Short answer
A traditional agency quotes a fixed scope, builds it, and quotes again when anything changes. The subscription charges one flat monthly fee and absorbs change through a queue. Fixed scopes suit projects with a fixed budget and a stable spec; the subscription suits live products where priorities move every week.
How the agency model works
You describe what you want. The agency scopes it, estimates the hours, adds contingency and quotes a price. Once work starts, anything outside the scope becomes a change request with its own quote. When the project ends, so does the team’s attention, unless you buy a support retainer measured in hours.
That model made sense when building was slow and expensive. With AI tools, the build itself has become much faster. What hasn’t changed is the cost of scoping every idea, negotiating every change and starting again for the next project.
How the subscription differs
| A traditional agency | Fluff, on subscription | |
|---|---|---|
| Getting started | Weeks of scoping and quotes | A short onboarding session, usually within a week or two |
| How you pay | A fixed quote, then change orders | One flat monthly fee |
| Changing your mind | Another quote | Add it to the queue |
| Senior judgment | Billed by the day | Built into every release |
| After launch | A support retainer measured in hours | The same queue keeps moving |
| When you need more | Start a new project | Add a track |
When a fixed-scope project is the right answer
Sometimes you need one number up front: a grant-funded build, a procurement process, a board that wants certainty. We still do that. It’s called a System Build, and it’s usually preceded by a Discovery Sprint so the scope is worth fixing.
Why we moved away from it ourselves
Changes that used to take hours now take minutes. Pricing by the hour undersold what clients get, and fixed-price MVPs started to look expensive next to something knocked up overnight in an AI app builder, even though making it production software is a different job. The full story is here.