Cognitive Bias Lives in the Gut
Venture capitalists are famous (and perhaps infamous) for basing investments on their “gut” intuitions, and they can be quick to resist the use of AI in their own decision making. VCs think everyone should adopt AI… except for them.
Daniel Kahneman's "Thinking, Fast and Slow" sheds light on why this happens, explaining how investors tend to rely on what he calls "System 1" thinking - intuitive, rapid judgments based on biases and heuristics. In contrast, “System 2” thinking involves more deliberate, analytical cognitive processing.[i] In the venture capital industry, the preference for intuitive System 1 decisions vastly outnumbers more quantitative approaches.
This would be fine if it worked, but unfortunately the venture capital industry has a poor accuracy rate; an estimated 75% of venture-backed companies fail.[ii] It’s time to do better and AI can help; if VCs have the guts.
Feast or Famine
Harvard Professor Clayton Christensen pioneered "Disruption Theory," which reveals how startups that initially target low end or niche markets can become market leaders that ultimately displace mainstream industry incumbents.[iii] This process of “disruption” has created so many market-leading companies that finding disruptors, as early as possible, is the driving fantasy of the venture capital industry.
Yet when it comes to identifying early disruptors, intuition-based decisions create blind spots. That’s because disruptors begin in less conventional markets or in niche sectors that usually appear unattractive at first, rendering them particularly counterintuitive in their early stages. For example, Uber began in the taxi industry, which had been more or less static for a century. Similarly, Google was initially dismissed by a long list of investors because Search was seen as “a feature, not a business.” VCs who rely heavily on intuition can, by definition, be expected to struggle with anything counterintuitive.
Later, as disruptors gain momentum and a consensus forms around their success, VCs also tend to overcompensate by shoveling money at these startups after-the-fact, often haphazardly and at inflated valuations. Relying on intuition creates an environment of feast or famine.
Leveraging AI and quantitative tools can help, because an ability to follow the data – even when it’s counterintuitive – can lead VCs to discoveries and points of view they never otherwise would have considered. In the VC funds we manage, for example, we almost never know which market or startup we’ll end up focusing on until after the AI is run. Then, once we’re confronted with the data, we try to wrap our minds around what we’re seeing and to make sense of it. This is where our biggest wins have tended to come from and, without AI and quantitative tools, we wouldn’t have been able to journey beyond our own intuitions, biases and preconceived assumptions.
Two Guts Aren’t Better Than One
The designer of the Mini Cooper in 1959, Sir Alec Issigonis, is often quoted as saying “a camel is a horse designed by committee.”
Along those lines, group or consensus-driven decisions by venture capital firms and their investment committees often exacerbate bias and System 1 thinking. Group processes essentially merge multiple intuitions into a compromise. They design a camel, so to speak, and can systematically weed out counterintuitive or non-obvious opportunities.
This contradicts the concept of crowd wisdom[iv], which describes circumstances where the collective insights and judgments of a diverse group of individuals tends to be more accurate and reliable than those of any single expert or small group. While true in many contexts, our empirical tests of crowd wisdom in the context of venture capital decisions haven’t shown statistically significant effects (crowd wisdom hasn’t beaten random chance).
Apparently venture capital isn’t one of those things crowd wisdom is especially good at. In fact, as VC investment committee decisions increasingly incorporate group dynamics, personal politics, and interpersonal compromises, often unrelated to the startup itself, consistency in VC decisions diminishes.
This is just another example of how accurate quantitative models can yet again provide a more consistent and reliable foundation for VC decisions.
A Hybrid Diet
When it comes to AI and its power to transform industries, it’s time for venture capitalists to eat some of their own dogfood.
To make better decisions and overcome biases, VCs need to more rapidly embrace quantitative tools, AI, and System 2 thinking. By using data-driven analytics, VCs can better spot disruptors that intuition might miss while increasing the accuracy of their predictions. This shift towards a more balanced diet of human judgment and technological tools is crucial for improving the venture capital industry. That is… if they can stomach it.
[i] Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
[ii] Shikhar Ghosh, The Venture Capital Secret: 3 out of 4 Start-Ups Fail, Wall Street Journal, Sept 19, 2012 < https://www.wsj.com/articles/SB10000872396390443720204578004980476429190>
[iii] Christensen, C. M. (1997). The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business Review Press.
[iv] For a deeper exploration of “crowd wisdom”, see: Surowiecki, J. (2004). The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations. Doubleday.