What SEO consulting is, and what separates it from delivery
SEO consulting is diagnostic and advisory work: establishing why a site does not earn the search visibility its content deserves, writing down precisely what has to change, and verifying that the change was made and that a search engine noticed. It is distinct from delivery. A delivery engagement produces content, links or code. SEO consulting services produce decisions — what to fix, in what order, and what to stop doing — and the specifications that let someone else act on them.
That distinction is the whole reason to buy search engine optimization consulting rather than headcount or an agency retainer. A team that both diagnoses the problem and sells the remedy has an interest in the diagnosis. A consultant who writes the specification and hands it to your developers does not: the finding stands or falls on the evidence attached to it, and you keep the ability to say no.
It also means an independent SEO consultant is frequently the wrong purchase, and this page says so in as many words further down. If you already know what is broken and simply need it built, you need a developer, not an SEO consultancy.
The disciplines that sit inside the work
"SEO consulting" is an umbrella over several disciplines that share a vocabulary and very little else. Most engagements draw on two or three of them, not all:
- Technical SEO — crawling, rendering, indexing, canonicalisation, site architecture and page speed. The work that changes what a search engine can reach and understand.
- Site migrations — the URL map, the redirects and the verification, on any move that changes addresses.
- Link profile and manual action work — diagnosing a penalty, and the remediation and reconsideration that follow.
- E-commerce SEO — faceted navigation, catalogue scale and product template problems, which behave differently from everything above.
- Local search — where the Google Business Profile, not the website, is the ranking asset.
- International SEO — hreflang, market boundaries and the return-link annotations that break silently.
- Paid search — the same measurement problems from the other side of the results page.
A consultant who claims equal depth in all of them is describing an agency's staffing chart rather than a person. Ask which two the engagement will actually turn on.
The first ninety days do not look like month ten
Early months are front-loaded with things that only need doing once. A full crawl of the site — a systematic machine fetch of every reachable URL — to establish what exists, what returns which HTTP status code, and what is being canonicalized away. An export of Search Console performance data before any change is made, so there is a baseline that predates the work. Field data for Core Web Vitals, Google's three measured page-experience metrics, at the 75th percentile of real page loads. A read of the analytics setup to find out whether conversions are even being recorded correctly, which they frequently are not.
Somewhere in month two or three the emphasis moves. The backlog of known problems is written and prioritized, and the month starts to be spent on the next tier of question: which templates earn impressions and which do not, where the content genuinely competes and where it never will, whether the internal linking reflects what the business actually sells.
By month ten the work is mostly judgment. What changed at Google, what changed on your site, what changed with your competitors, and which of those three explains the graph.
Where a typical month actually goes
A month in a standing engagement generally breaks into four kinds of work, and it is reasonable to ask for a rough split before you commit.
- Diagnosis. Re-crawling, reading Search Console's Page Indexing and Crawl Stats reports, checking status-code drift, watching for new duplicate or parameter URLs appearing, and reading query-level performance rather than aggregate traffic.
- Specification. Turning findings into instructions a developer can act on without a follow-up meeting: the exact rule, the exact URLs, the expected response, and how to test it.
- Verification. Checking what was built against what was specified. This is the step that gets dropped by providers who bill by deliverable rather than by outcome, and it is where most of the value leaks out of an engagement.
- Judgment. Deciding what not to do. A list of forty findings and no ranking of them is a way of transferring the hard part back to you.
Between monthly reviews there is contact when something needs a decision — a release going out, a drop appearing, a redirect rule that will do damage if it ships as written.
The situations this fits
A retained engagement earns its keep when the search problem is continuous rather than discrete. Some concrete shapes:
- The site publishes or changes regularly, so new problems are created faster than a one-time audit can describe them.
- There is a development team or an agency who can implement, but no one internally who can decide what should be implemented or check that it was done correctly.
- There is an in-house marketer who understands the business well but has never had to argue with a developer about response codes, and would benefit from having someone to check their reasoning against.
- The organic channel is commercially significant enough that a bad release is expensive, and someone should be looking at the site the week it ships rather than the quarter after.
The common thread is that the constraint is decision-making capacity, not knowledge of a single fix. If you already know exactly what needs doing and just need it specified once, you are describing a project, not a retainer.
What drives the number on the invoice
Cost in this work is driven by a small number of structural things, and you can estimate the direction of your own quote before asking anyone.
- Site size and template count. A thousand URLs across four templates is a different job from two hundred thousand across forty. Crawl time, analysis time and the number of distinct fixes all scale with it.
- Platform. Some hosted platforms restrict access to server logs, robots directives or redirect rules. Where a control is unavailable, the work becomes finding a legitimate way around it, which takes longer than using the control would have.
- Who implements. Handing specifications to a competent development team is cheaper than coordinating three vendors who each own part of the stack.
- Number of markets and languages. Every additional locale multiplies the validation work, particularly around
hreflang, the annotation that tells a search engine which language and region version of a page to serve. - Reporting and stakeholder load. A monthly written review for one person costs less than a recurring meeting with four departments.
Retainers are the most common way this work is bought, with fixed-scope projects second and hourly advisory third. None of the three is more legitimate than the others; they answer different questions.
