Economics and benefits ledger

Metric ID / version / Finance owner:
Customer entity / cohort / currency / tax treatment:
Horizon and cohort age / data cutoff / outcome lag:
Numerator / denominator / inclusions / exclusions:
Acquisition cost scope and allocation convention:
Contribution costs already included:
Costs explicitly outside contribution:
Source records / quality and reconciliation checks:
Estimate type: observed / attributed / forecast / incremental effect
Identification method and assumptions if incremental:
Uncertainty, missing data and sensitivity:

Prefer direct contribution per assigned eligible account to an approximate retention-rate-times-value calculation when data supports it. Keep trial treatment count separate from a projected rollout. If the incremental-customer denominator plausibly spans zero, report effects and economic ranges instead of a precise CAC ratio.

Option / cohort / tranche:
Comparable horizon and customer definition:
Observed quality: activation, retained contribution, returns, service cost:
Forecast value and out-of-time validation, if used:
Incremental customer/treatment effect and interval or range:
Incremental expense and extra delivery/system cost:
Net contribution: downside / base / upside:
Cash timing and payback:
Minimum feasible amount / upper cap / capacity / strategic commitments:
Alternative use of the same money and staff:
Decision, owner and condition that reverses it:

A CAC/LTV comparison is acceptable only when cost scope, horizon, causal status and customer entity align. Use cohort aggregates first; a forecast ratio alone does not authorise scale.

Create a column for A current practice, B repaired conventional work, C B plus AI, and D C plus additional integration or autonomy. Each row contains local amounts and sources, not an assumed industry benchmark.

Year 0: configuration/build, relevant data repair, evaluation, procurement,
        security, training, transition and dual run
Years 1–5: licences, infrastructure, model use, maintenance/API changes,
           review/rework, measurement, support, incidents and retraining
Exit: export, migration, contract exit, retained records and retirement
Staff capacity/opportunity cost: separately identified by role and year
Adoption and outcome-maturation assumptions:
Benefit rows: non-overlapping incremental contribution and realised cash savings
Cash-flow convention, Finance discount rate and tax/working-capital treatment:
No-change and downside scenarios:

For each option calculate net benefit as incremental contribution plus realised savings minus costs not already included. If calculating cash NPV, use signed net cash flow by year and the Finance-approved discount convention; year 0 is undiscounted. Report B−A, C−B and D−C separately. Allocate shared cost transparently and count it only once. Do not count an incident allowance both as cost and as a second expected-loss deduction.

Benefit ID / intervention ID / owner / Finance reviewer:
Current process cost / new process cost, including review and rework:
Accepted output standard and observed volume:
Capacity saved: hours, role, period and evidence:
How capacity will be used: remove expense / avoid approved expense /
                           named redeployment / unmonetised capacity
Revenue effect, causal status and horizon:
Contribution or margin effect, net of variable costs:
Risk/quality effect and evidence (do not monetise without a defensible basis):
Implementation cost / recurring cost / exit cost:
Cash action: exact budget, contract or approved-hire change and date:
Payback and uncertainty:
Overlapping benefits excluded:
Forecast / observed / realised status and review date:
Stop or restatement condition:

60-minute workshop: Finance defines scope for ten minutes; the owner maps A–D for ten; the team prices costs and benefits for twenty; it challenges downside adoption and double counting for fifteen; the final five assigns the expense or delivery action. If inputs are missing, commission those inputs rather than manufacturing ROI.