Every unit says its case is the best return. Whose assumptions are you actually comparing?
The questions behind the question:
- Four sites, four business units, four cases for the same envelope. Which ones, and on what basis?
- 180 brands, 15 categories, 65 markets. How should I balance marketing and sales focus?
- Our decarbonisation trajectory commits us to a target. What is the cheapest path to it, and what does each year of delay cost?
- Our R&D pipeline has more projects than engineers. Which ones advance, and what does delaying the others cost?
- We are refitting a plant and launching a programme in the same budget year. Can we do both?
Why it is hard today
Each scenario is built in isolation
Four directors, four models, four sets of assumptions, four different definitions of payback. The arbitration compares advocacy, not options.
The chain of hypotheses is not traceable
You are asked to defend an allocation without being able to trace the assumptions that produced it — and the person who built the case is rarely the person who has to defend it.
Re-forecasting is unaffordable
A change in the cost of capital, a tariff or a raw material price means rebuilding the chain. Teams spend time reconstructing instead of finding better options.
The opportunity cost is never quantified
You know what you chose. You rarely know what the second-best option would have delivered, and by how much you beat it.
Resources are allocated once a year and never revisited
The split is set in the planning cycle and becomes a fact for twelve months. When a market moves in March, nobody reopens it, because reopening it means rebuilding everything.
What Paradygma changes
One model, one set of assumptions, every option
Each case is evaluated against the same constraints, the same costs and the same horizon. What you compare is options, not advocacy.
Every number traces back to its hypothesis
Change the cost of capital and the whole portfolio recomputes in front of you, with the delta visible. Your board can ask “and if?” and get the answer during the meeting.
The efficient frontier, not a shortlist
You see the whole set of allocations that are not dominated by another, and where your candidate sits on it.
Auditable by construction
Nothing is inferred by a model you cannot inspect. Every result is the consequence of a rule and a value, both of which you can read.
Interactions between allocations, not a ranking
Options are rarely independent. Funding one delays another, two share the same engineering team, a marketing push needs the capacity to serve it. The model carries those links, so what you get is a coherent allocation rather than a ranked list you then have to reconcile by hand.