How I Read a PR Agency Retainer
I read a PR retainer as a risk document. The monthly number tells me who pays if nothing lands, what "done" means in writing, and whether the price still describes the labor inside the work. A quote that cannot answer those three things is not a decision yet.
The price bands, model by model, are in AI PR Agency Pricing 2026: What Brands Actually Pay. That page is the deeper comparison. This essay is how I would use it as a founder: the questions I put next to a quote, the scope I would accept, and the evidence that would change my mind.
What I put beside the quote
Most PR agencies in 2026 charge between $5,000 and $25,000 per month on retainer. That range has held roughly steady for a decade. What changed is what the month buys.
Thirty-nine percent of CMOs plan to cut agency budgets this year, according to Gartner's 2025 CMO Spend Survey. The same survey puts marketing budgets at 7.7% of company revenue, with 59% of CMOs saying that share is insufficient, and finds that 22% of CMOs say generative AI has already let them reduce reliance on external agencies for creative and strategy work. The money is being redirected. The top actions named are eliminating unproductive relationships and renegotiating contracts.
Forrester's 2026 predictions project a 15% reduction in agency headcount this year, on top of an 8% cut in 2025. One global holding company CEO said the operating plan out loud: "By 2028, we'll double profits and halve the people." At the same time, 75% of marketing agencies now fund their own AI capabilities without passing costs to clients, an 83% jump from 2024. The same reporting describes agencies getting up to 80% faster to market on AI-assisted projects and cutting production costs by 40 to 50%, while absorbing the gain internally. IPG posted a 4% decline in US revenues. S4 Capital saw double-digit drops. WPP informed US employees that up to 45% will be affected by restructuring.
I use those figures as a reading of the invoice, not as a reason to stop buying coverage. The retainer was built to sell labor by the hour. When the production cost of that labor falls and the invoice does not, I am paying for a cost structure the agency has already left behind. Forrester also recommends that B2B marketers reallocate at least 15% of content or digital spend toward improving AI search visibility, including modular content, schema markup, and expert profile optimization, and reports that expectations for agency investment in brand development dropped 11 points year over year. The budget is moving from impressions toward discoverability. A retainer that still prices impressions is answering last year's question.
The four shapes of quote
| Model | Monthly range | What the contract names | Where the risk sits |
|---|---|---|---|
| Traditional retainer | $7,500 to $25,000/mo | Strategy, media lists, pitching, reporting | I pay whether a placement lands or not |
| AI-augmented retainer | $5,000 to $15,000/mo | AI-assisted pitching, monitoring, faster turnaround | Lower invoice, same misalignment |
| Fixed-budget / project | $3,000 to $10,000/project | Campaign-scoped deliverables with a hard cap | Scope creep, and no ongoing relationship |
| Performance-based | Per placement | A live placement | The agency has to have the relationships to deliver |
The first three leave the outcome risk with me. Performance terms are rare because they only work when the agency can actually place the story. A pitch template is not that capability. A figure that shows up in this market is eight hundred pitch emails a day hitting a top tech editor. The agency that can call is in a different market from the agency that adds to that pile.
How I read a vertical quote
The aggregate range hides the contract I am actually signing. I put the quote against the band for the work, then I ask whether the term matches the calendar of the work.
Sports PR typically sits at $5,000 to $20,000 per month, tracking athlete tier and campaign scope. A regional brand or mid-tier athlete may land a boutique at $5,000 to $8,000 a month. National work through firms like PMK or DKC can run $15,000 to $20,000 or more. The event calendar already draws the project boundary: a launch, a draft cycle, a sponsorship announcement. I would accept a fixed budget scoped to that event. I would push back on a 12-month retainer for seasonal work when another firm will scope the season.
Travel and hospitality PR sits at $4,000 to $15,000 per month globally. Australian travel PR retainers, as one benchmark in this range, typically run AU$5,000 to AU$12,000 a month, roughly $3,200 to $7,700 USD, with destination campaigns sometimes higher. A leading Australian or UK travel firm often charges 30 to 40% less than a comparable US firm for equivalent scope. Tourism brands need seasonal coverage. I read a year-round travel retainer against that season, not against a habit of monthly billing.
