customer-patternslisted

# Deal Pattern Analysis ## Pipeline contract **Stage 1 of 20.** Reads `00-customers.csv` (`domain` required; `company_name, deal_size, close_date, champion_title, source, notes` optional). Writes `01-customer-patterns.md` and `01-seeds.csv` (`domain`). - With only ~5 customers and no loss data, run Analysis 1 and 5 only and say so in the output. Five rows is a hypothesis, not a pattern — label every finding with how many customers support it. - Research each customer's website before filling the 12 variables; leave a variable blank rather than guessing. - Done when `01-customer-patterns.md` ends with an "Ideal Deal Profile" the human has read. Files live in `runs/<run>/`. Column definitions: [`/prospecting-system`](../prospecting-system/SKILL.md). Standalone use: ignore this block. ## When to Use - Quarterly business reviews when analyzing sales performance - Refining your ICP based on which deals actually close (vs. which you thought would close) - Building predictive models for deal scoring and pipeline forecasting - Training new sales hires on what "good" looks like at your company - Diagnosing a pipeline problem: "We have meetings but can't close" ## Framework ### The Deal Pattern Hypothesis Every company has hidden patterns in their deal data. The companies that find and exploit these patterns close 2-3x more efficiently than those that don't. **The core question:** What do your best deals have in common that your worst deals don't? If you can answer that ques