How to Track Keyword Rankings Properly
There is no longer a single position for a search. What appears depends on where somebody is, what they are using and what they have done before. So a tool reporting one number is reporting one sample, presented as a fact.
There Is No Single Position Any More
Two people searching the same words at the same moment see different results. Different order, sometimes entirely different pages. So the question of where you rank has no single answer, which means every report claiming one is answering a question that no longer exists in that form.
What varies it. Four things at least.
Where the searcher is standing. What device they are using. What they have searched and clicked before. And the shape of the results page itself, which differs by search.
Why location matters most. Especially for trades.
For anything a person needs performing at their property, results shift substantially across a town. Two customers a few streets apart can see a different order.
What the results page adds. Structural variation.
Maps, direct answers, product listings and question panels all occupy space. Being in a given position means something different depending on what sits above it.
What that does to old advice. Retires it.
Guidance built around a fixed ladder of positions assumes a page that no longer exists. The ladder varies by person and by search.
What checking yourself proves. Very little.
Searching your own term shows one personalised result from one place. A private window helps slightly and does not remove location, so the answer is still a sample.
Why A Tracker Reports A Number Anyway
A rank tracker performs a search from a chosen location, in a standardised way, then records what it found. That is a reasonable thing to do. The number it produces is accurate about that check and it is not a statement about where you sit generally.
What it does well. Consistency.
Checking the same way each time means changes are comparable. The direction of movement is more trustworthy than the number itself.
What it cannot do. Represent everybody.
One check from one place cannot describe what thousands of different people see. That is not a flaw in the product, it is the nature of the thing being measured.
The location setting. Worth getting right.
For a local business, checking from the wrong place produces a number describing a market you do not serve. Set it to where your customers actually are.
Why the reports look so precise. Presentation.
A clean column of positions with arrows beside them reads as measurement. Shown as a range with a note about method, nobody would treat it the same way.
Why we name none of them. The same reason as everywhere.
We do not name tools anywhere on this site, since recommendations date and naming one implies an endorsement we cannot maintain. Our tools guide sets out the position in full.
What to ask a supplier. Where they checked from.
If a report shows positions, ask which location the checks used. The answer tells you whether the numbers describe your market at all.
What Your Own Data Shows Instead
Your own search reporting shows your typical position across everybody who actually saw you, rather than one checked example. That is a genuinely different measure and for most purposes it is the better one.
Why it is different. It is an average of reality.
Real people, real searches, real placements, summarised. Nobody stood in one place and checked. The variation is included rather than eliminated.
What that means to read. A summary, not a spot.
The figure describes where you typically sat. Somebody nearby may have seen you higher and somebody further away lower. Both are inside the same number.
What else it gives you. Everything around it.
How often you appeared and whether anybody chose you, alongside the position. Those together answer questions a position alone cannot.
Its limitation. Only where you appear.
Searches you have never appeared for are invisible, which is precisely where a tracker still earns a place.
How to read it properly. Our guides.
Using it for research covers finding opportunities. the performance report guide covers what each figure means.
The cost. Nothing.
Free, plus already available on most sites. That alone should settle where reporting starts.
Position Is A Proxy, Not A Result
No business is paid for positions. Position is a proxy for being found. Being found is a proxy for visits. Visits are a proxy for enquiries. Each step away from the enquiry makes the number easier to measure and less worth knowing.
Why proxies get reported. They are easy.
Positions can be checked without any access to the business. Enquiries require somebody to know what actually happened, which is harder and considerably more useful.
The report that flatters. Improvements everywhere.
A document full of upward arrows on searches nobody buys from looks like progress. Nothing in the business changed, which is why the owner is unconvinced.
The report that alarms wrongly. Drops with no effect.
Positions slipping on terms that never produced anything reads as a problem. It is noise on searches that were not working anyway.
Where the proxy is genuinely useful. Early evidence.
Before enough enquiries have accumulated to judge anything, movement in position is the first sign something is happening. That is a real use.
The rule we apply. Never lead with it.
Position can appear in a report and never at the top of one. What happened commercially goes first. The proxies explain it.
What that changes in practice. The conversation.
An owner asked about enquiries engages. An owner shown a table of positions nods and stops reading.
