How to Tell if SEO Is Actually Working
Usually asked three or four months in, usually while working out whether to keep paying. It deserves a direct answer rather than a lecture about metrics, so this includes the questions to put to whoever is doing the work and what to do if the answer turns out to be no.
You Are Asking Because You Are Not Sure
Nobody searches this while things are obviously going well. The question usually arrives a few months into an arrangement, after several reports that were perfectly pleasant and did not answer it. That is a reasonable position to be in and a reasonable question to ask.
Why it arrives when it does. The novelty wears off.
The first reports are interesting because everything is new. By the third or fourth, the charts look similar and the underlying question surfaces.
Why it feels awkward to ask. It sounds like an accusation.
It is not one. A supplier confident in their work will welcome it, since it gives them a chance to show what has moved.
How a good firm responds. Directly.
With what has changed, what has not and why. Deflection, jargon or a longer report in response to a simple question all tell you something.
What this guide will not do. Tell you to switch.
We do this work, so there is an obvious interest in you concluding your current arrangement is failing. What follows is written to be usable against us as easily as against anybody.
What it will do. Give you tests.
Signals that are real, signals that deceive, questions with checkable answers and the four possible explanations when the answer is no.
One thing to settle first. What you were promised.
Before judging anything, recall what you were told to expect at the outset and roughly in what order. If nothing was said, that absence is itself part of the answer and it shapes everything below.
What Should Be Moving By Now
We will not tell you how long any of this takes. It depends on how competitive your trade is, where the site started and how much was wrong with it, so any figure would be a guess dressed up as expertise. What can be described accurately is the order things happen in.
First. Things get fixed.
Pages that could not be found become findable, obvious faults get corrected and the site becomes properly readable. Invisible from outside, though everything else depends on it.
Second. The range widens.
The site starts appearing for more different searches, mostly specific ones nobody deliberately targeted. Block three explains why this comes first.
Third. Something specific starts ranking.
A narrow page built deliberately begins appearing properly. Usually not the term the business cares most about. It is the first visible win.
Fourth. The first enquiries.
Somebody gets in touch having arrived through search. Frequently from a page nobody expected to produce anything.
Fifth. The commercial terms move.
The searches the business actually wanted. These come last because they are the most contested. For some businesses they never fully arrive.
How to use this. Locate yourself.
Work out which stage you are at rather than which month you are in. That is checkable. A date on a calendar tells you nothing about whether the approach is sound.
What should be true at every stage. Something changed.
Regardless of results, work should be visibly happening to the site. That is block six.
The Earliest Real Signal
Appearing for more different searches almost always comes before appearing higher for the ones that matter. That means the earliest genuine evidence is not a position at all, it is the number of separate searches your site is being seen for.
Why it works that way. Breadth before height.
As a site becomes better understood, it starts being shown for a wider range of things. Improving on any single contested search takes considerably longer.
Where to see it. Your search reporting.
The count of distinct searches your site was shown for over a period. Free, plus it sits in the same place as everything else, per our measurement guide.
What a healthy pattern looks like. Steady widening.
More distinct searches month by month, mostly ones nobody chose. That is a site being understood better rather than a site being pushed.
Why it is trustworthy. Hard to fake.
It cannot be manufactured by tracking a flattering list, unlike a ranking report. It reflects what search engines actually did with your pages.
The qualification. Relevance still applies.
Appearing for a wider range of irrelevant searches is not progress. The widening should be in the direction of what you sell.
How long to watch it. Longer than a month.
A single month of widening could be ordinary variation. The same direction holding across several periods is the pattern. That is what makes it evidence rather than a hopeful reading.
What it does not prove. Money.
It is a leading indicator rather than an outcome. It tells you something is moving. Block five is where the actual question gets settled.
Signals That Look Like Progress And Are Not
Some numbers rise convincingly while nothing improves. These four appear in reports constantly and none of them means what it appears to mean, so knowing them protects you from being reassured by the wrong thing.
Traffic from somewhere unrelated. The commonest.
A page attracts visitors interested in something adjacent to what you sell. The total rises, the visitors were never customers and nothing about the business changes.
Rankings for terms nobody searches. The flattering one.
Appearing well for phrases with no demand behind them. Easy to achieve, impressive in a table and worth nothing, which our rankings guide covers.
Impressions with no clicks. The subtle one.
Being shown far more often while nobody chooses you. That usually means appearing for things you do not serve. Or appearing too far down to be seen.
A rise that matches the season. The one everybody misses.
Most trades have a shape to their year. A rise arriving exactly when it always does is your trade rather than your supplier. Comparing against the same period last year settles it.
The branded search version. Worth knowing.
Searches for your business name count as search traffic even though those people already knew you. A rise driven by your name is a rise in recognition rather than in reach.
The test that catches all five. One question.
Did more people who had never heard of you get in touch. If a rise cannot survive that question, it was not progress.
Why they get reported anyway. They are true.
Every one is an accurate number. That is what makes them hard to challenge and why knowing the patterns matters more than doubting the figures.
