Erik Torenberg sat down with investor Keith Rabois in front of a live audience at the LAUNCH conference earlier this year as part of Product Hunt's ongoing "Maker Stories" series of podcasts.
Keith was an early executive at LinkedIn, PayPal, and Square and has invested in YouTube, AirBnB, Quora, Palantir, and Yelp. He is considered a member of the infamous PayPal Mafia.
Keith started out as a lawyer and litigator before jumping out of the legal world at age 30 "sort of without a parachute" in early 2000 (the height of the Internet bubble). He joined PayPal right after Elon Musk had been fired as CEO and replaced by Peter Thiel.
What PayPal was like when he joined:
He didn't know PayPal was going to be a huge company at the time. It had ~4.5 million customers then, which was a lot, but it was "bleeding money", losing $10 million per month (a huge burn rate for the time, though now that's "almost common"). "It was a complete mess." PayPal was not obviously a rocket ship. It was a brand people had heard of, though. The company had gone through three CEOs in six months.
Peter Thiel started fixing things. The company went from very unprofitable to profitable, then filed an S1 to go public the day after 9/11. PayPal ended up going public in early 2002, then soon sold to Ebay. It was two years from "complete mess" to IPO to acquisition.
At PayPal, Keith was working at the intersection of business and regulatory problems. Payments intersects with all kinds of regulations. His job was to "keep all these evil companies like MasterCard and Visa from killing us and Ebay from killing us". He describes wielding the law and regulation to his advantage. He had 17 full-time lobbyists in D.C. working for him offensively and defensively. At one point the Treasury Department wanted PayPal to collect a social security number from every buyer, which would have really killed the company. "I had to stop that."
Keith describes how the PayPal Mafia became influential:
Peter Thiel and Reid Hoffman were the only ones investing in consumer internet companies during the "nuclear winter" in the Bay Area after the Internet bubble. Every consumer-focused entrepreneur went to them and it turned out some of those ideas were decent. Some were potentially interesting but crashed and burned, like Friendster.
What happened after PayPal:
Elon tried to recruit Keith to join SpaceX, to help with regulatory complexity, but Keith didn't want to move to LA. Instead he got involved with LinkedIn. LinkedIn was growing fairly linearly and did not look like a rocket ship at the time. After a year and a half it had 1.5 million users "which for a social product, a viral product, it is very very low." But, it was always growing and had a "monopoly characteristic" despite not having an exponential curve. It was the only professional database for normal people. It was obvious that it could be "pretty valuable" but not totally obvious that it was going to be massively valuable.
Keith describes his hiring philosophy:
Keith says he was bad at interviewing people at PayPal. He was 50/50 on bringing in good new people. His poor hiring record was interfering with is career. He decided to focus on recruiting stars within the company to join his team. This later lead to lots of good investments, as these people went on to start companies. He tried to look for common denominators in these people that he could look for in interviews. "Almost all of the best people I've hired went to fairly random universities with non-technical backgrounds that had never worked in the startup world before, all 21 to 23 years old." But it's a lot like drafting athletes, you're going to be wrong. It shouldn't be "zero-defects". If anyone tells you that, they're not hiring fast enough.
Peter Thiel taught him that you can't recruit people who have already established themselves and accomplished things. Everyone is going after these people.
To describe the qualities he looks for in candidates, Keith cites this post by Paul Graham, writing about people who are "relentlessly resourceful". He recounts a story about an intern at Square who managed to do something dozens of people at the company had tried and failed at: delivering cold, refreshing, tasty smoothies to the engineers who were working late at night. The task was mundane, but it was complicated to solve because the smoothie shops were all closing before 9PM and the drinks were always arriving late and at room-temperature.
Keith says, start your employees off with a mundane task. For those who thrive at it, increase the complexity and sophistication of tasks over time until they actually show they can't handle something. You'll find talented people in pretty random places and "those are the people you build companies around."
Keith talks about his approach to job search:
Keith and Erik discuss the advice about joining high-growth companies that Sheryl Sandberg has become known for. To paraphrase: When you get an offer to board a rocket ship, don't ask which seat. Keith says, at these high-growth startups there are always new problems and you can't hire people fast enough. There are so many opportunities for career growth. At PayPal, "new things Peter was annoyed with, he'd just throw them at me."
The hard part though is picking a rocket ship. Keith says it's not a good career decision to try to pick a rocket ship before it's a rocket ship. The best investors in the world at that stage are right only 30% of the time. That's a bad approach for most people deciding where to work. If you want to work at an early-stage firm, Keith says, pick based on your boss. Ask whether this is someone you can learn from and assume that your equity value is zero ("like literally zero").
Keith explains what his new company, OpenDoor, is all about:
OpenDoor is a company which let's you sell your home online instantly. If you input your address, his company will buy your home, sight unseen in about three days. This contrasts with 84-90 days for most people to sell their home in the US. Today 20-25% of people who list their homes end up not selling at all. The site is live in Phoenix and is hiring for data, engineering, and business roles.
Keith says the residential real-estate market is still mostly unaffected by technology. Zillow has had only a small effect. Trillions of dollars exchange hands in residential real-estate and 5.75 million Americans sell their home each year. Real-estate agents earn $76 billon in fees in the US. The market is massive and illiquid and painful. "Anytime you can add liquidity to a massive market, it's quite successful."
Some final thoughts:
On tech becoming more mainstream: It's probably better that people spend their time in software fields than financial services or on Wall St., for example. However, the celebrity culture and the Hollywood version of startups, "where everything's going to be easy and we're going to make a lot of money", attracts people with the wrong motives. "This stuff's actually really really painful."
On tech in government and education: There are long arcs of history working in this direction. Tech and innovation will eventually apply in these areas, but there is a lag. You can't accelerate that necessarily. Education will be radically different. Credentials matter more than people realize. It's not just substance. Most companies attack only one side. Some company will cleverly fuse together a new credential that conveys certain things with the reality that this person is smarter, more thoughtful, or more capable. Many people go to Stanford for the perception, not what they learn there.
On investing: Keith is a "bottom-up investor." He wants entrepreneurs to explain the future to him, so he can decide whether he believes that to be a plausible outcome. Others (including his partners) have stronger views of the future and then go looking for consistent opportunities.
On trends: Math will replace humans in every field. OpenDoor replaces your view or your real-estate agent's view of what your home is worth with an algorithm. Doctors are just as fallible as every other human expert. In the future you won't get a diagnosis by talking to a doctor. A machine will calculate what you should do. Keith has an investment along these lines in the legal profession. One day algorithms might start drafting legal briefs for you based on what rebuttal is most likely to be successful. Humans will do 1% of the work and machines and math will do 99% in all parts of society. Self-driving cars are the most obvious example. On what to do about the corresponding decline in jobs, Keith is unsure. He cites this post by Sam Altman about how technology tends to impact jobs and what can be done.