A pro-rata 10% reduction looks fair but ignores marginal returns, commitments and thresholds.
A pro-rata 10% reduction looks fair but ignores marginal returns, commitments and thresholds. The relevant object is not annual average return: it is contribution lost from the last removable tranche, considering unavoidable costs, synergies and risk of breaking a capability. Rank genuinely actionable tranches, protect floors and show the economic cost of each scenario.
Decision and method.
Remove tranches with the lowest marginal contribution cost under constraints, preserving capabilities, tests and commitments whose break would create disproportionate loss. Split each lever into avoidable tranches with date, commitment, exit cost and operational floor; exclude savings that do not materialise within the period. Apply response curve, margin and interactions to estimate contribution lost per tranche and show a range outside historical variation. Rank by marginal loss, then apply floors, synergies, minimum coverage and reactivation capacity; compare with explicit pro-rata. Set leading signals, stop thresholds and review cadence, retaining a small reserve for loss above the expected range.
Worked example.
A €1m portfolio must save €100,000. Final available tranches lose €0.45 contribution per euro on A, €0.90 B, €0.30 C and €0.60 D; D can decline only €30,000. Marginal plan removes €50,000 A, €40,000 C and €10,000 D: loss 22,500 + 12,000 + 6,000 = €40,500. Pro-rata removes €30,000 A, €25,000 B, €20,000 C and €25,000 D: loss €57,000. Retain marginal plan with sensitivity range: it avoids €16,500 expected loss and protects B, whose final tranche is most productive.
Checks and limits.
Each tranche is truly avoidable within reduction horizon; loss is calculated on marginal rather than average return; floors, synergies and rebuild costs are explicit; plan has monitoring and reactivation rule. Historical curves extrapolate badly for large simultaneous cuts. Competitors can react to lower presence. Brand and learning effects are more uncertain in the short term.
Resources and sources.
Download the reduction scenario. Related: calculate marginal returns, limit regret under uncertainty, compare response functions. Sources: Little (1970); Doyle and Saunders (1990); Keeney and Raiffa (1976).