Retainer, Project, or Sprint: Which One to Actually Buy

Nine months ago a founder signed a $2,500 a month retainer with an agency that came recommended. The calls happen on schedule. The reports arrive. The invoices clear. Last week she tried to answer one plain question, what did I actually buy, and could not name a single asset she owns that did not exist before the contract started. No landing page in her name. No tracking she controls. No list she could export tomorrow. Twenty-two thousand five hundred dollars of spend, and the only durable thing the money produced was a habit of paying.

She bought the right thing in the wrong order. A retainer is a subscription to ongoing operations, and she had nothing operating. She was paying rent on a machine nobody had built, and every month the agency stayed billable was another month it could keep meaning to build it.

There are three ways to buy agency work, and they are not interchangeable. This post maps all three to a single sentence you can finish yourself before you sign anything. Get the sentence right and the model picks itself. Get it wrong, and you default to the one option that quietly bills you forever for a thing that was supposed to end.

Should you start with a retainer or a project?

Start with a sprint or a project, not a retainer, unless you already have a working system that only needs someone to run it. The test is one word: are you building, fixing, or maintaining. Building and fixing are finite jobs with a finished state, so you buy them as fixed-scope work; only maintaining, operating a machine that already exists and already produces, is the recurring job a retainer was designed to do.

Almost every first-time buyer is building or fixing. You do not have a funnel yet, or the one you have leaks, or the site loads in four seconds and the tracking reports numbers nobody trusts. Those are projects. They have a start, a defined deliverable, and a day they are done. A retainer has none of those things by design, which is exactly why it is the wrong container for finite work. Pour a build into a retainer and the build has no incentive to finish, because the finished state is the day the recurring revenue stops.

The confusion is understandable. The retainer looks like the safest choice from the outside. A predictable monthly number, a team on call, one throat to choke when something breaks. But predictable spend is not the same as productive spend, and a team on call for a machine that does not exist yet is just a standing invoice with meetings attached. The question is never how do I lower risk. It is what am I actually buying, and does that thing have an end.

What do a sprint, a project, and a retainer each buy?

A sprint buys one shipped deliverable in two weeks at $5,000 flat. A project buys a complete build, a website from $8,000 or a brand and site together from $15,000. A retainer buys the ongoing operation of a system that already exists, $2,500 a month for growth or $5,000 for infrastructure, month-to-month after the first term. All three sit in the open on the pricing page, and each one answers a different version of the build, fix, or maintain question.

Here is what each one is actually for.

• The sprint, $5,000 flat, two weeks, one deliverable. This is the surgical option. You know the exact thing that is broken or missing, and you want it fixed and handed over, not adopted into a standing relationship. A landing page that converts. Server-side tracking that finally reconciles to real revenue. A review engine wired up and running. The scope is one thing, the timeline is fixed, and the output is yours the day it ships. If you can name the deliverable in a single sentence, you want a sprint.

• The project, website from $8,000, brand and site from $15,000. This is the build. You do not have the machine yet, and a project manufactures it: the site, the identity, the acquisition infrastructure, built once as an asset you own outright. A project is the correct first purchase for most businesses that have never worked with an agency, because you cannot operate a system you have not built. The number is larger than a monthly retainer looks, but it buys a finished, owned thing rather than a rented activity that renews.

• The retainer, $2,500 or $5,000 a month, month-to-month after the first term. This is operation once the machine already exists. Growth at $2,500 runs acquisition on top of infrastructure that already works. Infrastructure at $5,000 keeps a larger owned system tuned, instrumented, and compounding week to week. A retainer is the right answer only when there is a running machine underneath it, and it is month-to-month on purpose, because a retainer you cannot leave in thirty days is a trap wearing a subscription's clothes.

Read those in order and the pattern is obvious. Two of the three have an end. The one that does not is the one you should buy last.

A retainer maintains a machine. It does not build one.

When does a monthly retainer actually make sense?

A retainer makes sense once there is a running system to operate: live tracking, a working funnel, an owned list, ad accounts with pixel history, a review engine that fires on its own. At that point the recurring work is genuinely recurring, someone tends the machine and improves it week to week, and the compounding accrues to you because you own the machine being tended. Before that state exists, a retainer is a subscription to good intentions.

Look at what a real operating machine is before you agree to pay monthly to run one. At Skin & Self, the build came first: a rebuilt site, server-side conversion tracking, a 40,000-contact CRM synced, an automated review engine that now sits at 4.9 stars across 757 reviews. Only once that infrastructure existed did operating it monthly mean anything, and operating it is what produced $1.3M in attributed revenue at 6.7x ROAS. Run the retainer before the build and there is nothing under the monthly fee but hope. The order carries the entire mechanism.

This is also where the honest retainer and the dependency-engineering retainer look identical on the invoice and behave like opposites underneath. An open-ended retainer with a vague scope, accounts registered in the agency's name, and reporting built to impress rather than inform is designed so that the machine never quite becomes yours and the monthly fee never quite ends. A healthy retainer operates a system you own, reports against real revenue, and could be canceled in thirty days without you losing a thing you paid to build. The tell is the same one from the build, fix, or maintain test: if you cannot point at the machine the retainer maintains, there is no machine, only the fee.

None of this means retainers are the villain. A machine that runs unattended still rots. Tracking drifts, ad accounts fatigue, lists decay, a funnel that converted last quarter quietly stops. Operating infrastructure well is real, ongoing, valuable work, and it is precisely what the $2,500 and $5,000 monthly tiers are for. The failure comes from buying that operating contract before the thing it operates exists, which is how a first-time buyer ends up nine months in with reports, invoices, and nothing owned.

Buy in the right order

The order is the entire lesson. If you are building or fixing, buy a sprint or a project, take the owned deliverable, and go run it yourself for a while. If it turns out you need a hand keeping the machine tuned once it is live, add a retainer then, month-to-month, with a machine you can point at. That sequence is the same one we lay out in what to buy in what order on a first budget, and it is the questions worth asking on any agency call, ours included, in how to hire without getting burned again.

If you already signed the retainer first, you are not stuck. You are one clear question away from renegotiating: name the machine it maintains. If the agency can point at owned, working infrastructure, keep operating it. If they cannot, you have found a build being billed as maintenance, and the fix is to scope the build, ship it, own it, and only then decide whether anyone needs to run it monthly.

Tell us which one you are, building, fixing, or maintaining, and we will tell you which of the three to buy, even when the honest answer is the smaller number. Book a call.