In 2002, American Major League Baseball team the Oakland Athletics were competing against opponents who could outspend them three to one. They had no way of competing for talent with the top teams. So, their general manager, Billy Beane, stopped trying.
Instead, as chronicled in Michael Lewis’s Moneyball: The Art of Winning an Unfair Game (2003), Beane asked a more fundamental question: what factors predict winning in baseball? The answer turned out to be different from what baseball scouts believed.
Baseball scouts were evaluating players using visible, intuitive signals like athletic build and physical appearance, pitching speed and the look of their swing. The problem is that these signals felt meaningful but were often poor predictors of actual performance. Scouts were essentially fooled by aesthetics.
Statistical analysis revealed that they undervalued metrics like a batter’s ability to get on base (their on-base percentage), which were far better predictors of scoring runs and winning games. Because scouts ignored these metrics, players with high on-base percentages but “unimpressive” aesthetics or appearances were being sold cheaply on the market. Oakland saw what others couldn’t and acted on it.
Beane proved that a disciplined buyer who understands the true drivers of outcomes can exploit mispricing to outperform far better-resourced competitors. Decades of social psychology research have established how easily even experts can be swayed by false heuristics, stereotypes, and confirmation biases.
Venture capital today faces a strikingly similar condition: the market is inefficient; the heuristics are entrenched and the mispricings are measurable. The institutional resistance to correcting them requires investors to override deeply held assumptions about what a successful founder looks like.
To make the Moneyball case for venture capital, you need two things: evidence that evaluators hold biased beliefs, and evidence that those beliefs produce measurable financial consequences. Recent research provides both.
In a study published in Venture Capital, Koch, Berger and Kuckertz surveyed 361 international venture capitalists (VCs) using a sensitive questioning technique called the Crosswise Model. This matters methodologically because it guarantees respondent anonymity and circumvents social desirability bias – the tendency of people to tell researchers what sounds respectable rather than what they actually think. The results are stark.
Among the 361 VCs surveyed:
These biases are not uniformly distributed. They are most pronounced among male investors, seed-stage funders, and corporate VCs. Male VCs reported a 30.0% prevalence of devaluing women’s participation, versus 16.2% for female VCs. Seed-stage investors – precisely where the most consequential and information-poor decisions are made – showed the highest bias at 35.9%.
The declining prevalence across the three statements (from bias through prejudice to discrimination) is itself analytically interesting. It suggests that many VCs hold biased beliefs, but that only some manage not to act on them, which echoes the Moneyball dynamic in which scouts could see the data, but most could not override their instincts. Koch and her colleagues’ study should be regarded as providing conservative estimates: even with anonymity protections, some social desirability effects likely persist.
Our findings reveal a clear blind spot in the way capital is allocated.
If Koch’s work provides the mechanism – documented bias in investor cognition – our own research considers the consequences: measurable financial mispricing in the market for startup capital.
We analyzed 22,345 startups from Crunchbase, including 165 unicorns, using regression analysis and cross-categorical comparison. We classified founding teams into three categories: male-led, female-led, and mixed-led (the latter defined as teams with an equal number of male and female founders). The dependent variable was the total funding raised. We also examined post-money valuations and constructed a proxy return-on-investment measure by dividing valuation by total funding.
Our findings reveal a clear blind spot in the way capital is allocated. They deserve to be set out in detail.
The current investment rate of 81.6% indicates that mixed-gender teams are not valued at their true worth, despite consistently outperforming single-gender teams.
The headline result is that mixed-gender founding teams – those with a balanced 50/50 split of male and female founders – outperform single-gender teams on a valuation-to-funding ratio, our proxy for return on investment.
The median proxy ROI for mixed-led teams is 5.244, compared with 5.001 for male-led and 5.000 for female-led teams. That 4.87% premium may seem modest, but in a high-stakes environment, small margins are economically meaningful.
The valuation data reinforces this. Mixed-led ventures reach a median post-money valuation of $170m, compared with $100m for male-led and approximately $19m for female-led ventures. Yet they receive only a median of $2m in funding – less than male-led ventures at $2.4m, though far more than female-led ventures at $600,000.
The investment rate for mixed-led teams is the highest of the three groups: 81.6%, versus 80.1% for male-led and 77.4% for female-led. This suggests the market sees these teams and funds them at comparable rates (frequency-wise and in proportion). It is also important to remember the absolute values – only 816 companies in our dataset were mixed-led, compared with 18,480 male-led and 3,049 female-led ventures.
