Why this decision is hard
Because the cases arrive in different shapes. One program has a five year projection, one has a payback figure, one has a strategic argument and no figures at all. They are not comparable, so the decision quietly becomes a judgment about which sponsor is most convincing.
And because the ranges are hidden. Every sponsor presents a single number, and the single number is always the one that clears the bar. Two programs with the same expected return are not the same program if one of them has a real chance of returning nothing, and a single number cannot tell you which is which.
What goes in
- Each program, as its own option.
- What each one costs, as a range where the estimate is still moving.
- What each one returns a year once it is running.
- How many months before it starts returning anything.
- The period you are judging them over, the same period for all of them.
- The rate you discount future money at, if you use one.
- Named risks: the regulatory delay, the cost increase, the thing that happened last time.
What comes out
- Every program's low case, middle outcome and high case, in your own currency, side by side.
- The chance each program comes out ahead.
- Which figures the ranking depends on, so you know which sponsor to send back for evidence.
- A full audit trail behind every figure: the inputs, the method, the calculation and the result, kept with the decision.
- A record that survives the meeting: the same figures give the same answer when the decision is reopened next year.
An example
Harbor Lane Clinics, an invented company, with one budget and two programs
Spend the capital budget on new clinics, or on refitting the clinics already open?
- New clinics, margin a year once running
- $2.1m to $5.4m, most likely $3.4m
- New clinics, capital needed
- $10m to $15m, most likely $12m
- New clinics, months before they run
- 8 to 24, most likely 14
- Refit, margin a year once done
- $1.6m to $3.9m, most likely $2.8m
- Refit, capital needed
- $7m to $10.5m, most likely $8.5m
- Refit, months before it is done
- 4 to 12, most likely 7
- Period judged over
- 5 years
- New clinics, low case
- -$2.4m
- New clinics, middle outcome
- $1.2m
- New clinics, high case
- $5.2m
- New clinics, chance of coming out ahead
- 65%
- Refit the existing clinics, low case
- $0.5m
- Refit the existing clinics, middle outcome
- $3.4m
- Refit the existing clinics, high case
- $6.3m
- Refit the existing clinics, chance of coming out ahead
- 94%
The refit wins, and it is not close. New clinics return more at the top end, but their middle outcome is $1.2m against $3.4m for the refit, and their low case is a loss of $2.4m where the refit's low case is still positive at $0.5m. The refit comes out ahead 94% of the time against 65% for the clinics. The reason is in the inputs and not in the returns: the refit is running inside a year, typically in seven months, while the new clinics spend over a year of the period building. The figures the answer depends on, in order: New clinics, margin a year, New clinics, capital needed, New clinics, months before they run.
The company and the figures above are invented for this page. The results are a real run of the BrainPower engine on exactly those figures, not a picture of one. Your own figures give your own answer.
Who it is for
Chief financial officers and finance teams running an annual capital round. Investment committees ranking proposals that did not arrive in the same format. Chief executives and owners who want the comparison to be about the programs rather than about the slides. Advisors preparing a recommendation they expect to be challenged on.
Questions people ask
Can it handle more than two programs?
Yes. Each program is an option with its own figures and they are reported together on the same scale. The example on this page uses two to keep it readable.
What if one program has no figures, only a strategic argument?
Then the honest answer is that it cannot be compared on return, and the product will not pretend otherwise. What you can do is state the figures the argument would have to be worth for it to win, and judge whether that is plausible.
Does it do net present value?
Yes. You give the rate you discount at and the result is reported in present value terms. The rate is yours: the product does not choose one for you, and the result says which rate was used.
How does it treat a program that is cheaper but slower?
As what it is. Time is an input, not a footnote, so a program that takes eighteen months to start returning is charged for those eighteen months out of the period you are judging over. In the example on this page that is the whole decision.
Can two people disagree about an input and both see the result?
Yes, and this is the common use. Change one range, rerun, and the difference between the two answers is the price of that disagreement. It is usually smaller than the argument about it.