Build a Deal-Risk Scoring Model

You are a sales operations lead at a digital marketing agency. You build simple, defensible scoring models that reps can actually understand and managers can defend in a forecast call.
Agency: [AGENCY_NAME] — [SERVICES]
Typical sales cycle: [SALES_CYCLE_DAYS] days
Risk factors that matter to us (and rough weight 1-5): [RISK_FACTORS_WITH_WEIGHTS]
Pipeline export:
[PIPELINE_EXPORT]
Apply a transparent weighted risk scoring model to every open deal in the export. Score 0-100 where 100 = highest risk of NOT closing. Show the factor breakdown per deal so the score is auditable.

- Use ONLY data present in [PIPELINE_EXPORT] and the weights in [RISK_FACTORS_WITH_WEIGHTS].
- If a factor cannot be measured for a deal (missing field), mark it 'n/a' and exclude from that deal's denominator — don't penalize for missing data.
- Show the math: score = sum(factor_score × weight) / sum(weights applicable) × 100.
- No subjective factors — everything must come from the input.

## Scoring Rubric (echo back so we can audit)
| Factor | Weight | How scored |

## Deal Scores (sorted high→low risk)
| Deal | Owner | Amount | Risk Score | Top 2 Contributing Factors |

## Notes
[Any deals where >2 factors were n/a — flag as 'low-confidence score']

• [AGENCY_NAME] — Your agency name
• [SERVICES] — Service lines
• [SALES_CYCLE_DAYS] — Average sales cycle
• [RISK_FACTORS_WITH_WEIGHTS] — e.g. days since last activity (5), stage age vs cycle (4), single-threaded (4), no next step booked (3), deal size <ACV (2)
• [PIPELINE_EXPORT] — CSV with the fields needed to score each factor

Agency: Northwind Digital — SEO, paid social
Cycle: 42 days
Risk factors:
- Days since last activity >14 (weight 5)
- Stage age >1.5x cycle (weight 4)
- Single-threaded / 1 contact (weight 4)
- No next step on calendar (weight 3)
Pipeline:
deal,owner,amount,stage,last_activity_days_ago,stage_age_days,contacts_count,next_step_booked
Acme,Jamie,6500,Proposal,4,21,3,yes
BlueTrail,Priya,4000,Discovery,16,74,1,no
Delta,Marcus,2500,Discovery,27,53,1,no

## Scoring Rubric
| Factor | Weight | How scored |
| Last activity >14d | 5 | 1 if true, 0 if false |
| Stage age >1.5x cycle | 4 | 1 if true, 0 if false |
| Single-threaded (1 contact) | 4 | 1 if true, 0 if false |
| No next step booked | 3 | 1 if true, 0 if false |

## Deal Scores
| Deal | Owner | Amount | Risk | Top Factors |
| Delta | Marcus | $2,500 | 100 | All 4 factors triggered |
| BlueTrail | Priya | $4,000 | 100 | All 4 factors triggered |
| Acme | Jamie | $6,500 | 0 | None triggered |

## Notes
All deals had full data — no low-confidence scores.

• Start with 4-5 factors max. A 12-factor model nobody trusts is worse than a 4-factor one reps can argue with.
• Run the model on closed-won AND closed-lost deals from last quarter to validate weights.
• Re-tune weights every quarter — the factors that predict risk shift as your ICP shifts.