Back in May, a web agency owner I coach spent most of our call on one question: should he build a monthly offer? He was having a great project quarter, with three website builds signed at margins most agencies would frame and hang on the wall. And the pull toward recurring revenue was there anyway (if you run a project shop, you probably know the feeling).
So we did the math together. He had roughly ten clients, current ones plus a few proposals out, who could plausibly say yes to a small monthly plan. But priced where he wanted it, the plan would have to be delivered in under an hour a month or it would earn less per hour than the project work it was competing with. Maintenance was already handled by a partner, so the monthly work would have been page additions and new components. That’s project work cut into slices and sold at a discount. We parked it, and he kept selling the thing that was working. (By July he was tracking well ahead of his annual goal, driven largely by repeat website work.)
He was about to change his pricing model for a reason that had nothing to do with his offer, and that’s the most common way I see owners pick one.
An agency’s pricing model should match how the client’s problem behaves: charge by the project when the problem ends once it’s solved, charge monthly when it keeps coming back, and only tie your fee to results you control.
What are the main agency pricing models?
Agencies use six common pricing models: hourly, project fee, monthly retainer, hybrid, performance-based, and value-based, with productized fixed fees as a seventh that most lists fold into project pricing. Google’s AI Overview for “agency pricing models” lists the same six, pulling mostly from agency software vendors like Scoro, Ignition, and Taskip, and it quotes retainers at $1,500 to $50,000+ a month and hourly rates at $50 to $400+.
Those lists are fine as a glossary. The trouble is they compare the models on the agency’s terms (cash flow, sales pressure, scope risk) and stop before the question an owner is asking. Which one fits my offer?
I sort them by two questions instead, and I’ll use them through the rest of this post:
- Ends or recurs? Does the client’s problem end, or does it keep coming back? That decides between a project price and a monthly fee.
- Who controls the number? Do you control the number you’d be paid on? That decides whether performance or value-based pricing is open to you at all.
Hourly is the odd one out because it answers neither question. It prices your time, and the client doesn’t care about your time.
Why hourly billing works against you
Most agencies should move off hourly billing, because hourly pricing caps income at the hours you can work and pays you less every time your team gets faster. It fits a narrow set of work: open-ended advisory time and scopes nobody can define yet.
AI made this worse. When a task that took a day takes an afternoon, an hourly agency just lost most of the fee for the same result, and the client keeps all of the savings.
I’ve coached owners off hourly rates onto tiered monthly packages, and the move usually helps both sides. The client gets a predictable bill, and the agency keeps the gains when it speeds up. The hard part is deciding what goes in each tier. You have to name what you deliver instead of how long it takes (work you should’ve done anyway).
Retainer or project: let the problem decide
Charge a retainer when the problem your agency solves comes back on its own, and charge a project fee when the problem ends once you’ve solved it. Paid media, SEO, content, and lifecycle email recur: the ad account needs managing next month because the auction moved, and the content calendar needs filling because last month’s posts are done. Websites, brand identities, strategy work, audits, and system setups end.
I check this with the R in PRISM, the problem test I run on every offer: Painful, Recurring, Impactful, Solvable, Marketable. I wrote it up in how to validate your agency offer with the PRISM model, and the R is the letter that decides how you bill. If the problem recurs, a retainer is the honest price of solving it continuously. If it doesn’t, a retainer is a subscription to a problem the client no longer has, and they’ll figure that out.
Try what I call the fired-tomorrow test: if your client fired you tomorrow, what would break in 90 days? If the answer is “their leads would dry up” or “their site would fall behind,” you’re selling something recurring. If the answer is “nothing, they’d just have the website we built,” you’re selling a project, and that’s a perfectly good business to run.
My own ongoing clients are the example I know best. The problem underneath is usually reliance on referrals, so we do their positioning and then help them build a marketing system, and a marketing system never really finishes. When a client stays on with me past the first engagement, it’s to keep working that problem, because it’s never solved entirely.
The web agency owner from the opening sat squarely on the project side. His work had an end, his margins came from the build, and the monthly plan would have taken his best hour and sold it for less. Project shops get told they’re running the riskier model, and in a sense they are, since next quarter’s revenue depends on next quarter’s sales. But a project agency with repeat clients and a working pipeline is a healthy business, and when its revenue swings, the pipeline is where I’d look first. The swing usually comes from delivery crowding out marketing, and a retainer does nothing about that.
Why retainers churn when the model doesn’t fit
Retainers churn when the problem that justified them gets solved, or when the client can’t see it being solved anymore. The first is a pricing-model mistake and the second is a visibility mistake, and they look identical in a cancellation email.
One owner I worked through an Advantage Intensive with had both. Small brands kept fizzling out of his retainers, and a lot of his community and social work couldn’t be tied back to revenue, and he named both as reasons clients left. Smaller clients whose problem wasn’t big enough to keep recurring, plus work whose results the client couldn’t see. The visibility half has its own post, why clients replace agencies that are doing great work.
Hybrid pricing, for an offer with two parts
A hybrid pricing model makes sense when your offer has two parts that behave differently, so each part gets the model that fits it. The common versions are a project in front of a retainer, and a flat base fee with a performance piece on top.
