How to Measure Whether Your SEO Agency Is Delivering Results
Is the campaign working, then is the agency doing its job. Two questions that get confused constantly, whose answers can differ in both directions. Separating them is the only way to decide anything sensible about an arrangement you are unhappy with.
Two Different Questions
Whether the campaign is working and whether the agency is doing its job are separate questions. They get treated as one. The answers genuinely can differ in either direction.
Why they get merged. Only one is visible.
You can see the numbers and you cannot see the work, so the numbers become the proxy for both. That substitution is understandable and it produces wrong decisions in both directions.
The cost of merging them one way. Losing a good firm.
A business in a difficult market dismisses an agency doing everything correctly, hires another, then receives the same numbers a year later having paid for the transition.
The cost of merging them the other way. Keeping a poor one.
Numbers looking acceptable for reasons unconnected to the work keep a firm in place while very little is done. This version costs more because it persists longer.
What this page covers. The agency only.
What the individual metrics mean and which are worth watching belongs in our measuring performance material. This page is about assessing the firm you are paying.
The Agency Can Be Fine While The Numbers Are Not
Four situations where results disappoint despite good work. Each is checkable by you rather than merely assertable by them, which is what separates an explanation from a defence.
A very competitive market. Check it yourself.
Search your main terms and look at who occupies the results. If they are national companies or large platforms, the difficulty is real and visible to you in a minute.
A worse starting position than anybody realised. Check the audit.
Sites carry accumulated problems that only surface once somebody examines them properly. Ask what the audit found and whether the early work was remedial rather than additive.
A site the business will not let them change. Check your own decisions.
Requested changes refused, delayed or blocked by whoever controls the website. Businesses forget these refusals and the agency rarely raises them twice.
The business has not supplied what was asked for. Check the requests.
Information never provided, approvals outstanding, access delayed. Look back at what was asked of you and whether it was answered, because this is the commonest cause of stalled work.
What makes all four legitimate. They are verifiable.
Each can be confirmed or dismissed by you. An explanation you cannot check is not on this list. That is the test to apply to anything else you are told.
The Numbers Can Look Fine While The Agency Is Not
Three mechanisms flatter a period in which very little was done. Each is common and none requires anybody to deceive you. Together they explain why acceptable figures are not proof of anything.
Inherited momentum. Somebody else's work.
Search results respond slowly, so improvement arriving now may come from work done before this firm started. A new agency's first months frequently benefit from the last one's efforts.
Seasonal movement. The calendar rather than the work.
Almost every business has a busy period, so figures rising into it look like progress. Comparing against the same months last year removes this. Reports frequently compare against the previous month instead.
A competitor's problem. Nothing you did.
A competitor closing, redesigning badly or stopping their own work improves your position without anybody at your agency doing anything. It reverses just as arbitrarily.
Why this matters more than it sounds. Comfort is expensive.
Acceptable numbers remove any impulse to ask questions, so a firm doing little can continue for years. The version of this problem where numbers are poor gets addressed far sooner.
How to test for it. Ask what caused it.
Ask your agency why a good period was good. A firm that did the work can attribute it to specific activity. One that cannot will describe the improvement rather than explain it.
Judge The Agency On Four Things
Four questions that assess the firm rather than the outcome. A business can answer all four without understanding any of the technical work.
Is work being produced. The easiest.
Something written, fixed or changed each month, which you could point at. This alone eliminates a surprising proportion of unsatisfactory arrangements.
Is it the right work. The hardest, with its own block.
Whether the activity connects to the objective you set. Block five covers what a non-specialist can reasonably assess here.
Is it being measured truthfully. The reporting question.
Whether you receive figures against a stated comparison, including in months that went badly. How an SEO agency reports on results covers what a report should contain.
Does the plan change in response. The one that separates firms.
Whether what they do next month reflects what happened this month. Block seven is the question that establishes it.
How to use the four. All must hold.
An agency failing any one of them has a problem worth raising. Failing two or more is a different conversation. None of this requires you to assess the work itself.
Is The Right Work Being Done
You cannot judge whether a technical decision was correct. You can judge two things that get you most of the way there. Neither requires knowing anything about search.
Does it connect to the objective. The one you set.
