On The Road to $10 Billion: 10 Key Lessons from Our Investment in Fastly
By Alexander Rosen and Gil Penchina, Ridge Ventures
In the beginning of 2010, a mere ten years ago, $1 billion valuations used to be exceptional, and $10 billion was, well…really exceptional — only seven public software companies were valued above this level: Microsoft, Oracle, SAP, Adobe, CA, VMWare, and Norton. There are now 452 private unicorns (not so special), but only 51 public software companies worth over $10 billion, plus four private ones. Still pretty exceptional!
While software has clearly eaten the world and winners have become huge in the process, there is still a big leap to get to $10 billion. This past week, Fastly reached a $10 billion market cap which makes it a member of that esteemed club. It has been a great ride, and we have learned a ton along the way (note: we are still personal shareholders and are not making any buy/sell recommendations). Gil and I have been involved since the early days. Gil helped CEO and Founder Artur Bergman start the business, wrote the first check into the company, and helped recruit initial team members. I led Ridge’s early investment in Fastly, and could have invested even earlier but thought the valuation was too high given gross margins at the time. (A mistake I swore never to repeat.)
Here’s what we’ve learned on this journey:
- 👩🏽💻 Follow Developers: Developers are the best early buyers and a key to early groundswell adoption. Developers loved Fastly’s product because every element was built with APIs, so it performed and felt like “your own CDN.” Fastly allowed developers to perform multiple tasks like instant purges and custom cache definitions by using Varnish, an open-source web accelerator. While Fastly “forked” its own version of Varnish, it continued to contribute to the project which also appealed to developers.
- 💵 Focus on Paying Customers: Starting from day one, Fastly focused on paying customers and offered no free trials. To encourage adoption, they offered the ability to use servers without even a credit card to start. Customers were then billed at the end of the month.
- 🧩 Embrace Complexity to be Enterprise-Ready: To get big, you need an enterprise sales force and $1M+ deals. Fastly transitioned early to enterprise selling to build a big business. Just like Twilio and many similar tool providers, Fastly closed its first million-dollar deal in year two. The product was largely “enterprise-ready” from the beginning, and Fastly’s team was correct in building custom features for early big customers, like control pane for Twitter. Most were features on a multi-year potential roadmap that got pulled forward to secure a deal sooner. For enterprise you can’t have a simple one-size-fits-all product; you need to embrace more complexity early.
- ⚙️ Build Hard Tech with Real Business Benefits: Many companies tried to compete with Akamai: Limelight, Cotendo, Instart, and countless others. Unlike prior efforts, Fastly built sophisticated technology that was not obvious from the outside but resulted in, well, faster performance and big cost savings. Instead of building lots of small POPs around the world, as was the CDN industry practice at the time, Fastly built a few extremely powerful POPs (even today Fastly only has 72 around the world) which allowed for really high cache ratios and minimal traffic to the origin server. Fastly optimized the full stack of transit, hardware and software, from the ground up, challenging many conventional assumptions. Instead of using expensive border routers with Internet routing tables, they wrote their own distributed routing agent and used cheap switches. This enabled them to achieve significant cost savings and control content delivery paths. They also built their own file system to reduce wear and tear on SSDs and deliver even greater performance. As an additional benefit, it is significantly easier to recruit talented developers when you’re building something innovative and have a technology thought leader like Artur.
- 🌊 Ride the New Technology Wave: Always curious and learning, the team knew file formats were changing and required different methods of transport. SSD usage and decreasing costs allowed an insanely unprecedented level of caching, if architected accordingly. Mobile dramatically changed POPs and CDNs requirements. Live streaming and social networks required a different approach to caching altogether. To be competitive, applications started demanding not just data but also computing at the edge. Always have your team checking the pulse on technology development and trends.
- 🏳️🌈 Recruit a Diverse and Distributed Team: Artur was a brilliant technical leader who engineered his way to becoming a great CEO. To fortify his team, he recruited not one but five former CEOs to his executive staff. At no time was he threatened by hiring people more qualified than he was. He knew talent mattered. Artur also understood the power of diversity early on. His executive team was split between men and women and over 50% of the engineering leads self-identified as women, people of color, or LGBTQ. The company continues to be committed to I&D. Further, Artur saw no problem building a distributed team from the start and didn’t believe that great talent could only be found in the Bay Area. In fact, he couldn’t find some of the talent he wanted here in the Bay Area. Fastly’s current CEO and long-time President, Joshua Bixby, is based in Vancouver; Tom Daly, early head of infrastructure, is in New Hampshire; Wolfgang Maasberg, EVP Global Sales and Field Operations, is based out of Austin. And the list goes on.
- 🌎 Pick a Big TAM: CDN and streaming presented a huge existing market with tons of adjacent markets. The core media delivery and web application acceleration market is close to $20 billion. Adjacent markets like Web Application Firewall, DDOS Prevention, Bot Detection, and Edge Computing represent another $15-$20 billion. Billion-dollar revenue businesses manifest in markets spending multiples of that.
- 😈 Compete with a Loathed Virtual Monopolist: In diligence we were shocked to hear how many customers hated Akamai, who took advantage of its (then) dominance with high prices, inflexible contracts, slow innovation, and poor service (but a great sales force). Those customers were begging to find an alternative. A market with a virtual monopolist that customers hate is a great market to go after.
- 💡 Capital Efficiency: Fastly only raised $220 million before going public—impressive for a business that requires a significant investment in physical infrastructure, with capital expenditures around $20 million per year leading up to the IPO. And last year’s data shows that companies performing above IPO raise, on average, 73% less capital than companies performing below IPO price (averaging $209 million, almost exactly Fastly’s level).
- 🧘🏻♂️ Don’t Worry About the Hype: For a long time, Fastly’s technology differentiation was poorly understood, as the company stayed largely under the radar. At times my friends at Cloudflare (also a great company) would refer to Fastly in diminutive terms. We never quite understood that. More importantly, the Fastly team never cared about external perception, and quietly built an awesome business, which the market eventually rewarded.
Some trends you can’t predict: sometimes a catastrophe turbocharges trends like remote work and digital transformation, and business takes off even faster, as it has recently. Seeing what Artur Bergman, an immigrant and high school dropout, built at Fastly is a lesson for both entrepreneurs and VCs.
