Agencies hide their pricing because the honest answer is uncomfortable: the number depends less on what you need than on what the agency costs to run. A retainer is mostly salaries. Once you understand that, the bands below stop looking arbitrary.
Rough figures for 2026, for B2B software specifically. Anything below $2,500 a month buys you a freelancer with an agency website. Between $2,500 and $5,000 you get real but narrow work, one channel done properly. Between $5,000 and $12,000 is where most funded SaaS companies land, and it buys a small team across two or three channels. Above $15,000 you are buying senior strategic attention and enough production capacity to move a competitive category, which usually means Series B and beyond.
Those are US and Western European numbers. Eastern European and Latin American agencies run roughly 40% cheaper for comparable senior talent, which is a real arbitrage and not a quality compromise, though it does cost you timezone overlap.
What each band actually buys
At $2,500 to $5,000, expect one channel and one person who is genuinely good at it, supported by junior production. This is enough to fix a technical foundation, publish four to six well-researched pages a month, and clean up your tracking. It is not enough to also run paid, do outreach, and produce a monthly narrative report. Agencies that promise all of that at this price are spreading someone across eleven accounts.
At $5,000 to $12,000 you should get a strategist who knows your business, a writer or two who can produce content a technical buyer will not laugh at, and someone doing outreach or digital PR. This band is where link acquisition becomes viable, which matters more in software than most categories because comparison and alternatives queries are almost entirely won on off-site signals.
Above $15,000 the difference is seniority and speed. You are paying for people who have seen your specific problem before and can skip the six months of learning your market. Whether that premium is worth it depends entirely on how competitive your category is. In crowded categories like project management or CRM it is the only thing that works. In a niche vertical with four competitors, spending $15,000 a month is lighting money on fire.
One number worth holding onto: agency gross margins run around 50 to 60%. On a $10,000 retainer, roughly $4,000 to $5,000 becomes actual labour on your account. At $150 an hour fully loaded, that is around 30 hours a month. Ask any agency to confirm the hours behind the number. The good ones will.
Pricing models, and which to refuse
Monthly retainers dominate for good reason. The work compounds, and month-to-month scoping wastes everyone's time. The main risk is that retainers rot quietly. Year two of a retainer often delivers noticeably less than year one because the senior people rotated off and nobody told you. Re-scope annually and ask who is on the account now.
Project pricing works for genuinely bounded work. A technical audit, a migration, a positioning overhaul. If an agency insists everything must be a retainer, they are protecting revenue predictability rather than serving you.
Then there is performance-based pricing, which sounds like perfect alignment and rarely is. The problem is measurement. Pay per lead and you will get leads, including students, competitors, and the same person filling the form four times. Pay per ranking and you will get rankings on keywords nobody searches. Pay on revenue share and you will spend the relationship arguing about attribution, because the agency will claim the deal that closed after a sales call and you will claim it was already in the pipeline. We have watched more relationships die over attribution than over results.
The exception is a small performance kicker on top of a fair base. Ten to twenty percent of the retainer tied to one metric both sides trust. That works because nobody is betting their payroll on it.
The costs that are not in the retainer
Budget separately for these, because they surprise people in month two.
- Paid media spend, which is paid to the platform and is not agency revenue. Management fees typically run 10 to 20% of spend, or a flat fee below roughly $30,000 monthly spend.
- Tooling. Ahrefs or Semrush, a rank tracker, an outreach platform. Somewhere between $400 and $1,200 a month depending on seats, and worth asking whether it is passed through at cost.
- Content production beyond writing. Design, custom graphics, video, and anything requiring a subject matter expert interview.
- Paid placements and sponsorships, if digital PR is part of the plan. This is where budgets quietly double.
A $10,000 retainer with $8,000 of ad spend and $1,000 of tooling is a $19,000 monthly commitment. Model it that way from the start.
How to tell if you are being overcharged
Price alone tells you almost nothing. A $15,000 retainer producing $400,000 in new ARR is cheap. A $3,000 retainer producing nothing is expensive. What you are looking for is whether the work matches the invoice.
Three signals that it does not. First, deliverables that would take a competent person under a week, arriving monthly. Second, reports that lead with impressions and sessions rather than pipeline, which usually means there is nothing better to show. Third, a strategist who cannot answer a question about your product without checking notes six months in.
The clearest test is to ask for the previous month's work log. Not a report, the actual list of what was done and roughly how long each thing took. Agencies doing honest work hand it over. The reaction to the question tells you more than the document.
What GEO does to the budget
Getting cited in ChatGPT, Claude and Perplexity has become its own line item, and most agencies have not priced it yet because most agencies are not doing it.
The work is unglamorous. It is placement and mentions on sources those models already trust, plus content on your own domain that answers the exact question a buyer would type. It looks more like digital PR than SEO, and it costs like digital PR, which means it is labour-heavy and slower to show a return. Expect it to add somewhere between $1,500 and $4,000 a month to a retainer if it is being done seriously rather than as a checkbox.
Whether that is worth paying depends on how much of your category's evaluation has moved. For developer tools and anything an engineer would research, it has moved a lot. For categories where procurement runs a formal RFP, less so, at least for now. Ask an agency to demonstrate the gap live before you fund the fix.
Setting your own number
Work backwards from ACV rather than from what you can afford. If your average contract is $12,000 a year and you close a third of qualified demos, then a $6,000 monthly retainer needs to produce roughly two extra qualified demos a month to break even in year one, and it will look profitable well before that once renewals stack. That arithmetic is worth doing before you talk to anyone, because it converts an anxious budget conversation into a straightforward one.
And if the number you arrive at is below $2,500, do not hire an agency. Hire a good freelancer for two days a month, or read the case for building in-house instead. A cheap agency is worse than no agency, because it costs you the year you spend finding out.