WE SAW IT EARLY
“If you see a constellation of new companies, tackling largely the same problem, being birthed AND you can’t understand why they are growing so popular, then chances are a new category is being born.” –Bob Kagle, Benchmark’s co-founder.
In early 2005 as an EIR at Benchmark I was scouring a list of the top #100-500 fastest-growing internet websites and methodically using every one of them. One trend stuck out: about half a dozen short-form user-uploaded video sites were growing extraordinarily fast and I couldn't understand why. “You've probably found a new and important category of company,” Bob Kagle told me. “‘Important’ because if consumers are flocking to them all at a super-fast rate, then these companies are by definition offering something of value that the consumers can't get somewhere else. And ‘New’ because it's different enough from things you've seen in the past that you can't easily analogize from what you know before.” He was spot on. Over the next 18 months, YouTube was born.
Roughly at the same time, a similar phenomenon happened to birth another category: digital worlds. Companies like Cyworld in Korea, Mobage (DeNA) in Japan, Habbo Hotel in Finland, and Second Life in the US all started and all of them were growing extremely fast with zero marketing. In 2005, Cyworld was used by virtually every South Korean in their '20s. DeNA launched the world's first $100M+ social game on Mobage in 2006. It was exceedingly difficult for outsiders to understand why they were gaining traction without using a strained analogy (World of Warcraft meets MySpace) to help contextualize. Were they social networks with games and embedded micropayments? Were they “sandbox-style” multiplayer role playing games? What distinguished them was:
A digital/virtual identity which was not necessarily tied to one's real world identity.
A social network where you could make new friends and hang out with existing friends.
Something fun to do; you didn’t simply look at or comment on someone's homepage/feed/video.Radically richer ways of self-expression than were available before, including the ability to buy virtual items for your avatar to either look cool or to power up in a game.
A metaphorical sense of a place where you were doing this expression and social interaction.
It wasn't a ‘webpage' but a 'room' or a 'home' in which you were doing it all.
Most called these new sites "virtual worlds” or 'The Metaverse,' a term popularized by Neil Stephenson in his profoundly prophetic 1992 book, Snow Crash. I labelled the category "online hangouts" because they were the equivalent of the last generation's mall or downtown: a place where you hang out with your friends. Sure, you might say you're going downtown to shop but on a deeper level the stores and crowds are just an excuse to have entertaining stuff to riff off while being with friends.
In their own ways, Snow Crash and Burning Man and Nick Bostrom's virtual world argument and MySpace and all MMOs pointed to something potentially huge–new and important on a scale that could dwarf MySpace. I jumped into one of the early players, GaiaOnline, as CEO in 2006. Over the next couple years, Gaia grew from $2M to $24M in sales and profitable, from a couple hundred thousand to 8M monthly uniques and pioneered to US teens virtual goods, prepaid cards and virtual movie theaters where you could co-watch YouTube videos, Sony, and WB movies with your friends.
In early 2008, the Benchmark partners asked me to meet Roblox founder David Baszucki and give them my take. I’d met all the other players in the space but in that meeting, I realized David might ultimately win the race, as he had the best and most practical vision for what a compelling Metaverse could be. At a speech at the IVS startup conference in Sapporo, Japan, in May 2009, I was asked what I thought was the most promising startup I'd ever seen: "Roblox," I said. Later that year I told Dave that my company, Gaia Online, would love to buy his company. We were a magnitude larger than Roblox at the time; Roblox wasn't gaining much traction yet; and I offered $20M of our $25M in cash. Dave declined politely. He wanted to give his vision a go. He believed. He was utterly correct in hindsight. Gaia and Habbo and Second Life all stopped growing substantially in 2010; Mobage and CyWorld lasted a bit longer. But like AltaVista was to Google, all of these precursor companies are now footnotes in Roblox's history.
In March 2011, I joined Meritech Capital and David was the very first CEO I introduced to my partners. Our team bought into David’s vision, but Roblox was far too early for Meritech's growth stage investment model. Truthfully it was still too hard for an amateur to make a great game on Roblox. But my partner Rob Ward and I flagged Roblox as one of the highest potential startups we'd seen and we met with David regularly over the next 5 years: we introduced job candidates, gave financial and strategic feedback and built a trusting friendship. We tried our best to leverage our learnings from both a previous gaming investment, PopCap, and a social networking investment, Facebook. Despite relatively low growth and little-to-no peer and Silicon Valley recognition, the Roblox team held steadfast and kept building better and better infrastructure and tools for the game developers and creators.
In the second half of 2015, nine years after Roblox’s inception, their heretofore-modest growth rate started to accelerate. Roblox is extremely complex: the physics engine, the marketplace, the server infrastructure, the studio tools, and more–and it took time for all of it to work well and to work well at massive scale. Roblox ended 2015 with $53M in bookings. They budgeted $67M in 2016 bookings but by April realized they’d beat that handily. We preemptively offered to invest in July but Dave wasn’t quite ready. In September, they upped their forecast for the year to $102M and David invited us to lead a round if we could meet his goal of a $500M pre-money valuation. Despite the momentum, many investors believed that Roblox’s addressable market was limited and its user retention rate was too low. We looked at the data with a different lens. My partner Max Motschwiller, who co-led the investment with Rob and me, had a novel insight, one that helped us see things differently from other investors and even people at Roblox. We felt comfortable writing the largest check in Meritech history. But we couldn’t speak for the whole round so David invited Neil Rimer at Index Ventures to split it with us. The company ended 2016 on a tear, doing $120M in bookings, and the flywheel just keeps spinning…
Why is Roblox so successful?
