How to scope an MVP: plan your first app release

Scope an MVP around one customer workflow. Work through a first-release example, decide what can wait, and write acceptance checks before hiring a developer.

By Chase Poirier · Updated · 8 min read

A feature list can grow forever. A first release needs an end. I would scope your MVP around a result one customer can finish, then make every feature justify its place in that workflow.

A first-release scope

  1. Customer result
  2. Essential workflow
  3. Acceptance checks
A first-release scope. A planning framework, not a client result.

Define the customer result

Write one sentence describing who uses the product, what they need to do, and what a successful result looks like. A consultant might want a client to submit information and receive a reviewed recommendation. That is more useful for scoping than "a dashboard with AI."

Next, walk through the whole journey. How does the customer get access? What information do they supply? Who reviews it? Where does the result appear? The first release needs enough of that journey to work without hidden gaps.

Turn an industry method into one complete transaction

Imagine a consultant who reviews supplier risk for small manufacturers. Today, the customer fills out a spreadsheet, emails it back, and waits for the consultant to produce a report. The proposed software business would let other consultants sell the same assessment to their clients. This is an illustrative product, not a past engagement.

The first release could serve one consultancy and one assessment type. A client receives an invitation, answers the questions, saves a draft, submits it, and returns to read the approved report. A reviewer sees submissions and publishes a result. That sequence exposes work a feature list can miss: invitations, draft recovery, review status, and permission to read the report.

Write the release boundary in those terms before adding a score builder or a marketplace for consultants. You can evaluate whether a customer completed an assessment. You cannot evaluate whether an undefined platform is complete.

A smaller release can still finish the job

Imagine a specialist who wants to sell an assessment app to other firms. The long-term vision includes automatic scoring, team dashboards, and integrations. The first customer needs to create an assessment, submit answers, and receive a result they can use.

The example below illustrates scope decisions. It is not a client project or a priced package.

One possible first-release decision
Include nowHandle manually or deferWhy
Customer login and assessment submissionSingle sign-onInitial customers can use individual accounts.
A review screen and delivered resultFully automated expert recommendationsThe founder can review results while testing the workflow.
A basic purchase or agreed billing processComplex billing plansCustomers need a clear way to pay, not every possible plan.

Choose what waits by naming the consequence

For each requested feature, ask what happens to the first customer if it is absent. If the answer is that the founder performs a manageable task, document that task. If the customer cannot receive the promised result, either include the work or change the release promise.

For the assessment example, the following choices keep the first transaction complete. The decisions would change for a self-service product sold to hundreds of firms, or for a customer whose procurement process requires organization-wide sign-in.

Illustrative assessment scope and the consequence of deferral
RequestFirst-release choiceConsequence to check
Custom question builderUse one agreed questionnaireThe founder must approve the questions before development.
Automatic scoringReviewer enters the approved resultA reviewer must be available within the promised turnaround.
Historical spreadsheet importBegin with new assessmentsPilot customers must accept starting without old records.
Team permissionsOne client user and one reviewer roleA client needing several reviewers may fall outside the pilot.
Branded PDF exportProvide a usable report in the accountCheck that the customer can use the result in their real process.
Multiple pricing plansOne agreed purchase arrangementSales materials must describe the same arrangement the product supports.

Make manual work an explicit decision

A manual step can save development effort when the founder has capacity to perform it and the customer still receives the promised result. Record who does it, how often, and what volume would make it impractical.

Do not hide essential work behind "we will do that later." If nobody can deliver the assessment result without the missing integration, that integration is a launch dependency rather than an optional improvement.

Calculate the capacity behind a manual step

Suppose each report needs 20 minutes of founder review. With 12 submissions a week, that is four hours of review before customer questions or corrections. At 60 submissions, the same task takes 20 hours. Those are scenario assumptions, not measured customer demand.

The first scenario may fit your available time; the second may not. Record a trigger such as reviewing automation when the weekly queue exceeds your review capacity for two consecutive weeks. Choose the threshold from your actual availability and service promise.

