Compare feature adoption across customer segments

Compare adoption of {{feature}} across customer segments.

Inputs:
- Usage data (per customer): {{paste CSV with segment, usage frequency, key actions}}
- Segment definition: {{small biz vs. enterprise / vertical / geography / use case / etc.}}
- Period covered: {{e.g. since feature launch / last 90 days}}
- The hypothesis we have about who would use this: {{paste}}

Output:

## Adoption table
| Segment | % using feature | Avg usage frequency | Time to first use | Retention impact |
|---|---|---|---|---|

## Where adoption is highest (and why)
The segment(s) with strongest adoption + the specific reason it fits their workflow (not vibes — the actual job-to-be-done it solves for them).

## Where adoption is lowest (and why)
The segment(s) where it hasn't landed + the diagnostic question: is it the segment doesn't need it, doesn't know it exists, or has tried and abandoned?

## Surprise findings
Segments that adopted unexpectedly or under-adopted unexpectedly. These are usually the most strategically interesting.

## Retention impact
For customers who adopted the feature: do they retain at higher rates? Expand more? Refer more? Be careful about causation — adoption might correlate with engaged customers who'd retain anyway.

## What to do per segment
- High adoption: amplify what's working — case studies, retention messaging
- Medium: probably an enablement gap — diagnose
- Low: either out-of-target (accept) or need redesign (specific recommendation)

## What this doesn't tell us
- Causation (high adopters might just be high-engagement overall)
- Whether the feature is actually solving the problem they think it is
- Whether segments that don't use it would WANT a different version

## Sources to add for the next analysis
The data you'd need (e.g. user research interviews, qual signal) to get from "adoption pattern" to "what to ship next."

Hard rules:
- Don't conflate "we built it" with "they wanted it"
- High adoption ≠ strategic value (some features get used a lot and provide nothing)
- Low adoption ≠ failure (some features are deliberately niche)