What Is Actually Worth Tracking
Tracking everything produces a document nobody reads. A short list of things that would change a decision is worth considerably more. It takes less effort to maintain.
Enquiries by source. The one that matters.
How many people got in touch and roughly where they came from. Everything else on this list explains that number.
Searches bringing people in. Second.
Which searches actually deliver visits, rather than which ones you hoped would. That is in your own reporting already.
The handful of commercial terms. Third.
Positions for the small number of searches that genuinely produce work. Not the full list. Not the ones chosen because they look impressive.
Pages appearing without being found. Fourth.
Near misses, which are the cheapest available improvements and the most actionable thing on any report.
What to leave out. Vanity terms.
Broad searches the business likes the sound of, which it does not rank for and would not profit from. Tracking them only creates disappointment.
What to record alongside. What you changed.
A dated note of every change to the site. Without it, no movement can ever be explained afterwards. That note costs nothing.
How Often To Look
We give no interval, because the right frequency depends entirely on what you would do with the answer. The useful rule is to look only as often as you would act. Most businesses check far more often than they act.
Why constant checking harms. Manufactured anxiety.
Reported positions move daily for reasons nobody controls. Watching that closely produces worry about noise and tempts people into changing things that were working.
The worst pattern. Reacting to a single day.
A drop appears, somebody rewrites a page, the position recovers on its own and the rewrite gets the credit. That teaches the wrong lesson permanently.
When to genuinely look. Three moments.
After a deliberate change. When enquiries have shifted. At whatever regular review the business already holds.
What to compare against. Like periods.
The same span a year earlier, so seasonal patterns in a trade do not masquerade as performance. Most trades have a shape to their year.
What a review should ask. Two questions.
Did enquiries change. Does anything in the search data explain it. That is a short conversation with a clear outcome.
What to do between reviews. The work.
Time spent watching positions is time not spent improving pages, which is the thing that actually moves them.
Movement That Means Nothing
Most reported movement is not a change in anything. Recognising which is which saves a great deal of wasted effort, since the commonest response to a wobble is to alter something that was working perfectly well.
Small daily shifts. Almost always noise.
Positions move constantly as results are recalculated and as different samples get checked. Movement of a place or two says nothing on its own.
A single bad day. Not a trend.
One check showing a drop is one check. Wait for it to persist across a period before treating it as real.
Changes affecting everybody. Look sideways.
If several unrelated searches move together, something changed in search rather than on your site. Check whether competitors moved too.
Seasonal movement. Frequently misread.
Demand for a trade changes across the year. So does who is actively competing. A quiet period is not necessarily a decline.
What does deserve attention. Sustained and specific.
A meaningful drop on a commercially important search that persists and coincides with something you can identify. That is worth investigating properly.
Where the diagnosis lives. Its own guide.
Recovering dropped rankings covers working out what happened and what to do about it.
What To Put In Front Of An Owner
A business owner wants to know whether this is producing work. A report that opens with a table of positions is answering a question they did not ask. It is why so many of them stop reading marketing reports entirely.
What goes first. Enquiries.
How many, from where, compared against the same period last year. That is the entire subject as far as the business is concerned.
What goes second. The explanation.
Which searches and pages produced them, plus what changed. Now the search data is answering a question that was actually asked.
What goes last. Positions.
For the handful of commercially important searches, as supporting evidence. Included, never leading.
What to state plainly. The uncertainty.
Say that positions vary by person and place, once, in a sentence. An owner who knows that will not panic at a wobble later.
What to leave out entirely. The long table.
Dozens of tracked terms nobody will read. Length is not evidence of work and it obscures whatever the report was meant to say.
Where reporting sits properly. Its own material.
Our measuring performance material covers building a report a business owner will actually read. Everything sits on the keyword research guide.
Two customers a
few streets apart
see a different
set of results.
There is no single position, so a report claiming one is answering a question that no longer exists in that form. It checked from one place, in one way, at one moment. Meanwhile position is a proxy for being found and being found is a proxy for visits. Each step away from the enquiry makes the number easier to measure and less worth knowing.
What we put in a report, in this order:
If a supplier shows you positions, ask where the checks were run from. The answer tells you whether the numbers describe your market at all.
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.
