The Founder Measurement Error

👋 Hi, welcome to Signal Lab where the team at Innate Language Processing digs deep into language & behavior, the limits of technologies, and the connection between what people say and do. 100% human authored, never with AI. For venture investors and founders.

How much is too much in a founder? Can they be too stubborn, too laser-focused, too “obsessed”?

VCs are often asked this question (usually on podcasts), and every time they demur. Because this is a quantitative question with a quantitative answer and VCs’ evaluation of founders is not quantitative but intuitive and personal, they have no answer. Instead of admitting this limitation or engaging with the nuance, they either mention that unicorn founders are “extreme people” or they shift the onus by adding more unmeasured qualities to the list, bloating demand on a founder’s soft skills. This plasters over the measurement problem entirely under the assumption that a founder with a multiplicity of “unicorn traits” will be less likely to quit or fail.

There is a measurement problem in VC, and the cost is being passed on to founders (like everything else).

So founders have to have it all…

… even to the point of embodying completely opposing qualities.

To some extent, the path of the founder is one of contradiction: in a matter of a few years, a unicorn founder has to go from being the hands-on jack-of-all-trades, to the hiring manager and then to corporate CEO, all the while leading. Founders have to reinvent themselves all the time.

But founders are expected to embody contradictions that go beyond the practical demands of scaling. Because investors cannot measure or weight communicative styles in a pitch, founders are encouraged to perform polarities, to put their exceptional abilities beyond doubt.

Mark Suster (Upfront Ventures) warns founders not to say things like “We hope to raise” or “If we could only” to avoid coming off as weak or uncertain. Founders are expected to exude confidence: “We are raising Z, our revenue will be X by Y.” But plenty of investors would rather hear thoughts in-process and will discount founders who front too much authority or certainty. Vlad Loktev (Index Ventures) wants to hear the raw thoughts, the challenges, the fears, what keeps a founder up at night. In his view, naming ten things you’re afraid of is what makes a conversation about building something great real.

Some investors insist on clarity about long-term plans and visions for the company. Others — Sequoia among them — look for abstraction as evidence of a founder’s willingness to adapt and change their mind.

Bill Gurley (Benchmark) says the best founders can hear hard truths from boards and executives, but Ben Horowitz says the great thing about the really great founders is they don’t take any advice.

Some investors (particularly deep tech and biotech) have noted that the most important companies are often the hardest to pitch early, because the founders are too close to a genuinely novel problem to have found the right language for it, yet. These investors tolerate early messiness without treating it as a signal in itself. Others believe messiness signals openness, and read it as a positive signal on its own.

Many investors believe founders need personal, identity-level conviction to sustain the long journey to exit: e.g., “I just have to build, there’s nothing else I want to do”. Peter Thiel views this unfavorably. He selects for specificity, for solving a particular problem rather than performing the identity of “entrepreneur.”

Keith Rabois (Khosla Ventures) has said repeatedly that the more often a founder zooms out to speak in terms of “whys” the better off the company will be. That level of reflection and big picture sense-making is close to the opposite cognitive skill from what Marc Andreesen (a16z) looks for: knowing the mechanics of absolutely everything about the company, “every knob and every button.”

Most investors expect polish and preparedness from a pitch. Jason Lemkin (SaaStr) thinks the opposite, that being too polished, especially as a first-time founder, can almost be a negative for seed and early stage investors.

The sharpest version of this contradiction comes from Sequoia. Michael Moritz has named clarity of thought and the ability to communicate clearly as defining traits of a world-class entrepreneur. But his partner, Pat Grady, pointed to clarity as the very reason he passed on two founders who went on to build unicorns (Peter Reinhardt of Segment and Nikolay Storonsky of Revolut). Both presented their companies clearly and concisely in five to ten minutes, leaving him with no questions. Total clarity delivered fast simply didn't register as conviction to him.

Two partners at the same firm, same trait, reaching opposite conclusions. If clarity is simultaneously the strongest predictor of greatness and a reason to pass, it isn't a predictor but a story told after the fact to explain decisions made on other grounds.

Some VCs attribute as much as 80% of their decision-making to their read on the founding team. Are they coachable? Do they have big vision? Are they self-aware? Are they engaging? All of these, weighed together, with no rubric for what “enough” or “too much” of any one trait looks like.

VCs are well-positioned to evaluate a founder’s execution ability from available data sources. That’s their domain. What they’re not well-positioned to evaluate are upstream qualities of expression. Qualities of expression get treated as proxies for quality of thought, which gets treated as a proxy for execution ability. Two layers of inference, stacked on intuition. Intensive founder evaluation makes sense at early stage, where there's little other data, or for first-time founders new to the scene — but it's applied everywhere, badly.

Reading people feels like a superpower, even when it’s done wrong. Traditionally this was the province of what used to be called wisdom. These days, expertise has pride of place. Regardless, in the complex subject of judging individual human beings, it is far from clear that VCs are qualified in either respect.

Market conditions change the bar, too. In hot markets, a lot of these expectations loosen, and VCs settle for less from their founders (usually without even realizing they’re doing it). For all their interest in precisely how unicorn founders discuss their company, many will fund a team even if they don’t know what they’re building.

The measurement error doesn't ever go away. But it does get cheaper to ignore when there's money to burn.

See you next month.