Imagine a small repair business planning its first customer portal. The wish list includes online bookings, live technician tracking, subscriptions, payments, reviews and a referral program. Each feature makes sense in isolation. Together, they create a first release that is difficult to build and even harder to learn from.
A useful first version makes a smaller promise and keeps it completely. For this business, that promise might be: a customer can request a repair and receive a confirmed appointment without a chain of phone calls.
Choose the question you need answered
Before choosing features, decide what uncertainty the release should resolve. Will customers submit the information staff need? Can the team confirm appointments quickly enough? Will a clearer booking process reduce avoidable follow-up?
These are different questions from whether customers like the color scheme. A first release should create evidence about the business workflow. Write down what you will observe and who will review it after launch.
Keep one journey complete
A booking screen alone does not deliver a booking service. Someone needs to receive the request, check availability, confirm the slot and handle changes. Customers need to know whether their request is pending or accepted. Staff need a way to recover when a notification fails.
This is where scope discussions often become uncomfortable. Operational screens look less exciting than customer-facing features, yet they make the promise reliable. Include enough administration to operate the service without inventing a separate spreadsheet on launch day.
Use manual work deliberately
Not every step needs automation immediately. A staff member could assign technicians manually while the team learns how assignments should work. Record that decision, including the expected workload and the point at which it becomes impractical.
Manual steps need owners and visibility. “We will handle it manually” is incomplete if nobody knows where requests appear, how often to check them or how customers receive an answer. A deliberate manual process can teach you something. An undocumented one creates confusion.
Keep a separate list for later
Place live tracking and referral rewards in a later-release list if they do not help answer the immediate question. Give each postponed feature a reason to return, such as repeated customer demand or a volume of bookings that makes manual coordination difficult.
This turns the roadmap into a set of decisions tied to evidence. It also gives the team permission to learn that an attractive feature is less useful than expected.
- Name the person and the job your first release serves.
- Write the promise in one sentence.
- Map the smallest complete journey that keeps it.
- Add the permissions, failure states and support steps it needs.
- Choose a review date and the evidence you will examine.
The first version does not need to demonstrate every ambition behind the product. It needs to work well enough, for a defined group of people, to make the next investment decision more informed.



