# How to Price Agency Work: Rates, Retainers, and Scope Without Guesswork

> A practical guide to pricing agency services: hourly vs fixed vs retainer vs value, how to set a floor with your effective rate, and how to defend scope so good projects stay profitable.

By The Zinx OS Team on 2026-06-28.

Pricing is where most agencies leave money on the table, usually because the number is a gut feeling dressed up as a quote. The good news is that you do not need a pricing consultant. You need a floor you will not go below, a model that fits the work, and the discipline to hold the line on scope.

## Start with a floor, not a guess

Before you pick a model, you need one number: the rate below which a project loses you money. That is your effective hourly rate, and you cannot price safely without it. If you have not read it yet, [agency metrics that actually matter](/blogs/agency-metrics-that-matter) covers how to calculate it. Whatever you charge, the math has to clear that floor once you count every hour the work really takes, not the hours you hope it takes.

## Pick the model that fits the work

There is no single best pricing model. There is the one that matches the risk in the work.

- **Hourly.** Honest and simple, best when scope is genuinely unknown, like discovery or ongoing support. The downside: you are paid for time, not outcomes, so getting faster pays you less.
- **Fixed fee.** A clean number the client can approve, best when the deliverable is well defined. The risk is entirely yours: if it takes twice as long, your effective rate halves. Only quote fixed when you can name the boundaries precisely.
- **Retainer.** A recurring fee for an ongoing relationship, the most stable revenue an agency can have. A clear [retainer agreement](https://en.wikipedia.org/wiki/Retainer_agreement) that defines what is included, and what is not, turns one-off projects into predictable monthly income.
- **Value based.** Price tied to the outcome rather than the hours. It can earn the most, but it requires real trust and a measurable result, so it is rarely where you start.

Most healthy agencies blend these: retainers for the base, fixed fees for defined projects, hourly for the unknown.

## Defend the scope, or the model does not matter

Every pricing model is destroyed by the same thing: [scope creep](https://en.wikipedia.org/wiki/Scope_creep), the slow growth of "just one more thing" until the project you priced bears no resemblance to the project you are delivering. A fixed fee with unbounded scope is just an hourly project where you forgot to bill the hours.

The defense is written, not verbal. A short [statement of work](https://en.wikipedia.org/wiki/Statement_of_work) that names the deliverables, the boundaries, and what triggers a change order is the cheapest profit protection you will ever buy. When a request lands outside the boundary, you are not saying no, you are pointing at the document you both agreed to.

## Measure whether the price actually held

The quote is a hypothesis. The only way to know if your pricing works is to compare what you charged against what the work cost, per client, after the fact. That is the report in [accurate profitability per client](/blogs/accurate-profitability-per-client), and it is where you learn that your favorite client is underpriced and the difficult one is your best margin. Pricing without that feedback loop is just guessing twice.

This is far easier when the quote, the tracked time, and the [final invoice live in one workspace](/blogs/lead-to-invoice-in-one-workspace) rather than three disconnected tools. The hours you logged become the evidence behind the next quote, and the next quote gets sharper instead of staying a gut feeling.

> Price from your floor, match the model to the risk, put the scope in writing, and check the result. Then raise your prices.

One more thing worth weighing before you discount to win a deal: switching costs and churn are real, and underpricing to land a client you then resent rarely pays off. The [real cost of switching tools](/blogs/real-cost-of-switching-saas) is a useful lens on that. When you are ready to put tracked time behind your numbers, [start free](/sign-in) and price your next project from data instead of a hunch.