The most expensive mistake in software is building the wrong thing and finding out on launch day. A website or app prototype and a small first version (MVP) let you test the idea while changes cost hours, not months. Here is what each one is, how they differ and how to decide what to build first.

What a website or app prototype is

A prototype is a layout of the main screens of a future website or app. It shows what goes where, which buttons exist and where each one leads. Colors, fonts and photos do not matter at this stage; the screens are often plain gray.

A clickable prototype can be opened on a phone or computer and “walked through”: tap “Book”, choose a service, reach the confirmation. There is no real logic or database behind it, but you already get a feel for the product.

Say you are ordering an app for booking at a beauty salon. In the prototype you see the home screen with a list of services, choosing a stylist, a calendar with open slots, a confirmation screen and a profile with past bookings. Even here you can notice that picking the stylist before the service feels awkward, or that there is no button to cancel a booking.

Why changes on screens cost less than changes in code

Moving a button or adding a step in a prototype takes minutes. The same change in a finished product can touch the design, the layout, the server logic, the database and the tests. The later a flaw in the concept is found, the more finished work has to be redone.

Back to the salon. Say it turns out customers want to book two services in a row, like a haircut and coloring. In a prototype that is one extra screen and half an hour of discussion. In a finished app it means changing how open slots are calculated, the confirmation screen, the notifications to stylists and the admin panel, and then retesting everything connected to them.

A prototype also aligns expectations. Words in a brief mean different things to different people: a “convenient catalog” may look nothing alike in the client’s head and the developer’s. A screen means the same thing to everyone.

If something can be tested on a prototype, test it on the prototype. Every change made before development saves money after it.

That is why many teams, NOVO included, start every project with a prototype: the client sees the main screens before development begins.

How to test a prototype on real people

Showing the prototype only to the owner is not enough. They know the business inside out and will guess things a regular customer will not. It is far more useful to sit down with three to five people from your audience.

A simple way to do it:

  1. Give the person a task, not a tour: “book a haircut for Saturday afternoon”.
  2. Stay quiet and watch where they hesitate, what they tap and where they go wrong.
  3. Do not help. If you had to explain, the screen needs to change.
  4. Afterwards, ask what was unclear and what was missing.

Five short sessions usually reveal the main problems. If three out of five cannot find the payment button, that is not carelessness, it is a problem with the screen.

What an MVP, or first version, is

An MVP (minimum viable product), or first version, is a real working product with the smallest set of features that solves the main task. It is not a mockup or a demo: real customers use it and real orders go through it.

The point is to reach people sooner and learn from how they behave rather than from assumptions. You often find that a feature that seemed essential is not used at all, while something nobody thought of gets requested every day.

To show how the stages relate, here are reference prices at NOVO for a website, web service or CRM:

Stage What you get From
Prototype Clickable main screens $300
First version A working product with the core features $1,200
Turnkey The full feature set and launch $3,000

For a mobile app the order is the same but the prices are higher. More in How much a mobile app costs.

How to decide what goes into the first version

The hard part of a first version is not deciding what to add, but what to postpone. Here is a method that works.

Find the main scenario

Put in one sentence why people will come. For the salon: “a customer books a stylist at a convenient time”. Everything that scenario needs goes into the first version. Everything else is open for discussion.

Sort features into three lists

  • Must have. Without it the main scenario breaks: choosing a service, the calendar, booking confirmation.
  • Nice to have. Useful, but can be handled manually for now: reminders, reviews, a loyalty program.
  • Someday. Ideas for later: a link to the cash register, several locations, gift cards.

The first version gets the first list and, if the budget allows, a couple of the most important items from the second.

Do by hand what can be done by hand

At the start, many processes can be handled by a person. An administrator can send booking reminders in WhatsApp, and a report can live in a spreadsheet. Automate it once it is clear the process really repeats and takes up time.

Decide how you will know it works

Decide in advance what you will look at after launch. For example, how many bookings came through the app in the first month, at which step people abandon a booking, what questions they ask the administrator. Without that, a first version is just a small product, not a way to learn something.

When starting small is a bad idea

Sometimes a product cannot be cut down. If you are moving a process the whole company relies on into a new system, such as warehouse accounting or document flow, a half-finished system only adds chaos: people end up working in two places. In that case you can still build in stages, but each stage should fully replace part of the old process.

Do not ship a rough version either. “Minimal” does not mean “buggy”. There are few features, but the ones that exist should work reliably and look tidy.

If you are just getting started, see the overall guide How to order a website, and before the prototype, prepare a plain-language brief: the prototype is built on it.

In short

  • A prototype shows the main screens before development, and changes to it take minutes.
  • Test the prototype on three to five people from your audience by giving them tasks, not a tour.
  • A first version is a working product built around the main scenario, not a demo.
  • Sort features into “must have”, “nice to have” and “someday”.
  • Decide in advance how you will know the first version is working.