How many partners can a consensus committee carry

What this check compares

This check reads the number of General Partners against the voting model those partners use. Some combinations of the two are not really voting arrangements at all, and the check exists to name them.

A consensus vote with a single General Partner is a formality, since there is nobody to reach consensus with. A simple majority with two partners is the same problem in a different shape, because a majority of two is either unanimity or deadlock and never a majority. At the other end, a consensus vote across many partners scales poorly for the familiar reason that every additional partner is another way for a decision not to happen.

Why it is worth checking

Firms usually write down a voting model at formation, when the partner count is whatever it happens to be, and then the partner count changes. A firm that adds a partner, or loses one, rarely revisits the voting language in the same motion. This is the check most likely to surface something a firm already half knows and has not gotten around to fixing, and it is also the cheapest to resolve, because changing the voting model is an internal decision rather than a negotiation with anyone outside the partnership.

A question that comes up on all three IC checks

Readers ask a version of the same question on each of the three checks that read the investment committee: does the norm against funding competing companies still hold, and does the model assume it does. The honest answer is that the norm is visibly eroding, and that the model does not depend on it either way.

Charles Hudson of Precursor Ventures has argued that as venture fund sizes keep growing, the tradition of firms holding a norm, if not a stated policy, against investing in competitive companies is likely to go away, and that this is a case where the business model for large funds is at odds with what most founders want. Chris Neumann, writing up that argument, notes that smaller funds and single-stage specialists are rethinking their own approach as well.

This is a long-simmering tension rather than a new development. Hunter Walk laid out the same mechanism in 2019, observing that large firms writing early checks would eventually find they had blocked themselves out of a category, and predicting pressure on the traditional principle that a firm will not back two competing companies. Two arguments seven years apart, reaching the same conclusion, is evidence of a slow structural drift, not of a break.

What this means for these three checks is narrow. The model reads the firm's declared voting structure, its declared scope, its declared economic rights, and its partner count, and asks whether those choices fit each other. It does not read the firm's conflict policy, and it takes no position on whether a firm should hold one. A firm that has relaxed its conflict stance has changed something real about how it operates, and the model will see that only where it changes one of the structural variables above.

• VCs are Changing Their Tune on ConflictsChris NeumannQuotes Charles Hudson of Precursor Ventures on why the norm is likely to go away as fund sizes grow.
• Coming Storms: Three Reasons That VC Firms May Start Overlooking “We’re Conflicted” and Make Competing InvestmentsHunter Walk, October 2019The same argument, made seven years earlier.

What this check does not say

The check reads General Partners, meaning partners who hold both economic and decision rights, rather than the full investment team. A firm with two General Partners and six other investment professionals is a two-partner firm for this purpose. That distinction matters because the check is about who votes, not about how many people work at the firm.

The check also does not say a small partnership is a weakness. A single-partner firm making decisions alone is a coherent structure and a common one, and the model treats it as such. What the check objects to is a firm describing that arrangement as a consensus process, which tells an LP something about the firm that is not true.