B2B and enterprise is where the retainer runs highest and where the fixed-budget shift matters most. A comprehensive B2B retainer covering PR, analyst relations, and digital marketing integration typically costs $10,000 to $25,000 per month, and enterprise programs can exceed $30,000. The word "comprehensive" is the trap. A founder gets quoted $7,000 for PR, then discovers separate content, SEO, and digital retainers that triple the total. I add those lines before I compare the quote to anything. A trade placement is worth more when a procurement team can meet it again inside an AI answer to "who leads this category." I still do not treat that hope as a line item unless the contract says how it will be observed.
When a fixed budget is the right shape
A retainer funds an agency's operations for a period. A fixed budget scopes the work, caps the cost, and ties payment to a deliverable. I want that shape when two conditions hold.
The agency has to be able to deliver inside the cap. Cold pitching at the volume above is the thing flooding inboxes. The second condition is a success metric written before the work starts. A media-impressions target is not that metric. A live placement in a named publication is. Performance pricing pushes the same logic further: payment follows the live placement. That term is only viable when the relationships exist. Most agencies still default to the retainer because the relationships do not.
The five lines I write in the margin
Before I sign, I want five answers in the proposal, not in a follow-up call.
- Placement mechanism. Direct relationships, or cold pitch volume. I ask which editors can be called, and I treat a media list as a list.
- Payment and outcome. A monthly fee regardless of result pays for continuation. A fixed budget is closer when the deliverable is named. A performance term aligns the invoice with a live placement.
- The citation question. A Forbes placement is one asset. A Forbes placement that shows up when a prospect asks an engine who leads the category is a different asset. If the agency does not say how it will look, it is optimizing the placement and hoping the second thing happens.
- The network, shown. Editors who answer are the asset. A pitch template is a process document. The network took years. It is not a subscription I can switch on in a quarter. I ask to see the relationship, not the template.
- The cost of waiting. Budgets are tightening on the brand-development line, down those 11 points. The window for earned coverage that machines can reuse is the thing I am buying time against. I still want the first four answers. Urgency does not replace them.
The decision
I would not sign a 12-month retainer whose success line is impressions, process hours, or "AI-assisted pitching" with no named publication and no way to see whether an answer engine reused the placement. I would sign a fixed budget scoped to a named publication and a named buyer question, or a performance term that pays when the placement is live, when the agency can show the relationships that make the promise real. For seasonal sports or travel work I would scope the season. For B2B I would add every adjacent retainer to the quote before I called it comprehensive, and I would ask which buyer question the placement is supposed to enter.
The scope of that rule is the 2026 bands above: traditional retainers from $7,500 to $25,000 a month, AI-augmented retainers from $5,000 to $15,000, fixed projects from $3,000 to $10,000, sports from $5,000 to $20,000, travel from $4,000 to $15,000, and B2B programs from $10,000 to $25,000 with enterprise work above $30,000. It covers the buying decision. It does not say coverage is useless, and it does not set a fair price for a human relationship that is doing a human job.
I would change my mind if a later Gartner or Forrester read showed CMOs increasing reliance on agency labor and the production savings being passed through as a lower retainer for the same named deliverable. I would also change my mind on a specific contract if the retainer itself priced a checkable citation outcome, with the query set, the engines, and the placement written into the scope. Until one of those is true, I read the monthly number as the price of labor, and I decide whether I want to buy labor.
The discipline I use for the citation half of that read is the one I laid out in Machine Relations: earned authority, entity clarity, citation architecture, distribution, and measurement, in that order. The names I need for the metrics sit in the glossary. The pricing page remains the comparison. The contract is still mine to read.
How I mark a quote that lands inside the bands
I do not need a fresh survey to mark up the next proposal. I need the bands already published, and I need the proposal to say which band it believes it occupies.
An $8,000 monthly B2B proposal sits near the $7,000 "PR alone" quote and under the $10,000 to $25,000 comprehensive band. I do not call that a saving. I list what it leaves out: analyst relations, content, search. If those come back as separate retainers, the true cost can triple, which is the jump these ranges describe. I would rather see one scope near the middle of the $10,000 to $25,000 band, with publications named, than an $8,000 scope that becomes an enterprise number by month three. Enterprise work that exceeds $30,000 is a different purchase, and it has to say so on page one.
