Agency pricing models: six ways to charge, and what each does to margin

Most agencies do not choose a pricing model. They inherit one. The first client wanted a monthly fee, so the agency sells retainers. The second wanted a website quote, so there is a project template somewhere. Five years later the agency runs three pricing models at once and nobody has checked which one makes money.

That check matters more than the price itself. Two agencies charging the same $8,500 a month for the same work end the year with very different margins, because one priced it as a flat retainer that drifted to 72 hours and the other priced it against a scope that held.

This guide covers the six agency pricing models in use in 2027, what each one does to margin, a worked example of the same work priced six ways, and how to choose between them for a marketing agency of 10 to 200 people.

What is an agency pricing model?

An agency pricing model is the rule that turns work into revenue: what the client pays for, how the amount is set and who carries the risk when the work takes longer than planned. The six common agency pricing models are hourly, project-based, monthly retainer, value-based, performance-based and hybrid.

The last part of that definition is the one that decides profit. Every model moves risk somewhere. Hourly billing puts the risk of slow work on the client. A fixed project fee puts it on the agency. Performance pricing puts the risk of the result on the agency too, including the parts of the result the agency does not control.

A pricing model is not the same as a billing model. Pricing decides how much and why. Billing decides when the invoice goes out and what is on it. A retainer priced on value still has to be invoiced on a schedule. For the invoicing side, read the four billing models agencies invoice on.

The six agency pricing models at a glance

1. Hourly pricing

The client pays a rate for every hour worked. The rate is set per person, per role, or as one blended rate across the team.

Where it earns. Hourly pricing is the only model where scope growth pays you. The client asks for more, the team does more, the invoice goes up. It fits work nobody is able to scope in advance: technical audits, crisis PR, consulting days, the discovery phase before a larger engagement.

Where it leaks. Revenue is capped by hours. A senior strategist who solves a problem in 40 minutes earns less than a junior who takes four hours. Clients also audit timesheets, question line items and ask for write-offs, so the hours worked and the hours paid drift apart.

The number to watch. Realization rate: billed hours divided by worked hours. An agency that works 60 hours and bills 54 has given away 10 percent of its revenue before the invoice goes out.

2. Project-based pricing

The client pays a fixed fee for a defined deliverable: a website, a brand identity, a launch campaign. The fee is usually built from an hours estimate times a rate, plus a buffer.

Where it earns. Clients like a number they are able to approve once. The agency keeps the upside when the team is fast. A build estimated at 200 hours and delivered in 170 earns the same fee at a better margin.

Where it leaks. The estimate. Most fixed-fee projects lose money in the last 20 percent of the work, in revision rounds nobody counted and in the two weeks waiting on client content while the team stays allocated. A $30,000 build estimated at 200 hours that takes 260 has lost 11 points of margin, and the agency usually finds out after delivery.

The number to watch. Hours used against the estimate, while the project is running. Knowing at 85 percent of the budget gives you a conversation with the client. Knowing at 130 percent gives you a write-off.

3. Monthly retainer pricing

The client pays a fixed monthly fee for ongoing work. There are three kinds, and they behave differently.

1. Hours-based retainer - the fee buys a block of hours each month. Clear to explain, easy to audit.
2. Deliverables-based retainer - the fee buys a set of outputs: four articles, two reports, one campaign. The client buys results, and the hours are your business.
3. Access retainer - the fee buys availability and judgment. Common in PR and senior advisory work.

Where it earns. Predictable revenue. Retainers make hiring decisions possible because next quarter is mostly known. They are also what buyers pay for when an agency is sold.

Where it leaks. Over-servicing. A retainer scoped at 60 hours that runs at 72 every month has cut its own price by 17 percent, and no invoice shows it. Account managers say yes to small requests, the team absorbs them, and the retainer that looked like the best client in the book turns out to be the worst.

The number to watch. Effective hourly rate: the fee divided by the hours the account took. Compare it against your rate card every month. The full method is in how to price a marketing agency retainer, and the retainer pricing calculator does the sum for you.

4. Value-based pricing

The fee is set by what the outcome is worth to the client, not by the hours it takes. A positioning project that helps a client raise a funding round is priced against the round, not against 80 hours of strategy time.

Where it earns. This is the highest-margin model an agency has. Price and cost are no longer linked, so expertise and speed are rewarded instead of punished.

Where it leaks. It depends on a discovery conversation most agencies skip. If the client has not said out loud what the result is worth, the price is a guess with a confident label. It also fits poorly with commodity work. Nobody pays value-based fees for monthly reporting.

The number to watch. Margin per client. Value pricing only works if you know your cost to deliver, because that cost is the floor under every quote. An agency without that number is pricing on instinct.

5. Performance-based pricing

The agency is paid on results: a fee per qualified lead, a percentage of revenue, a share of ad spend, or a bonus for hitting a target. Percentage of ad spend is the most common version in paid media.

Where it earns. It closes deals. A client who doubts agencies will sign when the agency shares the risk. When a campaign works, the upside has no ceiling.

Where it leaks. The agency carries two risks: the work taking longer than planned, and the result not arriving for reasons outside its control. A slow sales team, a broken checkout or a price increase on the client side all cut the agency's pay. Attribution arguments are common, and cash arrives late.

The number to watch. The base fee against your cost to deliver. Pure performance deals are a bet with payroll. A base fee that covers cost, with performance pay on top, keeps a bad quarter from becoming a loss.

6. Hybrid pricing

Hybrid pricing combines models inside one client relationship. In practice this is what most agencies past 15 people run:

• A fixed-fee onboarding or audit project, followed by a monthly retainer
• A retainer for core work, with hourly billing for anything outside the scope
• A base retainer plus a performance bonus
• A retainer for the work and pass-through billing for media spend

Where it earns. Each part of the work sits in the model that suits it. Defined work is fixed, ongoing work is retained and unknown work is hourly.

