Guide7 min readUpdated
MVP Design: What to Build First in Your Startup App
MVP design explained for founders: how to cut a long feature list down to a first version of your app, with a worked example and a three-pile method.
By Tithi Mondal, UI/UX Designer
Suppose a founder in Siliguri wants to build a tiffin subscription app. Her list has 34 features: daily menu, meal plans, UPI payments, delivery tracking, loyalty points, chef profiles, reviews, a referral code, dark mode, and more. She asks me, "Which of these do I build first?" The honest answer is: far fewer than 34.
This guide shows how to cut a feature list down to a first version of an app. That first version is called a minimum viable product, or MVP. I will use the tiffin app as a worked example.
What MVP design actually means
A minimum viable product is the smallest product that lets a real person get real value. "Minimum" does not mean ugly or broken. "Viable" means it works well enough that people trust it.
MVP design is the design work for that small product. You design fewer screens, but you design each screen with care. The goal is to learn from real users quickly, before you spend months on features nobody wants.
Many founders confuse a first version with a rough draft. A rough draft has errors. A first version has a narrow scope. These are different things. A narrow product can still have clean layouts, clear words and a working payment step.
Step 1: Write one promise
Before you touch the feature list, write one sentence that finishes this line: "Using my app, a person can ___."
For the tiffin founder, the sentence is: "A working person in Siliguri can order a home-cooked lunch for the next day in under one minute."
This sentence is your filter. Every feature must help this sentence come true. If a feature does not, it waits.
A good promise has three parts: one kind of user, one task, and one clear result. If your sentence has the word "and" three times, you have three promises. Pick one.
Step 2: Sort every feature into three piles
Write each feature on a separate card or sticky note. Then sort the cards into three piles.
- Pile 1, the promise breaks without it. If this feature is missing, the user cannot complete the task at all.
- Pile 2, the promise works but feels weak. Users can finish the task, but with effort or doubt.
- Pile 3, nice to have. Users will not notice for the first few months.
Be strict. Most founders put 20 cards in Pile 1 on the first try. Ask for each card: "If I removed this, could one real user still get lunch tomorrow?" If yes, the card moves down.
Your MVP is Pile 1 plus a very small part of Pile 2. Pile 3 goes into a written list called "later". Keep that list. It is not lost work. It is a plan.
A worked example: the tiffin app
Here is how the 34 features could sort. This is a hypothetical example, not a real client.
| Feature | Pile | Reason |
|---|---|---|
| Phone number login with OTP | 1 | The app must know who is ordering |
| Tomorrow's menu (one or two dishes) | 1 | The user must see what they are buying |
| Place order and pay by UPI | 1 | No payment, no business |
| Order confirmation on WhatsApp | 1 | Users already check WhatsApp all day |
| Saved delivery address | 2 | Typing the address every day is tiring |
| Weekly meal plan | 2 | Useful, but one-day orders prove demand first |
| Live delivery tracking on a map | 3 | The founder can send a phone call or message instead |
| Loyalty points and referral code | 3 | Needs many users before it matters |
| Chef profiles and reviews | 3 | Needs content and moderation |
| Dark mode | 3 | Pleasant, not necessary |
The first version now has about 6 screens: login, menu, order summary, payment, confirmation and order history. One designer can finish that in a few weeks, and one developer can build it without chaos.
Notice what the founder gives up. There is no live tracking. Instead, the delivery person sends a WhatsApp message when he leaves. This is a manual step, and that is fine. I come back to this idea below.
What a first version still needs
Cutting features is easy. Cutting too much is the danger. Some things look optional but are not.
- A clear first screen. The user should understand what the app does in 5 seconds.
- Error and empty states. What does the user see when the menu is not yet posted, or when payment fails? Blank screens lose trust. See my notes on app onboarding.
- A way to get help. A WhatsApp button or phone number is enough. A user who paid ₹120 and has a problem must know who to ask.
- Payment that feels safe. In India, users expect UPI. Show the amount and the merchant name clearly. More detail is in UPI payment UX.
- Fast loading on a mid-range Android phone. Many of your first users will not have a flagship phone.
These items protect trust. A first version that is small but trusted will teach you more than a big one that people abandon.
What to do by hand instead of building
A useful MVP trick is to do the work manually behind the screen. The user sees a simple app. Behind it, a person does the job that software will do later.
In the tiffin example, the founder can post tomorrow's menu from a simple admin screen, or even from a shared sheet, instead of building a full chef dashboard. She can confirm orders by phone before she builds automatic routing.
This is not cheating. It is a way to test the idea before paying for software. When a manual step starts to hurt, because there are too many orders, that is the signal to automate it. Your own pain tells you what to build next.
How design helps you cut the list
Founders cut better when they can see the product on paper. Here is the order I follow on a first-version project.
- Understand. I ask who the user is, what they do today instead of using your app, and what must be true for the first release to count as a success.
- Map the flow. I draw the path from opening the app to finishing the task. Any feature outside that path is a candidate for "later".
- Wireframe. I make plain boxes-and-text screens. These are cheap to change. The founder often says, "Now I see that we do not need this screen."
- Design the interface. Only after the scope is agreed do I design colours, type and spacing. If you want the difference between these stages, read wireframe vs mockup vs prototype.
- Prototype and hand off. I make a clickable prototype that you can test with 5 or 6 real users before development starts.
Testing before development matters. A change on a prototype costs an hour. A change after coding costs days. If money is tight, usability testing on a small budget shows how to run these tests cheaply.
Mistakes I see in first versions
- Building for everyone. A tiffin app for office workers, students, families and hostels at once has no clear screen layout. Choose one group.
- Adding features to avoid a hard question. Founders sometimes add a feature because they fear the main idea is not enough. Test the main idea first.
- Copying a large app. A big app has years of added features. Do not copy its current state. Copy its first version, if you can find it.
- Skipping the "later" list. If cut features are not written down, the founder feels she is losing them, and adds them back.
Cost follows scope. A smaller first version is cheaper to design and build. Prices vary by the number of screens and user roles, and you can read what drives them in mobile app design cost in India. To understand your user better before you cut, start with user research for startups.
If you have a long feature list and are not sure where to cut, I can help you shape it through mobile app design, or you can send me a message.
Frequently asked questions
How many features should an MVP have?
There is no fixed number. Keep only the features needed for one user to finish one important task. Many first versions have 4 to 8 core screens.
Is an MVP the same as a prototype?
No. A prototype is a clickable model used to test an idea, and it has no working code behind it. An MVP is a real product that real users can use.
Should an MVP look polished?
It should look clean and trustworthy, but it does not need every visual detail. Clear layout, readable text and working buttons matter more than animations.
When should I add the features I cut?
Add them when real users show a need, for example when they ask for the same thing repeatedly or when a manual step becomes too slow to continue.
Working on something like this?
Tell Tithi what you are planning. You get a clear scope, a fixed price and a timeline before any work begins. I reply to every enquiry within 24 hours.