The Compact Room

Part of The Build. Start with the guide for the full map.

If I had to bet on one organizational-design claim holding up across every future presidential campaign, it would be that the campaigns that function well keep their true decision room small, somewhere around five to eight people, and the ones that don’t, don’t. I’ve cited this number elsewhere in this series as one of the field’s clearest invariants. Here’s the case for it, tested rather than just asserted, including the specific number’s weakest point.

Is there a general relationship between the size of a campaign’s decision-making group and its ability to function under pressure? And if so, is 5-8 a threshold or a number that happens to describe a few convenient cases?

Every 2020 campaign that let its effective decision-making group grow past roughly eight people, whether by formal design or informal drift, produced documented confusion about who actually had authority to decide something in real time, contested calls that either stalled past the point where a fast decision mattered, or got made by whoever was loudest in the room rather than whoever formally held the authority. Every campaign that protected a tighter boundary didn’t show that same pattern.

The canonical case for this, fully worked through elsewhere in this series, is Kamala Harris’s 2019 campaign. It was a formally compact org chart that concealed an oversized, unresolved decision room. This produced a resource dispute that got answered too late before the campaign wound down. Full detail is in the decision-rights piece and the Contingency Table. The point that matters here is the number isn’t about headcount on a chart, but it’s about how many people actually have a real vote on a contested call, formally or otherwise. Harris is a clear case of those two numbers diverging.

Harris isn’t the only one. Warren’s campaign, despite building the most functionally elaborate organization in the 2020 field, kept its actual senior decision room compact even as the functions around it grew. That’s worth noting specifically because it rules out an easy alternative explanation that a small decision room is really just a symptom of a small, under-resourced campaign that hasn’t grown enough functions to need a bigger one. Warren’s campaign had every reason, by that logic, to need a larger room. But it didn’t build on and its documented organizational failure (a burn rate around 158% of inflow in a critical 2019 quarter) traces to resource sequencing, not decision-room confusion. Two campaigns, two different failure modes, and the one that kept its room compact didn’t show the authority-confusion pattern the size claim predicts, even under real organizational complexity.

Coordination cost. The standard organizational-behavior explanation is that communication overhead among a group grows roughly with the number of possible pairwise connections, not linearly with headcount. So a group of fifteen has to manage nearly four times the coordination burden of a group of eight. This explains why bigger rooms get slower and more confusing, but it doesn’t fully explain why 5-8 specifically, rather than, say, 10-12 is where campaigns seem to hold the line.

Functional coverage. A narrower, campaign-specific explanation is that 5-8 is roughly the minimum needed to cover the functions that need a seat. Campaign manager, political/strategy, communications, finance, and usually the candidate’s closest formal advisor, without leaving an obvious gap. Three is probably too narrow to cover those functions; the number grows from there only as fast as new functions need direct representation rather than information flowing in through the existing seats.

A confound I can’t fully rule out. Campaigns that keep their decision room small might simply be better-run campaigns generally for reasons unrelated to room size. It could be better initial hiring, a candidate with clearer instincts about delegation, or a campaign manager with more authority to enforce discipline in the first place. If that’s true, room size could be a symptom of good management rather than a cause of it and the fix wouldn’t be “shrink the room” so much as “hire and empower a campaign manager capable of keeping it small,” which is a related but distinct intervention.

The instinct to expand the room is usually well-intentioned. More expertise, more perspectives, and more buy-in from more parts of the campaign. But every function granted a permanent seat is a function implicitly told it gets to weigh in on everything, not just its own lane, which slows every contested decision down regardless of whether the individual added real judgment to that specific call. That’s a mechanism distinct from generic “good management” and it predicts something the confound above doesn’t. A well-managed campaign which nonetheless lets its room grow past eight should still show the confusion pattern. I don’t have a clean test case of a well-managed-by-every-other-measure campaign with an oversized room in this dataset which is exactly why I’m not calling this established.

Here’s the discipline this series tries to apply to itself elsewhere. The direction of this claim—smaller, clearer decision rooms outperform larger, murkier ones—is well-supported. The specific boundary of 5-8 is a real pattern in the cases available, not a number derived from a controlled test across enough campaigns to say confidently that seven works and nine doesn’t. A campaign with a disciplined nine-person room where every seat has exercised authority and no informal parallel channel exists might function fine. I don’t have a documented case of that in this dataset to confirm or rule it out.

The core claim—org charts are a poor predictor of how a campaign actually functions and the real decision room’s effective size (not its formal size) predicts functional confusion—is supported-but-provisional. It’s well-evidenced in the clearest cases, consistent with a real coordination-cost mechanism, but resting on a small sample rather than a systematic survey. The specific 5-8 number is directionally right and should be read as a useful range. It’s closer to context-dependent than established per the discipline this series tries to apply to its own borrowed numeric precision elsewhere.

What I’d tell a campaign building its structure from this is that the goal isn’t hitting a specific headcount. It’s making sure the number of people who can actually cast a vote, in the moment, on a contested call is small enough that everyone in the room knows exactly who the others are and exactly how much authority each of them actually has, and that this real number matches the org chart’s stated number rather than diverging from it the way Harris’s campaign did. The moment those two numbers split is the moment a campaign starts making its worst mistakes without realizing it’s making them.