Foldwren

Designing a minimum viable product

How to design a minimum viable product by choosing one job, cutting features without regret, launching small and learning from real use.

A minimum viable product, or MVP, is the smallest version of your idea that real people can use to get something done. Small enough to build quickly. Complete enough that it is actually useful. Its purpose is to learn whether you are on the right track before you spend the full budget.

The hard part is not understanding the idea. It is having the nerve to leave things out.

What does "minimum viable" really mean?

Both words carry weight, and people tend to honour only one.

Lean too far towards minimum and you ship something broken or pointless: a booking app that cannot send a confirmation. Nobody learns anything from that except that broken things are annoying.

Lean too far towards viable and you build the full vision under a different name. Loyalty points, referral codes, a chat feature, dark mode. Eight months later you launch, and only then discover which half of it nobody wanted.

Choose the one job

Every MVP should be built around a single job the user is hiring it to do. Write it as one sentence:

"A [type of person] can [do this one thing] so that [this result]."

Some examples:

  • "A regular at our cafe can order and pay before arriving so they skip the queue."
  • "A client of our accounting practice can upload their tax documents so they do not need to email them."
  • "A site supervisor can complete the daily safety checklist on a phone so it is filed before leaving site."

If you find yourself writing "and" in the middle, you have two products. Pick one. The other can wait for version two, when you will know far more than you do now.

How to cut features without regret

Start by listing everything you imagine the product doing. Be generous. Then sort the list into three piles.

  1. Without this, the one job cannot be done. These are in.
  2. This makes the job nicer. These are out for now.
  3. This is a different job. These are out, and may never come back.

Be strict about pile one. For the cafe ordering example, it might hold: see the menu, choose items and options, choose a pick up time, pay, get a confirmation, and a way for staff to see the order. That is it.

Order history, favourites, loyalty stamps, gift cards and a store locator all sit in pile two. They are good ideas. They are not needed to find out whether regulars will order ahead.

A few questions that help when you are stuck on a feature:

  • Would a customer refuse to use it without this?
  • Can we do this by hand for the first fifty users?
  • Can an existing tool handle this part, such as Stripe for payments?

That second question is underrated. In an MVP, plenty can be manual behind the scenes. If refunds are rare, process them yourself. If onboarding a new business customer takes a phone call, make the phone call. Automate once the volume proves it is worth automating.

Small does not mean sloppy

Cutting scope is different from cutting quality. The features you keep should work properly and feel considered.

For the handful of screens in an MVP, we would still insist on:

  • Clear wording on every button and heading.
  • Sensible error messages when something fails.
  • A design that works on small and large phones.
  • Readable text and good contrast.
  • A way to contact a human.

People forgive a product that does only one thing. They do not forgive one that does that one thing badly. First impressions of a rough product are hard to undo, particularly with your own loyal customers.

Launch to a few, not to everyone

An MVP does not need a grand opening. A quiet release to a small group is safer and teaches you more.

Say you run that cafe. Invite thirty regulars by name. Tell them it is new, tell them you want their honest reaction, and make it easy to reply. You can watch every order come through, fix problems the same day, and ask people in person how it went when they collect their coffee.

What should you measure?

Decide before launch what would tell you the idea is working. Keep it to a few plain signals.

  • Do people complete the job? Of those who start an order, how many finish?
  • Do they come back? One use is curiosity. A second and third suggest real value.
  • Where do they drop out? A step where many people quit is your first fix.
  • What do they ask for? Keep a simple tally of requests. Patterns emerge quickly.

Deciding what comes next

After a few weeks you will be in one of three positions.

  1. It is working. People use it and return. Take the most requested item from pile two and build that next.
  2. It half works. People try it and stall at a particular point. Fix that point before adding anything new.
  3. It is not working. Hardly anyone uses it, even when prompted. That is disappointing, but you found out for a fraction of the full cost. Go back to your customers and ask what problem they actually have.

All three are useful results. The only bad outcome is the one an MVP is designed to prevent: spending everything on the complete version and learning the same lesson at the end.

Your order