An MVP, or minimum viable product, is the smallest version of your product that is enough to test the most important assumption behind your idea with real customers who use it or pay for it. Build first the one path that solves the core problem from start to finish, postpone everything else, and let the budget follow the size of that path, not the length of your wish list.
This guide explains what makes a first version truly "viable", gives you steps to decide what to build first, what should not be postponed even in the smallest version and what you can safely postpone, then covers what drives MVP cost, the mistakes that waste it, and an example from HBS work.
What is an MVP, and why is it not a poor version of your product?
Eric Ries, the author of the Lean Startup method, defines it in his guide to the minimum viable product as the version that lets a team collect the maximum amount of validated learning about customers with the least effort. The key word is "learning": the goal is not to launch a small product but to get an answer to a specific question. So an MVP is small in scope, not in quality, and the path it offers must work well, so the customer judges the idea rather than the bugs.
The term is often confused with two close concepts:
| Stage | What it is | Who uses it | The question it answers |
|---|---|---|---|
| Prototype | An interactive design or a model that does not fully work | A few potential customers in test sessions | Does the customer understand the idea and know how to use it? |
| MVP | A product that actually works for one core path | Real customers in real conditions of use | Does the customer use it, come back to it, or pay for it? |
| Full product | A product with many features, plus operations and marketing | The whole target market | How do we grow and keep customers? |
Before any of these stages comes validating the idea itself: is the problem real, who has it, and will they pay to solve it? We covered the tools for that stage in our guide to starting a company with AI, from idea validation to launch.
Does an MVP have to be custom software at all?
Not always. The question you want answered decides the shape of the first version, and sometimes the answer costs far less than development:
- A landing page with pre-orders. It describes the product and its price and asks visitors to reserve or pay a deposit. It measures real interest before anything is built.
- A manual service at first. You deliver the result the product will deliver to a few customers through your team's effort, so you learn exactly what they need before automating it.
- A working front with manual work behind it. The customer sees a working product while your team carries out some of the steps by hand behind the scenes.
- Ready-made tools stitched together. Online forms, a ready-made store and spreadsheets are sometimes enough to test demand before you pay for development.
- A custom product with a single path. When the question is about the user experience itself, or when the value cannot be delivered by hand.
How do you decide what to build first?
Write down your answers to these steps before you talk to any developer:
- Write down your riskiest assumption. The one that, if wrong, brings the whole idea down. For example: "Owners of small clinics will pay a monthly subscription so their patients can book online."
- Pick one user and one job. Not "all small businesses", but one type of customer and one job they want done.
- Map the core path from start to finish. From the moment the customer arrives to the moment they get the result, step by step.
- Cut every step the path does not need to complete. For each feature, ask: does the test fail without it? If not, it is postponed.
- Decide what you will do by hand behind the scenes. Many steps, such as reviewing requests or sending reports, can be done manually by your team at first, until they prove they are worth building.
- Set the success metric and its deadline before building. Such as how many customers complete the path or pay within a set period. Without a number written in advance, every result looks "encouraging".
Once you finish these steps, write the resulting scope in a short document you can send to more than one development team. You will find a ready template in our guide to the software requirements specification (SRS).
What should not be postponed, even in the first version?
An MVP is small in features, but some foundations cost a lot to add later. Do not postpone these:
- Separating customer data if the product serves more than one company. Changing the data structure once you have real customers is costly and risky.
- Basic permissions. Who sees what, even with only two roles at the start.
- Payment, if your question is "will they pay?". Verbal interest proves nothing; actual payment is the answer.
- Arabic and right-to-left layout from day one. If your market is Arabic-speaking, converting a product built in English to Arabic later is much harder than building it in Arabic from the start.
- Security basics and backups. The data of your first customers matters as much as the data of thousands of customers later.
- Usage tracking. If you do not measure which steps customers complete and where they stop, you will learn nothing from the launch.
- Ownership of the code and accounts. Even for a small product, the code, the servers and the domain should be in your name.
What can you safely postpone?
- An advanced admin panel. A spreadsheet or a simple panel is enough in the first weeks.
- The mobile app. A web app that works well on phones is usually enough to test the idea before paying for two store apps.
- Automation. What one person can do by hand for a few customers does not need code yet.
- Many integrations. A file export or manual entry is enough until you know which integration is worth building.
- Detailed reports. Start with the metric you set for success, and add reports as questions come up.
- Multiple payment methods and complex plans. One payment method and one plan are enough to test willingness to pay.
- Additional languages. Start with the language of your first market and add others when customers need them, as long as the product's structure is ready for that from the start.
How much does an MVP cost, and what drives the cost?
There is no single right figure for MVP cost, because the term describes an approach, not a size. An MVP for a platform that connects buyers and sellers is a completely different project from an MVP for an internal tool with one type of user. These are the factors that set the cost:
| Factor | Pushes cost up | Brings cost down |
|---|---|---|
| Platforms | Web plus Android and iOS apps | A web app that works well on phones |
| User types | A platform with two or more sides, each with its own interface | One user type and a simple admin panel |
| Integrations | Payments, shipping, invoicing and messaging from day one | One integration the test needs, the rest done by hand |
| Design | A fully custom identity and interfaces from scratch | A ready design system with a touch of your brand |
| Real-time and AI features | Live chat, live tracking or custom AI models | Simple notifications and ready services through APIs |
| Clarity before building | An open scope that changes every week | A written scope and a defined success metric |
The practical way to manage the budget is to flip the question. Instead of "how much will all of this cost?", ask "what can be built well with this budget and in this time?", then keep part of it for changes after customers start using the product, because the first version will not be the last. For price ranges for apps in the Egyptian market and what moves them, see our guide to mobile app development cost in Egypt. And if your product is a subscription platform serving other businesses, you will find its steps in our guide to building a SaaS platform from idea to launch.
Which mistakes waste the value of an MVP?
- Building it for everyone. When it tries to serve every kind of customer, it learns nothing clear from any of them.
- Polishing before launch. Extra weeks on cosmetic details delay the one answer you need.
- Testing it only with friends. Politeness measures nothing. A stranger is the one who tells you the truth.
- Choosing a technology only one person knows. If the only developer who understands the code leaves, the product stops. Choose common technologies that are easy to find people for.
- Spending the whole budget on the first version. Nothing is left for changes when the feedback arrives.
- Ignoring an unwelcome result. If the metric is not met, that is valuable information that deserves a change to the idea, not to be ignored.
How does HBS approach a first version?
At HBS we start with one question: what do you need to learn from this version? Then we propose the smallest scope that answers it, and we separate the features that can wait from the foundations that cost more to postpone than to build now. We do not treat launch as the end of the work, because the weeks after it decide what is worth building next.
One example of that balance is Waslah, a micro-systems marketplace where companies install small workflow tools instead of a heavy ERP. The platform did not launch with a full catalogue of tools. It launched with one tool, EstateNexus, running in a real operational environment as practical proof of the idea. At the same time, the shared layer was built once from the start: separate workspaces for each company, role-based permissions, notifications, audit logs, and a common framework that makes adding the next tools faster. In other words, the features started small, while the foundations that are hard to change later were built properly from day one.
If you have a product idea and are not sure what to build first and what to postpone, tell us about the idea and the customer you are targeting. Our team will sit down with you in a free discovery session and help you define the smallest version that answers your question, and you can learn more about HBS custom software and digital product development.








