A Minimum Viable Product (MVP) is the smallest version of your product that delivers real value to real users and teaches you something important. Done well, an MVP saves months of effort and a large share of your budget. Done badly, it becomes either a half-finished prototype nobody trusts, or a “version 1” that took a year to build.
1. Start with the problem, not the features
Write one sentence describing who your user is and what problem you solve for them. Every feature in the MVP must directly support that sentence. If it doesn’t, it goes to the backlog.
2. Define what you want to learn
An MVP is an experiment. Decide upfront which assumptions you’re testing. Will customers pay? Will they use it weekly? Will it replace their spreadsheet? Then define measurable success criteria for each.
3. Prioritise ruthlessly
Sort every feature into three buckets:
- Must have: the product doesn’t work without it.
- Should have: important, but users can manage without it for a few weeks.
- Could have: nice to have. Revisit after launch based on feedback.
Only the “must haves” go into the MVP.
4. Choose architecture that can scale
You don’t need microservices on day one, but you do need clean code, a sensible data model and automated deployments. These decisions are cheap at the start and very expensive to fix later.
5. Plan the feedback loop
Add analytics, collect user feedback and schedule regular review sessions. The goal of launching is to learn quickly and decide what to build next with evidence, not guesswork.
Need help planning your MVP?
BP Tech Consulting runs focused MVP planning workshops that turn your idea into a clear scope, roadmap and estimate. Get in touch to book a free consultation.