Define the test before the feature list

An MVP is a minimum viable product: a usable first release built to test a specific business assumption. “Build a marketplace” is not a test. “Can a customer request and pay for one service, and can an operator fulfill it?” is. Write the user, the action, the success signal and the point at which you will change direction.

For a founder building a booking product, the first release may need discovery, a booking flow, payment confirmation and an operator view. Loyalty points and advanced analytics can wait. The operator view cannot be ignored simply because customers do not see it.

Identify the largest cost drivers

  • Discovery: research, process mapping and decisions that remove ambiguity.
  • Design: real user flows, accessibility and responsive states.
  • Integrations: payment, identity, messaging or data from other products.
  • Admin work: review queues, permissions, content management and support tools.
  • Launch: testing, analytics, deployment, documentation and handover.

Each external integration adds dependency and failure handling. An apparently simple feature can be expensive if it must work with inconsistent data or strict operational rules.

Ask for a scope you can compare

Request a written list of included user journeys, exclusions, acceptance criteria and assumptions. Ask which decisions are still open and how changes will be priced. Compare proposals on the same scope, including who owns design, quality checks, deployment and post-launch fixes. A fixed number without those details is difficult to evaluate.

Set aside a separate operating budget for hosting, third-party subscriptions, support, security updates and learning from real users. A launch is the start of product feedback, not the end of spending.

Sequence the first release

Build the smallest end-to-end journey first, test it with actual users, and keep an operator able to repair exceptions. Add secondary features only when evidence shows they unblock the core outcome. If you are also deciding who should own architecture and vendor choices, read our fractional CTO versus agency guide. For the broader build-versus-buy choice, see custom software versus SaaS.