Most pricing pages for SEO answer a different question than the one you asked. You want to know what it will cost you. They tell you what a package contains.
This explains what actually moves the number, so you can read a proposal and judge whether the price makes sense for the work described — and recognise the situations where the honest answer is that you should not buy SEO at all.
Why nobody quotes a price up front
SEO is sold as a service but priced as labour. That distinction matters, because labour is the only part anyone controls.
An agency cannot sell you a ranking. Search results are determined by a third party that publishes no guarantees, changes its behaviour continuously, and ranks you relative to competitors who are also working. Any agency that guarantees a position is either targeting a term so obscure nobody competes for it, or lying.
What can be sold is a defined quantity of skilled work: a technical audit and its remediation, a content architecture, a specified number of pages produced to a standard. The price is the cost of that labour. So the real question is not "what does SEO cost" but "how much labour does my site need" — and nobody can answer that before looking.
The four things that move the number
1. How much technical debt the site carries
This varies more than anything else, and it is the least visible from outside.
A site built on a modern framework with clean URLs, correct canonicals and a sensible information architecture may need almost no technical work. A site that cannot be crawled properly, duplicates every page across three URLs, or renders its navigation only in JavaScript needs weeks before content work is worth starting.
The difference is not marginal. Publishing content onto a site search engines cannot index properly is spending money to produce pages nobody will see.
This is why an audit precedes a quote rather than following one — until someone has established what the site can and cannot do, any number is a guess. It is also why
the technical SEO work that usually comes first.2. How much content has to be produced
Content is the largest recurring cost in almost every engagement, and its volume is set by how much ground you have to cover.
A business with one service and one audience needs a handful of genuinely good pages. A business with eight service lines, each with sub-services, needs a structured cluster for each — and the honest number of pages is dozens, not hundreds, because pages that exist only to hold a keyword do not work any more.
The cost driver within content is depth, not word count. A page that a practitioner would find useful takes research; a page assembled from what already ranks does not, and it produces the same result as everything else assembled that way.
3. How competitive the terms are
Competition sets how much work a page needs to rank, not whether it can.
For a low-competition term, a well-structured page that genuinely answers the query may be enough. For a term where every result is a large publisher with years of accumulated authority, ranking requires either substantially better content, genuine authority signals, or a narrower angle nobody has taken.
The strategic response to high competition is usually to stop competing. Ranking third for a narrow term that converts beats ranking fortieth for a broad one. A good proposal will tell you which terms it is deliberately not chasing, and why.
4. How much you absorb internally
The single largest lever you control.
| You provide | Effect on cost |
|---|---|
| Subject-matter interviews | Large reduction — research is the expensive part of good content |
| Draft content your team writes | Large reduction, if the drafts are genuinely usable |
| Developer time for implementation | Moderate reduction — the agency specifies, your team ships |
| CMS access and fast approvals | Moderate — waiting is a real cost |
| Nothing; agency does everything | Highest cost, fastest delivery |
The middle row is where most of the disappointment happens. "Our developer will implement the fixes" reduces cost only if that developer actually has capacity. Otherwise the recommendations sit in a document and the engagement produces nothing while still being paid for.
The three engagement models
| Model | Fits | Fails when |
|---|---|---|
| Retainer | Ongoing work that compounds — content production, iterative improvement, monitoring | The work is actually bounded. You keep paying after it finished |
| Project | A defined, finishable task — migration, technical remediation, content architecture | The work genuinely needs to continue and the relationship ends mid-way |
| Hourly | Advisory, second opinions, unpredictable scope | Almost always. It rewards slowness and makes budgeting impossible |
Hourly is the one to be most careful with. It aligns the incentive against you: the supplier is paid more for taking longer, and neither side can forecast the total. It has a legitimate use — buying a few hours of senior advice — but as a model for delivery it suits neither party.
The common mistake is a retainer for what is really a project. A technical remediation is finishable. Once it is done, the retainer continues and the work quietly becomes reporting.
How to read a proposal
Four questions that separate a real scope from a package:
What does the first ninety days produce? Ask for deliverables, not activities. "Ongoing optimisation" can be performed forever without producing anything. "The crawl issues in section 3 of the audit, remediated, plus six pages covering the service cluster" is a thing that either exists in ninety days or does not.
What is deliberately not being done? A proposal that covers everything has not prioritised. One that says "we are not chasing these four terms because the results are dominated by publishers we cannot outrank this year" has thought about it.
What happens if it does not work? Not a guarantee — nobody can offer one. But a supplier who has thought about the downside will tell you when they would recommend stopping.
Who does the work? The person in the meeting is often not the person writing the content. That is not automatically a problem, but you should know.
When SEO is the wrong purchase
An agency that only ever says yes is not giving you advice.
-
Your conversion rate is the constraint. If the traffic you have does not convert, more of it multiplies the problem. Fix the page, the offer, or the follow-up first — the same traffic will be worth more, and then SEO compounds against a working funnel instead of a leaking one.
-
You need revenue inside sixty days. SEO does not do this. Paid acquisition does. Buying SEO under short-term revenue pressure produces an expensive few months and a cancelled engagement just as it starts working.
-
The offer does not sell on a call. Search traffic is colder than a referral, not warmer. If a warm prospect on the phone does not buy, a stranger from a search result will not either.
-
Nobody internally can approve anything. Content that cannot get approved does not publish. This kills more engagements than any technical factor, and it is entirely predictable in advance.
What the money actually buys
Stripped of packaging, an SEO engagement buys three things:
- Diagnosis — what is preventing this site from ranking, in priority order
- Production — the technical work and content that fixes it
- Judgement — deciding what not to do, which is most of the value
The first is cheap and often the most valuable. The second is where the money goes. The third is the difference between an agency and a content mill, and it never appears as a line item.
If a proposal contains only the second, you are buying production capacity. That is a legitimate purchase — but price it as such, and do not expect it to include the thinking.
Want this handled properly?
You now know what the work involves. If you would rather not do it yourself, that is what we do.