Every SEO retainer proposal contains the same line item, and almost nobody asks what it means.
“Monthly reporting and analysis.”
It sits there between the keyword research and the link building, taking up a slot, and it is the single most reliable indicator of what the rest of the engagement will feel like. In a good retainer it costs about two hours a month and describes a call where someone explains what changed. In a bad one it is the deliverable — the thing that actually arrives every month, dressed up as insight, generated by a tool with a logo swapped in.
You can tell the two apart before signing. Here is what to look at.
What a retainer is for
An SEO retainer buys continuous work on a system that keeps changing: your site, your competitors, your rankings, and Google itself.
That is a genuine reason for an ongoing arrangement, and it is worth naming, because plenty of SEO work is not like this at all. A technical audit is a project. A site migration is a project. A one-time content architecture is a project. They have a start, an end, and a definition of done.
The retainer exists for the part that doesn’t finish: watching what moved, publishing against a plan, earning links, fixing what breaks when someone ships a new template on a Tuesday.
If a proposal puts a project inside a retainer, ask why. Sometimes the answer is reasonable — spreading a large technical fix over three months to match your developer capacity. Sometimes the answer is that a project priced monthly, indefinitely, earns more than a project priced as a project.
What is actually in a working retainer
Roughly, and the proportions matter more than the list:
Content production. Usually the largest share. Briefs, drafts, edits, publishing, and the internal linking that connects new pages to existing ones. If content is in scope, ask how many pieces a month and who writes them.
Technical maintenance. Crawl monitoring, fixing what breaks, checking that new templates ship with correct titles and canonicals. Small in a good month, large in a bad one.
Link acquisition. Digital PR, resource outreach, unlinked-mention recovery. This is the slowest and least predictable line, and the one most likely to be quietly dropped when a month gets busy. It is also three different disciplines sold under one word — SEO, digital PR and link building are not synonyms, and a proposal that treats them as one is telling you something.
On-page optimisation. Existing pages that rank on page two and need the title, the intro, or the internal links they never got.
Measurement and decisions. Not the report. The half-hour where somebody reads the report and changes the plan.
Notice the shape: four lines of doing, one line of reporting. If a proposal inverts that, you are buying a subscription to a dashboard.
The deliverable list that signals a padded retainer
Some line items appear in proposals specifically because they sound like work and take almost none.
“Monthly ranking report.” Automated. Costs minutes. Fine as a byproduct, a warning sign as a deliverable.
“Ongoing keyword research.” Keyword research is front-loaded — a real piece of work in month one, then small revisions. Billed as a constant monthly line, it is padding.
“Competitor monitoring.” Genuine as a quarterly review. As a monthly line item it usually means someone glances at a tool.
“Google algorithm update monitoring.” Reading industry news. Every agency does this to stay in business; charging for it as a deliverable is charging you for their own professional development.
“Search Console and Analytics management.” Setting these up is real work. “Managing” a property that needs no management is not.
None of these are frauds exactly. Each one is real in small quantities. The tell is proportion: when the list is six items long and four of them are monitoring, the retainer’s actual output is a PDF.
The counter-question that sorts it out, and it is worth asking exactly this way:
In an average month, how many hours go to making or changing something on my site, and how many go to reporting on it?
You will learn more from how the question is received than from the number.
How to structure the first 90 days
Front-load, then even out. A retainer that runs at the same tempo from month one is usually not doing the diagnostic work at all.
Month one is mostly audit and setup. Technical crawl, content inventory, keyword mapping to existing URLs, tracking verification, a prioritised backlog. Very little publishing. This month should feel like it produces documents rather than pages, and that is correct. Priced separately, this is what an audit engagement buys, which is a reasonable way to test an agency before committing to twelve months of anything.
Month two is fixing and shipping. The technical items from month one, the highest-value on-page fixes, and the first new content against the plan.
Month three is the first honest read. Whether you can even have that read depends on tracking that was correct before month one, which is a measurement question rather than an SEO one. Not results — three months is too early for most competitive terms, and any agency promising otherwise is managing your expectations badly on purpose. What you should get is evidence the machine works: pages indexed, positions moving in the right direction, crawl errors falling.
If month three arrives and nobody can show you what changed on the site, that is your answer.
What should come out of the retainer
Two categories, and pulling them out usually saves money.
Large technical projects. A replatform, a site-wide migration, a rebuild of the internal linking architecture. These have a defined scope and a defined end. Priced as a fixed-scope project, you know the number. Absorbed into a retainer, they consume the months they take and the retainer’s ongoing work quietly stops.
Anything with a hard finish line. A schema implementation. A Core Web Vitals remediation. A one-off content architecture. If you can describe the state where it is done, it is a project.
Keeping the retainer for genuinely continuous work has a side effect worth having: the retainer gets easier to evaluate. When it contains only ongoing work, “is this worth it” is a simpler question than when it contains a half-finished migration.
The boundary nobody agrees on until it’s a problem
Three areas sit at the edge of every SEO retainer, and the arguments always happen in month four rather than in the proposal.
Development work. SEO produces recommendations; someone has to implement them. If the retainer includes the recommending but not the doing, and your engineering team has no capacity, the retainer produces a backlog rather than results. Settle this before signing: does the agency ship code, do they hand tickets to your team, or is there a budgeted development allocation? All three are workable. Assuming without asking is not.
Content volume versus content quality. “Four articles a month” and “four good articles a month” are different prices, and the gap is mostly research and subject-matter access. An agency writing about your industry without talking to anyone at your company produces text that ranks for nothing, because it says what every other article says. If nobody has asked for interview time with your people, that tells you which version you bought.
Digital PR and link earning. Sometimes in the retainer, sometimes a separate line, occasionally not covered at all while everybody assumes it is. It is also the slowest thing to show results, which makes it the easiest to quietly deprioritise. Ask specifically how many hours a month go to it, because “link building included” covers everything from real outreach to submitting your site to twelve directories.
None of these has a correct answer. They have an answer you agreed on, or an argument in month four. Most of the questions worth asking before signing anything are the same across disciplines — how to choose a marketing agency collects them.
The cancel-any-time question
Ask about notice period early, and read the answer as information about the business model rather than as a term to negotiate.
Twelve-month lock-ins are sometimes defended on the grounds that SEO takes time. That is true about SEO and irrelevant to the contract: a good agency does not need a contractual barrier to stop you leaving, because leaving mid-work is against your interest anyway. A lock-in protects the agency’s revenue, not your results.
Thirty days notice is the normal shape, and it puts the pressure where it belongs — on the agency showing you something every month.
The related question that matters more: what do you keep if you leave? Content on your domain, obviously. But also the keyword map, the backlog, the technical documentation, the tracking configuration. If the answer is vague, or if the working documents live in the agency’s tooling rather than yours, you are renting the work rather than owning it.
Short version
- A retainer is for work that doesn’t finish. Projects should be priced as projects.
- Count the doing-to-reporting ratio in the deliverable list. Four to one is healthy; inverted is a subscription to a PDF.
- Ranking reports, ongoing keyword research, and algorithm monitoring are padding in quantity, real in small doses.
- Ask how many hours a month change something on the site.
- Expect month one to produce documents, month two to produce changes, month three to produce evidence — not results.
- Take large technical projects out of the retainer, so what’s left can be judged.
- Thirty days notice. And ask what you keep on the way out.
If the question underneath this is what the number should be rather than what’s in it, SEO agency pricing has the 2026 US market ranges alongside our own rate card.