Is algorithmic VC investment compatible with due diligence?

In algorithmic investing, investors use a company’s metrics to decide whether to participate in a deal. But when the art of choice is factored out, it becomes more difficult to perform deep due diligence on founders who may be about to receive millions of dollars via a wire transfer.

In practice, attempts to remove bias can create newer, blind spots that are harder to identify.

In theory, algorithmic investing hedges against investors’ preconceived notions and pushes emotions to the side. Fintech unicorn Clearco and venture firm SignalFire have spent years implementing data-focused investment processes, joined more recently by AngelList and Hum Capital. While this approach isn’t new, the movement against solely emotion-based decisions feels louder given the proliferation of dollars out there.

Metrics, even in the earliest stages, are becoming more mainstream.

AngelList’s recently closed early-stage venture fund is basing all of its investments off of one key metric that AngelList has been tracking for years: a startup’s ability to hire.

When I spoke to Abraham Othman, head of the investment committee and data science at AngelList Venture, he told me they win deals because they are less adversarial to portfolio companies than other firms.“Our approach? This is our data set — let’s see if we can put money into them,” he said.

No further due diligence? No problem.

It’s no small set. About 2 million individuals use AngelList Talent to apply to startups each quarter. About 35,000 companies per quarter are contenders for AngelList talent, but only half of those companies are investable early-stage businesses.

Othman said that early in the process, he was surprised to see a solid percentage of applicants on the platform flocking toward startups based in India. His venture operation is now investing in India alongside the United States, underscoring how hiring demand helped guide interest.

There are challenges with leaning on such signals to make investments. Some metrics can be biased; for example, while an investor may decide to set aside a founder’s personality, their personal brand may be the factor that encourages applicants to consider applying in the first place. Additionally, a startup could get a ton of applicants due to pay, location or even recent coverage in a Well Known Tech Blog — which can bode well for success, but could also just be a result of great marketing.

Clearco co-founder Michele Romanow has bet her career on the algorithmic investing space, but she doesn’t think hiring demand on its own is a strong enough signal.

“We don’t believe any one signal is enough to holistically look at the success or future growth of a business. We try to look at as much data as possible, and while hiring can be one aspect, on its own, [it’s] just too subjective and open to manipulation,” Romanow said. The co-founder built a fintech unicorn that offers businesses a certain amount of non-dilutive capital based on a select few metrics. The company, formerly Clearbanc, started with growth capital for marketing expenses and recently expanded to inventory and payroll.

“Every little while, we hear more rumblings that the industry will shift to lean in this direction, but candidly, it’s hard,” she said. “It requires deep technical expertise and a product to match, so while it sounds nice on the surface, people generally revert to what they know, which is the traditional gated system with humans making decisions based on intangible factors.”

Mat Sherman, the founder of Seedscout, a database of early-stage startup information, thinks that the best metrics tend to be a bit vague.

“By the time a startup is ramping up hiring, access matters more than ever, which they probably have,” he said. “ As a fund, it makes sense. [But the] real alpha is, can you predict a company will be hiring a lot before it starts?” Sherman declined to share how he specifically vets investments, adding that it’s a secret, but did say “the metric is more of a triangulation of different metrics that result in a feeling…the feeling is the metric.”

At its best, algorithmic investing isn’t a bet against emotions — it’s a bet despite them. Is idealistic to assume that a startup’s success can be decided off of a few numbers, especially in a bull market when everything is up and to the right.

But, when used as a tool, algorithms can help influence our decisions in ways our emotions can’t account for. We know we have inherent blind spots, so maybe it’s time to rely more heavily on data so we can reach a clearer conclusion.