The project-to-retainer version sells the piece that has to happen first, then moves the client onto monthly work that piece makes possible. My own business runs this way. The Advantage Intensive is a one-time engagement at $1,997, and behind it sits a 90-day coaching program at $2,997 a month. If someone joins within 60 days, the full Intensive fee comes off their first month. Plenty of people stop after the Intensive, and that’s fine, because they bought a finished thing. An agency I coach that builds websites and SEO for contractors talked through the same shape: a small website sprint as the way in, with ongoing SEO behind it.
This only works when the project is a true prerequisite for the monthly work. I broke down how to find that seam in how I structure a first engagement that sells the next one.
The base-plus-performance version is covered in the next section, because it has the same catch as every performance model.
Does performance-based pricing work for agencies?
Performance-based pricing works when the agency controls the number it’s paid on, and it turns into a bad bet when that number depends on things the client controls. Cost-per-lead, revenue share, and percentage of ad spend all fall under this heading, and each one fails the control test differently.
Take percentage of ad spend. A paid ads agency I took through an Advantage Intensive, selling to industrial companies, prices on flat monthly tiers set by how many platforms they run and the client’s spend band. They chose that on purpose, to stand apart from percentage-of-spend pricing, because a fee that grows with spend rewards the agency for spending more of the client’s money.
An Amazon ads agency I coach went the other direction and made it work. They decided to publish a flat base fee plus a percentage of sales on their site, with a second option that waives the base for a higher percentage. That fits the control test, because an Amazon ads agency runs most of the levers that move the sales number. Publishing the prices was a differentiator by itself, because their competitors don’t publish theirs.
Pay-per-lead came up with the contractor agency too, as a lower-risk way in for contractors with small budgets (their cold calls kept hearing “call back in spring”). It lowers the barrier, and it has the sameness problem I warn every lead gen owner about: so many lead gen agencies already say “you don’t pay until you get a lead” that the promise stopped setting anyone apart. And a lead is the one number in that chain the agency controls. What happens after it (the sales call, the close) belongs to the client, so the lead is the thing to price on.
Value-based pricing runs on the buyer’s own math
Value-based pricing is worth it when you can show the buyer, in their own numbers, what the result is worth to them. If you can’t do that math with them on a call, a project fee with a clear scope is safer.
The best version I’ve seen came from an agency owner who works with accounting firms. Her clients balked at her monthly fee because it was more than they charged their own clients. Her answer was arithmetic the buyer already understood: how many new clients she’d need to bring them to pay for herself, and that everything after was theirs. I wrote that call up in which agency differentiators AI just made worthless. She kept her normal fee and anchored it to the buyer’s own math. Most agencies could do the same thing on their next sales call.
Productized fixed fees only pay when delivery repeats
A productized fixed fee is one price for one defined deliverable, the same for every client, usually published. The Advantage Intensive is one. It’s sold at a single price, runs the same two sessions for everyone, and ends in the same kind of report.
It only pays if the delivery behind it repeats. A fixed fee on work you scope from scratch every time is bespoke work with a flat invoice, and the margin goes wherever the scope does. Most agencies don’t need to go all the way to a productized service. Standardizing the repeatable parts is usually enough to make a fixed fee or a retainer profitable.
How do I choose a pricing model for my agency?
Choose your pricing model from the shape of your clients’ problem first, your delivery second, and your cash flow last. Most owners run that order backward, starting from the revenue they want and looking for a model that promises it.
Run these on your main offer:
- Does the problem end or recur? Ask the fired-tomorrow question. Ends means project; recurs means monthly.
- Is there a prerequisite? If the ongoing work needs a strategy, a build, or a setup first, sell that piece as a project and build a hybrid.
- Do you control a number worth being paid on? If yes, a performance piece on top of a base fee is open to you. If the number depends on the client’s sales team or budget, leave it out.
- Does the work repeat? If every month or every project gets scoped fresh, standardize before you sell a fixed fee or a retainer.
- What does the hour math say? Divide your proposed monthly price by what your project work earns per hour. That’s how many hours you can spend before the new model pays worse than the work you already have.
Then remember the model talks to the buyer before the price does. A monthly retainer says ongoing partner, a project fee says defined outcome, and month-to-month in a market of twelve-month contracts says you expect to earn the renewal. I went through that in how to price your services so you’re not competing on cost (it’s the Economic Model lever in the 11 differentiation levers).
And if you’re still building your revenue engine, predictability might matter more than margin for a while. I’ve argued that revenue earns you the right to optimize profit, and a few steady retainers at a thinner margin can be the right trade at that stage. Just make it on purpose.
If you do change models, move new clients first and existing ones at renewal, and only where the new model fits their problem. A client who needed a website and now has one doesn’t need a monthly plan, and selling them one just books a cancellation for later.
If you’ve been feeling guilty for running a project shop because somebody told you grown-up agencies have retainers, you can drop that. My client with the great project quarter ran the numbers, saw a monthly plan would pay him less than the work he already had, and went back to selling websites. Good for him.
If you’re going back and forth on how to price your offer, bring your numbers to the Dynamic Agency Community and we’ll work through it with you.