Evan Reiser and Sanjay Jeyakumar do not come from cybersecurity backgrounds. The pair linked up at TellApart, the advertising technology startup that was sold to Twitter for $479 million in the social media giant’s largest acquisition to date.
After quitting Twitter in 2018, Reiser and Jeyakumar bet that their marketing mindsets could translate to the world of email security. Their idea to use behavioral profiling—the method by which Facebook and Twitter can display ads tailored to you—to predict socially engineered email attacks has gained steam, surpassing the valuation of TellApart in two years thanks to a lofty fundraise this week.
Abnormal Security announced Wednesday a $50 million Series B funding round led by Menlo Ventures with participation from prior investor Greylock. The funding values the San Francisco-based company at more than $500 million, according to a source familiar with the deal, more than tripling its $140 million valuation following a Series A last November.
“A lot of people in security come from a threat intelligence background with antivirus and antispam, where you’re trying to recognize signals when they come in,” says Reiser, who is CEO. “For us, it’s a different kind of mental model. It turns out this style of approach is very effective for this next generation of cybersecurity attacks, which are about impersonation.”
That impersonation often comes in the form of unassuming emails with large sums of money on the line. In the common example, a company’s head of finance receives an urgent message from the CEO: “Please wire $10,000 to this bank account.” Only, it’s not the CEO, but a scammer spoofing the email account. Lately, the attacks have evolved to innocuous emails from fake employees, from requests to HR reroute their direct deposit to emails asking for a copy of their tax form. An FBI report found such email attacks to be far and away the top cause of financial crime on the internet, accounting for $1.8 billion in losses in 2019.
To keep up, Xerox, Alliance Data and dozens of other enterprise companies are installing Abnormal Security onto their cloud. Abnormal integrates into Office 365 and G Suite, allowing companies to begin relying on its artificial intelligence-based threat detection system within days. Competing against established, billion-dollar email security firms such as Proofpoint and Mimecast, Abnormal bills itself as being able to stop a new form of cyberattacks the older platforms cannot catch.
Just as marketing technology looks at a user’s data to serve up suggestions or ads based on their clicks and interests, Abnormal Security uses machine learning to learn the behavior of an organization’s email account holders. “We can build this relationship graph of everyone inside the company. Who are they talking to, about what topics, at what time of day?” Reiser says. Together with other AI-minded technological functions—like using computer vision to detect whether an invoice looks legitimate and if it matches with other emails—the software makes the call on whether incoming emails are legitimate or fraudulent.
The technology comes from Reiser, chief technology officer Jeyakumar and a team of more than one dozen former colleagues from TellApart and Twitter, who have spent the last decade building machine learning for adtech. The strength of the team, plus information on the evolving nature of cyberattacks gleaned from conversations with dozens of CIOs and chief information security officers, helped Reiser convince Greylock to go in on a seed round in April 2018 despite having no product or code.
The startup has grown rapidly since its inception, with the team now made up of about 100 employees, according to Reiser. Business for Abnormal, which uses a “per user, per year” model, began to surge around May, he adds. Encouraged by the progress on the product and customer base, Reiser began seeking more capital and received immediate interest from Menlo Ventures.
“We’ve been tracking this company for a long time, and we had been given the Heisman—the whole time, we could not get a meeting,” says Venky Ganesan, a partner at the firm who is joining the Abnormal board as part of the round. On a Friday last month, Reiser finally said he was ready to meet. After a four-hour dinner and a weekend spent modeling the opportunity, Ganesan was ready to invest, and called an emergency meeting with his partners Sunday night to finalize the offer.
Reiser says the cash injection will go towards expanding the team to 250 employees by the end of 2021. As the product scales, the CEO envisions its technological core—behavior profiling combined with anomaly detection—to be applicable beyond securing the cloud, to areas such as data loss protection.
“Our approach right now is working, so we want to just double down,” he says. “We want to be the premier machine learning factory in Silicon Valley.”
-By Kenrick Cai, Forbes Staff