Founders rarely struggle to think of features. They struggle to say no to them. Every investor meeting, customer call and competitor launch adds another item to the list, and soon the minimum viable product is neither minimum nor launched. Months pass, budget runs out, and the one question that mattered, whether people want this, is still unanswered.
This guide is about the cutting. It shows how to find the single assumption your MVP must test, which features most founders can safely leave out of version one, what you should never cut, and how to handle the rest manually until users give you a reason to build it.
Quick answer: Keep only the features needed for one user to complete the core journey that tests your riskiest assumption, plus security, payments, analytics and backups. Leave out extra roles, admin automation, both native apps, multiple languages, advanced search, loyalty schemes and most integrations. Handle edge cases manually and build them only when real usage shows they are needed.
Start with the one question your MVP must answer
Write down, in one sentence, the assumption that would sink the business if it were wrong. For example: small clinics will pay a monthly fee to send appointment reminders, or dealers will place repeat orders through an app instead of calling the sales rep. That sentence is your filter. A feature belongs in version one only if a user needs it to reach the moment that proves or disproves the assumption.
Next, draw the core journey as five to eight steps, from sign-up to the moment of value. Anything outside that path is a candidate for later.
Features most founders can leave out of version one
- Detailed admin dashboards: early on, a simple admin list with filters and an export is enough. Charts can wait until you know which numbers you watch.
- Many user roles: start with one or two roles. Fine-grained permissions matter when teams of customers use the product, not before.
- Two native mobile apps: a responsive web app or a single cross-platform app usually tests demand just as well as separate Android and iOS builds.
- Several languages: launch in the language your first users actually use and design the text so translations can be added later.
- Advanced search and filters: with a few hundred records, basic search works fine.
- Referral, loyalty and gamification: these grow an audience you do not yet have.
- Integrations with every tool: pick the one or two that the core journey depends on, such as a payment gateway or SMS, and skip the rest.
- AI features that are not the product: if the core value does not depend on AI, a chatbot or recommendation engine is a distraction in version one.
- Settings for everything: make sensible defaults and remove options nobody has asked for.
What you should never cut
Cutting scope is different from cutting quality. These items protect your users, your data and your ability to learn, so they stay in even the smallest MVP:
- Security basics: safe login, encrypted connections, role checks on every request and no secrets in the code.
- Correct payments: if you charge money, payment confirmation, failed-payment handling and refunds must work reliably from day one.
- Analytics and event tracking: without tracking each step of the core journey, you cannot tell whether the MVP succeeded.
- Backups and error logging: losing early users' data costs trust you will not get back.
- A clear privacy notice and consent: say what personal data you collect and why, and collect only what you need.
- A way to reach you: a support email or in-app feedback form, so early users can tell you what is broken.
- Clean, readable code: a small, well-structured codebase can grow. A rushed one often has to be rewritten.
Do it manually first: workarounds for version one
Many features can be replaced by a person or a simple tool for the first few months. This is often called a concierge approach. It is slower per user, but it teaches you exactly what to automate.
| Feature | Version one workaround | Build it when |
|---|---|---|
| Matching or recommendations | The founder or an operations person does it by hand | Volume is more than one person can handle well |
| Reports for customers | A weekly export sent by email | Several customers ask for self-service reports |
| Multiple pricing plans | One plan, with special cases handled through payment links | Users repeatedly ask for different plans |
| Onboarding flow | A short onboarding call or video | Calls become the bottleneck to growth |
| Notification centre | Email or SMS for two or three key events | Users miss updates that matter to them |
| Content management | The team edits content directly through a simple admin form | Non-technical staff need to publish often |
A simple scoring method for feature requests
- List every requested feature in one sheet, with who asked for it and why.
- Mark whether it sits on the core journey. If not, it moves to the later list straight away.
- For the rest, score value to the riskiest assumption, build effort and risk on a simple 1 to 3 scale.
- Sort them into must have, should have, could have and will not have now, the MoSCoW method.
- Draw a line where your budget and timeline end. Everything below the line waits.
- For each parked feature, write the evidence that would bring it back, such as five paying users asking for it.
Once the MVP is live, the same list becomes your roadmap. Our article on the technology roadmap from MVP to enterprise covers what comes after launch.
Signs your MVP scope is still too big
- You cannot describe the core journey in under a minute.
- The estimate keeps growing, but the riskiest assumption has not changed.
- More than one user type must be fully served before anyone gets value.
- Features are justified with what a competitor has, not what a user asked for.
- Launch depends on a partner integration you do not control.
If two or more apply, cut again before you start building. A focused MVP launched sooner teaches you more than a larger one launched late.
How EthicsComputer can help
Our MVP development work begins with a discovery sprint that defines the problem, users and smallest valuable scope, and we suggest what to cut when a budget needs to fit. Most MVPs typically launch in 8 to 14 weeks after a 1 to 2 week discovery sprint, with analytics set up from day one and full source code handed over to you. If you need help deciding scope, stack or vendors before any build, see our startup technology consulting.
Is your MVP feature list growing faster than your budget?
Send us your feature list and the assumption you want to test, and we will help you find the smallest version worth building.
About the author
Vinay Grover founded EthicsComputer in Rohtak, Haryana in 2018 and leads its cybersecurity, software and AI work for clients across India and abroad. EthicsComputer is ISO 9001, ISO/IEC 27001 and ISO 20000-1 certified.