Watching the birth of Flickr co-founder's gaming start-up
Stewart Butterfield and his friends are back at it with a new company. CNET's Daniel Terdiman was given exclusive, behind-the-scenes access as they built it from scratch.
SAN FRANCISCO--Stewart Butterfield and his business partner Cal Henderson stared at the MacBook Pro in front of them.
For nearly a year, they'd been struggling to figure out what to call the game their start-up was building. Any time a team member loaded a working version, they'd sit through a few seconds of a splash screen with nothing on it but a generic title featuring little more than the name and logo of their company.
But now, the group had finally given their baby an official moniker: Glitch. And this was one of the first times the two had sat through the splash screen since plunking down a low-five-figure sum to buy glitch.com.
Butterfield and Henderson, dressed casually, were hovering over the computer in the bright, east-facing front room in a beautiful Victorian vacation rental that they'd been using for a four-day company off-site in mid-January. Everyone else had already left. Energized from an intense four days of brainstorming (and maybe a coffee run to a local hot spot called the Mercury Cafe) they were running a demo of their game. Watching the bland screen load as they had countless times before, Henderson's eyes lit up.
"I guess we could replace that with the title of the game now, couldn't we," Henderson deadpanned. "Yeah," said Butterfield. "We don't have to call it 'The Game Being made by Tiny Speck' anymore."
Indeed, on Tuesday, CNET is reporting exclusively, Tiny Speck plans to officially unveil their new game, Glitch, and for the first time, people will be able to see what this start-up has been working on since last March. The game will be in private alpha for now, and is expected to launch publicly in the second half of the year.
If Tiny Speck doesn't sound familiar, that's because the company has been under the radar since its founding. But if Butterfield's name rings a bell, it should. He and his former wife, Caterina Fake, were the co-founders of Flickr, the hugely popular photo-sharing site they eventually sold to Yahoo for a reported $35 million in 2005.
By 2008, Butterfield and the other original pre-acquisition Flickr team members who had joined Yahoo celebrated "Vestfest," a shindig honoring the full vesting of their windfall from the sale, and he . His resignation letter was true to form for the 36-year-old geek with a Cambridge degree in philosophy who grew up in a Canadian hamlet once popular with well-educated Vietnam draft dodgers.
"As you know, tin is in my blood," the cheeky letter to his boss began. "For generations my family has worked with this most useful of metals. When I joined Yahoo back in '21, it was a sheet-tin concern of great momentum, growth, and innovation. I knew it was the place for me....Please accept my resignation....I don't need no fancy parties or gold watches (I still have the one from '61 and '76). I will be spending more time with my family, tending to my small but growing alpaca herd and, of course, getting back to working with tin, my first love."
Today, Butterfield explains that the letter was little more than a way to enliven the mundane lives of Yahoo's human resources department, and contained no metaphors. Still, reading the letter now, it's obvious that "tin" refers to Web-based entrepreneurship.
Indeed, despite his Master's in philosophy and an original plan to get a Ph.D., Butterfield was lured away from a life of academia by the excitement of the late-1990s dot-com boom. He consulted for a while, and then helped start, and quickly sold, a small company along the lines of Classmates.com called GradFinder.
True Flickr devotees will remember that before that service blossomed, the team behind it had first been operating as a start-up known as Ludicorp, which was working on an online social game called Game Neverending. Wikipedia defines it as an "atypical role-playing game primarily based on social interaction and object manipulation. [It] was lighthearted and humorous; indeed there was no way to win, nor even any definition of success." Despite developing a passionate audience, they abandoned the game after it became obvious that Flickr was more viable commercially. Now, years later, Butterfield had reassembled a cadre of the earliest members to do what they didn't last time: build a game and a real business around it.
Their rationale? That the lessons they learned about how to build a passionate, actively-involved community while making Flickr a world-famous company, along with their ideas about how to construct a compelling social game, position them uniquely at the potentially highly-lucrative intersection of World of Warcraft and Facebook.
I'd written about Flickr since 2004, so when I ran into Butterfield last March at a soul food restaurant in San Francisco during the, I told him I'd heard about his new start-up and made him an offer I was sure he'd refuse: In return for determining when I could publish, but no other editorial control, he'd give me regular, exclusive access for a behind-the-scenes story about the new company.
Butterfield and his partners had built one of the flagship Web 2.0 companies and had become household names among the geek set. Who wouldn't want the very rare opportunity to see behind the curtain as he and his crew tried to do it again? I know I did.
But companies almost never let the press see the sausage works, so I expected a polite no. Instead, after about a month of intermittent e-mail exchanges discussing the idea, a message popped up on my screen that began simply: "Hi Daniel - let's do it!"
Tiny Speck off and running
In an early back story that Butterfield, Henderson, and fellow co-founders Eric Costello and Serguei Mourachov came up with for their new game, an 11-year-old Japanese girl loved drawing in different styles. But she was never satisfied with her work and would crumple up her compositions and toss them out the window. There, said Butterfield, they would "get sucked up into this wind and each one comes to life (like) tiny specks on motes of dust."
