction.
A business under pressure wants visible activity. An agency under pressure wants to demonstrate it. Holding still while improvement is assessed is the professional choice and it is rarely the popular one.
What Not To Do
Four things account for most of the harm done after a core update. Each of them is widely recommended, each feels decisive and each is usually a response to a punishment that was never issued.
Do not remove content indiscriminately. Judge page by page.
Deleting everything published before a date removes the pages bringing enquiries alongside the ones bringing nothing. Where a page genuinely serves nobody, removing it is reasonable. Bulk pruning by date is not.
Do not disavow links reflexively. It was not a link judgement.
A broad reassessment says nothing about your links, so disavowing after one treats a diagnosis nobody made. It can also discard references you earned. Our risky practices material covers the narrow cases where it applies.
Do not rebuild the site. It changes the wrong thing.
A rebuild alters the presentation and introduces migration risk while leaving the substance untouched. Businesses regularly emerge from one with the original problem and a new set of technical faults.
Do not buy a guaranteed recovery. Nobody owns the outcome.
A guarantee is a claim about something the seller does not control. Our material on hiring an agency sets out what a credible proposal looks like instead, which starts with somebody willing to tell you what may not come back.
When It Was Not The Update After All
A fair proportion of these investigations end somewhere else entirely. If you are some way into this page and the evidence has stopped fitting, that is worth taking seriously rather than pushing through.
The signs you are in the wrong place. Three of them.
The fall arrived in a single day. It is confined to one page or section. Your competitors sit exactly where they always did. None of those describes a broad reassessment.
Where to go instead. Back a step.
Our guide to telling whether an update caused the drop runs the tests properly. Our measuring performance material owns the prior question of whether the drop is real at all.
Why the mistake is so common. An update is the available story.
It has a name, a date and an industry discussing it. A tracking change nobody noticed has none of that, so it stays invisible while the update takes the blame.
Where this leaves you. With the durable question.
Once you are through this, the useful work is making the site the sort that reassessments leave alone. Preparing for future updates covers that. The full series sits on the Google algorithm updates guide.
We will not
guarantee your
old position
back.
Nobody controls where you rank, so a guarantee would be a promise about somebody else's decision. What we will do is compare your pages against what now outranks them, tell you what the difference is and be straight with you about the parts that may not return.
What the work actually involves:
If the standard has moved past the page you want restored, we will tell you that rather than bill you for a year.
Every guide.
One position.
How updates work, the history of what each one changed, whether an update caused your drop, what recovery genuinely involves, what helpful content means now and how to build a site that rollouts leave alone.
