Building an in-house SEO team without burning two years

Three conditions justify a full-time hire, and all three have to hold. Organic search has to be a primary acquisition channel rather than a supporting one, because a role that matters to nobody in the leadership meeting is a role that gets cut in the first difficult quarter. The site has to change often enough that transferring context to an external partner is genuinely expensive — a marketing site with quarterly releases does not qualify, a marketplace shipping daily does. And there has to be enough senior-level work to fill the week.

That last one is where most hires go wrong. Below roughly two full days of genuine SEO work a week, a hire is under-used, and the role quietly becomes something else — reporting, content coordination, whatever the team needs. Nobody plans this; it happens because a competent person with spare capacity gets given work, and eighteen months later the company has an operations manager and no SEO programme.

The honest comparison is not salary against retainer. It is loaded cost per productive hour against agency cost per delivered hour, including the ramp, the tools and the breadth one person cannot cover. The cost comparison simulator runs all three models side by side, and the answer it usually returns for mid-sized companies is a hybrid rather than a pure model.

Salary is between sixty and seventy per cent of the real cost. Employer contributions add twenty-five to forty-five per cent depending on the country. Recruitment costs a fee or several weeks of internal time. Equipment and workspace are annual. And the tool stack does not shrink because there is only one person using it: a suite licence, a crawler and whatever specialist tool the role needs still runs to several thousand a year.

Then subtract productivity you will not get. A full-time role delivers roughly 1,600 productive hours a year after holidays, meetings, training and administration. Of those, the first two to four months run at partial output while the person learns the product, the market, the site and the internal politics that decide whether anything ships. Budgeting a full year of output from a start date is the most common planning error in SEO hiring.

None of this argues against hiring. It argues for hiring with the real number in front of you, because the comparison that matters — what could the same budget buy externally, and would it deliver more breadth — is only meaningful once both sides are loaded honestly.

The first SEO hire should be a generalist with real technical credibility. Not necessarily a developer, but someone who can read a rendered DOM, understand why a canonical is wrong, write a ticket an engineer can implement without a translation layer, and hold their ground in a sprint planning. The most common failure of a first hire is not a lack of SEO knowledge; it is being unable to get anything shipped.

Specialists come second, and which one depends on what limits growth. A content-led business needs an editor who can brief and hold a quality bar; a large catalogue needs someone comfortable with logs, crawl budget and templates; a competitive market needs digital PR capability that is closer to journalism than to marketing. Hiring the specialist first, before anyone owns the whole picture, produces excellent work on the wrong problem.

Interview for judgement rather than for trivia. Useful questions: describe a time you decided not to do something and why; walk me through diagnosing a thirty per cent traffic drop; which of your recommendations was wrong and how did you find out. Candidates who have never been wrong have either never shipped or are not telling you. The same principles apply to evaluating external partners — our questions to ask an agency transfers almost directly to an interview.

An SEO with no path into the deployment queue produces documents. The single strongest predictor of whether an in-house programme works is not the person's expertise but whether their tickets get prioritised, and that is an organisational question settled long before the hire. Agree in advance where SEO work sits: a standing allocation in each sprint, a named engineering contact, or a quarterly technical debt slot that SEO can claim.

Write tickets that engineers can act on without a meeting: what to change, in which template, why it matters, how to verify it worked, and what the expected impact is. "Improve internal linking" is not a ticket. "Add a related-products block to the product template, six items, server-rendered, linking to canonical URLs" is one, and it will ship.

Give the role visibility into what is being built, not just into what broke. The cheapest SEO work in any company is being in the room when a new template is designed; the most expensive is retrofitting indexability into something already live. If the SEO hears about a redesign at the same time as the customers, the structure is wrong regardless of who was hired.

In companies between roughly thirty and three hundred people, the arrangement that outperforms both pure models is one internal owner plus external specialists. The internal person holds context, priorities and access; the external side brings depth in whichever discipline currently limits growth, and absorbs the peaks. It costs less than a full agency retainer and removes the single-point-of-failure risk of a lone freelancer, because the knowledge lives inside the company.

Two conditions make it work. The internal owner needs a mandate, not just a job title: the authority to decide what ships and what waits. And the external scope has to be specific — "technical audit and template fixes for the catalogue", not "SEO support" — because vague external scope combined with an internal owner produces two parties each assuming the other is handling something.

Whichever way you go, evaluate external partners on evidence rather than on pitch. Providers listed here who connect their Search Console publish anonymised verified performance, and you can filter agencies and freelancers by market and specialty. If the decision is still open, size the workload first with the budget estimator: knowing how many hours a month the programme needs settles the hire-or-outsource argument faster than any debate about principle.