Retainer or project pricing: which should a freelancer use?

Use a retainer when the client's need is ongoing and its monthly volume can be measured, and a project fee when the work has a defined result and an end. Neither model pays more by nature: a retainer buys steady income at the cost of capped upside, a project buys a higher possible rate at the cost of gaps and scope risk, and the rollover, overage and change terms you write decide what you actually earn per hour.

Most advice on this choice stops at "retainers are predictable, projects are flexible". That is true and not enough to decide. The useful questions are how predictable, at what price, and under which clauses. This article describes the five common variants of the two models, compares them on the same five criteria, and then runs a full year of income under each model for the same freelancer, with the gaps between projects, the unused retainer hours and the overruns counted. Every figure in the worked example is illustrative: the assumptions are stated where the example starts, and you can replace them with your own.

The article applies wherever you work. Contract law and tax treatment of prepaid fees vary by country, so the clauses described here are commercial terms to adapt, not legal advice.

What is the difference between a retainer and project pricing?

A retainer is a recurring fee for a recurring commitment, paid for a period; a project price is a single fee for a defined result, paid for an outcome. The retainer client buys your time, your availability or a fixed monthly output, month after month, until one side gives notice. The project client buys one thing, described in a scope, and the relationship ends when it is delivered and paid.

The difference that matters in practice is where the uncertainty sits. With a project, the fee is fixed and the hours are uncertain: if the work takes longer than you estimated, you carry the difference. With a retainer, the monthly fee is fixed and the usage is uncertain: the client may use less than they paid for, or ask for more than the fee covers. In both cases a number is fixed and another floats, and the contract says who pays when the floating one moves.

A second difference is the selling. A project has to be won each time: a brief, a proposal, a negotiation, often a wait. A retainer is sold once and then renews until notice is given. The hours you spend selling are unpaid in both models, but a project practice spends many more of them.

How a retainer is billed is a subject of its own, and the short version is enough here: the fee is invoiced in advance, on the same date each month, for the month that follows, and anything beyond the agreed allowance is added to the next invoice. Work for the month starts once that invoice is paid.

What are the main types of retainer and project fee?

There are three common retainers (an hours bank, an access retainer and a deliverables retainer) and two common project structures (a fixed project and a phased project). They are often discussed as two models, but the five variants behave differently enough that the choice is really between them.

An hours-bank retainer is the most common form. The client pays for a block of hours each month, usually at a rate a little under your standard one in exchange for the commitment. It is the easiest retainer to explain and the one where the rollover and overage clauses do the most work.

An access retainer pays for availability rather than output: a guaranteed response time, a place in your schedule, advice on call. Its effective rate depends entirely on how much the client calls. On an illustrative $1,500 monthly fee, 6 hours of use works out at $250 an hour and 20 hours at $75. An access retainer therefore needs a ceiling: a number of hours, or a list of what is included, beyond which work is quoted separately.

A deliverables retainer fixes the output, for instance four articles, two campaigns or one report each month. The client knows exactly what arrives; you carry the time risk on every item. On an illustrative $2,400 a month for four articles, 5 hours per article gives $120 an hour and 7 hours gives $85.71. It suits work you have done often enough to know how long it takes.

A fixed project is one fee for one result. A phased project splits that result into stages and prices them one at a time, usually starting with a paid discovery or scoping phase. Phasing exists because the first estimate of a large project is made when you know least about it.

How do retainers and project fees compare?

On five fixed criteria, the retainer leads on income predictability and on selling time, the project leads on upside and on freedom to reprice, and scope risk exists in both in a different form. The table compares the plain versions of each model: an hours-bank retainer and a fixed project.

Two of these rows deserve a closer look, because they are where the common advice goes wrong.

Predictability is a property of the client list, not of the word "retainer". A retainer can be ended with notice, usually 30 days, and a freelancer with one large retainer has an employer's concentration risk without an employee's protections. Three smaller retainers are more predictable than one large one of the same total.

The effective rate of a retainer is not the rate printed in the agreement. It is the fee collected divided by the hours actually worked, and both the unused hours and the overruns move it. The next section computes it.

How much does a freelancer earn in a year under each model?

In the worked example below, the same freelancer earns $69,000 from seven fixed projects or $80,300 from three hours-bank retainers, and the gap comes mostly from weeks without paid work, not from the hourly rate. The example is illustrative; what matters is the method, which you can run on your own year.

