Every guide to hiring a GEO agency lists what the agency should deliver. None list what you have to supply for any of it to work, which is unfortunate, because that is where engagements more commonly stall.
The failure is rarely dramatic. Content sits in review for three weeks. The schema change waits on a developer sprint. The subject matter expert who was going to be interviewed keeps rescheduling.
Six months later the programme has produced less than it should have, and the post-mortem blames the agency. Sometimes correctly. Frequently not.
Key Takeaways
- Review speed is the most common bottleneck: content in approval is content not working.
- Technical access needs settling upfront: not at the first schema deployment.
- Expert time is what makes content citable: specificity cannot be outsourced entirely.
- Outreach needs pre-cleared approval: or every placement waits on legal.
- Budget roughly a day a month internally: less than that and scope should shrink to match.
What Actually Gets Supplied From Your Side
| Input | From whom | Realistic time |
|---|---|---|
| Content review and approval | Marketing owner | 2-4 hours monthly |
| Subject matter interviews | Product, engineering, sales | 1-2 hours monthly |
| Technical deployment | Engineering or web team | Sprint capacity, front-loaded |
| Outreach and quote approval | Comms or legal | Pre-cleared, then minimal |
| Sales and support input | Revenue teams | 1 hour quarterly |
| Decision on positioning questions | Whoever owns positioning | Front-loaded, then rare |
Roughly a day a month across several people, concentrated in the first quarter. That is a small commitment relative to the retainer, and it is the difference between a programme that ships and one that accumulates work in progress. It is also the input least visible in standard visibility reporting, which records what was delivered rather than what was waiting.
The Review Bottleneck
Content waiting for approval is doing nothing. Third-party sources published in month four influence what models retrieve over the following months, so a three-week review delay is not a three-week delay in outcome.
This is the single most common cause of underperformance, and it is almost always structural rather than deliberate.
The usual pattern: content requires approval from someone whose primary job is not marketing, who reviews in batches when they have time. Each piece waits, and the cumulative effect over two quarters is substantial.
Three fixes, in order of effectiveness. Name a single approver with authority to publish rather than a committee. Agree a turnaround commitment in writing, typically five working days. And define in advance which content types need review at all, since not everything does.
Technical Access
Schema deployment, crawler configuration, and template changes all require someone with access to your site, and in most organisations that person does not report to marketing.
The failure mode is predictable: work is specified in month one, enters an engineering backlog, and lands in month four. Everything downstream shifts accordingly.
Settle three things before the engagement starts. Who can deploy changes, what their lead time is, and whether marketing has any direct capability through a CMS or tag manager. If the honest answer is that all changes take a full sprint cycle, the technical portion should be scoped and sequenced around that rather than assumed away.
Expert Time
The content that gets cited is specific. Named systems, actual numbers, real constraints, and the kind of detail that comes from someone who does the work rather than someone writing about it.
An agency can structure, edit, and place content. It cannot invent the specificity that makes a passage worth citing, which means your engineers, product people, and salespeople have to be available in some form.
The efficient version is short and structured: a thirty-minute recorded conversation produces enough material for several pieces. The inefficient version is asking experts to write drafts, which they resist and do badly. Buyers' actual questions, drawn from sales calls and support tickets, make these sessions considerably more productive than open-ended interviews.
Outreach and Quote Approval
Third-party placement is where most of the citation value sits, and it requires your people to be quotable.
Editorial opportunities have short windows. A journalist working on a piece this week needs a comment this week, and an approval chain that takes five days means declining most of them.
What works is pre-clearance rather than per-instance approval. Agree in advance which spokespeople can be quoted, on which topics, and what requires escalation. In regulated categories this needs a proper conversation with legal at the outset, and it is worth having once rather than repeatedly under time pressure.
This is the input most directly tied to outcomes, since third-party mentions and authority carry considerably more weight than anything published on your own domain.
