SEO agency vs consultant vs in-house: which problem does each actually solve?

The question is usually asked as a budget question: what do we get for $3k a month versus $8k. That framing produces bad decisions, because the three models solve genuinely different constraints and only one of them is likely to be yours.

Here is a more useful way to decide, written by people who sell one of the three options — so weigh it accordingly and check the reasoning rather than the conclusion.

First, name your actual constraint

One of these is usually true. Be honest about which.

"We do not know what is wrong." You have data, reports and possibly an agency, and still cannot say which two things are costing you money. This is a diagnosis constraint.

"We know what is wrong and cannot get it shipped." The findings exist. They sit in a backlog behind product work. This is an execution and prioritisation constraint, and it is often an internal-politics constraint wearing a technical costume.

"We know what to do and lack the hands." Content production, page building, tag implementation, ongoing housekeeping. This is a capacity constraint.

"Nobody owns this at all." Search performance is split across marketing, engineering and an agency, so no single person notices when something breaks. This is an ownership constraint.

What each model is actually good at

Agency

Best for capacity. A retainer buys a team, a process and volume: content production, link outreach, reporting cadence, someone to chase things.

Bad at diagnosis across systems. Most agencies are structurally organised around search. When the real problem is in your analytics, your CRM mapping or your rendering layer, it is outside their remit — not because they are incompetent, but because they were not hired for it and often cannot see it.

Watch for: reporting built on clicks, impressions and rankings without a business number attached. Also a scope where the deliverable is activity — four posts and two audits a month — rather than an outcome. Activity scopes are easy to fulfil and hard to evaluate.

Independent consultant or small specialist firm

Best for diagnosis, specification and hard technical judgement. Small, senior, cross-system, and able to say "the SEO is fine, your tracking is wrong". Usually the right answer when the constraint is not knowing.

Bad at capacity. A consultant is not going to write forty articles or build sixty pages. If you need volume, this is an expensive way to buy it.

Watch for: the audit that ends at the audit. A specification nobody implements is worth nothing, so ask specifically how implementation happens, who validates it, and what happens if it does not work.

In-house

Best for ownership and speed of iteration. Someone who knows the product, sits near the engineers, and notices when something breaks in week one rather than quarter two.

Bad at breadth, early on. One person cannot be senior in technical SEO, analytics implementation, content strategy and AI search measurement. It takes years, or a team.

Watch for: hiring one generalist and expecting all four. The common failure is a capable in-house hire spending their first year on reporting and firefighting because nobody scoped the role around a constraint.

The combinations that work

In practice, most companies past a certain size end up with a mix, and there are three that work well:

In-house owner + specialist consultant. The owner holds priority, context and internal relationships. The specialist is brought in for diagnosis, migrations, measurement rebuilds and hard technical questions. This is the most cost-effective structure we see.

In-house owner + agency for capacity. The owner sets priority and quality standard; the agency produces. Works well as long as the owner is genuinely senior enough to reject work.

Diagnosis first, then decide. If you do not know your constraint, buying twelve months of anything is a guess. A one-off diagnosis is a much cheaper way to find out which of the three you actually need — and a good one will tell you if the answer is "hire in-house" rather than "keep paying us".

The combination that does not work

Agency with no internal owner. Nobody on your side can evaluate the work, prioritise it against product, or get it shipped. The agency reports activity because activity is the only thing they fully control. Twelve months later nothing has changed and both sides are frustrated for structural reasons rather than bad intentions.

If you cannot name the person internally who owns organic performance, fix that before you sign anything.

Questions worth asking any of the three

• What would you need to see before you would tell us to stop doing something?
• Which of your recommendations from last quarter did not work, and how did you find out?
• Show me a finding and the evidence behind it.
• What business number will move if this works, and when would you expect it to?
• What is outside your scope, and who should own that?

The last one is the most revealing. Anyone who says nothing is outside their scope is either much larger than they appear or has not thought about it.

Our position, stated plainly

We are the second model — a small specialist firm — so we are good at diagnosis, technical judgement, migrations and measurement, and we are the wrong choice if what you need is forty articles a month. We say so on our own is this for you section, and we decline work that is really a capacity problem.

That is not modesty. Taking a capacity brief and delivering diagnosis makes everyone unhappy and does not solve the constraint.