High Level, there are Two Fundamental Reasons for Roblox’s success:
Roblox boasts two network effects. All great marketplace businesses have one powerful network effect (think Facebook, YouTube, DoorDash and AirBnB). Roblox has two:
First: More (and better) games → drives more users → drives more (and better) games. This is just like YouTube (but far harder to replicate because it's always been easy to upload videos to YouTube while it's far harder to create software to allow a non-coder to build a game).
Second: The more friends on Roblox at any given time makes it more fun to be on Roblox so more users → leads to a better experience → attracts more users. This is just like Facebook, Instagram, and Snapchat. A gaming (or more precisely: "entertainment co-experience") platform + a social network.
Roblox rides two unstoppable, ever-growing cultural trends:
Digital Worlds: Ever-more powerful devices + faster internet speeds mean people spend more time online and value their digital life more and more. People want ever-more entertainment and more social connection. Games provide the former; Snap and Zoom the latter. Roblox is one of the few places that provide both. Parents, perceiving ever more risks in the real world, have come to restrict their children's physical freedom. 50 years ago, it was more common for young people to hang out outside and explore their neighborhood and beyond with their friends unsupervised until dinner. Roblox is not only becoming the most fun place to hang out with your friends–it's becoming the safest.
User Generated Content: As software enabled amateurs to make good-enough content, users fueled the creation of most of the entertainment in the last decade: YouTube made it easy to create videos; then MySpace, Facebook, Instagram and Snap made it easy to post ever-more entertaining (and easier-to-make) updates; Apple and Google for podcasts, and more. Today, Roblox provides an unparalleled toolset for a person to create something awesome and share it with friends
But there’s more to it than that. Let’s dig deeper:
A Game-Centric Metaverse
Hangouts, online or off, need to be fun. Most early attempts at a Metaverse, including the one I ran, were just virtual spaces that were just not fun enough for most people. On the other hand, because Roblox started with games, it's like Disney World. You go with your friends or family for the day to go on the rides. The rides are awesome in and of themselves.
But It’s A Metaverse First, Not Just Another Gaming Platform
Disney recognized those very rides were not just awesome fun–they also served as the vehicle through which you hang out with and 'play' with your friends and family. Disneyworld’s characters, castle, music and more, create a space far more rich than the sum of the rides. And Roblox, for all its game-centricity, is not a game. It's far more ambitious than that. David really does want to build the Metaverse - a Third Place (vs. home or school). This is a place where, by creating an infinitely customizable avatar, you can safely be who you really are, or be who you really want to be; you can create nearly anything you can imagine. You can express yourselves more fully than you can in the real world through your creations and your behavior. Roblox is a place where you can find others with shared interests, and then share tens of thousands of experiences. You can hang out with your best friend, no matter what time of day, regardless of physical distance. If you want, you can make an experience and sell items and make a living in Roblox. Or you can simply interact with others' experiences. You can try new things. And you can be safe. Roblox calls that Human Co-Experience. Neil Stephenson in Snow Crash and Ernest Cline in Ready Player One gave us examples of what this might look like. Dave is showing us a third.
User-Generated Architecture vs. Traditional Measures of Graphic Fidelity
For years, most gaming execs, gaming investors, and gamers didn't get Roblox because the games looked amateur. Things have changed, but still most people underestimate it. A big reason is that the games themselves are still primitive on a few dimensions vs. Triple-A games that generate reliable billion-dollar annual franchises: first, the graphic fidelity is relatively low; and second, most games are not deep. Both dimensions have improved dramatically in the last 2 years. And they will continue to improve for at least the next decade.
There's a good reason that many Roblox Games have lower graphic fidelity than games on other platforms. Roblox's north star is to make it easy for a non-programmer with little or no funding to build a creative game and that the game can be played across nearly any device including old phones. To do that, Roblox comes at everything from a different angle than a well-funded game studio. Imagine a game where your character gets to run around a large wall into a castle. What's the nature of that wall? In a Triple-A game made by a well-funded game studio, some team designed the layout of the map and the castle. Another group, artists, "paint" that wall to look near photo-realistic. Another group, developers, design the location-boundaries of the wall such that a character cannot run-through it. All this is done over many months and costs lots of money. In contrast, at Roblox, a kid just builds the wall like one would in the real world: she takes bricks and puts them on top of each other to make a wall.
