How to Recover Rankings for Keywords That Have Dropped
Most people arrive here wanting to change something immediately. Do not. The commonest outcome of acting quickly is removing whatever was working alongside whatever was not, so everything here is diagnosis until the sixth section.
First, Confirm It Actually Dropped
A great many reported drops are not drops. Positions move constantly and results differ by person and place. A single check showing something lower is one sample rather than a finding. Establishing that something real happened is the first job and it frequently ends the investigation.
What is not a drop. A single check.
One look showing a lower position proves nothing, because there is no single position to begin with. Our tracking guide covers why.
Also not a drop. A quiet fortnight.
Fewer enquiries can mean seasonal demand rather than lost visibility. Check whether you are still appearing before assuming you are not.
What does count. Three things together.
Sustained rather than momentary. Specific to identifiable searches. And commercially important, meaning the searches concerned actually produced work.
Where to confirm it. Your own reporting.
Compare a period against the same period earlier and look at whether impressions and clicks fell alongside position. If only the position number moved, be suspicious of the source.
What to establish while you are there. The date.
When did it start. That single fact does more diagnostic work than anything else on this page, because it lets you ask what else happened that week.
Why slowing down pays. Reversibility.
A change made in a panic is hard to undo once several more follow it. Half an hour of checking protects months of work.
Was It You Or Was It Everybody
Something happening to your site alone and something happening across the results are completely different problems with completely different responses. This question splits the investigation in two and it takes minutes to answer.
The test. Look sideways.
Did several unrelated searches move together. Or did one page slip while everything else held. Broad movement points away from any single page.
The second test. Look at competitors.
Search the term and see who is where. If the whole order rearranged and new kinds of site appeared, something changed in search rather than on your site.
What broad movement suggests. An update.
Search engines change how they assess pages regularly. Our algorithm updates guide covers what those are and how to respond.
Why that changes your response. Patience.
Reacting immediately to broad movement is how businesses damage themselves, since some of it settles on its own and none of it is fixed by editing one page.
What single page movement suggests. Something specific.
A change on your site. Or a competitor overtaking you on that particular subject. Blocks three and four take those in turn.
The check worth doing either way. Nothing formal.
Confirm there is no notice waiting in your search reporting. Rare, quick to rule out and it changes everything if there is one.
What Changed On Your Site
Most genuine drops follow something somebody did. The difficulty is that nobody remembers doing it, which is why the date from block one matters so much. Take that date and ask what happened to the website that week.
The obvious causes. Edits to the page.
Somebody rewrote it, shortened it, changed the heading or removed a section that was doing more work than anybody realised.
The less obvious ones. Structural.
A redesign, a change of address, a menu reorganisation or internal links removed. All of them affect pages nobody touched directly.
The one nobody suspects. A new page of your own.
Publishing something covering the same subject divides the site against itself. Our cannibalisation guide covers spotting it.
The technical ones. Accidental.
A page blocked during a rebuild and never unblocked, a broken redirect, a slow server. Your developer can rule these out quickly.
Why nobody remembers. No record.
Small changes get made constantly by several people. Without a dated log of what was altered, the connection is invisible.
What to do about that permanently. Start the log.
A dated note of every change to the site, kept from today. It costs nothing and it makes the next investigation straightforward.
What Changed On Theirs
Your page can slip without anything happening to it. Somebody else published a better answer. Being overtaken feels identical to being demoted. This is a genuinely common cause and almost nobody looks for it, because the instinct is to examine your own site.
How to check. Read what is above you.
Search the term and read the pages now ranking higher. Ask what they cover that yours does not, plainly rather than defensively.
What you usually find. Something specific.
A question answered that yours skips. Better detail. Clearer pricing. Or simply a more complete treatment of the same subject.
Why it feels unfair. Nothing broke.
Your page is as good as it was. The bar moved, which is how a competitive market behaves rather than a fault in anything.
The other version. A larger site arrived.
Occasionally a national brand or a large publisher decides to cover your subject. That is a different problem and it may not be winnable back.
What actually recovers it. Being better.
Not more words. Answering the search more completely than the page above you, which usually means adding the specific knowledge only somebody doing the work has.
The advantage you have. Real experience.
A large site writing generally cannot match detail from actual jobs. That is the ground on which a small business wins the position back.
When The Search Itself Changed
Sometimes a search comes to mean something different. No amount of improving a page written for the old meaning brings it back. This is the diagnosis nobody wants and recognising it saves months of effort aimed at an unwinnable target.
What it looks like. Different kinds of page.
The results are no longer the sort of page yours is. Guides where service pages used to be. Listings where guides used to be.