The Question Behind The Question
Whether SEO is working means whether more of the right people are getting in touch. Everything else in this guide is a step towards that. It is the measure that settles the argument. It is also, in most businesses we see, not being recorded at all.
Why it goes unrecorded. The phone.
For most trades the valuable outcome is a call. A call is invisible to anything measuring a website. The chain between the work and the result breaks at that point.
What that produces. Arguments about traffic.
With no outcome data, the conversation stays at visitor level permanently. That is why so many of these relationships never resolve.
The simplest fix. Ask every caller.
Recording how each enquiry arrived, consistently, gives better attribution than most technical solutions. It needs a habit rather than a budget.
What to count. New people.
Enquiries from people who did not already know the business. Existing customers finding you by name are welcome and they are not evidence of anything working.
What to weigh. Not all the same.
Counting enquiries without regard to value hides the difference between a small repair and a full installation, which matters enormously in the trades.
Where this is covered fully. Its own guide.
Connecting traffic to leads and revenue covers the attribution problem properly, including why search is usually underreported rather than overreported.
Ask Your Agency These
Six questions. A firm doing the work will answer all of them easily and will probably enjoy being asked. A firm that cannot answer them is telling you something the report cannot. That holds regardless of whether the numbers look good.
What did you actually do this month. The essential one.
Specific changes to specific pages. If the answer describes monitoring, reporting and analysis without naming anything that changed, the work has become observation.
What was that expected to achieve. The follow up.
Every change should have had a purpose before it was made. Work with no stated expectation cannot be evaluated afterwards by anybody.
Did it do what you expected. The one that reveals most.
Sometimes the answer should be no. A supplier who reports that everything always worked is either lucky beyond belief or not looking.
What happens next if it did not. The plan.
A specific next step rather than continuing as before. This is the question most reports never address.
Which searches actually produce enquiries. The commercial one.
Not which ones rank. Which ones bring people who get in touch. A supplier who does not know is not connected to your business.
What would make you tell me to stop. The revealing one.
Ask what circumstances would lead them to advise ending the arrangement. Somebody who has thought about it has your interests in view. Somebody who says nothing would has not.
How to use the answers. Look for specificity.
The content matters less than whether the answers are concrete. Vague replies to concrete questions are the finding.
What If The Answer Is No
Four possibilities, needing completely different responses. Assuming the first one is how businesses lose a year. Assuming the last one is how they lose the progress they already made.
One. Not enough time has passed.
The commonest explanation and the one we list first, despite it being the answer least convenient for anybody selling this. If the sequence in block two is progressing, the work is probably fine.
Two. The wrong things are being measured.
The work may be producing results nobody is capturing, particularly where enquiries arrive by phone. Fix the measurement before judging the work.
Three. The work is being reported rather than done.
Reports arrive, nothing on the site changes. Block six catches this. It is the failure this whole series is built around.
Four. The approach is wrong for the business.
Sometimes the work is real, competent and aimed at the wrong searches or the wrong audience. That is a strategy conversation rather than an effort problem.
How to tell them apart. Sequence, then activity.
Check the stages against block two. Then ask block six. Between them they separate the first three. The fourth usually emerges from the answers.
What each needs. Different things.
Patience, better recording, a direct conversation or a change of plan. Applying the wrong remedy to the wrong cause makes things worse.
The one to rule out first. Measurement.
It is the cheapest to fix and it is frequently the whole answer.
When To Stop Waiting
There is a point at which patience becomes avoidance. We cannot tell you when that is, because it depends on your trade and where you started. Any figure we gave would be quoted back at somebody unfairly. What we can give you is the conditions.
Condition one. Nothing is changing on the site.
If you cannot identify specific work done over recent months, the question is no longer about results.
Condition two. No stage has been reached.
If none of the sequence in block two has occurred, not even the widening in block three, something is wrong beyond patience.
Condition three. Nobody set expectations at the start.
A firm should have told you what to expect and roughly in what order. Without that, there is nothing to judge against. That is itself a finding.
Condition four. The answers stay vague.
Concrete questions producing evasive answers repeatedly. That pattern rarely improves.
What is not a condition. A slow month.
Nor a quarter that flattened, nor a competitor doing well. Those are ordinary and they are not evidence of failure.
What to do before ending anything. Say it out loud.
Tell the firm plainly that you are unconvinced and ask what they would change. A good one responds with a plan and a bad one responds with a longer report.
Where to go next. Selecting a firm.
Our guides on hiring an agency cover what to look for and what to avoid. Everything here sits on the measuring performance guide.
Ask what was
actually done
this month.
If the answer describes monitoring, reporting and analysis without naming a single thing that changed on the site, the work has become observation. A firm doing the work answers that easily and usually enjoys being asked. The content of the answers matters less than whether they are concrete, because vague replies to concrete questions are themselves the finding.
Put these to whoever is doing your SEO:
We will not tell you how long it takes. It depends on your trade and where you started, so any figure would be a guess dressed up as expertise.
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.
