James Lindenbaum learned his lessons the hard way. When he co-founded Heroku in 2007, life wasn’t nearly as easy as it is now for startups targeting application developers as their end-users. He and his team had no real choice but to host their application platform on Amazon Web Services and to learn the ins and outs of that cloud service as if they were real, dyed-in-the-wool systems engineers. Only, they weren’t.
They were app developers; the transformation into systems engineers was just a necessity of growing the business. A website that looked pretty and worked smoothly was little more than veneer if the real product — the AWS virtual servers running Heroku’s customers’ applications under the covers of the hip samurai-themed web service — didn’t work. At one point, Lindenbaum joked during a recent call, Heroku had to do a UX refresh “and we literally didn’t have anyone in the company who could build web apps anymore.”
After leaving Heroku and ultimately its parent company Salesforce.com (s crm), Lindenbaum is back in the public eye with a new effort called Heavybit Industries that aims to save other startups from Heroku’s early growing pains. The idea came after so many fledgling companies came to Heroku looking for help, and after Lindenbaum got personally involved with some as an adviser or investor. He eventually realized that there’s a lot of institutional knowledge out there about how to build business that serve developers, but there’s also a lot of duplicated effort because the people starting these businesses often don’t know their peers exist, much less what they’re up to.
How Heavybit works
Heavybit is like a co-working-space-meets-incubator-meets fraternity, and Lindenbaum has recruited some of the biggest names in venture capital, application development and developer-focused startups to make sure Heavybit delivers on its promise. It works like this: Companies that have raised some money, gained some traction among developers, and now have to deal with the difficult problems of scaling their code or monetizing their business come to Heavybit. Once accepted, they’re in the “active period” for nine months, which includes a curriculum of weekly talks on technology or entrepreneurship; office hours with experts and investors; and meeting/work space in a building in San Francisco’s SOMA neighborhood.
Membership in Heavybit is in exchange for equity, and the community has space for between 10 to 15 companies at a time, Lindenbaum said (although it’s really a “membership for life” situation). The first batch of startups includes some more-established ones — Zencoder, Stripe, PagerDuty and Meteor — that will serve as mentors as well as receive mentorship. Other inaugural Heavybit members include Treasure Data, Kodowa, Iron.io, CircleCi, CloudConnect, Keen IO, Codenvy and Backlift.
Early Heavybit expert advisers include Derek Collison (Apcera/VMware/Google), Adam Gross (CloudConnect/Salesforce.com), Jesse Robbins (Opscode/Amazon), Javier Soltero (Acompli/VMware/Hyperic), and Lindenbaum’s Heroku co-founders Adam Wiggins and Orion Henry. Among Heavybit’s early investor partners are Ping Li (Accel Partners), Chris Sacca (Lowercase Capital), John Connors (Ignition Partners) and Matt Ocko (Data Collective).
Bringing “bad-ass” engineers back from mobile apps
One of Lindenbaum’s goals when putting together the advisers — and one of his continuing goals with Heavybit — is to to put member companies in touch with people who really understand the business and architectural complexities of distributed, multitenant applications. “The reason [these developer-focused] products are so great is because they’re built … by app developers for app developers,” Lindenbaum said. But, like the Heroku team early on, the founders aren’t always skilled in building systems designed to scale.
“The scale curve is much steeper for these companies,” he explained, because the way it usually works is these businesses attract customers who also have their own customers to serve. So rather than handling data for one company, they might be handling data for 20 of that company’s clients, as well.
In order to put his companies in touch with the best of the best of distributed systems engineers, though, Lindenbaum first has to walk those guys back from the Instagram edge. Once you prove yourself at a place like Google or Facebook, Lindenbaum said, “[E]veryone thinks you’re a bad ass. As soon as you say you’re going to build a thing, VCs line up to give you money.”
Unfortunately, he added, many of these people are chasing the past rather than future, trying to cash in their lottery tickets on building the next photo-sharing app rather than on the hairy enterprise-grade systems problems where their skills would really be valuable.
But he has a plan to bring them back to the enterprise side. He intends to push the message of how hard these problems are and how much the cloud services and developer-focused products industries are becoming analogous to traditional heavy industries in terms of the complex but mature supply chains involved.
Essentially, he wants to do the “Got Milk?” ads for the cloud services industry, educating the market so the individual companies don’t have to. It’ll be like a trade association, he joked, “only the non-evil version of that.” Once you shine a light on the difficulty of the problems, Lindenbaum said, the really good engineers come running back to solve them.