Scope MVPs by mapping user decisions, not features

Why map decisions, not features

Traditional lists of features lead teams to pack a first release with assumptions. Instead, treat an MVP as an experiment that tests a value-proposition and a solution hypothesis: what decision do you want a user to make, and what minimal experience will let you observe that decision? This framing keeps scope tightly focused on measurable learning rather than on a feature checklist. (Sourced fact: MVP as a learning tool). 1

Turn discovery into a hypothesis pipeline

Start discovery by naming the business decision you need to take (for example: "will local customers pay a premium for same‑day delivery?"). Break that decision into user decisions, the micro‑choices a visitor makes on the path to conversion (noticing the offer, understanding benefits, starting checkout, completing payment). Each micro‑decision becomes a hypothesis you can test with a lightweight experiment, from a prototype to a live A/B test. This sequential approach separates testing of value proposition (do users see the offer as useful?) from testing of the solution (does this checkout flow actually convert?), reducing wasted effort. (Sourced fact: separate value‑proposition and solution hypotheses). 1

Use prototypes and usability testing to avoid premature coding

Choose the cheapest format that answers your hypothesis. For questions about comprehension or usefulness, paper or clickable prototypes and moderated usability tests are fast and low‑risk. For questions about behaviour in market conditions (discovery, acquisition, retention), a live code MVP that collects analytics will be necessary. Keep prototypes usable enough that feedback reflects the idea, not poor interaction design. Include accessibility checks in early tests so usability failures don’t mask value. (Sourced facts: when to use prototype vs live‑code MVP; usability testing differences and accessibility tooling). 1 2

Concrete steps to scope and run your MVP

1) Define the business decision and the one metric that will inform that decision (a single conversion or retention metric). 2) Map the user journey and identify 3, 5 micro‑decisions that lead to that metric. 3) For each micro‑decision choose the cheapest valid test format: guerrilla prototype, Wizard‑of‑Oz, or live‑code. 4) Write a brief experiment plan: hypothesis, format, sample size or timebox, and success criterion. 5) Run usability sessions first to check comprehension, then deploy a live test if you need behavioural data. 6) Use browser devtools, simple analytics and accessibility checks during the live test to understand failures. These steps turn discovery into repeatable experiments, which helps keep scope minimal and evidence‑driven. 1 2

A small realistic task to try this week

Task: Pick one page that matters (e.g., a product page or local‑service booking page). Map the micro‑decisions a visitor makes there (headline comprehension, trust signals, pricing clarity, CTA action). Build a one‑page prototype (paper or Figma) that isolates one micro‑decision and run five 20‑minute moderated usability sessions with real or proxy users. Record whether users make the target decision and why they hesitate. Use those notes to tighten the page or decide whether a live test is needed. This task takes a day or two and gives concrete learning to scope an MVP. 2

Where this helps small businesses in London and beyond

For owners asking "how much does a custom website cost?" or "website designer in London", this approach reframes the question away from price lists and towards which decisions the site must support first. A smaller, evidence‑driven MVP scope reduces initial build time and lets you commission only what you need to validate value, before committing to a larger design or development programme. Suggestions here are practical steps; cost and supply options will still vary by supplier and local market, so use the experiment outputs to brief designers or builders concisely. (Observation: reframes cost question to decision‑driven scope). 1 2

Sources

Checked 2026-10-16.

• Minimum Viable Product (MVP): Definition, Nielsen Norman Group
• Testing and usability guidance, MDN / developer.mozilla.org