What SEO Metrics Matter Most to Business Owners?
One test settles it. If a figure could move substantially in either direction and nobody would do anything differently, it is decoration. Accurate decoration, frequently. Decoration all the same.
The Test For Any Metric
Take any figure in your report and ask what you would do differently if it doubled. Then ask what you would do differently if it halved. If both answers are nothing, that measure is not earning its place, however precisely it has been calculated.
Why this test works. It ignores accuracy.
Most arguments about measures are about whether they are correct. This one asks whether being correct would matter, which is a considerably more useful question.
What it catches. Things nobody acts on.
Plenty of accurate figures describe something real and connect to no possible decision. Those are the ones that fill reports without informing anybody.
What it lets through. Imperfect but actionable.
A rough figure somebody would act on beats a precise one nobody would. That reverses how measures are usually judged and it is the right way round.
Why owners find this easy. They already think this way.
Anybody running a business is used to asking what a number changes. The habit is only unusual in reports about this subject.
What the page does with it. Applies it.
Which measures pass, which fail, why the failing ones persist and when one of them is nevertheless worth looking at.
What Actually Matters To An Owner
Four questions. Every one of them passes the test comfortably. Notice that none is a search measure. That is not an oversight, it is the point.
Are more people getting in touch. The first.
Enquiries, calls, forms, bookings. The measure everything else exists to influence. Also the one most reports somehow omit.
Are they the right kind. The second.
More enquiries for work you do not want is not progress. An owner distinguishes these instinctively and no search measure can.
Are we reaching people who did not know us. The third.
The half of the work that is actually bought. Searches for your own name were already going to find you, which is why the split matters so much.
Is it worth what it costs. The fourth.
The question underneath the other three. Our guide on return covers why it is harder to answer than it looks.
What all four share. Consequences.
Each one, moving substantially, changes what a sensible business would do next. That is what separates them from everything in block three.
The Measures That Only Ever Rise
Some figures cannot fall. A measure that would have risen whatever happened during the period cannot tell you anything about what happened during the period. This category is larger than most people realise.
Anything cumulative. The clearest case.
Total pages published, total content produced, total anything since the start. These go up by construction, so an increase reports the passage of time.
Total searches you appear for. Close behind.
The count grows as a site accumulates pages, regardless of whether any of the new searches produce work. It rises during a bad quarter as readily as a good one.
Impressions in isolation. The respectable version.
Appearances rise as coverage widens and say nothing about whether anybody chose you. Genuinely useful beside clicks and misleading on its own.
Total traffic, unsegmented. The one that fools people.
It looks like a real measure. It hides the branded split, which means it can rise while reach is falling and nobody would know.
What they have in common. Comfort.
Every one of them is pleasant to receive. That is precisely why they keep appearing, which is block four.
Why They Persist
These measures survive because they suit both parties. The supplier gets something that reliably rises and the client gets reassurance. Nobody is being deceived and everybody is being misled, which is exactly why this is so difficult to dislodge.
What the supplier gets. A good month, always.
A figure that only rises produces a positive report in a difficult quarter. That is useful to somebody with a monthly conversation to have. It requires nothing underhand whatsoever.
What the client gets. Relief.
Somebody spending money every month wants to see something improving. A rising line satisfies that want. Wanting it is entirely reasonable.
Why nobody objects. Both sides are served.
An arrangement where everybody is content with the reporting has no internal pressure to change it. The correction has to come from outside, which is why it usually arrives as a cancellation.
What it is not. A conspiracy.
We are not suggesting suppliers set out to mislead. Most of this is habit, inherited templates and the ordinary human preference for good news.
Why that framing matters. It is fixable.
A problem caused by incentives can be corrected by changing what gets asked for. A problem caused by bad faith cannot, which makes this the more useful diagnosis.
Who has to move first. Whoever notices.
Frequently the client, since the supplier has less reason to. One question in one meeting is usually enough to start it.