Where it leaks. Complexity. When the CRM, the time tracker and the invoicing tool do not talk to each other, a hybrid account becomes three partial pictures. The retainer looks healthy in one place while the out-of-scope hours sit unbilled in another.

The number to watch. Margin by fee type inside the same client. A client is often profitable on the retainer and losing money on the projects, or the reverse.

The same work, priced six ways

Take one engagement: SEO and content for a B2B software client. The plan is 60 hours a month. The team's fully loaded cost is $55 an hour, so the work costs $3,300 a month to deliver. The rate card says $150 an hour.

Same team. Same client. Same 60 hours of planned work. The margin runs from 34 percent to 73 percent depending on the model and on whether the scope held.

Two things stand out. The over-serviced retainer earned an effective rate of $118 an hour against a $150 rate card, and nothing in the invoice history shows it. And the gap between the best and worst outcome is wider inside the performance model than between any two other models. That is what carrying the risk means.

How to choose a pricing model for your agency

There is no best agency pricing model. There is a best fit for each kind of work. Run every service through four questions.

1. Is the scope known? Defined deliverables suit a fixed fee. Open-ended or exploratory work suits hourly. Ongoing work with a stable rhythm suits a retainer.
2. Is the work repeated? Monthly SEO, paid media management, social and PR are recurring by nature. Price them as retainers and review the scope every quarter.
3. Do you control the result? Only take performance pricing where the agency controls most of what drives the number, and where both sides trust the tracking. Keep a base fee that covers cost.
4. Do you know your cost to deliver? Value-based and performance pricing both need a cost floor. If you do not know what an account costs in loaded hours, start with retainers and hourly work, measure for a quarter, then move up.

By discipline, the usual fit looks like this:

• SEO and content - deliverables-based retainer, with a fixed-fee audit up front
• Paid media - retainer or percentage of spend, with media as a pass-through cost
• PR - access or hours-based retainer, with project fees for launches
• Web and creative - fixed-fee projects, with a support retainer after launch
• Full-service - hybrid, with one retainer per department so pacing shows which part of the account is running hot

Stage matters too. An agency of 10 people often sells whatever the client asks for. By 25 people, the pricing models need to be a short written list, because five account managers quoting five different ways makes margin impossible to compare. Put the models next to the rates in your agency rate card.

How to change pricing models without losing clients

Moving a client from hourly to retainer, or from retainer to value-based, is a sales conversation. Three rules keep it calm.

1. Change at renewal, not mid-term. Renewal is the moment clients expect terms to move.
2. Bring the data. Twelve months of hours, deliverables and results makes the case. "Your account took 68 hours a month against a 50-hour retainer" is a fact, not a negotiation.
3. Start with new clients. Sell the new model to every new client from a set date. Move existing clients over one renewal at a time. In a year, most of the book has shifted without a single hard conversation.

A proposal that offers two or three priced options also does a lot of this work for you. The client picks a tier instead of arguing the number. Coinbound and Clickstrike tripled their proposal close rates after moving their proposals into Verbial.

The model matters less than seeing it mid-month

Every pricing model in this guide fails the same way: the agency learns the margin after the month closes. The hourly write-offs, the project overrun and the over-serviced retainer are all visible in the hours long before they are visible in the accounts. The trouble is that the hours live in one tool, the fees in another and the costs in a spreadsheet.

Verbial is the operating platform we built after nearly a decade of running our own agencies on tools that did not talk to each other. It runs retainers, fixed-fee projects, hourly work and pass-through costs for the same client, and shows the margin on each while the month is still running.

• Retainers pace themselves. You see an account running hot in week two, not at renewal.
• Fixed-fee projects warn you early. At 85 percent of the estimate, not at delivery.
• Margin per client is live. Fully loaded team cost against the revenue the client pays, whatever model it is priced on.

See how it works on the agency profitability page, or read the guide to client profitability for marketing agencies for the method behind the numbers.

Frequently asked questions about agency pricing models

What are the most common agency pricing models?

The six most common agency pricing models are hourly, project-based fixed fee, monthly retainer, value-based, performance-based and hybrid. Most marketing agencies past 15 people run a hybrid: a fixed-fee project to start, a monthly retainer for ongoing work and hourly billing for anything outside the scope.

What is the most profitable pricing model for a marketing agency?

Value-based pricing has the highest margin potential, because the fee is set by the outcome instead of the hours. In practice, the most profitable model is the one the agency measures. A well-scoped retainer checked against its hours every month usually beats a value-based fee nobody tracks.

Is a retainer or project pricing better for an agency?

Retainers are better for recurring work such as SEO, paid media, social and PR, because they make revenue predictable. Project pricing is better for work with a clear end, such as a website or a rebrand. Many agencies use both: a project to start the relationship and a retainer to continue it.

What is value-based pricing for agencies?

Value-based pricing sets the fee by what the result is worth to the client, not by the time it takes to deliver. It needs a discovery conversation where the client states the value of the outcome, and it needs the agency to know its own cost to deliver so every quote sits above a floor.

Should a digital agency charge hourly?

Hourly pricing fits work with no clear scope, such as audits, consulting and urgent requests. It is a weak main model for a digital agency because revenue is capped by hours and faster work earns less. Most agencies keep hourly rates for out-of-scope requests and price core services as retainers or fixed fees.

How do agencies price performance-based work?

Most use a base fee plus a variable part: a fee per qualified lead, a percentage of revenue or ad spend, or a bonus for hitting a target. The base fee should cover the cost to deliver. Pure performance deals leave the agency paying its team while waiting on results it only partly controls.