Custom Software
SaaS and MVP development: each stage answers one question
Three stages, three questions
Built properly from the first real user
01
Accounts and roles
Who's who
People sign in as themselves, belong to a firm and have a role, from the first day.
02
The core record
The thing the product is about
For this product, the maintenance request: its fields, states and links, designed to last.
03
History
What happened, when
Every change is kept, so questions about real use can be answered later.
04
Separation
Each customer's data apart
Every record belongs to one firm, so the second customer never sees the first one's data.
05
Measurement
The number that answers the question
The stage's question is counted from the start, in the product itself.
Three ways a stage ends
Continue
- Evidence
- the answer is yes, by the rule written at the start
- Next
- build the next stage on the same foundation
- Kept
- everything, including the measurement
Change
- Evidence
- “not yet”, with a reason, like the MVP's 54%
- Next
- change one thing and test again: here, a link for tenants
- Kept
- the foundation and most of the screens
Stop
- Evidence
- the answer is no
- Next
- stop spending; keep what was learned
- Kept
- the lesson, the data and the code, in case the question changes
What product development hands off
In
Out
Worth doing when you can name the question
Good fit
- You have a product idea and a first group of customers to try it.
- You can say what result would make you stop.
- You expect to keep building if it works.
Not the right fit
- The software is for your own team. An internal tool or platform fits better.
- Nobody outside the team has heard the idea yet. Talk to customers first.
Questions about SaaS and MVP development
Shouldn't an MVP be cheap and thrown away?
The screens may be rebuilt; the foundations shouldn't be. Accounts, the core record and history cost little more to build properly once, and rebuilding them later means moving real customers' data.
How do we know when to move to the next stage?
Write the rule before you start: what result means continue, change or stop. Then the evidence decides.
Do you work with non-technical founders?
Yes. We write the question and the decision rule with you, and say plainly what's being built for real and what isn't.
Is the prototype code?
Often not. It can be design files or a clickable mock-up, and it's meant to be thrown away once it has answered its question.