Call her Ines, a freelance content strategist with a standard rate of $100 an hour. The assumptions for each year are these.

• Project year. She wins seven fixed projects, each quoted at $100 an hour on her estimate: $6,000, $9,000, $12,000, $7,500, $15,000, $9,000 and $10,500, a total of $69,000 for 690 estimated hours. Each is paid half at the start and half on delivery. The actual hours come to 66, 104, 126, 75, 186, 95 and 118, a total of 770. She writes 14 proposals to win the seven, at 5 unpaid hours each.
• Retainer year. She holds three retainers of 25 hours a month at $90 an hour, a 10% reduction on her standard rate, so $2,250 a month each. Two clients stay all twelve months. The third leaves after month 5 and is replaced in month 8, which gives 34 client-months and 850 hours sold. Unused hours lapse at the end of each month, and hours beyond the bank are billed at the standard $100. Over the year, 790 of the 850 hours are used, 60 lapse, and 38 overage hours are billed.

The project year first. The fees are fixed at $69,000, but the 770 hours worked are 80 more than the 690 estimated, an overrun of 11.6%. Her income per hour of client work is $69,000 divided by 770, or $89.61. Adding the 70 unpaid hours of proposals, she worked 840 hours and earned $82.14 for each of them.

The retainer year. The base fees are 34 client-months at $2,250, or $76,500. The 38 overage hours add $3,800, for a total of $80,300. She worked 790 hours inside the banks and 38 beyond them, 828 in all, so her income per hour of client work is $96.98, above the $90 printed in the agreements. The 60 lapsed hours, 7.1% of the hours sold, are what lifts it. Her unpaid time is smaller too: four proposals at 5 hours and one hour of usage reporting per client-month, 54 hours, which brings the all-in figure to $91.04 over 882 hours.

The third row is the same retainer year with two clauses changed. Unused hours roll over without limit, so all 850 hours sold are eventually worked, and overruns are absorbed rather than billed, so the 38 extra hours earn nothing. Income falls to $76,500 for 888 hours, or $86.15 an hour, below the project year. The model did not change. The terms did, and they moved the result by $10.83 an hour.

Read the first two rows with care as well. The retainer year earns $11,300 more, but it also contains 58 more hours of client work, and the project year's lower figure reflects two idle stretches that a busier pipeline would have filled. With seven projects and no overrun, Ines would have earned exactly $100 an hour, more than any retainer row. The project model has the higher ceiling and the lower floor.

How uneven is project income from month to month?

In the same example, monthly income under fixed projects ranges from $0 to $12,000, and under retainers from $4,500 to $8,150. The annual totals hide this; the month-by-month view is what determines whether you can pay yourself a regular amount.

The project line touches zero twice, in months 5 and 9, while a long project is under way and nothing is due. Those are not idle months: Ines is working on the $12,000 and $15,000 projects during them. The payment schedule, not the workload, produces the gap, which is why a long project is usually better billed in milestones than in two halves.

The retainer line has a dip of its own. When the third client leaves after month 5, income drops by a third for two months. A retainer practice is exposed to notice periods the way a project practice is exposed to gaps, and the answer is the same: a reserve of a few months of expenses, and attention to the pipeline before it is needed.

What rollover and overage terms should a retainer have?

A retainer needs one sentence on what happens to unused hours and one on what happens beyond the allowance, and the workable defaults are a limited rollover and overage billed at your standard rate. Without those two sentences the client will assume the most generous reading of both, and the example above shows what that costs.

No rollover is the simplest clause and the hardest to sell, because the client sees $5,400 of hours paid for and not received. Unlimited rollover is the easiest to sell and the most dangerous to hold: the unused hours accumulate into a debt of work that the client can call at once, in the month you have least room for it. A limited rollover, capped in hours and in time, answers the client's fairness concern without building that debt.

For overage, the question is the rate and the trigger. Billing beyond the bank at your standard rate rather than the reduced retainer rate is logical: the discount paid for a commitment, and the extra hours were not committed. The trigger matters as much. Write that you will tell the client when a set share of the month's hours is used, for instance 80%, and that work beyond the bank needs their approval before it starts. A client who is surprised by an overage invoice disputes it; a client who approved it pays it.

