What retainer structure did I settle on after six years?

What retainer structure did I finally settle on?

A fixed monthly fee for a named set of recurring outcomes, with an explicit list of what is not included and a thirty-day exit on either side. No hour banks, no rollover, no vague availability. It took me most of six years to get to something that simple.

Everything before that was some version of selling time, and selling time to a client who is paying monthly quietly converts you into staff without the salary or the protections. The structure is what stops that, not your discipline and not their goodwill.

This is the honest version, including the parts that still do not work well. I am not going to pretend I found a perfect model, because I have not, and anyone telling you they have is selling a course.

Why did the hour-bank retainer keep failing?

Because it makes the client's incentive to extract hours and my incentive to conserve them, which puts us on opposite sides of every conversation. The moment you are both watching a counter, the relationship is about the counter rather than the work.

It fails in a specific way too. Early months underuse the bank and the client feels cheated. Later months overrun and I feel cheated. The unused hours become a debt in the client's mind, and I end up doing a month of free work to discharge a feeling rather than an obligation.

Worst of all, an hour bank never says what it is for. So it absorbs whatever arrives: a quick change, an opinion on a vendor, a favour for someone's colleague. I was busy and nothing improved, which is the exact failure mode a retainer is supposed to prevent.

What does a retainer actually buy, if not hours?

A set of recurring outcomes with names. Something like: the site stays fast and accessible, new content ships to a standard, the automations keep running and get fixed when they break, and someone competent is thinking about this every month instead of nobody.

The test I use is whether the client could tell, at the end of a month, that they got it. Hours are not that kind of thing. "Two pages shipped and the lead routing did not break" is that kind of thing. If neither of us can tell whether the month succeeded, I have written a bad retainer.

This also changes what the work feels like. When the deliverable is an outcome, I get to choose how to achieve it, including by automating something so it stops taking time. Under an hour bank, automating my own work is financially irrational, which tells you everything about that model. It is the same logic as being careful about when a spreadsheet beats an automation: incentives decide what gets built.

Why does the exclusion list matter more than the inclusion list?

Because scope disputes are almost never about what was promised. They are about what the client assumed was obviously included. An inclusion list ends where you stopped writing; an exclusion list ends where the argument would have started, which is the only place worth defending.

Mine names the things that sound adjacent and are actually projects: a new page template, a rebrand, a migration, a new tool integrated from scratch, anything requiring a third party's cooperation. Those are quotable work, and saying so in advance means I can say yes to them cheerfully rather than resentfully.

The conversation feels awkward to write and it prevents a far worse one later. I would rather lose a retainer at the proposal stage over an exclusion than spend four months with someone who believes I owe them a rebrand. The list is not defensive, it is clarifying, and it is the same instinct behind knowing what to do when a fixed-fee project goes over scope.

How should you price a retainer against project work?

Against the value of continuity, not as a discount on your project rate. A retainer buys the client responsiveness and accumulated context, both of which are genuinely worth something, and it buys me predictability. Neither party is doing the other a favour.

The discount framing is a trap I fell into for a long time. If a retainer is simply your project rate minus twenty percent, then every month you are hoping for a quiet month, which means you are hoping not to deliver value. That is a business model at war with itself.

My fixed-fee habit carries over here. Most of my projects are between $1,000 and $10,000 and are priced on outcome rather than effort, and a retainer is the same idea stretched over months. If you want the reasoning behind that pricing posture, I have written about why I publish my pricing as a solo consultant.

Why does every retainer need an exit on both sides?

Because a retainer without an exit is a hostage situation for whoever wants out. Thirty days, either direction, no reason required. It sounds like it makes the arrangement fragile and it does the opposite, because it removes the only real reason to feel trapped.

It changes my behaviour too, in a way I have come to value. Knowing a client can leave in thirty days keeps me honest about whether this month was worth the fee. That is a healthier discipline than a twelve-month lock-in, which mostly protects revenue from the consequences of mediocre work.

The symmetry is the important part. If I can leave and they cannot, that is exploitative. If they can leave and I cannot, I will end up serving a relationship that has stopped being good for either of us. Equal notice is the only version I am comfortable signing.

What kind of work actually suits a retainer?

Work that recurs and degrades. A site that keeps growing, content that keeps shipping, automations that run unattended and eventually break. Anything where the value is in something staying good rather than in something being built once, and where nobody else is watching it closely enough to notice the day it stops.

The automation work is the clearest case. I run automations in production for Ajust on Airtable with WhaleSync, and for Kismet Health on HubSpot through Zapier. Systems like that do not finish. APIs change, volumes grow, edge cases arrive, and someone has to be the person who notices. That is a retainer whether you call it one or not.

What does not suit a retainer is a redesign, a launch, or anything with an end. Putting finite work on a monthly fee just makes the client pay in instalments for something that should have been scoped, and it leaves both of you unclear about when it is done. Scope it as a project and be glad when it ends.

What still does not work well about my retainers?

Quiet months. A month where nothing breaks and nothing ships is exactly the month the retainer is working, and it is also the month a client is most likely to wonder what they are paying for. I have not fully solved that and I am not sure it is solvable.

My partial answer is to report on the invisible work: what I checked, what I found, what did not happen because of it. It helps. It does not fully overcome the fact that prevention is unsatisfying to buy, which is the same reason people resent insurance until the week they need it.

The other unsolved part is knowing when to end one. A retainer that has become comfortable for both sides but stopped producing anything is a slow harm, and I am worse at spotting that in my own work than in a client's. Usually the honest move is to propose ending it before they do.

What should you do next?

If you are running hour-bank retainers, rewrite one as outcomes before your next renewal. Name three things that will be true each month, name five things that are not included, and set equal notice periods. One page, and it will be the most useful page in your contract.

If you are buying a retainer, ask the person selling it what you will be able to tell at the end of a month. If the answer is a number of hours, you are buying availability rather than progress, and you should decide whether that is genuinely what you want.

Six-plus years, 70+ projects and 25+ clients later, the retainer is the part of my business I have rewritten the most and am least certain about, which is precisely why I would rather write the honest version than the confident one. If you are restructuring yours and want to think it through with someone who has got it wrong several ways, reach out.