How AI can make the venture capital market more equitable

In the summer of 2020, COVID-19 continued to ravage the United States. Cases spiked in southern states. Companies underwent rounds of layoffs as retail businesses either limped along or collapsed under the weight of government restrictions. The country was suffering with no end to the pandemic in sight.

Yet venture capital seemed to move into comeback mode, particularly in the case of late-stage dealmaking. In Q3 of 2020, US companies raised a record number of $100M+ rounds, representing 54% of total venture funding. Meanwhile, first time financings fell to a record low of 5.5% as investors shifted their focus to existing portfolio companies or established brands with proven track records. In a period of extreme uncertainty, familiarity became a proxy for safety. While underrepresented entrepreneurs have long been overlooked in the venture capital ecosystem, this so-called “flight to quality” accentuated the divide between the haves and the have-nots.

But more large checks do not necessarily mean better quality – or greater opportunities. Even before the world went remote, investors plucked from a narrow universe of private companies for two key reasons: they rely on asymmetric information (lacking transparency) and highly manual processes (lacking efficiency). The inevitable outcome is decisions based on personal networks, heuristics and unintended bias. This is a structural problem that not only creates a massive opportunity gap for investors; it often denies otherwise talented entrepreneurs the resources they need to grow their businesses.

Funding trends throughout the pandemic underscore this opportunity gap. In the first three quarters of 2020, total deal value for female-founded companies dropped 31%, vs. the 18% decline across all US venture rounds. Further, female-founded companies received only 13% of all venture dollars deployed during that time, down from 15.5% in 2019. And despite the fact that Black people and Latinos make up 32% of the US population, companies founded by members of those demographic groups raised only 2.6% of US venture funding over the same period.

These statistics cannot be separated from the lack of diversity in venture itself. In the US, women comprise just 12% of decision-making roles at firms with $25M+ AUM, while only 3% of venture investment partners are Latino and another 3% are Black. The composition of decision-making teams is critical because human nature involves a predisposition to associate with similar people, known as homophily.

Investors tend to back founders to whom they relate and trust, often because they share social or physical characteristics. While largely unconscious, the result is a funding environment that tilts toward one gender and one race. Location and pedigree matter too: almost half of all investments are concentrated in Silicon Valley, and more than a quarter flow to founders from the Ivy League.

These barriers to capital are not just inequitable – they are economically unsound. Studies show that diverse founding and investment teams actually outpace their competitors. One observed that diverse founding teams eventually outperformed their counterparts with a 30% greater return at IPO or acquisition.

A different study linked non-diverse investment teams to lower investment performance: the comparative success rate of an investment at IPO or acquisition dropped by 26% when investment partners shared the same ethnicity. Another revealed a trillion dollar opportunity gap caused by the lack of representative funding for women and people of colour. Who knows what investment opportunities have been missed because brilliant ideas could not get off the ground?

We are on the cusp of an industry transformation we have seen before. Technology has already brought transparency and efficiency to several markets, from Zillow in real estate to Google with search ads to Amazon with price discovery. In fact, the public markets were transformed by the likes of Bloomberg and Renaissance Technologies decades ago. Our world has become increasingly digital since then, equipping us with new tools for accessing, storing, and manipulating data. Yet, there has been limited innovation in private investing to date.

This lag is understandable. Systematically acquiring and normalizing information on private companies is a huge technical challenge. Unlike public market data, which is widely available, structured, and consistent, private company data is fragmented, opaque, and largely ephemeral. CircleUp is harnessing big data and artificial intelligence (AI) to tackle this problem. The end goal is a complete mosaic of the private company landscape, and the initial wedge is early-stage consumer brands.

CircleUp’s technology ingests, maps, and organizes terabytes of unstructured data across over 200 sources. It then identifies and assesses brands based on several factors predictive of future growth, such as distribution, product differentiation, and consumer sentiment. These capabilities power data-driven investing strategies that apply the quantitative techniques of public market investors to access the long tail of emerging consumer businesses.

The objective of this technology is to shine a light on entrepreneurs that deliver outstanding business performance, regardless of their location, identity or experience. By introducing transparency and efficiency to the investment process, our strategies have uncovered opportunities that have historically gone unnoticed. There is financial incentive for this data-driven model: objective insights drive smart investment decisions, which yield strong fund performance. But if these smart decisions are truly merit-based, the same technology that optimizes for profit should optimize for greater representation in parallel.

Early fund performance demonstrates that technology can power alpha-generating investment strategies that are more inclusive by nature. Within CircleUp’s equity fund, marked in the top quartile for its vintage, 55% of portfolio companies have a female founder, and 35% have a non-white founder (based on a self-reported survey). Geographic proximity to venture hubs also becomes irrelevant with a data-first approach. Across strategies, CircleUp’s portfolio companies span 19 US states.

These initial results indicate data-driven investing can provide a scalable solution to the funding gap. Technology enables more efficient, effective, and repeatable decision-making at each stage of private investing. Investors can source top brands without attending industry conferences or tapping into personal networks. For diligence, data supplants preconceived notions that often lead investors to a “no”, even when a product deeply resonates with its customers. And for post-close support, AI can help answer an entrepreneur’s strategic questions – advice that has historically relied on pattern-matching and gut instinct.

We must use these capabilities wisely. While AI can be a force for positive change, it can also perpetuate the status quo. Capital allocators that use technology to generate strong returns must regularly assess whether their techniques are driving the more inclusionary outcomes we would expect.

With sufficient data and computational power—coupled with an eye toward reducing potential bias in AI – innovators can help to level the playing field for underrepresented entrepreneurs. A data-driven approach to investing will empower businesses the world over with a fair shot at success, while generating attractive returns for capital allocators.

While COVID-19 may not have upended the private investing industry, change is assuredly coming. And it will be a welcome sight.