Crucially, the allocation is not proportional to performance. Given their superior ROI, the mixed-led investment rate should be around 84% (considering the 4.87% premium discussed earlier). The current investment rate of 81.6% indicates that mixed-gender teams are not valued at their true worth, despite consistently outperforming single-gender teams.
Investors misread educational credentials as a linear signal of entrepreneurial capability and reward a proxy that does not predict exceptional outcomes.
Our regression shows that each step up in educational attainment adds approximately 21.7% to funding. Yet when we focused exclusively on unicorn founders, the pattern reversed: unicorn founders predominantly hold undergraduate degrees, not doctorates. This suggests that investors misread educational credentials as a linear signal of entrepreneurial capability and reward a proxy that does not predict exceptional outcomes.
IMD’s Venture Capital Asset Management program is designed to accelerate your journey by providing the practical tools and insights needed to navigate the risks and opportunities of venture capital investing.
Research-based startups (AI, biotech, pharmaceuticals) represent 6.2% of our overall population but receive a significant funding premium: our regression estimates a 70% increase associated with being in a research-based field. However, only 3.7% of unicorns are research-based – nearly half the overall proportion. The most-funded sector (Software & IT) is not the sector that produces the most unicorns (Finance & Security). This suggests that investor capital is swept along by crowd signals rather than tracking the sectors where breakout performance actually occurs.
We define a “genius investor” as one who has participated in two or more unicorn funding rounds. Their involvement is associated with a dramatic increase in both funding (a 3,334% premium) and median post-money valuation ($1bn for genius-backed ventures versus $10m for non-genius-backed companies). But genius investors invest on average 9.3 years later than typical investors. Strikingly, no genius investor in our dataset participated in a funding round within the first three years of a company’s existence. This raises important questions about whether the most celebrated investors are identifying exceptional companies or riding the momentum of companies that have already proven themselves.
Non-unicorns have a median time-to-first-funding of one year; unicorns have a median of six to seven years.
Contrary to the prevailing assumption that exceptional ventures attract capital quickly, we find that a longer lag between founding and first funding is positively correlated with both higher subsequent funding and higher post-money valuation. Non-unicorns have a median time-to-first-funding of one year; unicorns have a median of six to seven years. This suggests that companies that bootstrap, build a solid base, and develop resilience may ultimately outperform those that scale rapidly on external capital – an insight with direct implications for how investors time their entry.
Perhaps the most striking structural finding is that at the unicorn level in our database, there are no female-led or mixed-led investor teams at all. The investors that have previously funded a unicorn are 100% male-led, and of the capital they have allocated, 91.3% goes to male-led ventures. The pipeline that produces the most celebrated outcomes in venture capital is, at the top, entirely homogeneous.
The underlying problem we highlight for venture capitalists is not the absence of evidence. It is the institutional difficulty of acting on evidence that contradicts entrenched assumptions. This is a leadership challenge as much as an analytical one.
Our findings on mixed-gender teams are the clearest illustration. The data says these teams produce higher median valuations per dollar invested. The data also says the market underweights them. The data further says that the investors at the top of the food chain – those funding unicorns – are 100% male-led.
Koch’s study clarifies the mechanism: implicit biases among VCs shape how they evaluate founding teams, and these biases are strongest at the earliest and most influential stage of investment, seed-stage funding.
Billy Beane’s baseball story is remembered as a triumph of data over intuition. But the data was available. The real challenge was convincing his own scouts and his own organization to act on what the data showed.
Venture capital’s version of that story is still being written. The evidence from Koch’s study combined with our findings suggests that the spreadsheet is ready. The question is whether the institution is.
he SpaceX IPO illustrates a key finding of the latest Future Readiness Indicator, today’s winners are no longer asset-light, but asset heavy. Does this mean that vertical integration should be top of...
AI excels at complex tasks, yet its effectiveness depends on strictly controlled boundaries...
Audio available
Explore Jan's journey of leadership self-discovery to enhance team trust and foster a collaborative workplace....
Are you caught in the insider trap? Learn to identify signs of leadership stagnation and explore effective ways to reset your approach for sustained success....
Explore first person business intelligence from top minds curated for a global executive audience