Claim and evidence
Interview people who recently switched to your product (or away) and walk forward through concrete events to learn why.
No dated example yet. Treat this tactic as a hypothesis.
Primary
The company or founder describes the tactic and its result themselves: own blog, talk, filing.
Working · 2026-11 Maintained by the method's authors; no outcome data on the fetched page.
What the numbers show
often enough for patterns to emerge for one job (method authors' rule of thumb, not a measured result)
jobstobedone.orgWhy it works
Recent switchers remember real events, so you learn causes rather than opinions or feature wishes. Patterns often show up within a handful of interviews.
How to run it
- Recruit people who switched recently
- Plan 45 to 90 minutes per interview
- Walk the timeline through concrete events, not opinions
- Capture the struggling moment and alternatives tried
- Look for patterns across 5 to 8 interviews
Ways to run it
Design options to test, not promised results.
- Option ASwitch-in interviews
Talk to people who recently started paying, to learn what pushed them away from the old way.
- Option BSwitch-out interviews
Talk to people who recently cancelled or moved to an alternative, to learn what pulled them away.
- Option CInterview pairs
Run interviews with two interviewers: one leads the timeline, the other captures events and quotes.
Dated examples
No dated example yet. Treat this tactic as a hypothesis.
Does it still work?
Maintained by the method's authors; no outcome data on the fetched page. source
Risks
- Recall bias if the switch was long ago
- Needs trained interviewers
Sources
- jobstobedone.org
“Plan for 45 to 90 minutes.”
✓ quote checked word-for-word on the page - jobstobedone.org
“Patterns start emerging surprisingly fast, often within 5 to 8 interviews for a given job.”
✓ quote checked word-for-word on the page
Change log
- 2026-11-20 — harvest v0.1 (ops-ai)