Analyze segment performance and surface what's actually driving the spread.
Inputs:
- Performance data by segment: {{paste CSV — segment, key metrics, time series if available}}
- The metric we care about most: {{paste — revenue, retention, LTV, etc.}}
- Period covered: {{e.g. last 4 quarters}}
- Definition of "segment" — be specific: {{geography / product line / customer size / acquisition channel}}
Output:
## Performance table
| Segment | Period metric | vs. average | vs. own prior period | Trend |
|---|---|---|---|---|
## Top performers (3)
For each: what's working (specifically — sales motion, product fit, talent), the lead indicator that suggests it's durable vs. a streak, what we'd do differently if we wanted to amplify.
## Bottom performers (3)
For each: what's broken (specifically, with evidence — not "execution"), whether the issue is fixable from inside the segment or requires structural change, the single action that would matter most.
## Misleading patterns to watch for
- Segments that look great but are mix-shifted (a few big deals masking real weakness)
- Segments that look bad but are timing-shifted (revenue recognition or seasonal)
- Segments that look stable but the underlying components are diverging (one product up, another down, average flat)
## What we'd watch monthly
The 2–3 leading indicators that would tell us if interventions are working — within 30–60 days, not at end of quarter.
## What we should NOT do based on this
The actions that would feel right based on the numbers but would compound the wrong problem (e.g. cutting investment in a low-performing segment that's actually high-LTV but slow-ramp).
Hard rules:
- Show absolute numbers, not just %. "Down 30%" on $1M is different from on $100M.
- Distinguish "underperforming relative to the average" from "absolutely small"
- Flag any segment with N too small to draw conclusions fromsegmentationperformancebusiness-analytics