At What Price Does a GEO Retainer Make Money?

The $75 an hour that is not a cost

Every GEO pricing guide currently ranking for this term makes the same arithmetic error, and it is not a small one.

DemandLocal's packaging guide prices a GEO strategist at $75 to $125 an hour and a content specialist at $50 to $85, then calls the result labor cost. Capston's agency model assumes 40 hours at $100 an hour blended for a setup engagement and never says where $100 came from. Those are billing rates wearing a cost label. If your cost per hour equals what you charge per hour, your gross margin is zero, which is presumably not the plan.

I checked nine pages ranking for GEO and AEO pricing before writing this. Every one of them is a survey of what other people charge. The Digital Elevator's guide contains zero labor cost, overhead, or break-even content. Digital Agency Network's version lists seven price bands and no hours. That is a market price, and a market price tells you what your competitors are willing to charge. It tells you nothing about whether you can afford to.

This post is the other calculation. I costed the scope in the GEO scope of work post at 26 hours a month for a real program: prompt set maintenance and multi-engine sampling, citation source analysis, page level retrieval work, off-site earned mention outreach, access monitoring, and reporting. Take that as the input. Everything below is what it costs.

Step one: what one hour of your delivery actually costs

Start with the only part of this that has a government dataset behind it.

The BLS Employer Costs for Employee Compensation release for March 2026 puts total employer compensation for private industry workers at $46.60 per hour worked, of which wages and salaries are $32.60 (69.9%) and benefits are $14.01 (30.1%). That ratio is the multiplier. Divide a salary by 0.699 and you get what the person costs the company before a single square foot of rent.

Then divide by billable hours, not paid hours. This is where most agency math quietly breaks. A full year is 2,080 paid hours. Nobody delivers 2,080 hours of client work. Define utilization as billable hours divided by the 2,080 you pay for, which folds holidays, PTO, sick days, internal meetings and sales support into one honest number.

Read the $85,000 row against the category's $75 an hour strategist. At 70% utilization that person costs $83.52 an hour. At 60%, which is closer to what most small agencies actually run, $97.44. The number the pricing guides use as a cost floor is below true cost for a competent mid level strategist, and it is being presented as the cost of the person doing the most senior thinking in the engagement.

Two honest caveats before you use this table. First, the BLS 30.1% is the private industry average and includes paid leave, insurance, retirement and legally required benefits. A two person shop with no health plan and no retirement match carries closer to 10 to 15%, so this multiplier runs high for a lean operation and I would rather you pull the real payroll expense line off your own P&L. Second, because paid leave sits inside the BLS benefit load and I am also dividing by paid hours, the model double counts leave slightly. It is conservative by a few percent. I would rather be conservative on the cost side than discover the error in December.

Step two: the two costs GEO adds that classic SEO did not

Here is where a GEO retainer stops being an SEO retainer with a new label, and it is not the part sellers talk about.

Per client platform cost is real and it does not amortize the way you expect. The published prices are not hidden. Ahrefs lists Brand Radar AI from $199 a month with custom prompt packages at $50, $100 and $250. Semrush includes AI search tracking from the $199 Starter plan upward. Alex Birkett, who runs Omniscient Digital, compiled published pricing across the category and lists Scrunch Agency Core at $500 a month and Peec AI at 89, 199 and 499 euros. Superlines cites an industry average GEO tool price of $337 a month, sourced to Rankability rather than to a survey, so treat it as a shape rather than a benchmark.

That is the trap in one line. A $500 a month agency platform across twelve clients is $42 per client. Across three clients it is $167. Most agencies adding GEO have three clients on it, price as if they have twelve, and never revisit. I use $150 per client per month in the model below, which assumes a mid tier platform spread across a modest book. Substitute your own.

Measurement is non-billable labor unless you build it in. AI answers are not stable, so a single run proves nothing. SparkToro and Gumshoe ran 2,961 prompts across ChatGPT, Claude and Google AI and found less than a 1 in 100 chance that two runs of the same prompt return an identical brand list. Repeated sampling across a fixed prompt set is the minimum defensible bar, which is why sample size discipline is a cost line and not a nice to have. Every additional engine, prompt and repetition is either platform spend or human hours, and if you did not price it you are absorbing it. The same applies to the prompt set itself, which needs maintenance rather than a one time build.