Why it happens. Behaviour shifts.
What people want from a phrase changes as circumstances change. Search engines follow that, so the results reflect a different need from the one your page was built for.
How to confirm it. Compare species.
Look at your page next to everything now ranking. If yours is a different kind of thing entirely, that is your answer, per our intent guide.
What the options are. Two, both legitimate.
Rebuild the page as the kind of page the search now wants. Or accept the search has moved away from you and redirect the effort.
What not to do. Keep polishing.
Improving a service page that a search no longer wants produces a better service page and no recovery. The work is not wasted and it is not solving this.
The related case. The results page changed shape.
New features can push everything down without your position changing at all. Fewer visits, same placement, nothing to fix on the page.
The Fixes Worth Trying, In Order
Now that you know what happened, fix that specifically. The order below runs from cheapest and most likely to most expensive and least certain. Stop as soon as the cause has been addressed.
First. Reverse the change.
If the date lines up with something somebody did, undo it. That is the fastest possible recovery and it is available surprisingly often.
Second. Fix the technical fault.
A blocked page, a broken redirect or a page that no longer loads properly. Quick to check, quick to correct and it happens more than people expect.
Third. Resolve the self competition.
If two of your own pages are serving one search, merge them or make one clearly the owner. The stronger page recovers as the division ends.
Fourth. Answer the search better.
Where a competitor overtook you, add what they cover and you do not. Specific and targeted rather than a general rewrite.
Fifth. Change what kind of page it is.
Where the search shifted meaning, rebuild rather than improve. Expensive, though the only thing that works in that case.
What comes with all of them. One change at a time.
Make one change, record the date and wait. Several at once means never knowing which worked, which matters more than speed.
How long any of this takes. We will not say.
It depends entirely on the cause, so any figure would be invented. Reversing an accidental change behaves nothing like rebuilding an outclassed page.
What Not To Do
These are the responses we see most and they consistently make things worse. All of them come from wanting to act rather than from any diagnosis, which is why the first six sections of this page exist.
Do not rewrite immediately. The commonest.
A full rewrite removes whatever was working alongside whatever was not. It also destroys your ability to identify the cause afterwards.
Do not change several things at once. Close second.
If something recovers, nobody knows which change did it. That teaches nothing and guarantees the same panic next time.
Do not delete the page. Rarely the answer.
A page that slipped still holds whatever standing it accumulated. Removing it discards that permanently for no benefit.
Do not add keywords. It was never that.
Working the phrase into more sentences does not recover a position and makes the page read worse, which our stuffing guide covers.
Do not buy anything hurriedly. A common reflex.
Sudden purchases of links or services in response to a drop are how ordinary problems become serious ones.
Do not assume the worst. Formal action is rare.
Most drops are ordinary competitive movement rather than anything punitive. Check for a notice, rule it out and stop worrying about it.
Do not act during broad movement. Wait.
If everything moved together, changes made now cannot be evaluated and some of it settles without intervention.
When Not To Chase It
Not every lost position is worth recovering. This is the question nobody asks during a drop, because losing something creates an urge to get it back regardless of whether it was producing anything.
The first case. It never produced work.
A search you ranked for that brought visitors and no enquiries. Losing it costs nothing and chasing it costs real effort.
The second case. The competitor is untouchable.
Where a national brand or a large publisher has moved in, the realistic assessment is frequently that this position is gone.
The third case. The search moved away.
Where a search now wants a kind of page your business has no reason to build. Rebuilding to chase it means writing something for the wrong audience.
What to do instead. Redirect the effort.
The same work spent on a search you can win produces more than a fight over one you cannot. That is the same reasoning as the two filters in our method guide.
How to decide. Ask about enquiries.
Did this search produce work. Could you realistically hold the position if you regained it. Two questions. Either answer being no means letting it go.
What to record. The decision.
Write down that you chose not to pursue it and why. Otherwise somebody revisits the same argument next year.
Where the rest sits. The hub.
All twenty-two guides are on the keyword research guide.
A full rewrite
removes what
was working
along with what
was not.
It also destroys any chance of identifying the cause afterwards. Most genuine drops follow something somebody did and nobody remembers doing it. That is why the date the drop started does more diagnostic work than anything else. Take that date and ask what happened to the website that week.
How we investigate a drop:
Some positions do not come back. Where a search now wants a different kind of page, the useful question is whether it still deserves your effort.
Every guide.
One subject.
What the words mean, what the numbers in a keyword tool actually are, how to find the searches worth targeting, how to turn a list into a publishing plan and what to do when a page you relied on slips.
