Contents
I sit on both sides of this transaction. I appoint and manage external SEO support as part of an in-house team, and I take on consultancy work independently. Those are different vantage points, and they disagree more often than you would expect.
What follows is what I look for before appointing someone, and by extension what you should expect from me.
The short answer
Judge a consultant on how well they diagnose your specific problem, not on how confidently they describe their process. Before you compare prices, establish four things:
- What they think is limiting your growth, and what evidence would change their mind.
- What they would deliberately not do first.
- What your team has to approve, produce or build for their plan to work.
- How they will separate work completed from progress made from money earned.
What kind of buyer are you?
Most guidance on this subject is written for one reader. The two situations are genuinely different, and so are the ways they go wrong.
| If you are | Your main risk | So the thing to test hardest is |
|---|---|---|
| A founder or small business | Buying a strategy you cannot afford to implement, or cannot implement at all | Whether the plan is sized to what you can actually build and publish |
| An in-house marketing lead | Hiring someone whose recommendations cannot survive your internal priorities, approvals and release schedule | Whether they have worked inside a business like yours, not only on sites like yours |
If you are the founder, the danger is a good plan you cannot act on. If you are the marketing lead, the danger is a good plan that dies in a backlog. Both are expensive, and neither is the consultant's fault alone.
Before you shortlist anyone
Write down what has actually happened before you speak to anybody. “Organic traffic fell 40% in March and has not recovered” is a brief. “We need SEO” is not. The first lets someone tell you within a conversation whether they have seen it before. The second invites a generic proposal, because you have given them nothing else to respond to.
Then decide what you are actually buying, because these are not interchangeable:
- A diagnosis. A bounded piece of work to establish what is wrong. The right start when nobody can explain the problem.
- Strategy. Deciding what to do and in what order, once the diagnosis exists.
- Implementation. Someone doing the work rather than specifying it.
- Ongoing consultancy. Continuous prioritisation as the site and the market move.
- Fractional leadership. Someone owning the function part-time, including managing your other suppliers. Covered in fractional SEO director versus SEO freelancer.
Buying strategy when you needed implementation is the most expensive mistake on this page, and the easiest to avoid.
What you need in place before anyone starts
Name a person against each of these. If you cannot, fix that first.
- Who ships technical changes, and how long their queue is.
- Who writes content, and who approves it.
- Who can authorise changing the navigation, retiring pages or altering templates.
- Who owns the Search Console and analytics accounts.
Saying this out loud in a first conversation is not a weakness. A consultant who adjusts their proposal in response is showing you how they will behave later.
The first conversation: five questions I use
These are the five I ask when I am the one appointing. They are chosen because the answers are hard to rehearse, and because the follow-up matters more than the answer.
-
What do you believe is limiting our organic growth, and what evidence could change your view?
Why I ask: it tests whether they hold a hypothesis or a script. Everyone has an opinion after ten minutes. The useful part is knowing what would falsify it.
Weak A list of everything that could be wrong, with no ranking and nothing that would change it.
Stronger “My first guess is that your category pages are competing with each other, based on what I can see from outside. If your Search Console data shows those pages holding steady impressions, I am wrong, and I would look at demand instead.”
My follow-up “How long would it take you to find out which it is?”
-
What would you deliberately not prioritise in the first 90 days?
Why I ask: anyone can produce a list of things to do. Prioritisation is the actual skill, and it only shows when someone names what they are leaving out.
Weak “We would look at everything.” That is a menu, not a plan.
Stronger “I would not touch link acquisition until the template issues are fixed, because you would be sending authority to pages that cannot convert it.”
My follow-up “What would have to change for that to move up the list?”
-
What must our team approve, produce or implement for your strategy to work?
Why I ask: this is the question that decides whether the engagement succeeds, and it is the one most often left unasked. A recommendation nobody can ship is a cost with no return.
Weak “We will hand over full recommendations and your team can take it from there.”
Stronger “I will write the specifications and the acceptance criteria. Your developers ship items one to four, and I need roughly two days of their time per sprint. Items five and six need a decision from whoever owns the navigation before anyone can start.”
My follow-up “What happens to the plan if we can only give you half of that?”
-
How will you separate completed activity, SEO progress and commercial impact?
Why I ask: most reporting conflates the three, which is how an engagement looks busy and successful while earning nothing. Keep them apart from the start:
- Delivery. Recommendations implemented, time to release, unresolved dependencies.
- SEO outcomes. Non-brand clicks, visibility on priority pages, eligible landing pages, organic conversions.
- Business outcomes. Revenue, contribution margin, qualified leads or sales.
Weak A monthly report of tasks completed, presented as performance.
Stronger Three tiers reported separately, with an explicit statement of what the numbers cannot yet tell you.
My follow-up “Which of those will move first, and roughly when?”
