Interviews produce a lot of material and most of it is noise. The skill is knowing which signals mean something and which are politeness wearing a useful costume.
Signals that are real
- They have already spent money on the problem, however clumsily.
- They built a workaround: a spreadsheet, a manual process, a hired temp.
- They can name when it last happened and what it cost.
- They get animated or irritated when describing it.
- They ask you, unprompted, when they can have it.
Signals that are not
- "That sounds useful." — the sound of politeness.
- "We would definitely consider it." — a hypothetical about a hypothetical.
- Enthusiasm without any past behaviour behind it.
- Praise for the idea rather than description of the problem.
The spreadsheet test
The single strongest signal in small business software is an existing spreadsheet. It means someone cared enough to build a tool by hand, which is a person who has already paid for the problem in effort.
Look for the pattern, not the anecdote
One person's complaint is a story. The same complaint from five of ten interviews, described in similar words, is a market. Resist building for the most articulate person you spoke to — articulate is not the same as representative.
Try it
Go back through your interview notes and mark every instance of money spent, workaround built, or time wasted. Those marks are your evidence; everything else is atmosphere.