Hire a GEO consultancy or build it in house?

Four criteria you can weigh in an afternoon, the real cost of each route, and the hybrid that usually wins: method and measurement outside, execution inside.

The honest answer is: it depends on four criteria you can weigh in an afternoon.

If you already have an SEO and content team, if your category has few buying questions that matter, and if you can wait half a year for results, building GEO in house is viable and comes cheaper. If the technical base is missing, if the category is contested, or if the competitive window is now, a consultancy shortens the path by months. And there is a third route that often wins in practice: method and measurement outside, execution inside. This piece gives the criteria, the real costs of each route, and the cases where each one loses.

What the work actually demands

Before deciding who does it, it is worth listing what there is to do. Serious GEO for a B2B SaaS is four distinct competences:

• Technical foundation: complete Schema.org, AI crawlers authorised and verified, llms.txt, HTML the engines extract without effort. It is technical SEO with a new layer.
• Citable content: pages that answer the category's buying questions with clean definitions and sourced data. Writing to be cited is a different craft from writing for the click.
• Entity and external authority: consistent presence in the sources the engines cite (press, comparison sites, communities, Wikidata), with the profiles linked by structured data.
• Measurement: running the same questions across the engines on a fixed cadence and measuring citation rate, share of voice and position, per engine and per mode (memory against web search). Without this, the rest is faith.

The central question of the decision is: how many of those four already exist in house, at what level, and with how much time actually available?

When building in house makes sense

The internal route works when three conditions come together. First: technical SEO and content competence already exist, and the GEO layer is an addition rather than a new foundation. Second: the category is narrow, with one or two dozen buying questions that matter, which keeps manual measurement bearable. Third: there is no competitive urgency, and six months of learning curve does not cost market share.

In that scenario the internal roadmap is known: technical foundation first, then the pages that answer the category's questions, and measurement from day one. We published the measurement methodology (citation rate, share of voice and mention rate) precisely for whoever wants to run this in house, and the visibility audit guide serves as the starting point. There is no secret, there is work.

When hiring pays more

The consultancy wins the account in three symmetrical situations. First: the technical SEO base does not exist, and building it by trial and error costs more than buying it ready made. Second: the category is contested and the window is now; in a market where AI answers have no owner yet, the first to occupy the questions keep the accumulated authority, and months of learning curve are share handed over. Third: measurement. Keeping a reproducible weekly measurement across several engines, with stable questions and history, is infrastructure; that is exactly why we built it as a product, Periscopy, instead of rebuilding it client by client.

The cost of the external route is the obvious one (fees) plus a less obvious one: if the consultancy executes everything, the competence does not stay in the house. That is a real cost, and it is the reason for the next model.

The hybrid that usually wins

The most efficient split we see: the consultancy brings the method, the technical foundation and the continuous measurement; the internal team keeps content execution and external authority, which demand knowledge of the product and the category that no outside partner has better than whoever lives there. The consultancy says where the gaps are and measures whether they closed; whoever writes the pages and talks to the press is the team. Over time the external dependency shrinks to what it should be: measurement and method review.

What neither route guarantees

In house or hired, nobody controls a language model's answer. What either route can guarantee is process: gaps identified, actions tied to each one, and measurement that shows the before and the after. Be suspicious of any proposal, external or internal, that promises positions; the honest currency of this work is dated numbers, not guarantees.

To close

Ask the four questions: what competences do we already have, how many buying questions matter in our category, what does waiting cost, and how are we going to measure. If the answers point inwards, the guides in this section are the map and you do not need us. If they point outwards or to the hybrid, the criteria for choosing a partner are in the guide to choosing a consultancy. On either route, start with the measurement: it is what turns a decision of faith into a decision of management.

Frequently asked questions

See also: when GEO is not worth it (the decision before this one), how to measure GEO, and a GEO agency in Lisbon for B2B SaaS (if the route is to hire).