Due diligence: Investors turn to tech support to identify red flags

Headline’s Deepdive technology helps the firm discover insights – and spot potential fraud – by analyzing a company’s raw financial and usage data.

After the SEC charged Theranos founder Elizabeth Holmes with fraud in 2018, her massive deception appeared to be an isolated incident. No other deceit on the same scale came to light in the four years before Holmes was sentenced to 11 years in prison in November. But just a month later, Sam Bankman-Fried was arrested and charged with fraud by federal prosecutors for alleged financial crimes at crypto exchange FTX.

Yet another alleged fraud made headlines in January, when JPMorgan sued Charlie Javice, the founder of VC-backed fintech company Frank. The Wall Street bank claimed she duped it into buying her company for $175 million by creating more than four million fake customer accounts.

The alarming spate of incidents should give venture firms pause. Do they have the proper systems in place to spot fraud before they write a check?

Headline, a tech investor based in San Francisco, relies on analytical software called Deepdive to catch anomalies in financial data. Partner Jett Fein tells Venture Capital Journal that the technology is the starting point for all the firm’s due diligence.

When considering an investment, the first thing Headline does is ask a company for basic financials and raw payments and usage data from sources such as Stripe and Google Analytics. Those numbers are imported into Deepdive, which analyzes and visualizes the data. “Essentially with very little information from a company, we’re able to produce 50-70 different charts to help us understand things like revenue retention, unit economics and capital efficiency,” Fein says.

The technology also does another neat trick. It uses forensic accounting to look for abnormal patterns in the raw data submitted by a company. Since it was developed about 10 years ago Deepdive has uncovered at least two cases of fraud. In one of those cases, Headline informed the board of the company, which ultimately shut down the business.

Fuel efficiency

Early fraud detection is just one of several benefits of using Deepdive. The main reason Headline uses the tech is to quickly separate strong investment prospects from poor ones. “Sometimes you can have a really shiny car, but it has really terrible fuel efficiency,” Fein says. “We’re looking for companies with really amazing fuel efficiency, or capital efficiency, that can turn a dollar of spend into $10 of revenue. That’s what Deepdive allows us to figure out in a really easy way.”

The insights that can come from the technology aren’t always obvious. Sometimes they become clear after data is input over time. For example, Fein says he gained insight into delivery service Gopuff back in 2016 that helped convince him to invest in the company, which today is reportedly valued at more than $15 billion.

“There are a number of different dimensions to this fact-gathering mission, but Deepdive tends to be something we can do very early in the process because it doesn’t require too much – just a couple pieces of data from a company”

Jett Fein 

When he analyzed the Deepdive dashboard to see how Gopuff performed in new markets, Fein was stunned to see that it grew faster in every new market it entered. “That was an incredibly positive sign to us because it meant that they were learning and getting better at launching new markets and getting to profitability faster,” he says.

The discovery was all the more powerful because it wasn’t something that was evident if you looked strictly at the performance of Gopuff’s first market in Philadelphia, which was three years old at the time. “It was definitely one of the aha moments,” Fein says. He is quick to point out that Deepdive is just one component of Headline’s due diligence process, which he describes as a big “fact-gathering mission.”

“There are multiple other sorts of facts we’re trying to gather about a company, like facts about the founding team,” he notes. “We try to learn how that founder has fared in adversity and chaos. We talk to people that have been around the founding team. And we spend a lot of time with the founding team. We are trying to judge whether this is a founder or a team that we would enjoy working with and that would be receptive to our help.”

In the case of Gopuff, Fein flew out to Philadelphia on multiple occasions, spending time in its warehouse and even going on a ride-along with a delivery driver. “On the very first delivery that I went on there was a customer that lived directly across the street from a 7-Eleven,” he recalls. “That’s when the light bulb went off.”

Headline also spends a fair amount of time trying out the products or services of potential portfolio companies. “There are a number of different dimensions to this fact-gathering mission, so it’s not just Deepdive, but Deepdive tends to be something we can do very early in the process because it doesn’t require too much – just a couple pieces of data from a company,” Fein says.

As for how long the due diligence process lasts, that really depends on the company. “Our Deepdive process takes 20 seconds,” Fein notes. “If the data analysis looks interesting, it can take anywhere from a number of days to a number of months [to complete the due diligence]. Maybe we’ve known the founders for a long time or maybe it’s a space we know really well,” which speeds up the process.

For Gopuff, the process took a full six months. “Throughout that roughly six-month period we got more data at multiple intervals and were able to see the company progress along a lot of the metrics we track,” Fein says. “We also got to know the founders and the rest of the team, so we really were able to uncover a lot about the company and really build our conviction before we made the investment. Contrast that to the last year when firms were making investments in companies that they met the prior day.”

For anyone thinking about developing their own Deepdive technology, you should be aware that it’s not a simple or inexpensive undertaking. “This has been conceptualized and built over the past decade,” Fein says. “We have a team of 20 full-time engineers that built Deepdive and one other product for the investment team and are exploring several others to build next.”

Fein won’t say exactly how much Headline has spent on the technology or disclose the ongoing cost, but he allows that “it’s quite expensive. We view it as a long-term investment. We believe it gives our investors real superpowers and generates real alpha for us.”