Why "We Do GEO Too" Is Not an Agency Offer

Adding a line to a services page is a claim. An offer is something a client can buy, scope, schedule and hold you to.

The difference is not semantic. It decides whether the work is deliverable at margin, whether it renews, and whether the person delivering it knows what done looks like. "We do GEO too" survives right up to the first client question: what exactly will I receive, in what order, by when, and how will we know it worked?

The claim, and what it costs

The claim is attractive because it requires nothing. No process, no evidence standard, no hiring, no decision about what you will not do.

Then a client says yes, and the bill for that sentence arrives.

No queue. The team knows GEO matters and cannot name the first task, so the account drifts into "ongoing optimisation" that nobody can describe.

No baseline. Six weeks later, a client asks whether anything improved. Without a stable prompt set and a stated method, the answer is a set of screenshots and a feeling.

Subcontracted writing with no brief. Content goes out to a writer who has never seen the citation evidence, so it repeats what everyone else says, which is precisely the content that loses a rerank.

Junior delivery. The pitch was made by a senior person. The work is done by someone learning what a vector search is, on the client's budget.

Reporting without classification. "You are at 22 per cent visibility" is not a finding and does not tell anyone what to do on Monday.

Each of those is a defect of the offer, not of the team's effort.

Six things a real GEO offer contains

1. Named deliverables with a definition of done. A prompt set of an agreed size and market. A citation audit listing the domains actually cited. A classified gap. A ranked queue. A named number of corrected passages or published assets per month. If it cannot be counted, it cannot be scheduled or renewed.

2. A stated prompt set and market. Which questions, in which country, on which engines, refreshed how often. This is the boring clause that makes every later report comparable.

3. A method you can show. Not a philosophy. The sequence: architecture, then passages, then corroboration, with the reason for that order. Clients do not need to understand reranking, but they do need to see that the work is ordered rather than improvised.

4. A cadence and an owner on both sides. Monthly rerun, monthly report, named counterpart at the client. GEO work touches content, site, PR and occasionally product, so an unowned plan stalls.

5. A price and a boundary. What is included, what is not, and what happens when the client's dev team is the bottleneck. A boundary is not a limitation, it is evidence that you have done this before.

6. Proof you can do it. At minimum, evidence from your own site or a client's: the prompt, the answer, the source, the fix, the change. One honest worked example beats a page of logos.

"After 27 years in search, I can tell you the biggest gap in SEO strategy isn't on-page optimisation — it's architects who only think in single pages. Topical authority changes that."

John Russell, Director, Digital Womble

Why GEO exposes weak offers faster than SEO did

Traditional SEO is forgiving of vagueness. Rank tracking produces a number every day, and a client can watch something move even when the strategy is thin.

The answer layer is less forgiving. Answers vary between runs, engines differ, and the citation set shifts. An improvised process shows up immediately as unexplained variance, whereas a defined one shows up as a trend line with notes.

It also punishes single-page thinking, because an engine reranks passages against a rewritten question rather than scoring a page. A service that can only deliver page optimisation will produce activity that never reaches the answer.

"Most SMEs think content strategy is a luxury, something only big companies afford. That is wrong. A proper content strategy costs £48-98 per cluster, not £5,000."

John Russell, Director, Digital Womble

That quote is about content strategy, and it applies directly to GEO pricing. The offer does not become credible by being expensive. It becomes credible by being defined.

Building the line without rebuilding the agency

Three decisions, in order.

Decide the first deliverable. Not a service line, a deliverable. An AI visibility audit with a classified gap and a ranked queue is enough to sell, and it produces the evidence for everything after it.

Decide build or buy. Building the audit means prompt capture, citation parsing and a repeatable report format. Buying it means a white-label partner and a margin, with your brand on the output. Neither is more virtuous; one is faster and one is more defensible. Pick deliberately.

Decide how delivery scales. The step that kills agency margins is treating every month as bespoke. A productised queue, produced from the same evidence base, is what allows a retainer to be both honest and profitable.

What the client is actually buying

Not visibility. They are buying a defensible story they can take to their own board: here is what engines say, here is why it matters to revenue, here is the work, here is what changed.

Deliver that story monthly and the retainer renews on evidence. Fail to, and the sentence "we do GEO too" becomes the reason they left.

If you would rather not build it

We run GEO as a white-label service for agencies: prompt set, citation evidence, classified gap, ranked queue and the content that follows it, delivered under your brand with your pricing on top. Start with the white label GEO service or see what a managed engagement includes.

Related: how to turn an AI visibility audit into a retainer and what an AI visibility report should actually show.