analyze-customer-profitability

# analyze-customer-profitability Make cost attribution and decision boundaries visible before ranking customers. ## When to use - Use for account planning, pricing review, capacity decisions, or segment economics. - Do not automate harmful treatment, deny essential service, or proxy for protected traits from a profitability score. ## Preconditions - Confirm decision scope, permitted customer data, reporting period, currencies, and approved cost-allocation policy. - Reconcile revenue, refunds, direct costs, support, fulfillment, and payment data. ## Procedure 1. Define customer identity and related-account mapping at a stable grain. 2. Calculate net revenue after discounts, credits, refunds, rebates, taxes, and revenue sharing. 3. Attribute direct product, fulfillment, payment, returns, onboarding, support, and customer-specific infrastructure costs. 4. Allocate shared costs only where a defensible driver exists; show unallocated and alternative views. 5. Calculate contribution dollars and margin by customer, cohort, product, channel, and tenure. 6. Adjust comparisons for contract stage, implementation period, seasonality, currency, and one-time events. 7. Reconcile aggregated revenue and costs to approved financial reports. 8. Test sensitivity to allocation, support-cost, churn, and renewal assumptions. 9. Combine financial evidence with contract, strategic, service, fairness, and capacity constraints before recommending action. ## Worked example A large customer appears unprofitable after equal allocation of all corporate overhead. A driver-based view shows the account has high onboarding cost but positive ongoing contribution. The report separates temporary implementation cost, shared overhead, and recurring service cost, then recommends a renewal and support-design review rather than an automatic termination. ## Done - A profitability model records identity, net revenue, cost attribution, allocation choices, contribution, segments, and lifecycle effects - Financial totals reconcile and recommendations are checked against sensitivity, privacy, fairness, contractual, and service constraints