You told them what work you wanted more of. Look at what has been produced and ask whether it aims at that. Content about services you do not want to sell is visible to you without any expertise.
Can they explain why it was chosen. In terms you follow.
Ask why this rather than something else. Somebody who decided has a reason connected to your business. Somebody following a template describes what the item is.
What that combination establishes. Direction and judgement.
Together those two tell you whether the work is aimed at your objective and whether somebody is choosing rather than processing. That is most of what matters.
What you are not expected to assess. Technique.
Whether a technical approach was optimal is not your job and never becomes your job. An agency implying you should have an opinion is transferring its own responsibility.
The signal to watch for. Explanations that never land.
If every question about why produces something you cannot follow, either it cannot be explained simply or somebody would rather you did not follow it.
Reporting Is Not Fixing
Consistent, accurate reporting of a decline that nothing is being done about is a failure of the agency rather than of the campaign. This is where that distinction does the most work.
Why it belongs on this page. It resolves the confusion directly.
A declining campaign might be nobody's fault. A declining campaign that is diligently documented and never addressed is a specific failure by a specific party. It is easy to mistake the documentation for a response.
What it looks like in practice. Three reports, one story.
Consecutive months describing the same fall, each accurate, each professionally presented, with no change to what is being done in between. That is a record of watching.
Why it persists. The reports feel like accountability.
A document arriving on time, acknowledging a problem, reads as a firm taking responsibility. Nothing about it commits anybody to acting. Clients regularly accept the acknowledgement as the action.
What should have happened instead. A changed plan.
A decline is information. The correct response is a different approach next month, explained. If the plan is unchanged after three declining reports, nothing has been learned from any of them.
How to raise it. Without accusing anybody.
Ask what they intend to do differently given the last quarter. That question is answerable by a firm that has been thinking and awkward for one that has not.
Ask What Changed Because Of What You Learned
An agency whose plan never responds to results is executing rather than working. One question establishes which you have. It works whether the numbers are good or bad.
Why it works in both cases. Learning happens either way.
A good period should teach something worth repeating and a poor one something worth avoiding. A firm that learned nothing from either was not paying attention to either.
What a good answer sounds like. Specific and consequential.
A named finding, what it implied and what they did about it. Somebody who reviews their own work has several of these available immediately.
What a poor answer sounds like. Continuation.
That they are continuing with the plan, else that these things take time. Both describe not changing. Neither answers what was learned.
Why this question is hard to fake. It requires a history.
Answering it means having actually reviewed something and altered course. That cannot be produced on the spot, unlike an assurance about commitment or effort.
What it connects to. The thinking time.
Reviewing, deciding and adjusting is the highest value part of what you buy and the first thing squeezed when a firm is busy. This question is how you check it is happening.
Give It The Time It Needs, Not More
This work needs months. That fact is also the industry's most convenient explanation for nothing happening. Distinguishing the two is the last judgement this page asks of you.
What reasonable patience looks like. Waiting for outcomes.
Work is being produced, the plan is coherent, the reporting is candid and the approach adjusts. In that situation, waiting is the correct response and impatience costs you.
What avoidance looks like. Waiting for activity.
Nothing produced, no plan, no response to results, with time invoked as the answer. That is not patience being asked of you, it is an excuse for the four things in block four.
The distinction in one line. Which thing is missing.
Slow outcomes justify waiting. Absent work does not, because no amount of time turns nothing into something.
The early version of this. Covered separately.
In the first months the picture is different again. What to expect in the first three months sets out what should be visible then.
If you conclude it is avoidance. Diagnose before deciding.
When to change your SEO agency covers doing it properly, including the cases where staying is the better answer. The full series is on the hiring an SEO agency guide.
What changed
because of what
you learned?
Ask it of us and of whoever you currently pay. It is hard to fake, because answering means having reviewed something and altered course, which cannot be produced on the spot. A firm that has been thinking has several answers available immediately.
What we commit to in writing:
Local £350 a month fixed. National £1,550 a month. Published rather than quoted on request.
Every guide.
One practice.
What an agency is, whether to buy at all, which model suits you, how to vet one, what things cost, what should be in a contract and how to work with them afterwards.