In March 2009, the four founders filed articles of incorporation in Delaware for their new company, Tiny Speck. The state of Delaware thought they were called Tiny Spec, and their checks read "Tiny Specks," but no matter. They were off and running.
From their home offices in San Francisco, Vancouver, and New York, Butterfield made the trains run on time; Henderson handled the front and back ends of their fledgling system; Costello built the client; and Mourachov--"the Russian mad scientist guy"--built the game server. (for more on the founders).
Their pitch to venture capitalists was simple: they would do for massively multiplayer online games [MMOs] what Nintendo's Wii did for video game consoles--bring a niche product to the masses. And make a lot of money doing it.
Certainly, they also had what countless entrepreneurs would kill for: a reputation for rewarding investors' faith in them. The team quickly scooped up $1.5 million in seed funding from the A-list VC shop Accel Partners.
Glitch, at its most basic, is a 2D Flash, Web-based, social MMO with a heavy puzzle-solving component. Or, as the team likes to call it, a "collaborative sim." Its back story centers on a great but very dark future, and a group of scientists who discover a path back in time to create the optimistic future everyone wants.
"The whole world was spun out of the imagination of 11 great giants," Butterfield, who often speaks in a soft, hard-to-hear voice, said. "So you have to go back into the past, into the world of the giants' imaginations and grow...the number of things in the world, grow it in terms of physical dimensions, to make sure the future actually happens. So all the game play takes place in the past inside the world of the giants' imagination."
Practically speaking, Glitch is a social game () about learning how to find and nurture resources, identify and build community, and proselytize to those around you. It's not about epic, bloody battles. "Rather than you and me fighting each other with swords," Butterfield explained, "it could be you and me having rival religious factions battling each other for converts."
Along the way, players level up by completing quests, gaining skills, growing all manner of things, and making their way through a sometimes Mario-esque world of different artistic styles, each of which can be thought of as being inside an individual giant's memories.
Even though Glitch is going into private alpha Tuesday, Tiny Speck already has thousands of people in a testing queue, and the company will likely be inviting about 100 people at a time into the game over the next few months. If you're not already on that list, and you want to play, sign up--but be patient.
In the minds of giants
Beginning last May, I met with Butterfield and Henderson every month or two, usually at the 29-year-old British-born Henderson's Hayes Valley loft here in San Francisco. Though he often walks around in T-shirts and shorts and eats Lucky Charms straight out of the box, there's no doubting he's accomplished. Henderson, who literally wrote the book on building scalable Web sites, is best known in geek circles for having designed and built the Flickr APIs, which influenced many developments on the Web, including OAuth.
During my first visit to his light-filled loft, located on the same block as Second Life publisher Linden Lab's first office (complete with a six-foot-tall plastic Ferris wheel, a shelf full of Make magazines, multiple computers, and the obligatory copies of Halo and World of Warcraft), the two partners weren't sure what to call this open-ended fantasia.
Among their considered, but rejected, ideas had been: Paper Moon, based on the Japanese cut-out collage motif of the 11-year-old girl and her crumpled paper, but another game company beat them to the name while Tiny Speck pondered; One Billion Daydreams, which was about the universe inside the mental wanderings of a billion people; and Inside the Minds of Giants.
During an October meeting with Butterfield and Henderson, they told me they'd settled on the latter. But moments into our January meeting in the Victorian vacation rental, Butterfield started things off by saying, over the sounds of traffic from the thoroughfare just outside the window of the front-room salon, "Yeah, so the name of the game is Glitch."
"We were going to call the game Billion Year Itch," continued Butterfield. "And then someone said 'Glitch' and [that] stuck. We just had really positive reaction from everyone we told it to."
On Glitch.com, Tiny Speck spells it out a little more: "It's called Glitch because in the far-distant and totally-perfect future, the world starts becoming less and less probable, things fall apart, the center cannot hold, and there occurs what comes to be called the 'glitch'--a grave danger of disemprobablization."
Tiny Speck's business plan, too, has morphed since its founding. In the beginning, Butterfield wanted to pursue a subscription-only model, but over time, he saw that a free-to-play game would probably attract a bigger audience. Money, then, would come from a combination of virtual item sales, subscriptions offering premium services, and the sale of mini games that unlock new, otherwise inaccessible skills, both for the Web and iPhones and Android devices.
With all those potential sources of revenue, Butterfield believes that Glitch can eventually earn between $30 and $40 per user per year, a range that, if large numbers of people sign up, could mean very significant profit margins.
Even if that's not possible, Tiny Speck's costs are fairly low. It has just eight full-time staffers, including the founders. Though the company will soon open offices in Vancouver and San Francisco, Butterfield said there's enough in the bank to get to launch, hopefully sometime in the second half of this year.
