Greylock’s Andy Johns on Web Growth: “You’re Training a Mentality”

As an authority on getting consumer Web companies to soar along the elusive curve of exponential growth, Andy Johns has succinctly epitomized his expertise in his Twitter handle, “@ibringtraffic.”

Last month Johns joined Menlo Park, CA-based Greylock Partners as the venture firm’s first “growth strategist in residence,” a title bestowed on the basis of his recent track record. According to his LinkedIn profile, Johns helped engineer a four-fold increase in traffic over one year at Quora; worked on Twitter user growth and engagement when the number of active users went from 40 million to 140 million; and helped lead Internet marketing for two years at Facebook, when the number of active users went from 100 million to more than 500 million.

To Johns, software has become the next industrial revolution. As he puts it, it’s about “taking code that allows a 14-year-old kid to create an app that can reach a couple million people in just a few weeks.”

Andy Johns

Andy Johns

With the help of San Diego Tech Founders, Xconomy invited Johns to elucidate the mysteries of online growth for the emerging cluster of Web entrepreneurs, investors, and CEOs in San Diego. He met with two dozen local tech leaders for a luncheon discussion organized by Xconomy Wednesday, and later spoke to about 170 people at a Tech Founders meetup.

Driving Web traffic is no simple matter. According to Johns, it means absorbing a Web startup’s core values—what he calls “the founders’ DNA”—while injecting some key business fundamentals to accelerate growth. At the same time, he talked about the seemingly contrary necessity of cutting through a company’s internal cultural resistance to change, especially as a growth team “begins to change the DNA fundamentals of a company.”

“You’re training a mentality,” Johns said. “You’re training a group of people to think in a particular way, and then you’re giving them the power to make changes.”

The discipline that Johns outlined is a wholly data-driven process. My colleague Wade Roush wrote about Facebook’s fixation on data earlier this year, and Johns said the time he spent at Facebook “was the best two years of business training I could have asked for.”

Johns said driving Facebook-style growth is a process that requires continual experimentation and a willingness to take risks. “Growth begins with the founders,” he said. “To go from being a college directory to the social fabric of the Internet, you have to take risks.”

So when do you start building a growth team?

“Building that knowledge should start from day one,” Johns said. When he left Quora earlier this year, Johns said eight of the company’s 50 employees worked on growth. At Facebook, the growth team had about 45 people when he left in 2010.

When Johns left Facebook to joinTwitter, he said Twitter was not a data-driven company. At that time, Twitter’s headcount was soaring, and the company was caught up in a discussion of new features that could be included on Twitter’s website. Many of the new hires wanted to contribute, “and the way they wanted to contribute was to build more features, which is bullshit,” Johns said. “They wanted to treat the signup page like YouTube.”

Johns viewed additional features as a distraction. He said Twitter’s greatest value is enabling users to create a personalized stream of information, and he wanted new users to be able to sign up quickly. So Johns devised an experiment that offered an alternate sign-up page to 10 percent of Twitter’s first-time visitors.

“It emphasized sign-ups,” Johns said. “It was either sign in and give me your e-mail address or get the fuck out.”

He ran the experiment for two days, and the alternate sign-up page increased Twitter’s registrations by 60,000 users a day. “It was the single biggest step change in the growth of Twitter,” Johns said, adding he implemented the changes immediately.

It also serves as an example of the type of experiments that Johns continually runs to measure new user registrations, user engagements, test new variables, and optimize growth. He said Web-based companies don’t need to run cohort-style analyses over two to three months. Merely diverting 1 percent of the new visitor traffic for a few days—assuming that site traffic already is a big number—can generate statistically significant results, he said.

Driving growth can be an intrusive process early on, Johns added. If a team has developed a new product or a new set of features, the growth team assesses how the new product would affect overall growth. Would it actually, meaningfully change anything?

For example, Johns said user engagement begins to suffer if a new feature takes more than a few milliseconds to load. “You have to respect the people, but you can’t give terrible ideas or opinions a long shelf life,” he said. At the same time, software developers have to embrace the fundamental concept that growth is their mission too, and to accept the fact that numbers rule.

“If our rate of following is down 30 percent, I don’t care if TechCrunch thinks some new feature is great,” Johns said. When it comes to driving Web traffic, he added, “TechCrunch can’t see what we’re doing. They don’t know.”

As the imperative for growth becomes more pervasive within a company, Johns said developers begin to coordinate their efforts with the growth team.

The challenge for a growth team, Johns said, is akin to maximizing the yield on an investment. “If I give you $100,000 to invest and get a 10 percent return, I’m pretty stoked,” he said. “That’s what growth is when you have a good product.”

The mission of a growth team, then, is to drive a company’s internal expansion by maximizing the increase in new users (and in reactivating dormant users) while minimizing the rate of user deactivation. It is a bit like modern portfolio management theory, Johns said.

“If your rate of growth is 1 percent a week, you won’t get [venture] funding,” Johns said. “You’ll only get funding if your net organic growth rate meets or exceeds 3 percent a week.”

Most of the Web companies out there are not even close to that kind of growth, he acknowledged.

Our luncheon discussion, part of a continuing conversation about the tech startup community in San Diego, was presented by Xconomy and San Diego Tech Founders. It was sponsored by the Cooley law firm, EY, the Kauffman Foundation, and the San Diego Regional Economic Development Corp.

In attendance were Mike Alfred of Brightscope, David Becerra of Mellmo/Roambi, Ryan Bettencourt of Blurtopia, Wolf Bielas of Wolfpack Ventures, Reid Carr of Red Door Interactive, Tom Clancy of Tao Venture Partners, Brant Cooper of San Diego Tech Founders (and co-author of The Lean Entrepreneur), Mark Fidelman of Evolve, Steve Gal of the UCSD Rady School of Business, Melani Hunter of Bevato, Austin Neudecker of Yealthy, Michel Kripalani of OceanHouse Media, Allison Long of Crescent Ridge Partners, Chuck Longanecker of Digital-Telepathy, Micha Mikailian of eBoost Consulting, Nik Souris of The Backplane, Jonathan Sills of Battery Ventures, Joaquin Silva of On-Ramp Wireless, Arie Trouw of Sambreel, Mitch Thrower of Bump Network, Matt Wickstrand of, Doug Winter of Seismic. Also in attendance were sponsors Eric Otterson and Steve Przesmicki of the Cooley law firm, EY partners Doug Regnier and Tim Holl, and Matt Sanford of the San Diego Regional EDC.

Bruce V. Bigelow is the editor of Xconomy San Diego. You can e-mail him at or call (619) 669-8788 Follow @bvbigelow

Trending on Xconomy