Failure Modes and Their Causes
| What you observe at month six | Usual cause |
|---|---|
| Less content published than planned | Review turnaround, not production capacity |
| Technical work still incomplete | Engineering access never scoped |
| Content is generic and uncited | No expert input; agency wrote from public sources |
| Few third-party placements | Quote approval too slow for editorial windows |
| Reporting disputed internally | No agreed baseline or method at the start |
| Positioning inconsistent across outputs | No single owner for positioning decisions |
Only two of those six are primarily agency failures. The rest originate on the client side and are usually invisible in monthly reporting, which reports what was delivered rather than what was blocked.
Decision Rights
The least obvious requirement and one of the most consequential.
GEO work surfaces positioning questions quickly. Which category do you belong in. Which segment are you actually for. How do you describe the difference between you and a specific competitor. These are not content questions and an agency cannot answer them.
When nobody owns the answer, outputs drift. Different pieces describe the company differently, external sources pick up whichever version they encountered, and the entity signal fragments. That inconsistency is itself a cause of poor performance, since models synthesise whatever they retrieve.
Name the person who decides positioning questions before the engagement starts, and give them enough authority that decisions stick.
If You Cannot Supply This
Not every company can commit a day a month and pre-cleared spokespeople, and pretending otherwise produces a programme that underdelivers against a scope nobody could have executed.
The honest response is to scope down rather than proceed at full price. A technical and entity-focused engagement, without heavy content production or outreach, requires far less internal input and produces correspondingly narrower results. That is a legitimate choice, and it is better than paying for earned media capacity you cannot feed.
It also affects the arithmetic. A reduced scope changes both cost and expected outcome, which is worth running through the same ROI calculation as the full version rather than assuming proportional returns.
Settling It Before You Sign
Five things to agree in the first conversation rather than discover in month three.
- Who approves content, and what turnaround they commit to.
- Who deploys technical changes, and what their lead time is.
- Which experts are available, how often, and in what format.
- Which spokespeople are pre-cleared for comment, on which topics.
- Who decides positioning questions when they arise.
An agency that does not ask these questions during scoping is either very confident or has not run enough engagements to know where they stall. Either way, asking them yourself costs nothing.
The Point of All This
Agency selection gets the overwhelming majority of attention in buying decisions, and it matters. But two companies hiring the same agency with the same scope routinely get different results, and the variable is usually internal.
Before comparing vendors, it is worth knowing honestly what your organisation can supply, because that determines which scope is realistic more than budget does.
If you want a read on where you stand before that conversation, the baseline diagnostic is the same either way: which sources produce your answers, where the gap sits, and what closing it would actually require from both sides. That is where Lureon starts.
FAQs
1. How much internal time does a GEO engagement require?
Roughly a day a month spread across several people, concentrated in the first quarter. That covers content review, subject matter interviews, technical deployment, and approvals. Less availability than that is workable, but the scope should be reduced to match rather than assumed.
2. What causes GEO engagements to underperform?
More often client-side bottlenecks than agency capability. Slow content review, technical changes stuck in engineering backlogs, unavailable subject matter experts, and approval chains too slow for editorial windows account for most of it, and none appear in standard monthly reporting.
3. Why do agencies need access to our engineers and product people?
Because specificity is what makes content citable. Named systems, real numbers, and actual constraints come from people who do the work. An agency can structure and place content but cannot invent the detail that gives a passage citation value.
4. Do we need legal approval for third-party outreach?
In regulated categories usually yes, and it is worth pre-clearing rather than approving each instance. Editorial windows are short, so an approval chain measured in days means declining most opportunities. Agree in advance who can be quoted and on what topics.
5. What if we cannot commit that much internal time?
Scope down deliberately. A technical and entity-focused engagement requires much less internal input and produces narrower but real results. That is preferable to paying for content and outreach capacity your organisation cannot feed, which produces work in progress rather than outcomes.
Time estimates describe typical patterns and vary considerably by organisation size, approval structure, and regulatory context. The requirements described assume a full-scope engagement including content production and third-party outreach.