observation
Survival depended on the category, in at least four ways
category-dependent-survivabilitystatus · Candidate2026-07-20
Evidence tier
Narrated · 1 of 6
Research components
Answers business questions
- Why did we disappear after follow-up questions?
- What must become true to survive evaluation?
- How do we measure whether our position improved?
- How do AI systems evaluate and recommend vendors?
Open founder decision
This object is gated on a decision only the founder can make. Its current content is provisional and marked as such.
Across five categories under the same requirement battery, the recommendation behaved in at least four distinguishable ways when requirements were added in sequence:
- Collapse to a single incumbent, in clinical trial management software (E-034).
- Survival with stable membership, in project management software (E-040) and video conferencing (E-045).
- Survival with near-total membership churn, in CRM software (E-041).
- Wipeout of the named field and migration to a higher tier, in accounting software (E-043).
These are recorded as observed classes, not as a settled taxonomy. How to code the accounting case, and whether the two survival modes are one class or two, are open judgment calls. Most runs used a single evaluator, and nothing here shows what property of a category produces which behavior.
What this means commercially
- Affected buyers
- Companies trying to anticipate how their category behaves under scrutiny
- Affected categories
- Five categories observed so far
- Potential product impact
- What survival looks like appears to depend on the category, so a general answer may not transfer between categories
- Current confidence
- Narrated. Five categories, one evaluator for most, no causal test of what makes a category behave one way.
- Evidence tier
- Narrated
This object references
Authors
Upstream Zero
Machine rendering