Three more terms complete a retainer: the notice period for ending it, a review date on which the fee and the hours are reconsidered (every six or twelve months), and a plain statement of what the hours can be used for. The last one keeps an hours bank from becoming an access retainer by habit.

How do you keep a fixed project from overrunning?

A fixed project stays profitable when its scope lists what is excluded, when changes are priced before they are done, and when a large or uncertain project is split into phases. Scope control is covered in its own right elsewhere; what belongs here is its effect on the choice of model.

In the example, the $15,000 project was estimated at 150 hours and took 186, so it paid $80.65 an hour and accounts for 36 of the year's 80 overrun hours. Priced in phases, the same work would have run differently: a paid discovery phase of 15 hours for $1,500, at the end of which the remaining work is visible and quoted at 171 hours for $17,100. The total becomes $18,600 for the same 186 hours, exactly $100 an hour. The client pays more than the first quote, but pays it knowingly, and keeps the right to stop after discovery with something usable in hand.

Phasing is not always accepted. A client with a fixed budget approved in advance may need one number. In that case the alternatives are a contingency built into the fee, a list of assumptions whose failure triggers a change order, or a smaller scope for the same price. Each of them moves some of the hour risk back to where the uncertainty comes from.

How do you convert a project client into a retainer client?

Convert a project client by measuring what they asked for after delivery, proposing a bank slightly above that average at a reduced rate, and limiting the first term to a three-month trial. The conversion works when the client already has a recurring need and is meeting it with one-off requests.

The moment to propose is shortly after a project is delivered and paid, when the client has seen the work and the next requests have started to arrive. Present the retainer as a way to keep the same access at a known monthly cost, not as a price change. If the usage figures do not show a recurring need, do not force it: a client who needs you twice a year is a project client, and a good one.

When does a retainer fit, and when does a project fee?

A retainer fits recurring work of measurable volume for a client who values continuity; a project fee fits work with a clear result, an end date and a scope you can describe in advance. The decision usually follows from four questions.

• Does the need repeat every month? Ongoing content, maintenance, reporting, advisory work and campaign management repeat. A rebrand, a migration, an audit or a launch does not. Pricing a one-off as a retainer leaves the client paying for months they do not need.
• Can the volume be measured? If you have three months of usage, an hours bank can be sized. If the output is standard, a deliverables retainer is possible. If neither is known, begin with a project or a paid discovery phase and convert later.
• How well do you know the work? Fixed fees reward experience. If you have delivered the same kind of project many times, your estimates hold and a fixed fee lets you keep the benefit of working fast. If the work is new to you, time-based pricing or phases protect you.
• What does your year need? A practice with no recurring income needs a base before it needs upside. A practice whose capacity is fully committed to retainers has no room for the larger project when it arrives.

Many established freelancers end up with both. In the example, Ines could keep two of the retainers all year and take the three largest projects alongside them. Two retainers bring 24 client-months at $2,250, or $54,000; with 560 of the 600 bank hours used and 26 overage hours billed at $100, they add $2,600. The $12,000, $15,000 and $10,500 projects bring $37,500 for 430 actual hours. The year totals $94,100 for 1,016 hours of client work, $92.62 an hour, and no month falls below the $4,500 the retainers pay. The total is higher mainly because more of the year is sold, which is the point of a base: it covers the months, and the projects fill the remaining capacity at a rate that can run above it.

What should you check before you sign either one?

Before signing, compute the effective hourly rate the agreement produces in a bad month, not the rate it names. For a retainer, that is the monthly fee divided by the most hours the terms allow the client to draw: the bank, plus any rollover, plus any overrun you would absorb. For a project, it is the fee divided by your estimate with a realistic overrun added; the 11.6% of the worked example is one freelancer's illustrative year, and your own past projects give you a better figure.

Then check the exit. A retainer should say how it ends, on what notice, and what happens to unused hours on the last day. A project should say what is owed if it stops halfway: the work done to date, at minimum, and whether a cancellation fee applies.

Finally, keep the records that let you choose better next time. A timesheet kept on fixed projects is not for the client; it is the only way to know your real rate per project type. A usage report on a retainer serves both sides: it shows the client what the fee bought, and it is the evidence you bring to the review when the bank needs to grow. The freelancers who price well under either model are rarely the ones with the better model. They are the ones who know their own numbers.