The case against the long discovery phase
The fastest way to find out what should be built is to put something real in front of the people who will use it.
Squircle · · 1 min read
In many software projects, the discovery phase is a polite way of saying “we are not ready to build, and we do not want to admit it.” It runs for months, produces a long document, and leaves the hardest questions unanswered.
What people can and cannot tell you
Customers and staff are very good at telling you what is wrong with the thing in front of them. They are much less good at describing, in the abstract, the thing that does not exist yet. So interviews alone, however careful, only get you part of the way.
A working prototype, put in front of the people who will use it, surfaces the real questions in days.
What we do instead
Our Review is short and specific. In two to three weeks we map how work actually moves through the business and put a cost on each problem. That gives a ranked roadmap, with each step priced.
Then, at the start of each Build phase, we shape the work with working prototypes. Problems surface while they are still cheap to fix.
When upfront research is right
Some work does need more research before anything is built: medical devices, heavily regulated processes, anything where being wrong in production is very costly. For most products and internal systems, the prototype is the research.