Search for SEO pricing and you will find a hundred pages explaining that it depends. They are not wrong. They are just not answering, and the reason they are not answering is that publishing a number invites comparison, and comparison is the thing agencies most want to avoid.
So here is the shape of it, stated plainly.
SEO is sold in roughly three tiers. There is a low tier, a middle tier, and a tier where the number stops being a retainer and starts being a headcount decision. What separates them is not quality of effort. It is how many hours of senior attention your account gets per month, and everything downstream follows from that one variable.
section 01why the cheap tier cannot work
The low tier is where most small businesses start, and it fails for reasons that are arithmetic rather than moral.
Work out what a monthly retainer buys once you subtract the agency's overhead, tooling, and margin. What is left is a small number of hours. Now consider what the work actually requires: a technical audit, a content plan, writing, internal linking, and monthly analysis of what moved. That is not a few hours. It is a job.
Something has to give, and what gives is always the same thing. The strategy work disappears first, because it is invisible. What survives is the part that produces an artifact: a report, a ranking spreadsheet, a couple of blog posts written to a word count. You receive evidence of activity every month. The activity is real. It is simply not connected to a decision anyone made about your business.
This is why cheap SEO so often produces the exact pattern clients describe afterward: months of deliverables, nothing to show, and no clear moment where it went wrong. Nothing went wrong. The math never worked.
section 02what you are actually paying for
At every tier, you are paying for three things in different proportions.
The first is diagnosis: figuring out what is actually limiting the site. This is the highest-leverage hour anyone spends on your account and the one clients most resent paying for, because it produces no visible output. A site can have every page written, edited, and paid for, and still have most of them sitting unindexed in Search Console because nothing internal links to them. No amount of additional content fixes that. Finding it takes ten minutes and is worth more than a quarter of publishing.
The second is production: content, technical fixes, the actual building. This scales with volume and is the part most agencies are quoting when they give you a number.
The third is judgment over time, which is the part nobody itemizes. Search changes. Competitors move. What worked last quarter stops working. Somebody has to notice and decide what to do, and that person needs to be senior enough to be worth listening to and close enough to your account to notice at all.
section 03three questions that reveal a real quote
You do not need to know the market rate to evaluate a proposal. You need to know whether the number is attached to anything.
Ask what the first ninety days produce, specifically. A real answer names deliverables and a diagnosis. A vague answer means the plan will be assembled after you sign, out of whatever hours are left.
Ask who does the work and how many hours a month you get. Not the team page. The person. If the answer routes to a strategist you meet once and never speak to again, you are buying the low tier at whatever price you were quoted.
Ask what happens if it does not work. The honest answer is not a guarantee. It is a description of how they will know it is not working, at what point, and what they will change. An agency that cannot describe its own failure mode has not thought about one. We wrote about the reporting version of this problem in your agency says it's working.
section 04the part that actually determines cost
Two businesses can get identical quotes and get wildly different value, because the cost of ranking has almost nothing to do with the agency and almost everything to do with the starting position.
A ten-year-old domain with existing authority, filling an obvious gap, is cheap to move. A domain registered last year, in a market with thirty serious competitors, is expensive, and no retainer changes that. Any proposal written before somebody looked at your domain, your competitors, and your current index coverage is a price for a job nobody has scoped.
The agencies that will not publish a number are not protecting a trade secret. They are protecting the ability to quote you based on what they think you will pay.
Ask for the scope before the price. If the scope arrives after the contract, you already have your answer.