Knowledge

What belongs in an MVP? Scope, omissions and learning

An MVP is a deliberately limited product that helps investigate an important business assumption through real use. Its value comes from a complete, useful journey for a chosen audience, not simply from having fewer screens. Define the learning goal, operational responsibilities and evaluation plan together before estimating implementation.

Planning an app screen sequence with paper sketches.

Choose an assumption that leads to a decision

Identify the initial audience, the situation in which they act and the difficulty the product aims to resolve. Ask an observable question: can these users complete the core task independently, and do they choose the service again when the need returns? Treat the release as a way to investigate uncertainty rather than a demonstration of a result already promised.

Specify what you will decide after examining use: continue with this audience, change the journey or investigate another approach. Distinguish business uncertainty from usability and technical uncertainty. An interface prototype may answer a navigation question, while a targeted technical exercise may reveal an integration constraint. Building a full product is not always the appropriate first experiment.

Build the smallest complete journey

Consider an imagined booking app. Its first journey could cover choosing a service, finding availability and receiving confirmation. This is an illustrative scenario, not a client reference. A loyalty programme might be deferred, but uncertainty about whether a booking exists must be addressed. Define completion from the user's perspective rather than stopping the specification at a button press.

Include empty states, interrupted requests and cancellation where relevant. A simple explanation and a safe retry may be sufficient for some cases, while others require more work. Also describe the organisational side: who manages content, answers questions and corrects records? Removing an action from the mobile interface does not remove the need to perform it somewhere.

Tie priorities to the purpose of the first release

For each item, ask whether it is necessary to investigate the assumption, complete the promised journey or provide required protection. Defer features aimed only at a later audience when that does not undermine the initial goal. Record missing information rather than assigning invented precision to uncertain priorities. Value, dependencies, uncertainty and effort are different considerations.

Keep a written list of omissions with reasons, but do not promise to implement every deferred idea. Real use may make some irrelevant. The initial agreement should distinguish committed behaviour, exclusions and items requiring another decision. This gives later requests a clear context and helps prevent scope changes from becoming invisible obligations.

Illustrative booking MVP decisions
ItemReason for the initial decision
Booking and confirmationNeeded to complete the core task
Uncertain result handlingUsers need to know whether the booking exists
Manual supportPossible with a named owner and workable process
Loyalty programmeDefer if it does not test the first assumption
Personal recommendationsRequire a separate product justification

Use manual operations deliberately

Manual approval or simple administration can replace some early automation. Identify the operator, how work arrives, how completion is recorded and what the user sees while waiting. Monitor the burden on that person. A hidden process known only to a developer is a fragile foundation even when the interface looks simple.

Do not treat a smaller feature set as permission to omit access controls, recovery planning or understandable error states. Determine the appropriate requirements for the actual product. If an error loses essential records or makes the result untrustworthy, it also damages the evidence you hoped to gather. Keep the system small enough to understand and complete enough to operate.

Investigate expensive uncertainty early

Use a clickable prototype to check terminology, navigation and missing information with intended users. Let them attempt the task instead of guiding every step. Record confusion and expectations. A successful prototype exercise supports usability decisions; it does not prove that a backend, hardware connection or production operating model will work.

If an external API or device connection is critical, investigate it through a targeted technical exercise before expanding the interface. Choose technology around required behaviour, maintainability and handover needs. We develop native iOS apps in Swift and Android apps in Kotlin, as well as Flutter, React Native and PWA applications. We also build backends and APIs, IoT/BLE apps and AI features. We choose the MVP implementation around its core user task. Also justify the initial platform coverage using the audience, acknowledging who is excluded and when the decision will be reviewed.

Measure what supports the next decision

Define the meaning of each event and the question it answers. Opening the app is not necessarily evidence of useful work. Distinguish a request being sent from a booking being created. Combine observed behaviour with conversations and support feedback so you can understand why a journey stops rather than merely counting incomplete actions.

Record the version and circumstances you examined. Avoid general market claims based on limited or unusual use. Agree which observations justify continuing, which require more investigation and which indicate that the core flow needs repair before expansion. Include operational effort in that review: a product can appear convenient to users while creating an unsustainable manual workload.

Plan the data handling too. In this implementation approach, optional analytics starts only after prior consent; have the actual requirements reviewed for the chosen setup. Avoid collecting detailed personal profiles when the question needs less information. This is general planning information, not legal advice. Test refusal behaviour as part of the release rather than judging it only by the appearance of a control.

Define done and assign the next review

Use concrete acceptance scenarios: the selected user completes the task with test data, understands the result and can receive help through the agreed route. The operator can handle related administration. Keep release instructions, known limitations and issue reporting information. A small product still needs an owner and a way to make a safe update.

After release, evaluate usage evidence and operating work together. The next step may be automation, a simpler core interaction or another experiment. Do not automatically execute the original wishlist. The purpose of the MVP is to make subsequent investment decisions better informed while maintaining the behaviour already promised to users.

Frequently asked questions

Is an MVP an unfinished app?

It is a product with deliberately limited scope and a complete agreed journey. State omissions and known limitations before release.

Which work can remain manual?

A background operation may remain manual if it has an owner, a clear process and manageable demand without making the user outcome unreliable.

How do we decide what to omit?

Check whether the item is needed for the first assumption or the promised operation. Other items may be deferred with a recorded reason and later review.

How long does an MVP take?

Estimate from scope, existing foundations, integrations and identified uncertainty. The label MVP does not imply a universal delivery deadline.

Discuss your project

Describe the audience, first use case and assumption you want to investigate. Include a journey sketch, existing systems and ideas you are willing to defer. These provide a basis for an agreed initial scope and a project specific implementation discussion.

Project enquiry form
Request a quote

Tell us what you need

A few details and a short description are all we need. We’ll get back to you with the details.

What happens next

  1. Describe the task, with a redacted sample if useful.
  2. An optional AI assistant asks clarifying questions.
  3. A person reviews your request, then we get back to you with the details.
  • Built around your workflow
  • With human review
  • Design, app, maintenance
[email protected]
0 / 4000