← Blog
SaaS Products
October 14, 2026 · 4 min read

Narrowing an MVP: what to cut from the list

A first release isn't your idea made smaller; it's a test of a single assumption. Here's how to prune the feature list with that in mind.

SaaS Products

Every SaaS project, meaning any software sold by subscription, starts with a long feature list. Building all of it takes months, and for that whole time not one real user sees the product. An MVP, the smallest first release that genuinely works, exists to shorten that stretch: it tests the riskiest assumption by the cheapest route.

Write the assumption down first

“Restaurants want to update their own menus and will pay monthly to do it.” That sentence is an assumption. The MVP contains only the features that confirm or refute it.

A good assumption has three parts: who, what they want to do, and what they give in return. “Small businesses want to go digital” isn't an assumption, because it can't turn out to be wrong. Once the sentence is written, ask yourself: if this is false, how will I know? If there's no answer, the sentence isn't narrow enough yet. If you have several assumptions, pick the riskiest: the one that ends the business if it proves false.

Choose a single metric

The assumption points to the number that tests it. In the restaurant example that number isn't how many restaurants sign up; it's how many move to a paid plan when the trial ends, because the assumption says “will pay”. Figures such as page views or sign-ups feel good but say nothing about the assumption. Choosing one metric doesn't mean ignoring the other numbers; it means deciding in advance which one makes the call.

Set the threshold in advance

Before launch, write down which result counts as confirmed and which as refuted. A threshold chosen afterwards bends to fit whatever result you have.

What you can safely postpone

  • Advanced roles and permissions: one admin role is enough for the first ten customers.
  • Detailed reports: an export button does the same job at the start.
  • A mobile app: a responsive web version covers the first months.
  • Automated invoicing: you can invoice early customers by hand.

The list varies by product; the test doesn't. If the assumption can be tested without a feature, that feature can wait.

What you can't postpone

The sign-up flow, taking payment, and the one screen where the product does its real job. If those three are flawed, you're testing poor execution rather than the assumption. We bring these parts from a ready foundation, so a narrow scope doesn't mean careless work, and the time goes into the one screen that sets the product apart.

If you're not embarrassed by your first release you probably launched late; but a release that doesn't work teaches nothing either.

A one-page scope and the phase-two list

We reduce scope with the client to a single page: one assumption, three screens, one metric. Everything else goes on the phase-two list. For the restaurant, that page looks like this:

  • Assumption: restaurants update their own menus and pay monthly to do it.
  • Screens: sign-up and payment, menu editing, the menu page the diner sees.
  • Metric: the number of restaurants that start paying when the trial ends.
  • Phase two: multiple branches, translation, taking orders, detailed statistics.

What you cut isn't thrown away. The phase-two list does two jobs. First, it puts the team at ease: the idea hasn't been lost, it is waiting its turn. Second, once the first users arrive it gives you something to compare against. What users actually ask for won't always match the list, so read it as a forecast rather than a promise. After the first months, put one question to each item: has a real user asked for this?

When has the MVP done its job?

An MVP isn't a product stage; it's a question, and its job is done when the answer arrives. There are three possible outcomes:

  • Confirmed: the metric passed the threshold you wrote down. You can now return to the phase-two list with usage data.
  • Refuted: users came, tried it and didn't pay. That is a result too; you change the assumption and test the new sentence.
  • Unclear: not enough people saw the product. The problem is reach, not the product, and adding features won't fix it.

The trap we see most often is starting to build new features in the third case. If ten people have seen the product, what's missing isn't a feature; it's the eleventh person.

Common mistakes

  • Building for imagined scale. Designing for thousands of users before the first customer is an investment in an untested assumption. Keep the foundation sound and let capacity grow step by step with users.
  • Copying a competitor's feature list. They reached that list over years by listening to their own customers; your assumption is different.
  • Doing a little of everything. Ten half-built features don't teach what one complete feature does.
  • Treating the MVP as a showpiece. The aim isn't to impress investors or rivals; it's to get an answer to one question.

Narrowing scope doesn't shrink the idea; it sets the order. Start with one assumption, three screens and one metric, and let users tell you the rest. We can write that single page together in the first call.

Work with us

Have a project in mind?

Describe your idea in a few sentences. We'll review it and come back shortly with scope, timeline and budget.