Still, Butterfield expects to secure a $3 million to $5 million A-round of funding in the next few months in hopes of not being distracted by fund-raising while in full production mode.
Ripping out the game mechanic
From its earliest days, the game had a core game mechanic--essentially a system of rules--in which players would initially choose one of five character classes--a dandy, a monk, a mastermind, a huckster, and a gumpteur--and then choose both a primary and secondary talent from among panache, flow, brains, gumption, and wile. A dandy would have panache and flow, while a monk would have flow and brains.
But when I visited with the Tiny Speck team in late October of last year at the tail end of a four-day offsite (see stop-motion video below of the Tiny Speck team's first design charrette--meeting--in Vancouver on April 29, 2009) in the Victorian vacation rental, Butterfield casually dropped a bomb: They were tossing that game mechanic onto the trash heap.
With the house full of engineers and all four of the founders, it was hard to hear as loud voices echoed through narrow halls and rooms with high ceilings. The team had been staying in the house for four days and working non-stop, morning to night, and it had the air of a geek slumber party. The password for the temporary Wi-Fi network they'd set up was "stewartisanass" because, co-founder Eric Costello said with a smile, "Stewart was being an ass the day we set it up."
According to the 39-year-old Costello, who lives and works in the Astoria neighborhood of Queens, N.Y., though the Tiny Speck team had bought into the original mechanic since the beginning, none had been totally happy with it.
"It seemed good enough," Costello, an early blogger and one of the first Ajax scripters, said, "but it was really foreign. Part of the motivation (had been) to make the upfront decision for a player about what type of character they're going to be. But (we decided) to let that come out of how they play the game and the choices they make while they play the game instead of something they choose before they're ever in the world and know anything about it."
From that point on, Tiny Speck began building a new core mechanic centered around skill trees in which players begin the game by choosing from certain low-level talents, such as animal kinship, green thumb, and the like. Then, as they build up skill points, they can spend them on new skills, which in turn open up "a bigger possibility space," Butterfield said.
From the get-go, the Tiny Speck team set out to craft their game using a series of five three-month development cycles, each of which would comprise two months of hard-core work, followed by a month of "cleaning up after ourselves," essentially testing and optimization, and each of which had a set of specific goals and milestones.
Each cycle had a name, and when I first started visiting with Butterfield and Henderson, they had just finished the first, which they called Happy Place. The goals for that three-month period? To get done "the minimum amount of stuff we needed to get done (for the game) to be fun."
The next cycle was to be called Bay of Pigs, and the objectives were to complete the addition of an in-game currency, auctions and stores, groups, object making, and a system for death. Following that was Water World, with mini-games, politics and elections, property taxes, and all things related to water; Macademia, which would augur a private beta; and finally, the "nice and ominous" Jonestown in time for the public launch.
Another big goal, of course, was to staff up. While a core team of four partners was, as Henderson called it, the "absolute minimum needed for developing," it was clear to the founders that they'd need to grow in order to meet their objectives.
Throughout my series of meetings with the team, Butterfield always made it clear that slow and deliberate growth was key. In June of last year, Tiny Speck was still just the four founders plus a small group of contract illustrators--"we don't have an accountant," he said, "instead we have several envelopes stuffed with stuff so when we get an accountant, they'll be busy for months."
But when I stopped by Henderson's place in August, they were romancing their eventual first new hire, then Digg creative director Daniel Burka. It was a secret that Burka was considering leaving Digg for Tiny Speck, though to hear the three of them talking that day, it seemed like a done deal. As Butterfield and Henderson talked to me about the status of the project, they were also filling in Burka on what was going on, and to me, it felt more like an orientation than an interview. If it was an interview, Burka was definitely the one with the upper hand. By October, Burka was on board as the director of design.
Now, Tiny Speck is up to eight full-time people and 12 contractors, including Vicki Wong--who goes by the name Meomi and who designed the mascot for this month's. Next up, said Butterfield, a vice president of operations and four other staffers.
An article of faith
When Flickr was in its earliest days and still independent, the start-up came up with a clever bit of psychological judo as a slogan. "Use Flickr, because why not use Flickr," it went. Last spring, at my first Tiny Speck meeting, Butterfield remembered that and said they might try that again this time around. After all, he said, "it's a very effective slogan. It's very difficult to argue with."
Now, though he chuckled and agreed with himself that it's hard to argue with that reasoning, he said Tiny Speck won't be borrowing that language. Instead, they're going with a much more succinct and to-the-point tag line: A game of giant imagination.
As anyone who's ever tried to start a company knows, it takes a lot of ambition, talent, luck and, yes, imagination, to make it work. But Butterfield and Henderson have never had any doubts that they will once again build a popular and financial success from the ground up.
"I'm just taking it as an article of faith that we'll make money," Butterfield said.
"If you have a good idea and it's well executed," Henderson added, "then you'll make money."
Come back tomorrow for part two of the Tiny Speck story, a closer look at the game's systems and platforms.