How the work is judged
Agree the measures before the work starts, because measures chosen afterwards are chosen to flatter whatever happened. What matters is movement in the queries and templates the engagement targeted, read at the page and query level rather than as aggregate traffic — and a dated baseline exported before anything changed, so there is something to compare against that nobody can revise later.
Every instrument available for this lies in a specific, knowable direction, and reading them honestly is its own discipline. That is covered separately rather than summarized badly here.
When you do not need SEO consulting
The honest cases, and they are common. A site under a few hundred URLs with no history of a manual action rarely has a technical problem worth a consultant's time — crawl budget is not your constraint and anyone selling you a crawl budget project is selling you something you do not need. A business whose search demand is a handful of branded queries does not have an SEO problem; it has a demand problem, and search engine optimization consulting will not create demand that does not exist.
Nor is a consultant the right answer when the real blocker is capacity. If findings from the last review are still unbuilt, another review produces a longer list and no change. That is a development scheduling problem wearing an SEO costume, and it is worth naming before anyone signs anything.
Which engagement model fits — a standing retainer, a fixed-scope project, or hourly — is a separate question from whether you need the work at all, and it is worked through on its own page rather than here.
What does not happen inside this engagement
No ranking positions are committed to in advance, in any phrasing. Google's published guidance for hiring an SEO says plainly that no one can guarantee a number one ranking, and warns specifically about claims of a special relationship with Google or a priority submission route. Both are worth knowing about before you read anyone's proposal, including mine.
No links are bought. No routine disavow files are filed. No content is published under someone else's name without their review. No changes are made to a live site without telling you what changed and why, which is also Google's advice — insist on understanding every change made to your site, and be careful with anyone who will not explain what they intend to do.
And no engagement continues because it is on a contract. If three consecutive months produce nothing worth reading, the right recommendation is to stop, and I would rather make it than invoice through it.
Frequently Asked Questions
How long before a retained SEO engagement shows results?
It depends on what is being fixed. A technical correction can take effect within days of the affected pages being recrawled. A migration takes a few weeks for most pages on a small or medium site by Google's own estimate. Content and authority work is measured in months. Google's documentation on traffic drops says some changes take effect in a few days while others could take several months, and declines to commit to any noticeable impact at all. The frequently quoted three-to-six-month figure comes from an opinion poll of practitioners, not a measurement of outcomes.What is the difference between SEO consulting and a managed SEO service?
Consulting directs the work; a managed service performs it. In a consulting engagement the diagnosis, the specifications, the prioritization and the verification come from me, and the implementation comes from your developers, your agency or your content team. A managed service takes execution in-house as well, which needs a staffed team. The distinction matters commercially: if you have no one able to implement, a consulting retainer will produce accurate recommendations that never ship. Ask any provider which of the two they are selling, because the words are used interchangeably in the market and the engagements are not interchangeable at all.Can I buy a one-time project instead of a monthly retainer?
Often you should. A retainer suits a continuous problem — a site that changes regularly, a channel that is commercially significant, no internal capacity to decide what to fix next. A single known event does not need one. A migration, a manual action, a slow site, or a first diagnosis of an unfamiliar site are all fixed-scope projects with a defined end. They cost less, they finish, and you can judge from the result whether ongoing work is worth buying. Starting with a project and converting later is a reasonable sequence and a common one.What do you need from my team each month?
Read access to Search Console and analytics, ideally at the property level rather than a filtered view. Access to server log files where the platform allows it, because they show requests that Search Console's Crawl Stats report cannot. A named contact who can either implement or route work to whoever does. Advance notice of releases, redesigns, platform changes and content launches — most damage arrives through a normal deployment nobody thought to mention. Roughly an hour a month of someone's time to review findings and decide priorities. Less than that and the engagement stalls at the specification stage.How is SEO consulting priced, and what makes one engagement cost more than another?
Three structures dominate: a monthly retainer, a fixed-scope project, and hourly advisory. Retainers are the most common. What moves the figure is structural rather than negotiable — site size and number of templates, the platform and whether it exposes logs and redirect controls, whether your team implements or the work must be coordinated across vendors, the number of languages and markets, and the reporting cadence you need. A site with four templates and one market on a platform with full server access sits at one end. A large multi-locale site with three agencies attached sits at the other.Do I still need a consultant if I already have an in-house marketer?
Frequently yes, and the reason is narrow. In-house marketers usually understand the business far better than any external party will, and are usually the right people to own content and commercial priorities. What they often lack is a second opinion on technical judgment calls and the standing to argue with a development team about response codes and rendering. A retained engagement can be scoped to exactly that — direction, review, and specification — rather than duplicating what someone in the building already does well. That version of the engagement is smaller and cheaper than a full one.What happens if my organic traffic falls during the engagement?
The first job is to establish which of the documented causes it is, because the responses are unrelated. Google lists seven: algorithmic updates, technical issues, security threats, spam violations, seasonality, changing user interests, and reporting glitches. A manual action appears in the Manual Actions report in Search Console and nowhere else. A technical cause shows corresponding spikes in the Crawl Stats and Page Indexing reports. An algorithmic drop is reported nowhere, correlates with a dated update window, and usually hits at template or topic level. Diagnosing that correctly before acting is most of the work.Published · Last reviewed