Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →mkurman/feedback-synthesis
When the user needs to analyze, categorize, or extract actionable insights from customer feedback across multiple sources, especially feature requests.
Passed 19 of 22 test cases with the skill, and 13 of 22 without it.
and 7 more in the full report
6 fixed13 passed either way3 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 6 test cases the skill fixed. Same model, same prompt, with and without the skill.
In Dan Olsen's product framework applied to Acme PM, users rate the importance of bulk export at 0.80 and current satisfaction with bulk export at 0.25. Users rate importance of custom themes at 0.90 and satisfaction at 0.85. Which feature area has the higher Opportunity Score and what are their respective scores?
| Checked | Without skill | With skill |
|---|---|---|
| Calculates the bulk export Opportunity Score as 0.60 using 0.80 * (1 - 0.25). | Missed | Passed |
| Calculates the custom themes Opportunity Score as 0.135 using 0.90 * (1 - 0.85), identifying bulk export as the higher opportunity. | Missed | Passed |
The instructions your model reads when the skill loads.
-|-----------|-------------------|-------------------|---------------------|----------| | [Theme] | [Count] | [Segments] | [Score] | [High/Med/Low] | [H/M/L] |
Actions that address frequent feedback with low implementation effort.
Themes explicitly deprioritized with reasoning.
Breakdown by source, segment, and time period.
## Frameworks & Best Practices
### Opportunities Over Features
Never let customers design solutions. Prioritize opportunities (problems), not features. When a user says "I want a Gantt chart," the underlying opportunity might be "I need to visualize project timelines and communicate status to stakeholders." Always dig for the job-to-be-done behind the request.
### Opportunity Score (Dan Olsen)
Score each theme: Opportunity Score = Importance x (1 - Satisfaction), normalized to 0-1. This surfaces problems that are both important and underserved. A high-importance, high-satisfaction area is already well-served and should not be prioritized over a high-importance, low-satisfaction gap.
### Signal vs. Noise Rules
- **One customer saying it is not a pattern.** Require 3+ independent mentions of a theme before treating it as a signal. Exception: if the one customer is a whale account citing it as a churn risk.
- **Recency bias check.** A flood of recent feedback about one issue can overshadow a persistent problem. Always compare against the prior period.
- **Loudest does not equal most important.** Power users and vocal customers generate disproportionate feedback. Weight by segment size and revenue contribution, not volume alone.
- **Praise is data too.** Track what users love. Knowing your strengths prevents you from accidentally breaking them during a redesign.
### Assumption Testing
For each top-priority opportunity, identify the highest-risk assumption and design the cheapest possible test. Do not build the full solution to validate an assumption that could be tested with a prototype, survey, or Wizard of Oz experiment.
### Source-Specific Guidance
| Source | Strengths | Watch Out For |
test cases passed, vs 13 / 22 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.