Answer
How many features should an MVP have?
Short answer
Three, in most cases. One feature that delivers the core value, and up to two that are genuinely required to support it. Below three the product usually cannot demonstrate anything; above five the build time, the bug surface and the cost all rise faster than the learning does.
Three is not a magic number, it is an observation about where the returns fall off. Features do not add time linearly, because every new feature interacts with the ones already there. Five features are not five-thirds the work of three, they are closer to triple.
A useful exercise: list every feature you think you need, then force-rank them. Not by importance — by what a user would do first. The top three are almost always the MVP, and the ranking exercise is usually more clarifying than the list was.
The counter-intuitive part is that fewer features make the product easier to sell, not harder. A product that does one thing is easy to explain in a sentence, and a product you cannot explain in a sentence does not get shared.
The numbers
- Recommended
- 3 features: one core, up to two supporting
- Why not more
- Build time and bug surface grow faster than feature count
- Why not fewer
- Below three, most products cannot demonstrate their value end to end
- How to choose
- Rank by what a user does first, not by what feels most important
- Common mistake
- Counting a dashboard or an admin panel as a feature. They are not — they are a consequence of having data.
Last reviewed .
Read next
Want this applied to your idea?
A 30-minute call turns a general answer into a specific one: whether your idea fits a 48-hour build, which three features matter, and what it would cost. You get a straight opinion, including when the answer is no.
Fixed $3,500 AUD. Scope agreed in writing before anything is built.