SEO pricing models compared: retainer, hourly, project or performance

Start with who carries the risk

A pricing model is a way of splitting risk between you and the vendor. A retainer puts the risk of drift on you. A project puts the risk of scope on the vendor. Hourly billing puts the risk of slow work on you. Performance pricing tries to put the risk of Google on the vendor, and that is exactly why it rarely works.

Read every proposal with that question first: if this goes badly, who pays?

Retainer: you buy a rhythm

A retainer pays for continuous work: content, fixes, outreach and reporting, every month. Agency retainers span $500 to $15,000 and above^market, and the average agency charges about $3,209 a month^market.

Project: you buy an ending

A fixed-price project fits a problem that ends. A migration, a technical cleanup, a service-page sprint, a one-time audit with a fix list. Typical prices run $2,500 to $30,000^market, and the most common band for one-off work is $2,501 to $5,000^market. A small local site sits near the bottom of that range and a large store migration does not.

It is the wrong buy for competitive rankings, because your competitors keep working after the project stops. The tell is a project pitched as "making you rank" instead of fixing something specific.

Hourly: you buy judgment

Hourly rates cluster at $100 to $250^market, with senior technical consultants in competitive verticals at $200 to $400 and up^market.

Hourly is honest for advice: a second opinion on a proposal, a strategy session each quarter, a plan reviewed before you sign. It is a poor fit for execution. Twenty hours a month at $150 is $3,000^estimate, with none of the accountability of a scoped retainer and no built-in reason for the vendor to get faster.

• Buy hours for a decision.
• Buy a scope for work.
• Do not buy hours for a monthly workload nobody is measuring.

Performance: you buy a promise

Pay only when you rank sounds like the fair deal. The trouble is that only Google controls rankings. A vendor who guarantees them is picking keywords easy enough to be worthless, defining success loosely enough to always collect, or using tactics that get you penalized after the invoice is paid.

Pure versions have mostly disappeared in 2026. The version that still shows up is a hybrid: a base fee of roughly 60 to 70 percent of a normal retainer^market plus bonuses tied to meaningful outcomes.

One business, four invoices

Take a single-location service company with a decent site, some technical debt and a $30,000^estimate budget for its first year of search.

The cash spread between the cheapest and dearest paid option is $4,200^estimate. The spread in risk is much wider. The cheapest model on paper is rarely the cheapest in outcome, and that is why the shape deserves as much scrutiny as the price.

Picking by situation

1. Known technical problems or a replatform in progress: sign a fixed project first, not a retainer that fixes plumbing slowly.
2. Competitive market where search already brings customers: a retainer with a named monthly scope.
3. An in-house marketer who can execute: hourly or quarterly consulting on top of their work.
4. Tight budget in a low-competition local market: a small local retainer, scoped tightly.
5. A guarantee of positions on the first call: walk away, no model fixes that vendor.

A project first and a smaller retainer after it is the structure we see work most often. You watch the vendor on a bounded job before committing to a monthly fee.