Manual work also needs an owner and an absence plan. If you are away for a week, can another reviewer deliver the result, or must customers wait? Put the answer in the operating plan and avoid selling a turnaround the first release cannot support.

Describe what finished means

For each essential workflow, write a check someone can observe. "A customer can submit an assessment and retrieve their own result" is testable. "A scalable assessment platform" leaves too much open.

Include failure cases: incomplete answers, duplicate submissions, and users attempting to open another customer's information. Your developer should help turn the agreed scope into appropriate acceptance tests.

First-release scope checklist

  • Name the initial customer and the result they need.
  • List the steps from first access to that result.
  • Mark essential, manual, and deferred work.
  • Record unknown integrations and data dependencies.
  • Agree on acceptance checks and founder responsibilities.

Replace a feature label with checks you can witness

An acceptance check describes a starting condition, an action, and an observable result. Use ordinary business language first; the developer can translate it into technical tests. Include the awkward path as well as the successful one.

For the assessment workflow, confirm that a client can save and return to an unfinished form, that submission places it in the correct review queue, and that publishing a result makes it visible only to the intended client. Decide whether a submitted answer can be edited and what happens to a result already under review.

This prevents a late disagreement over what the word assessment included. It also gives you a way to review a milestone without inspecting code. A demo with prepared records should be followed by your own attempt using the agreed starting conditions.

Given a client has an unfinished assessment
When they sign out and return
Then their saved answers remain available

Given client A has a published result
When client B tries to open its address
Then client B cannot read the result

Scope the route into the product

For a product you plan to sell, include the steps before the main app workflow. How does a prospective customer understand the offer, sign up or book a demo, and reach the first useful result? A working dashboard does not answer those questions by itself.

List the launch steps beside the product scope and name who owns each one. In Vision to Launch, the launch funnel is agreed in the quote. Recruiting customers and running ongoing acquisition remain the founder’s responsibility.

See the scoped product and launch offer

Keep a place for the larger vision

Deferred work belongs in a separate list, with the evidence that would make it worth building. A request from a pilot customer may justify investigation; it does not automatically justify adding it to the current release.

At TheShop, we use these decisions to scope the build and explain the tradeoffs. If a material technical unknown prevents a reliable quote, we discuss bounded discovery before committing to the build.

For Magnet, Chase worked with Alexander Jones to translate a fundraising method into workflows and wireframes, then built the first product with donor tasks, contact actions, team administration, and billing limits. The case study shows how a specific professional method became a release scope.

Keep a change from quietly becoming a second product

If a pilot customer asks for multi-office reporting, record the need behind it. They may need a comparison once a month, which could initially be supplied manually. They may instead require daily access for several regional managers, which changes permissions and reporting scope.

Ask the builder for the effect on the current release before agreeing to the request. A scope swap should compare effort, dependencies, and risk. Removing a simple settings screen does not automatically pay for organization-wide reporting. Preserve the original release boundary until both sides agree to a replacement.

See the Magnet first-release work

Common questions

How many features should an MVP have?

There is no useful universal feature count. Choose the smallest set that lets your intended customer complete the agreed workflow, including the administrative work needed to deliver the result. A request form and an administrator response may belong together; a reporting dashboard may wait. Describe the finished customer journey before counting features.

Can part of an MVP be handled manually?

Yes, when the manual step can deliver the promised result at the expected initial usage. Record who does the work, how long it takes, and what happens if they are unavailable. Keep the customer informed where the service depends on human review. Use early usage to decide which step deserves automation next.

What should I give a developer before scoping an MVP?

Bring the customer problem, your current way of solving it, and any evidence from prospective users. Explain your available budget and relevant deadlines. You can leave technical choices open. Ask the developer to describe a complete first workflow, its exclusions, and the observable checks that would show it is ready for use.

Put the release scope in writing

Use the product brief to record the customer, essential workflow, manual steps, and what you are leaving for later. You can share that document with any shortlisted builder.

Write your product brief