These two different approaches lead to large differences. If anyone at the game studio decides last minute that they want to move the wall, they must go back to the drawing board. In contrast, if the creator wants to move her wall in Roblox she can grab it and shift it in seconds. This is profoundly empowering. She doesn't have to plan a year in advance, and mistakes are easily fixed. Making a game is easier and faster by an order of 2 magnitudes. On the other hand, for the players, the resultant wall is not as photo-realistic as a AAA game. Roblox made a conscious choice for over a decade to focus on empowering creators - it took the far harder route. Recently, Roblox started dedicating resources to higher fidelity too, and it's starting to pay off:
TYPICALROBLOX GAME in 2008
A 2019 ROBLOX GAME
Another limiter-to-date for Roblox games is their relative lack of depth/complexity vs. Triple-A games. That is changing as DAUs increase and the payoff for a developer to build bigger, deeper games grows. Also, true to Roblox's North Star, the company is investing in tools to make it ever easier for groups of developers to work together on a game.
Unsurprisingly, as games get more graphically rich and complex, older users are flocking to the platform. Whereas Roblox used to be more popular among younger kids, users over 13 are now growing at a faster rate than under 13, thanks to these two foundational improvements. The growth rate of users over 17 is fastest of all Roblox user cohorts.
Brilliant Team Focused on 20-Year Building Blocks
David is not only brilliant, but unconventional in two important ways. First, he architects systems vs. assigning projects. Building the Metaverse is extraordinarily complex. Few CEOs could conceptualize, much less manage, this complexity. This is Dave’s superpower: like the founding fathers of the US, Dave is the ‘ultimate systems architect.’ I’d like to say we at Meritech invested because we’d ‘figured this out’ but the truth is that he has surprised us over and over with his sheer intellect and systems-related rigor. For example, today there are over 40 autonomous product-development teams working on different elements of the ‘Roblox 'system.’ And they are growing more efficient, not less, as Roblox scales. And that number will grow.
Second, David prioritizes multi-year plans vs. quarterly objectives. He eschews the short-term optimization option in favor of the long-term foundation-strengthening option every time. It manifests in dozens of ways including building an employee hiring ‘system’ to identify and evaluate talent vs. hiring more recruiters, and creating incentive structures that encourage creators to make games that delight users vs. making a quick buck. I’ve never worked with a CEO who more consistently passes on short-term opportunities in favor of users’ long-term experience; who will build the best software or organizational solution even if it takes 3 times longer to build than the quick-fix.
Dave also has a huge heart and that, coupled with his brilliance, makes for a magnetic leader and company culture. In the last 5 years, David has leveraged that to augment himself and his core team dramatically. In fact, the only company in our portfolio we’ve seen hire such high quality people across so many hundreds of positions is Facebook. Dave's built a strong council of advisors, too.*
It’s StIll Early Days
In most entertainment categories, there’s a power-law and only a few properties “win.” If there's a social network element, the reason is obvious–if all your friends are in one place, you hang out there. That’s why it’s very hard to unseat Facebook or Instagram. And Roblox is just scratching the surface of what it could ultimately become. It's easy to imagine the number of users scaling by 3-5x. As Roblox experiences look better and get more complex, older teens and people in their twenties and thirties will use Roblox more. And the company will attract more users of all ages in Europe and Asia. Similarly, tripling the revenue per user looks straightforward by getting some more of its users to pay a bit, without the aggressive tactics used by most game companies. Unlike Zynga and other high-monetization focused game companies that depend on "whales," Roblox's bookings are more evenly distributed across paying users. Most of this will happen naturally as Roblox ages-up.
But some growth trajectories take imagination. Today, virtually everything on Roblox is a game. But that will change over time. Lil Nas X hosted a concert in Roblox in December, and millions came to it. Once video is made more available inside Roblox, people will watch a TV show or YouTube video with their friends. Once voice is added (today everyone talks through text chat), people can hold business meetings, stage plays, convene book groups, offer talk therapy, and learn languages. Once notifications get added, people will opt for one-off events. Soon you’ll be able to buy name-brand clothes (virtual and real). Games will likely always be the core of Roblox, and they'll always beat (cooler-looking) games off-of-Roblox because they come with an identity and social layer. But Roblox has from the start been about something much, much bigger than games. Tellingly, inside the company people talk about “experiences” not “games.” It’s just a matter of time before we all do.
If you’re open minded to this last assertion, then you have a taste for the true scope of the Roblox opportunity. 10 years ago, people underestimated Roblox because they couldn’t imagine what it could become. We suspect there are many people who will do the same today, thinking it’s merely the next-generation Gaming Platform–a “YouTube for Games.”
Of course we could be wrong (it happens quite frequently!) about the Metaverse and about Roblox being the one to create it. Dave Baszucki has proven prescient so far, but we can’t predict the future. To be sure, for Roblox to fulfill its promise it must remain safe for its users. Roblox also must remain nimble. And there is always the possibility that someone takes over the space from an orthogonal direction–perhaps a game or social app that adds the other Metaverse elements in just the right way and 'voila!', just strikes lightning in a bottle. There’s no doubt that the Metaverse itself is a bold, ambitious idea… but we think it's inevitable. It's only a question of how soon and whether there’s one or many. Dave Baszucki and Roblox is our bet to build it.
*Anthony Lee at Altos Ventures stands out. He believed in and supported Dave through the challenging years before the Roblox network-effect flywheel started to accelerate and has earned the role as Dave’s most trusted advisor. I’ve learned from him.