The practical consequence is that GEO carries a higher fixed cost per account than SEO at the same hour count. That is the actual reason the arithmetic lands where it does.

Step three: run the fee formula, not the market survey

The formula already exists and it is better than anything the GEO category has produced. Marcel Petitpas of Parakeeto publishes it plainly:

Agency Fee = (Delivery Cost / (1 - Margin Target)) + Pass-Through Costs

Note why the project target is higher than the agency wide one. Parakeeto's benchmark for the P&L is a delivery margin of 50% or higher, with project level targets 10 to 20 points above that, because the gap absorbs utilization shortfalls, time off, turnover and training. Price every project at the agency wide target and you will hit the agency wide target only if utilization is perfect, which it never is. Parakeeto also puts shared delivery expenses at 4 to 8% of adjusted gross income, which is roughly where per client GEO tooling lands if you are disciplined about it.

Now build the delivery cost. Take the 26 hour scope split as 10 senior hours and 16 mid hours, using the 70% utilization column above.

Run the formula. At a 60% target the fee is $6,418. At 65%, $7,334. At 70%, $8,557. Blended cost across the scope works out to $92.97 an hour.

Use this graphic on your site

Free to republish with a link back to this page. Copy the embed code:

<a href="https://josephtimpson.com/insights/geo-retainer-pricing"><img src="https://josephtimpson.com/assets/infographics/geo-retainer-pricing.svg" alt="Bar chart of monthly gross margin on a single GEO retainer across six price points, starting at zero at the break-even price of $2,567" width="1200" style="max-width:100%;height:auto"></a><p>Graphic by <a href="https://josephtimpson.com/insights/geo-retainer-pricing">Joseph Timpson</a></p>

That is the whole calculation. It took four numbers and none of them were an industry benchmark.

The uncomfortable finding

Compare $7,334 to what the market actually charges and the problem is obvious.

Ahrefs polled 439 SEO providers and found the single most common monthly retainer band was $501 to $1,000 at 20.4% of respondents, with agencies averaging $3,209 and an average agency hourly rate of $98.90. That page is dated August 2024 and is a practitioner poll rather than a market sample, so treat it as the 2024 shape. Backlinko's survey of 300+ professionals puts typical monthly SEO cost at $1,000 to $2,500. The GEO specific guides put entry level AEO and GEO retainers at $1,000 to $2,500 and mid market at $2,000 to $10,000.

Set that against a $2,567 delivery cost and the entry tier of this category is at or below break-even before overhead. Not thin. Negative.

Invert the formula and it gets worse, because now you can see what a price actually buys. Hold the 65% delivery margin, subtract $150 of platform, and divide by the $92.97 blended cost.

A $2,500 GEO retainer, priced to a healthy project margin, buys under eight hours a month. Eight hours does not cover a defensible measurement cadence, let alone content work and outreach. Either you deliver eight hours and the client gets a monitoring subscription with a human attached, or you deliver twenty six and eat the difference.

That is the honest read on the whole category. Most GEO retainers currently sold are not underpriced by a little. They are priced at roughly a third of what the advertised scope costs, and the gap is being closed by the agency owner's unbilled evenings. Which is also why so much of what gets sold as GEO reduces to an llms.txt file and a schema audit. Those are cheap because they do nothing, and cheap is what a $2,500 price point can afford.

Scope creep, priced

The second series in the chart above is the same model at 38 hours instead of 26, which is roughly where DemandLocal puts a tier two GEO engagement at 38 to 58 hours. Call it five extra senior hours and seven extra mid hours a month. Nothing dramatic. One extra content revision cycle, one unplanned competitor question, one call that runs long.

That costs $1,125 a month. $13,500 a year. Per account.