A $6,000 sports retainer sits in the boutique band of $5,000 to $8,000. I ask which event it covers. A launch, a draft cycle, or a season can live as a fixed project in the $3,000 to $10,000 project band. The same $6,000 stretched across twelve months, with impressions as the success line, is the contract I return. National sports work at $15,000 to $20,000, the PMK and DKC range, is a different product. I do not compare it with the boutique number as if the only difference were confidence.
A $9,000 travel quote from a US firm goes beside the $4,000 to $15,000 global band, beside the Australian benchmark of AU$5,000 to AU$12,000 (roughly $3,200 to $7,700), and beside the 30 to 40% gap between US firms and leading Australian or UK firms for equivalent scope. If the US quote is not buying US media the others cannot reach, I am paying geography. Tourism follows a season. I ask which season, and I look for a cap.
On every one of those quotes I write the same margin from the Gartner and Forrester figures. Thirty-nine percent of CMOs plan to cut agency budgets. Marketing spend is 7.7% of company revenue, and 59% of CMOs say that share is not enough, so the cut is a reallocation. Twenty-two percent say generative AI has already let them rely less on agencies for creative and strategy. Seventy-five percent of agencies now fund their own AI capability, an 83% jump from 2024, with speed gains up to 80% and production costs down 40 to 50%, and the client invoice is not where the saving appears. Forrester's 15% headcount reduction this year, after 8% in 2025, and a holding-company line about doubling profits and halving people by 2028, describe a labor model the retainer still prices. I will pay for access that can place a story. I will not pay a monthly fee that prices headcount the industry is already taking out.
Forrester's recommendation to move at least 15% of content or digital spend toward AI search visibility, and the 11-point drop in expectations for agency investment in brand development, are why a citation question belongs in the proposal. That is the direction of the reallocated money. A proposal that cannot say how a named placement will be looked for in an answer is a proposal aimed at the budget line being cut.
FAQ
What should a founder look at before the monthly fee?
Who holds the risk if nothing publishes, what the contract counts as done, and whether the fee still matches the labor. The ranges on the pricing comparison are the context. The signature is the risk.
Are fixed budgets available for sports and travel, or only for tech?
Sports and travel already have natural project edges. Boutique sports work in the source ranges sits around $5,000 to $8,000, national sports work around $15,000 to $20,000, and travel retainers around $4,000 to $15,000 a month, with Australian benchmarks at AU$5,000 to AU$12,000. A season, a launch, or a destination campaign can be a fixed scope. A 12-month retainer is a choice, not a property of the category.
Why do B2B quotes jump after the first number?
A $7,000 PR line often excludes analyst relations, content, and search. Comprehensive B2B programs in these ranges run $10,000 to $25,000 a month, and enterprise programs can pass $30,000. I add the missing lines before I compare agencies.
Does a cheaper AI-augmented retainer fix the model?
It lowers the invoice into a $5,000 to $15,000 band and it can speed production. The Forrester figures say agencies are funding that capability themselves, an 83% jump from 2024, and keeping the efficiency. A lower retainer with the same "pay regardless" term leaves the risk where it was.
When would I accept a classic retainer anyway?
When the deliverable is a standing relationship I have already seen work, the success line is a named placement rather than impressions, and the fee is the price of access I cannot get another way. I would still write the citation question next to it. I would not treat the retainer as evidence that the question has been answered.
About the author
Jaxon Parrott founded AuthorityTech in 2018 and coined Machine Relations in 2024. He writes about earned authority, AI discovery and founder judgment.
Continue reading
The Category Ownership Trap
Owning a category definition does not mean owning the buyer journey. How I connect Machine Relations to the questions people ask before they know its name.
How I Judge a Pay-Per-Placement Offer
I would pay for a placement only when it makes an engine more able to name the company, and I would treat the invoice model as secondary.
How I Read a Page That Is Writing Our Shortlist
I would read a vendor comparison page as an input to the shortlist, and I would earn a place on the pages an engine already retrieves.