Five respondents mention a spreadsheet, but they may be using it for five different tasks. Repeated words do not necessarily establish a repeated customer problem.
For founders reviewing discovery responses, group evidence by the task, context and consequence. The useful pattern should survive a return to the original examples rather than depend on an attractive theme label.
What counts as a repeated discovery problem?
A repeated problem involves sufficiently similar circumstances and unmet needs to support a shared product question. The respondents can use different language while describing the same difficulty, or the same language while describing different ones.
Define the unit you are counting. One participant repeating a concern in several answers should not automatically become several independent examples.
Build an evidence matrix before naming themes
Use fictional accounts from people handling service enquiries. The proposed columns make important differences visible during analysis.
| Evidence field | Why it matters |
|---|---|
| Participant and relevant role | Preserves the source and context |
| Recent situation | Distinguishes an event from a general opinion |
| Current workaround | Shows how the task is handled today |
| Consequence | Explains the practical effect |
| Supporting source reference | Makes the interpretation reviewable |
| Contradicting detail | Prevents selective synthesis |
For example, one spreadsheet may track unresolved calls while another calculates installation estimates. Calling both a need for a better dashboard would discard the task distinction.
Count carefully and preserve exceptions
Report how many relevant participants described the pattern and how they were recruited. A small purposeful study can reveal a promising question without establishing its prevalence across a market.
Keep cases that do not fit. If some offices already solve the task cheaply, inspect what differs in their process rather than removing them from the summary. That difference may help define the audience or show that a product is unnecessary.
The faithful research-note guide helps preserve qualifiers before grouping responses.
Convert the theme into a decision
Write a bounded statement: in this study, certain office managers described difficulty identifying unresolved enquiries after a busy period. Then state what remains unknown, such as frequency, buying authority or the value of an alternative workflow.
My StackFlight case study describes original-response collection and founder review. It does not claim that a generated theme proves demand or that every repeated phrase represents an opportunity.
Choose the next test from the uncertainty that matters most. Another interview, a manual workflow trial or a prototype task can each produce different evidence. More responses from the same easy channel may not resolve the missing context.
Should every repeated phrase become a product theme?
No. Compare the underlying task, circumstance and consequence before grouping the accounts.
Can a small study measure market prevalence?
Keep the sample and recruitment limits explicit rather than presenting qualitative frequency as a market estimate.
What should we create first?
Build a source-linked evidence matrix before discussing a product discovery system.