The gap between the two series is constant across every price point, which is the useful part. Scope creep does not scale with your fee. It is a fixed subtraction, so it hurts a $3,500 account nearly three times as much in percentage terms as it hurts a $10,000 one. At $3,500 the base scope earns $933 a month and the crept scope loses $192.

PMI's Pulse of the Profession found 52% of projects experienced scope creep in its 2018 edition, up from 43% five years prior. That is enterprise project management rather than agency retainers, and I am citing it as directional. The agency specific signal is sharper anyway: AgencyAnalytics surveyed 494 agency professionals between February and April 2026 and found 44% say clients now expect faster turnaround than a year ago, while 66% report increased client demand for AEO and AI search work. Faster expectations plus a service category clients cannot yet evaluate is the exact condition under which scope quietly expands.

Price the creep or govern it. Governing it is cheaper, and it starts with a reporting format that shows what was delivered against what was scoped rather than a dashboard screenshot.

Four exits when the arithmetic says no

If your current price does not clear the model, there are exactly four moves and only one of them is comfortable.

The third exit is where the arithmetic points more often than agency owners expect, because the fixed costs in GEO are unusually high relative to the hours. Platform spend, prompt set maintenance and the sampling discipline required for a defensible number do not get cheaper at one client. They get cheaper at forty. That structural fact is what makes white label GEO delivery a real option rather than a lazy one, and it is the first thing I would model before building the service line internally.

One cost saving move that is genuinely cheap and genuinely works: verify crawler access before anything else. An access audit is an hour or two of work with the highest certainty of any lever in this category, and running it first stops you from billing twenty hours of content work into a site that no engine can fetch.

What the client's side of the ledger has to survive

Your break-even is only half the trade. The other half is whether the account lives long enough to matter.

AgencyAnalytics found 62% of clients stay two or more years and 42% name budget cuts and economic pressure as the top churn reason, which has overtaken performance. That matters for pricing because it says your account dies from the CFO, not from the rankings. The CMO Survey, fielded January 2026 with 308 US marketing leaders, found marketing budgets at 9.0% of revenues and marketing expenses cut 45.4% of the time when profits fall short, more often than any other category. Gartner's 2025 survey found 39% of CMOs planned to cut agency budgets specifically.

So the pricing conclusion is not simply charge more. It is charge enough to deliver something whose value survives a budget review, which means the retainer has to produce evidence a finance person recognizes. That is a measurement problem before it is a pricing problem, and it is why AI search attribution belongs in the scope you are pricing rather than in the upsell you never get to.

Where this model is weak

The candid section, because a pricing model presented as certainty is how people end up underwater with confidence.

The 26 hour scope is my estimate, not a benchmark. No published survey measures the hours a GEO retainer actually consumes. I built the figure from my own delivery and cross checked it against DemandLocal's 38 to 58 hour tier two range, which brackets it. If your scope is different, the entire output changes. That is the point of the model rather than a flaw in it, but do not treat 26 hours as a standard.

Utilization is the input with the widest error bar. The benchmark numbers circulating for agency utilization come almost entirely from software vendor blogs with no stated methodology, so I did not cite any of them. I used 60, 70 and 80% as a range because it spans what I have seen, not because a study says so. Your real number is in your time tracking, and if you are not tracking time you do not have a pricing model, you have a hope.

Promethean and AgencyAnalytics measure agencies broadly, not GEO delivery. Promethean's finding that studio agencies under ten people average 19% net margin against 8% for firms above fifty is useful context and says nothing about this service line specifically.

The Ahrefs pricing poll is stale and self selected. August 2024, 439 practitioners who chose to answer. I used it because it is the most detailed public dataset on SEO retainer distribution and because Backlinko's independent survey lands in the same territory. Two vendor polls agreeing is weaker than one good sample.

The test takes twenty minutes. Pull last month's payroll expense from your P&L, divide by the billable hours your team actually logged, multiply your scope hours by the result, add your platform spend, and divide by 0.35. If the number that comes out is above what you are charging, you have one of the four exits to pick and a decision to make this quarter rather than next year. The rest of the method is about making the number defensible once you have set it, and that is the work I do.