The Ones That Look Serious And Are Not
Some measures can fall as well as rise, which makes them look like proper indicators, while still failing the test. These are harder to spot precisely because they behave like real measures.
Third party authority scores. The commonest.
An estimate produced by a company that sells software, describing your website rather than your business, used by no search engine. Our guide covers it fully.
Average position across a whole site. Genuinely misleading.
It moves for reasons unconnected to performance and frequently worsens while things improve, which makes acting on it actively harmful.
Engagement measures reported as results. A category error.
Time on page and pages per visit describe behaviour rather than outcomes. Useful as prompts and wrong at the top of a report.
Rankings for searches nobody makes. Technically true.
Appearing well for a phrase with almost no demand is a real achievement of no commercial consequence. The position is genuine and the search is empty.
The tell for all four. Apply the test.
Ask what you would do if any of them halved. The answer is usually to investigate, which means they are diagnostics rather than results.
What To Track Instead
Four measures. Between them they answer the four questions in block two. All are harder to produce than what they replace, which is the main reason they are less common.
Enquiries from search. The headline.
How many people got in touch having arrived from a search. Connecting the two takes setting up, which our guide on leads covers.
Non-branded visits. The reach measure.
People arriving from searches that did not contain your business name. The half the work is bought to affect. It is invisible in any combined total.
Visits to your commercial pages. The intent measure.
Traffic to the pages describing what you sell rather than to guides and articles. It distinguishes interest from readership.
Enquiry quality. The one nobody tracks.
Whether the enquiries are work you want. A rough note against each one is enough. It is more informative than any figure in a search tool.
What makes these harder. They need you.
Every one requires something from the business rather than from a tool, which is precisely why they cannot be automated away.
When One Is Legitimately Useful
Almost every measure criticised above is genuinely useful to somebody doing the work. The distinction is between a diagnostic and a result. The failure is one of placement rather than of the measure itself.
What a diagnostic does. Points somewhere.
It suggests where to look rather than telling you how things are going. That is a real contribution and it belongs in working notes.
The clearest example. Impressions.
Useless as a headline and genuinely valuable as an early sign that a site is starting to be understood more broadly.
The second. Engagement measures.
One page behaving unlike everything similar to it is worth investigating. That is a prompt for the practitioner rather than news for the owner.
The third. Authority scores, narrowly.
As a rough way of sorting a list of sites, they do a job. As something to report to a business, they do not.
The rule that follows. Placement decides.
The same figure can be useful in one document and misleading in another. Nothing is wrong with the number and something is wrong with where it was put.
Ask What Would Change
You do not need to know anything about this subject to apply the test. Point at any figure and ask what would happen if it doubled or halved. The quality of the answer tells you what you needed to know.
The good answer. Specific.
If that halved we would look at these pages, because it usually means this. Somebody who has thought about the measure can say what it connects to.
The revealing answer. Vague.
It would show progress is slowing. That describes the figure rather than naming a consequence. It is what you hear when nobody has asked the question before.
What to do with a bad answer. Suggest removing it.
Ask whether it needs to be in the report at all. A shorter report with fewer measures is easier to produce and easier to act on.
What replaces it. Something from block six.
Not nothing. The space freed should go to a measure answering one of the four questions an owner actually has.
Where this lands. The document.
Our reporting guide covers what a report should contain. Everything sits on the measuring performance guide.
Nobody is being
deceived and
everybody is
being misled.
A measure that only ever rises produces a positive report in a difficult quarter, which is useful to somebody with a monthly conversation to have. Meanwhile a business spending money every month wants to see something improving. Wanting that is entirely reasonable. Both sides are served, so nothing inside the arrangement pushes back. That is not a conspiracy. It is habit, inherited templates and an ordinary preference for good news.
What we report instead:
Point at any number in your report and ask what you would do if it halved. If the answer is nothing, it is decoration however accurate it is.
Every guide.
One question.
What each number actually means, how to read the data without being misled, what belongs in a report and what does not, then how to connect any of it to the enquiries the business is actually paying for.
