Start with the decision you need to make
An MVP is not “the cheapest website.” It is the smallest product that helps you learn something important: will users pay, does the workflow stick, can ops handle volume, or does the integration path work?
Write that learning goal in one sentence before you open Figma or a ticket board. If the goal is unclear, scope will inflate.
Define the primary user and one job
Pick one primary user and one job-to-be-done. Supporting roles can wait. Admin panels, multi-role permissions, and “nice to have” dashboards are where timelines quietly double.
- Who must succeed on day one?
- What action proves the product works?
- What can stay manual for 30-90 days?
Separate must-ship from can-fake
Rule of thumb: Use spreadsheets, Stripe Checkout, or email workflows until volume proves you need a custom billing engine or in-app notifications.
Must-ship usually means: core flow, basic auth, clear empty states, and enough admin control to operate without engineering for every change.
Timebox discovery, then lock the cut list
We typically run a short discovery: goals, users, constraints, success metrics, and a cut list. Then we lock scope for the first release. New ideas go into a backlog labeled “after we learn.”
That discipline is what protects budget - not cheaper hourly rates.
Ship, measure, then expand
Plan instrumentation early: events for the core job, funnel drop-offs, and qualitative feedback. The second release should be funded by evidence, not optimism.