DEV Community

Elora Digital
Elora Digital

Posted on Originally published at eloradigital.com.au AI-assisted

What is an MVP, and what goes in the first version?

Key takeaways

  • An MVP, a minimum viable product, is the first version of a product: one core workflow, built well enough for real users to rely on and pay for.
  • It is not a prototype. It has sign-in, real data, an admin panel and the quality to grow into version two.
  • The hardest part is deciding what to leave out. Everything that does not serve the one workflow goes on a list for later.
  • With us, an MVP costs $15,000 to $40,000 ex GST and takes 8 to 12 weeks, fixed in a written scope before the build starts.

An MVP, or minimum viable product, is the first version of a product with one core workflow, built properly enough that real users can rely on it and pay for it. The point is to learn from real use before spending money on features nobody has asked for, so the scope is deliberately small: sign-in, the one workflow the product exists for, an admin panel so you can run it, and notifications for the moments that matter. Everything else waits for version two. With us, an MVP costs $15,000 to $40,000 ex GST and takes 8 to 12 weeks, with a written scope from a paid discovery workshop and a fixed price before the build starts. This article explains what goes in a first version, what to leave out, and how to keep it small without making it flimsy.

What is an MVP, and what is it not?

The word "minimum" causes most of the confusion. It refers to the scope, not the quality.

  • An MVP is a working product. Real people sign up, use it with real data, and you can see what they do. It is built on the same foundations a larger product would use, so version two extends it rather than replacing it.
  • An MVP is not a prototype or a clickable demo. Those are useful for testing an idea with a few people, and AI app builders now make them quick to produce, but they are usually thrown away.
  • An MVP is not a cheap version of the whole product. Cutting quality across every feature produces something that is slow, insecure and has to be rebuilt. Cutting features and keeping quality produces something you can grow.

What goes in the first version?

Start from one sentence: who the product is for, and the one thing it lets them do. Then build the shortest path from sign-up to that thing.

Usually in:

  • Sign-in with email, Google and passkeys.
  • The one core workflow, built properly and tested.
  • An admin panel, so you can see and fix what users do.
  • Email notifications for the moments that matter.
  • Deployment with continuous integration, documentation and a 30-day warranty.
  • Stripe billing as a priced add-on, if people pay from day one.

Usually out, for now:

  • Native mobile apps, when a responsive web app will do. If the phone's own features are the point, the MVP is a React Native app instead; our mobile apps cover that case.
  • Every integration you can imagine. One or two that matter, at most.
  • A custom design system. A clean, consistent interface comes first.
  • Features for the customer you do not have yet.

The list of what is out is as important as the list of what is in. Writing it down in the scope stops it creeping back during the build.

How do you decide what to leave out?

Three questions settle most arguments about scope.

  1. Does the first user need it to get the one result? If not, it waits.
  2. Will you learn something from building it now that you could not learn later? Payments often pass this test, because whether people pay is the most important thing to learn.
  3. Is it expensive to add later? Some things, such as how accounts and permissions are structured, are cheaper to get right now. Most features are not.

This is the work of the discovery workshop. It names the one workflow, the users and the shortest path to value, and it ends with a scope that says what is in, what is out and why.

A worked example

Scentena, a product our founder built before ELORA DIGITAL, shows the shape. It is an AI perfume formula generator for hobbyists, students and indie perfumers: a brief in plain English goes in, and a buildable formula comes out. It has one core workflow (brief to formula), sign-in with Google or GitHub, a free plan and a paid plan on Stripe. The shape is the same as most AI products founders ask about: one workflow, sign-in, and billing if people pay from the start.

How much does an MVP cost in Australia?

With us, an MVP costs $15,000 to $40,000 ex GST: our MVP package is the same package as our web apps. The low end is one core workflow with sign-in, one kind of user plus an administrator, email notifications and an admin panel. More kinds of user, a second workflow and more custom design move it up the range.

Stripe billing and other integrations ($1,500 to $5,000 each), AI features (from $3,000) and a design system in your brand (from $2,500) are priced add-ons on top. The service page shows three example scopes, made up to show what sits at the low end, the middle and the top of the range. Every price is on our pricing page.

How long does an MVP take?

Eight to twelve weeks, in the same five stages as every project on our process page:

  1. Discovery: a paid workshop ($500 to $1,500, credited to the project if it goes ahead) that ends in the written scope and the fixed price.
  2. Design: wireframes and designs you approve before any code.
  3. Build: two-week sprints, each ending with a working demo. From the first sprints, you have something to show users and investors.
  4. Launch: on hosting, domain and code accounts in your name, with documentation and a 30-day warranty that sits alongside the consumer guarantees under the Australian Consumer Law and does not limit them.
  5. Care, then version two: a care plan from $300 a month, and the next release scoped as a second fixed-price phase, based on what real users did with the first.

How long does it take to build an app? breaks the weeks down further.

How do you start?

Write the one sentence: who the product is for and the one thing it lets them do. Then tell us about it. We reply with an estimate range within one business day, and if a smaller build, or no build at all, is the right move, we say so.


Originally published on the Elora Digital blog.

Top comments (0)