What GEO costs in 2026: DIY, tool, agency, or sprint, and how to choose

GEO costs time before it costs money. The honest range runs from zero dollars and two to three hours a week for DIY, through $20 to $800 a month for a tracker that tells you a number, up to $1,500 to $25,000 a month for an agency retainer, with a fixed-scope sprint sitting between the last two and ending cleanly. Which route fits your team depends on one question: can you ship a site change yourself this week?

Here is what each route actually buys, where each one stalls, and how to decide without an analyst.

• DIY costs nothing in cash but roughly 2 to 3 hours a week, and it stalls when you hit the off-site corroboration phase that a founder cannot clear alone.
• Trackers run from about $20 to $800 a month for self-serve tiers, verified against every vendor’s own pricing page in August 2026, and return a visibility score, not a shipped fix. They are right for teams that already have someone to act on the number.
• Agency retainers run $1,500 to $25,000 a month; anything under about $3,000 a month is almost certainly relabeled SEO, not genuine GEO work.
• A fixed-scope sprint ships a defined set of changes, measures the delta, and ends. Collimer’s sprint is fixed price, scoped in one call.

What are you actually paying for: a number, a plan, or a shipped fix?

Every GEO product sells one of three things. A tracker sells a number: your AI visibility score today versus last week. An agency or sprint sells a plan plus execution: pages rewritten, structured data added, third-party mentions placed. DIY sells neither, because you are the product.

The distinction matters because a number without a fix does not move the score. A plan without execution does not move the score. The only thing that moves the score is a shipped change that AI crawlers can read and that third-party sources corroborate. Before you spend anything, be clear which of those three things you are buying.

What does DIY cost, and where does it stall?

DIY costs zero dollars and roughly 2 to 3 hours a week, sustained over 90 days. In that window a founder who can edit their own site can realistically complete the first two phases of a GEO plan: auditing what AI engines currently say about them, rewriting on-site content to answer the questions buyers actually ask, and adding structured data so crawlers can extract clean answers.

The third phase is where DIY stalls. Third-party sources dominate evaluative queries, and placing those mentions, getting cited in roundups, and earning coverage in publications that AI models treat as authoritative is not a task a founder can clear in an hour. It requires outreach, relationships, and time that compounds. Most solo DIY efforts plateau around week 10 to 12 for this reason.

DIY is the right starting point if you can ship a page change this week and you have 2 to 3 hours a week to own the work. It is not the right endpoint if your category is competitive or if you need to show a board a result inside a quarter.

One thing you can do today: run the free scan and write down which of its top three recommendations you could ship yourself this week. If the answer is all three, you do not need an agency yet.

What does a tracker cost, and what does it not do?

AI-visibility trackers span roughly $20 to $800 a month for self-serve tiers, verified against each vendor’s own pricing page in August 2026. That is a forty-fold spread for products that all return a visibility score. Three reference points from that verification:

• Otterly: entry plan at $29 a month, top published plan at $489 a month.
• Athena HQ: free tier available, paid plans up to $295 a month.
• Profound: entry at $99 a month, mid-tier at $399 a month, enterprise pricing not published.

The full vendor-by-vendor table, including who hides pricing entirely, lives in GEO tool pricing compared, re-verified the same week this guide was written.

Profound is the right tool for enterprise teams that have an analyst to interpret the data and a content team to act on it. Its reporting depth is built for that workflow.

What no tracker does: ship a fix. A tracker returns a number, and that number needs a confidence interval to be useful, because AI engine outputs vary by phrasing, model version, and retrieval context. A score that moves a few points week over week may be noise, not signal. A tracker is right for a team that already has someone to own the “now what” after the number arrives.

Collimer’s own plans run $150, $250, and $500 a month with unlimited seats on every plan, plus a one-time $99 audit and a free scan. The difference in kind: the product is the ranked fix and the re-check that verifies it moved, not the score.

What does an agency retainer cost, and what should you ask before signing?

Agency retainers for GEO work run $1,500 to $25,000 a month per our August 2026 pricing research. The low end of that range is almost always relabeled SEO: keyword research, meta-tag updates, and a monthly report that does not mention AI engine behavior. Genuine GEO retainers, meaning work that directly targets how large language models retrieve and cite your brand, start closer to $3,000 a month and typically require a 3 to 6 month commitment. One-time audits from agencies run $1,500 to $5,000 and are worth considering before a retainer if you want an outside read on your current state.

Before signing any retainer, ask three questions:

1. What ships in the first 30 days? A credible answer names specific deliverables: pages rewritten, structured data added, a target list of third-party placements. A vague answer (“we’ll develop a strategy”) is a warning sign.
2. How is progress measured, and against what baseline? The agency should be able to name the prompts they track and show you a before score. If they cannot, they are not measuring GEO.
3. Who owns the site changes? Some agencies write the copy and hand it to you to publish. Others need CMS access. Know which model you are buying before the contract is signed.

An agency retainer is right for a team that cannot own 2 to 3 hours a week internally, needs to move faster than DIY allows, and has budget above $3,000 a month to spend on work that is actually GEO.

What does a fixed-scope sprint cost, and when is it the right size?

A sprint is a bounded engagement: a defined set of changes ships, the delta is measured, and the engagement ends. It sits between a tracker (which ships nothing) and a retainer (which runs indefinitely).

Collimer’s 90-day sprint is fixed price, scoped in one call. The exact figure is not published because scope varies by site; the call takes under 30 minutes and produces a written scope before any commitment.

A sprint is right when you need a verified, measurable result inside a defined window, you do not want an open-ended retainer, and you have a specific gap, such as zero AI visibility on a product category, that a bounded set of changes can close. It is also the right format when you need something to show a board: a sprint produces a before-and-after delta, not a monthly score.

A sprint is not right if your site has foundational technical problems that require ongoing work, or if your category requires sustained off-site placement that a one-time engagement cannot maintain.

Which route fits which team?

Decide by working through three questions in order.

Can you ship a page change this week? If yes, start with the free scan and DIY the top three recommendations. If no, you need either a sprint or an agency.

Do you have someone who can own 2 to 3 hours a week for 90 days? If yes, DIY is viable through the first two phases. Add a tracker at $150 a month or less if you want a score to watch. If no, a sprint or retainer is the right structure.

Do you need a result you can show a board inside a quarter? If yes, a sprint gives you a fixed scope, a fixed end date, and a measurable delta. A retainer can do the same but costs more and runs longer. DIY cannot reliably produce a board-ready result in 90 days if your category is competitive.

The 90-day plan maps the DIY route week by week and is consistent with the effort estimates above.

The one route that costs nothing

Start with the free scan. It reads your current AI visibility, surfaces the top gaps, and tells you which of those gaps you can close yourself. If the answer is “all three,” you do not need an agency yet. If the answer is “none of them without help,” you know what kind of help to buy.

The scan takes under 10 minutes and costs nothing. It is the only route on this list where that is true.