What actually changes a follow-on decision

Three elements, every time

Every scenario the module runs returns a rationale with exactly three parts: the strongest signal in favour, the strongest signal against, and the binding constraint. Three, always, including when the inputs were too thin to produce a confident recommendation, in which case the three parts say so.

The fixed shape is deliberate. A rationale that lists whatever happened to be notable produces long explanations for easy decisions and short ones for hard decisions, which is exactly backwards. Forcing three parts means the module always names something that argues against its own recommendation, which is the element a reader is most likely to want and least likely to be given.

The strongest signal in favour, and the strongest against

The first two elements name what is pulling hardest in each direction in the evidence about this company. They are drawn from the same reading that produced the recommendation, so they are not commentary added afterwards; they are the parts of the picture that moved it most.

A yes with a named signal against it is more useful than a yes on its own, because the named signal is the thing to go and check. If the strongest argument against following on is something a General Partner already knows to be stale, or already has private context on, the recommendation should be read differently. That is the whole purpose of surfacing it.

When the evidence genuinely runs one way, an element can report that there is nothing meaningful on that side. That is an honest output rather than a gap, and it is a stronger statement than a weak counter-argument dressed up to fill the slot.

The binding constraint

The third element answers a different question from the first two. It names not what the evidence says, but what is actually limiting the recommended allocation. Four possibilities.

• Reserves

  The fund does not have the remaining follow-on capacity to fund the allocation the evidence would otherwise support.

• Concentration

  The firm's existing exposure to this company, counted across every fund it runs, is what limits the number.

• Timing

  The round's position relative to what is about to become knowable is the limiting factor rather than money.

• None

  Nothing external is binding. The allocation is what the evidence supports, unmodified.

The binding constraint tells a General Partner what would have to change for the number to change, which is more actionable than the number. An allocation limited by reserves is a fund-level problem with fund-level answers: recycling, a rebalanced reserve plan, or accepting the dilution deliberately. An allocation limited by concentration is a firm-level position question that has nothing to do with whether this company deserves more money. An allocation limited by timing is not a money problem at all.

Two funds can receive the same allocation for the same company with entirely different constraints binding, and they are in entirely different situations. Reading the number without the constraint loses that completely.

How to use it

Read the third element first. It tells you what kind of decision you are actually making. Then read the second, the signal against, because that is where your own context is most likely to change the answer. The signal in favour is usually the one you already knew, since it is why the company is on your desk.

The rationale is the part of the output designed to be disagreed with. A General Partner who reads it, decides the module has weighted something wrongly, and follows on anyway has used the module correctly. The point is not to be deferred to; it is to make the reasoning explicit enough to argue with.