Define one core outcome, map the full journey and separate essential dependencies from optional improvements. Test uncertain assumptions early and agree a deliberate list of things the first release will not do.
01 / Decide what the release needs to teach you
An MVP is a first working product used to learn from real customers. A proof of concept instead investigates whether a technical idea is feasible; a prototype explores an experience before the full implementation exists. Choose the right experiment for the uncertainty. You may not need a production application to test whether anyone understands the offer.
Write a hypothesis with an audience, problem and expected behaviour. For example: people coordinating a small team will use a shared task view to assign and complete work. Identify what would support or challenge that hypothesis. A download alone may say little about whether the product actually helps.
02 / Map one complete customer journey
Describe the path from arrival to a useful outcome. For a simple task product: create a task, understand its state, complete it and find it again. Include the business side where necessary. Someone may need to administer content, investigate a failure or remove an inappropriate record.
Use a journey map to distinguish a dependency from a convenience. If the task cannot complete without a data store, it is a dependency. If a configurable dashboard merely makes the experience more flexible, it may wait. Avoid shipping a collection of partially finished capabilities that never deliver the promised outcome.
Reduce the number of promises in the first release, not the care with which you deliver them.
03 / Give each feature a reason to exist
For each proposed feature, record the customer task, evidence, expected value, implementation effort and dependency. Ask what breaks if it is removed from the first release. Stakeholder enthusiasm is worth hearing, but it should not automatically turn an idea into a launch requirement.
Use simple prioritisation to make trade-offs visible. Our MVP Scope Builder separates essential items and ranks other work against relative impact and effort. Its points are planning units, not engineering hours. If essentials exceed the available capacity, the answer should expose the conflict rather than hide part of the promised experience.
- Essential: the core journey cannot work without it
- Useful: improves the journey but has a workable alternative
- Later: valuable after evidence or a larger audience
- Unknown: investigate before committing
04 / Keep a clear quality floor
Minimum scope still needs understandable feedback, appropriate access controls, error recovery and a reliable release process. If people can submit information, decide how validation works and what happens when a connection fails. If money changes hands, verify payment state and recovery rather than only the happy-path button.
Quality requirements should match the consequence of failure. A private pilot can have a narrower operating arrangement than a broad public launch, but the limits need to be explicit. Do not call missing essentials “MVP” merely to preserve a delivery date.
What happens when a save fails?
- SubmitCheck the required fields.
- Connection lostTell the user their changes are still here.
- Keep the draftPreserve everything they entered.
- Retry safelyConfirm the save once, without duplicates.
05 / Test the most expensive assumptions first
If the main risk is comprehension, put a prototype in front of representative users. If it is an integration, run a technical investigation against its real documentation and constraints. If it is demand, test the proposition and commitment required from a customer. Choose the smallest test that can change the decision.
Document what you learned and update the scope. A screen that looked straightforward may require a different permission model; a planned feature may prove unnecessary. Treat changes as decisions with a cost and a reason, not as invisible additions to the original estimate.
06 / Write a practical first-release agreement
Capture the included journeys, acceptance criteria, exclusions, dependencies and customer responsibilities. Add a pilot audience, a feedback method and the measures you will inspect. State who provides content, approves changes and makes the final launch decision. These details reduce ambiguity on both sides.
Keep a later-release list and define what would bring an item forward. After launch, look at whether people reach and repeat the useful action, not simply whether the roadmap shipped. Use that evidence to decide the next increment rather than automatically building every deferred feature.
What to take away
A focused MVP makes one valuable promise, delivers the whole journey and creates a way to learn. Keep the launch small through deliberate exclusions, while preserving the quality the task requires.
Turn the feature list into a first release.
Compare impact, effort and capacity in the scope builder. Keep or download your result without an email gate.
Build your scope