The right measure also depends on the problem. A technical recovery is fairly judged first on restored indexation; a growth programme on non-brand revenue. Insisting on revenue from month one for a project whose first job is making pages indexable sets up an argument nobody needs.
-
Tell me about a recommendation that stalled internally. What did you change?
Why I ask: everyone with real experience has had work blocked by something that had nothing to do with SEO. How they responded tells you whether they will work with your constraints or complain about them.
Weak A story in which the client was the problem and nothing was learned.
Stronger A specific case where the recommendation was re-scoped to fit what could actually be released, with an honest account of what was lost by doing so.
My follow-up “What would you ask us now that you did not ask them then?”
How to assess the proposal
If three quotes differ by a factor of four, they are almost certainly not quotes for the same work. Normalise before you compare price.
| Compare on | The question that reveals it | What a vague answer is hiding |
|---|---|---|
| Scope | What exactly is delivered, and what is explicitly excluded? | Room to reinterpret the deliverable later, in their favour. |
| Seniority | Who does the work day to day, and how many of their hours do I get? | The person pitching is not the person you get. |
| Implementation | Does this price include doing the work, or only specifying it? | A second, unbudgeted cost landing on your development team. |
| Commitment | What is the minimum term, and how do I exit? | You are locked in past the point where you can tell it is failing. |
A proposal worth accepting states the problem being investigated and the assumptions behind it, what is included and excluded, who completes and who implements each item, how priorities will be decided, how success will be measured, and the fees, dependencies, review points and exit terms. If it reads as a list of activities with a monthly price attached, you are being sold capacity rather than judgement.
How price relates to scope, seniority and complexity is covered on how much an SEO consultant costs.
Verifying what they claim
Screenshots of traffic graphs prove very little. Traffic rises for many reasons, including seasonality, paid activity and competitors making mistakes. What you are testing is attribution.
- What was the starting position? A percentage with no baseline is meaningless. Going from 200 to 600 visits is a 200% increase, but you cannot judge what it was worth without knowing which visitors increased and what they produced. In a high-ticket market those 400 visits might matter a great deal.
- What did they personally do? On a team of eight, “we grew organic revenue 40%” may be true and have very little to do with the person in front of you.
- What else was happening at the time? A new site, a TV campaign, a competitor collapsing. Someone who volunteers the confounders unprompted is telling you something useful about their honesty.
What to ask for depends on the problem:
| Your situation | Evidence to request |
|---|---|
| Traffic decline | A timeline, which templates were affected, the diagnosis reached and what was done about it |
| Migration | The redirect plan, before-and-after crawl data, and how launch was monitored |
| Ecommerce growth | Category strategy, how faceting and indexation were handled, and the commercial outcome |
| Manual action or unnatural links | The cleanup method, what was removed versus disavowed, and the risk controls applied |
| International | Hreflang approach, market architecture, and performance by region |
Note that “penalty” is a loose word. Ask whether they mean a manual action, an algorithm-related decline or an unnatural-link problem: different situations, different remedies, and someone who uses the terms interchangeably may not distinguish them in practice either.
Checking a reference properly
“Can you provide references” produces a name that will say something positive. Ask that person better questions. Did the consultant communicate clearly when results were slow? Were the recommendations practical given their team and platform? How often did the scope change, and who raised it? Did the person who pitched the work stay involved afterwards?
The last one is the most predictive, and the least likely to be volunteered.
What to agree before signing
- The problem, in your words, with their response to it specifically.
- Deliverables listed explicitly, and exclusions listed too.
- Who implements each item.
- Success measures, in writing, separated into delivery, SEO and commercial.
- Minimum term, and how to exit.
- Who you will actually deal with each week.
- What happens if implementation is delayed on your side.
- Who owns the documents, accounts, data and work produced during the engagement.
- Nothing guaranteed that cannot be guaranteed. Nobody controls Google's ranking systems, so a promise of a particular position is either meaningless or a reason to walk away; the same now goes for promises about appearing in AI answers. Google says as much in its own guidance on hiring an SEO.
Access and account ownership
Give the minimum that lets the work happen, and keep what is yours:
- Read-only to begin with. Search Console and analytics access should be read-only during an audit. There is no diagnostic reason to need write access.
- You own the accounts. Search Console, analytics, the CMS and any paid tools stay in your name, with the consultant added as a user. If a supplier owns your accounts, changing supplier means losing your history.
- Changes get documented before release. So that when something moves, you can tell whether it was the work or the market.
- No third-party tool is “Google approved”. Nobody has that status.
For larger organisations
If you are buying inside a business that has a procurement function, three things will shape the engagement more than the SEO conversation does. Raise them early rather than discovering them in week six:
- Supplier onboarding often takes longer than the first piece of work. Security review, data processing agreements and purchase orders can add weeks before anyone starts. Ask whether the consultant has been through it before; the answer changes how you sequence the first month.
