SEO consulting pricing models compared
Retainer, project and performance-based SEO consulting models compared: which one works when, and where each of them breaks.
By Roozbeh Nazari · CEO
Pricing discussions in SEO consulting usually start with "how much", and they start in the wrong place. The figure is the output of the model; the model itself depends on how the work is defined. The same work can be priced as a retainer, as a project or on performance, and in all three cases the risk each party takes on is different. This piece gives no figures; it sets out which model works in which situation and where each one breaks.
Our reason for not giving figures is a principled one: market prices vary by scope, sector, number of languages and team seniority, and a published average is misleading for your own situation. If you want to talk about your own scope you can look at our services or send a brief directly through the contact page. For questions to ask when choosing a consultant, we would also point you to our list of twelve.
Retainer: buying continuity
A fixed monthly fee is the most common model in consulting. The client buys a certain capacity for a certain period; the consultant applies that capacity in priority order. The strength of the model is continuity: in SEO most results come not from a single intervention but from interventions repeated in the same direction, and a retainer is what makes that repetition possible.
Where the model breaks is when the scope is left undefined. The phrase "monthly SEO support" says nothing; with no shared expectation of what gets delivered by month end, the client thinks too little was done and the consultant thinks out-of-scope requests keep arriving. Both can be right, because there is no yardstick.
What keeps a retainer working is a definition of output, not of hours. Not how many hours will be worked per month, but which kinds of work are in scope, at what rhythm it will be reported, and how out-of-scope requests are handled, needs to be written down. With those three in the contract, the model is the most efficient one over the long run.
Project-based: the price of a defined piece of work
In the project model a piece of work with a beginning and an end is priced: a technical audit, migration support, setting up a multilingual architecture, redesigning the content structure. From the client’s side it is the most predictable model; the cost is known up front, the deliverable is defined, and when the work is done the relationship ends naturally.
That is also where its limit lies. In SEO most work does not end at delivery; what is delivered is a list of recommendations, and it produces its value only when implemented. If the client team will take on implementation and has the capacity, the project model is efficient. Without implementation capacity, what emerges at the end of the project is a report on a shelf.
The arrangement that works well in practice is to define the project together with implementation follow-up: the audit is delivered, then a time-boxed follow-up window opens in which implementation questions get answered. That addition lowers the risk of the report being shelved without turning the project into a retainer.
Performance-based: the attractive-looking and riskiest model
Performance-based pricing ties payment to the result and looks risk-free from the client’s side. In practice it carries three separate problems, and none of them is solved by negotiation.
- Defining the result. If payment is tied to rankings, then because no ranking guarantee can be given, the parties have agreed on a variable neither controls. Google explicitly lists consultants who guarantee rankings as a warning sign.
- The attribution problem. If payment is tied to organic conversions, how much of the organic conversion came from the consultant’s work cannot be separated out. A product launch, a campaign or a seasonal effect in the same period changes the result.
- Incentive distortion. A consultant whose payment depends on a short-term metric is under pressure to reach for methods that lift the metric in the short run but are risky in the long run. That pressure sits inside the model, independent of either party’s good faith.
Is there no defensible version of it? There is, but a narrow one: hybrid arrangements where the result is tied to a single measurable output the consultant directly influences, the term is kept short, and there is a fixed fee at the base can work. Arrangements tied entirely to the result with no base fee, on the other hand, in practice lead either to the consultant selecting against the riskiest clients or to them leaving the work early.
Four questions when choosing a model
Choosing a model is a diagnosis, not a negotiation. The answers to the four questions below usually determine on their own which model fits.
- Who will do the implementation? If the client team has developer and content capacity, the project model is enough; if not, continuity is required.
- Is the end of the work defined? For something with a clear beginning and end, such as a migration, a project; for a continuously contested area, a retainer.
- Who is the decision-maker and how often are they reachable? In structures where decisions are slow, the project model leads to the consultant spending time waiting.
- Is the measurement infrastructure ready? If measurement is not in place, no performance-linked model can be set up; the first job is to build the measurement.
There is a fifth question, and it usually goes unasked: when this work is finished, what remains inside the organisation? The lasting output of consulting is not only rankings or traffic; it is the client team being able to make the next decisions on its own. An arrangement that treats knowledge transfer as out of scope turns the relationship into open-ended dependency, and that is costly for both sides in the long run.
The answers to these four questions change over time. In most relationships the healthy order is to start with a defined project, clarify the measurement and the priorities, and then move the part that requires continuity onto a retainer. Entering a long retainer from the outset means both parties committing before they know each other.
A pricing model is ultimately an agreement about who carries which risk. On a retainer the client carries the continuity risk, on the project model the implementation risk, and on the performance model the measurement and attribution risk. Knowing which risk you can carry comes before knowing which figure is reasonable.
Clauses to look for in the contract
Whichever model it is, there are a few clauses the contract needs to carry. They are not about the price but about what the price buys, and most disputes come out of exactly their absence.
- Access and ownership. Who owns the Search Console, analytics property and tag manager accounts should be written down. A property the consultant opened under their own account means data loss when the relationship ends.
- Rights over produced content and documents. Who keeps the audit report, content drafts and technical documents after the relationship should be stated.
- Reporting rhythm and format. Monthly or quarterly; which metrics, against which comparison period. If the report format is not settled up front, every report gets renegotiated.
- Exit terms. Notice period and handover scope. A well-defined exit clause makes the relationship easier to continue, because it stops both parties feeling locked in.
What these four have in common is that none of them is about the fee. Spending part of the time given to price negotiation on these clauses produces more value over the long run than the price does; because the price is discussed once, whereas these clauses come up in every month of the relationship.