MVP (minimum viable product)

Eric Ries defines the MVP as the version of a new product that lets a team collect the most validated learning about customers with the least effort. In practice the word now covers anything from a landing page with a sign-up form to half of a roadmap, so it pays to say which one you mean.

We use it for a working product. Its intended users can sign in and finish a meaningful task from start to end, including what happens when a step fails. The scope can be narrow: one kind of customer, one integration, a pilot for a few accounts, or a step a person handles by hand behind the scenes. The parts it does have must hold up in real use, because people depend on them.

Some questions only real use can answer: whether a team comes back every week, whether a customer will pay, whether the workflow survives a busy month. A demo cannot show those. An MVP is the cheapest way to get that evidence, and it stops a company spending a year building features nobody has asked for. Cut too far and it fails for a different reason: people try a product that cannot finish the job, and the team learns nothing from their leaving.

An MVP, or a build of that size, usually takes us eight to twelve weeks from a settled scope. We agree who it is for, the task it has to complete and what waits for later, then design the whole task, error and empty states included. The product is put together from design system components written in code, and it launches to the people it was scoped for, with the code in the client’s accounts.

The product we have built end to end is our own, Prism, which gives service businesses one view of the work across their organisation. We started with how owners, managers and employees each move through it. It had to be useful the moment someone opens it, showing what needs attention first.

Say a startup wants to help physiotherapy clinics fill cancelled appointments from a waitlist. The first version serves one kind of clinic and one booking system. When a slot frees up, it texts the next patient on the list. If nobody replies within an hour, the receptionist gets a prompt to phone instead, and clinics are told that step is manual. The team counts slots filled each week, and that number decides what gets built next.