- Development capacity is usually the real constraint, not budget. A recommendation competing with a replatform or a peak-trading code freeze will not ship, however good it is. Say so up front and let the plan reflect it.
- Someone needs to own escalation. When SEO work is blocked by another team's priorities, there has to be a named person who can resolve it. Without one, the engagement quietly becomes advice nobody acts on.
What should happen in the first 90 days
A reasonable engagement front-loads diagnosis and produces decisions rather than findings.
-
Understand the business
Goals, margins, which products matter, and what the site currently does well and badly.
Output: agreed goals and priority commercial areas.
-
Establish the baseline
A crawl, a Search Console and analytics review, a look at the platform and templates, and a competitive gap analysis. Depth depends on the site, the access available and the scope agreed.
Output: a baseline, and an explicit note of what the measurement cannot tell you.
-
A prioritised plan, not a list
Quick wins separated from short, medium and long-term work, ordered by likely commercial impact against effort required.
Output: each action with an owner, an effort and impact estimate, its dependencies and a timeframe.
-
Ship something
Ninety days with nothing released is a warning sign, even when the diagnosis was sound. Something should have gone live, and you should be able to name it.
Output: implemented changes, and a record of what was blocked and why.
-
Correct course
The first plan is a hypothesis. By day 90 there should be evidence that changes part of it, and a consultant willing to say which part.
Output: a revised plan, with the reasoning for what changed.
How long results take depends on what is wrong. Technical fixes can move things in weeks; recovering from a mishandled migration, or building authority in a competitive market, takes months. Be careful not to confuse early movement with commercial impact: rankings and impressions often move before clicks do, and clicks before revenue.
Frequently asked questions
Should an SEO audit be free or paid?
A free audit is a sales document. It is produced quickly, from tools rather than judgement, and its job is to find enough problems to justify a proposal. A paid audit is the deliverable itself, and the consultant is accountable for whether its conclusions were right. If you would not act on the findings without a second opinion, you did not buy an audit.
What access should I give an SEO consultant before hiring them?
Usually none beyond what is publicly visible. A consultant can form a view from a crawl of your public site and from what they can see of your competitors. Search Console and analytics access, read-only, belongs at the start of a paid engagement rather than during a pitch. Keep ownership of every account in your own name.
Should I choose an SEO consultant with experience in my industry?
It helps, but it matters less than experience of your type of problem. Someone who has handled a large ecommerce migration will be more useful on your migration than someone who happens to share your sector but has only worked on small brochure sites. Ask about comparable problems, not comparable logos.
Is it a red flag if an SEO consultant will not name their clients?
Not on its own. Plenty of good consultants work under NDAs, particularly with larger brands. What matters is whether they can describe the problem, their specific contribution and how the outcome was measured in enough detail to be credible, and whether they can offer a reference even if they cannot publish a name.
How long should I commit to an SEO consultant?
Commit to the diagnosis first, not the programme. A short paid engagement that establishes what is actually wrong gives you a far better basis for deciding on a longer commitment, and a consultant confident in their work will not object to being judged on it.
Assessing my consultancy by the same standards
It would be inconsistent to publish all of that and then ask you to skip it. Roles change; this reflects the position at the time of writing.
Frasers Group, 2019 to present
Head of SEO, leading the group's SEO team across Sports Direct, House of Fraser, Flannels, Studio and GAME. It is a leadership role over a team and a set of agency partners, so the outcomes are the team's rather than mine alone. What is mine is the strategy, the roadmap and the decisions about where effort goes. It is also where the buyer's half of this page comes from.
New Look, 2014 to 2018
SEO Marketing Manager, managing global SEO performance on a six-figure budget and leading both in-house and agency resources. In 2018 New Look sat top of the UK fashion leaderboard for commercial search terms. That is a third-party index rather than my own reporting: it scores brands on visibility across a fixed basket of the highest commercial-value terms in a sector, which is a harder measure to move, and to flatter, than keyword count or traffic. A team delivered it. I set the direction and owned the channel.
Wrappz, 2013 to 2014
I led an SEO recovery after a previous agency's black-hat link building had cost the site rankings, traffic and sales. The 35,000-plus figure is the count of unnatural inbound links identified during that audit, not a count of links removed: some were removed at source, the rest disavowed. I am not going to quote a recovery percentage, because I do not have a documented one, and inventing one here would fail my own test above.
Things you can check without asking me
I build and publish SEO tools, which is verifiable evidence of how I work rather than a claim about it: the AI Visibility Prompt Generator, Organic Trends for Search Console and GA4 analysis, and the AI Product Feed Readiness Checker. There are more on my profile. Use them, break them, judge the thinking behind them.
My consultancy work is undertaken independently of my role at Frasers Group, and is limited to a small number of suitable engagements.
Not sure whether you need an audit, a consultant or ongoing support? Describe the problem in a few sentences and send it over. I will tell you what I would investigate first, what information I would need, and whether I am the right person to help.