SIFTEO Cubes are an award winning interactive gaming platform that respond to your gestures and sense each other. Sifteo got its big break back in 2009 with a TED talk that has received over 1.3 million views. To date, the company has raised $13.5 million in venture financing and launched the second generation of the product last fall. We sat down with Sifteo’s Co-Founder David Merrill to discuss the company’s journey from the MIT Media Lab back in 2006 to today.
I grew up in a family of teachers. My mom was a early childhood development teacher her whole career. My dad also spent a few years as a teacher and most of their friends are teachers. I have a love for explaining things, and tools for learning. Also, my dad was very handy. He was a hands-on, weekend project guy who always be pouring concrete to make new paths around our house or building a new fence. He once constructed a lofted bed in the bedroom that I shared with my brother. He built that from scratch in a single day. Growing up watching him build things gave me a certain fearlessness about building. I grew up playing with construction-oriented toys like Legos, Lincoln logs, tinker toys and a lot of matchbox cars.
There’s a great picture of me in the garage when I was a kid and I was playing around under the table saw where my dad had been sawing and there’s this big pile of sawdust. I carved out little roads for my matchbox cars to make a little neighborhood of sawdust with a network of streets.
I grew up feeling empowered to build things, surrounded by a family who loved to build things and my parents who both had very high expectations for what their kids would do – expectations for achievement. I always felt like I could do what I wanted to do with my life, and that I should achieve. That was just expected. That was the background that formed me.
As an undergrad at Stanford I studied computer science and psychology. I was in a major called Symbolic Systems, which is like human computer interaction, human cognition and computer science and is concerned with how these things fit together. I learned to program but I never even really thought about building interactive physical things until I took a class about musical instrument design.
The class was about building music controllers with electronics and that’s where I got my first taste of building interactive objects with microcontrollers, wireless sensing and embedded computation. I got totally hooked. I loved this idea that I could build tools that extended beyond the confines of my laptop screen into the physical world. To create physical things with behavior was way more fun and exciting.
I took that love of building with me to MIT, where I created more musical instruments and wearable systems for information gathering. It was at the Media Lab where I built the prototype of what has become Sifteo’s product with my friend Jeevan, who is now co-founder of Sifteo. We had known each other ever since we were 19 years old. We were undergrads together and roommates and we were in a rock band together our senior year of college. One day we were sitting in the kitchen at the MIT Media Lab thinking about people’s interactions with computers and how that could be better. We noticed how good our hands are at interacting with physical stuff and how the computer still wasn’t taking advantage of that skill very well. We thought, what if interacting with the computer was like jamming your hand into a pile of Legos? You’d just push pieces of data this way and push other pieces that way and scoop some of them towards you and it could really become more like having a pile of Legos or alphabet blocks in front of you. What would that be like? It was such an exciting idea that we got to work immediately to find out what it could become.
That was the summer of 2006 — a while ago! We got so fired up about this idea that we thought, “OK, let’s map out what this could be” and so we wrote a whitepaper about what we called “The Siftable Computer” and we started building non-functional prototypes that were made out of lasercut wood and acrylic. They were little tiles with photographs in them. We began to prototype, to think together and to show these mockups to other people to get feedback and ideas about how we’d use this kind of tool.
I spent a year with a group at the Media Lab that worked on sensor networks. The idea of systems comprised of a bunch of little interactive, smart things that could be embedded into the environment was—and still is—very pervasive at MIT. We wanted to turn this kind of system in a more human-centric direction and so we thought, what if instead of building an interface to monitor a sensor network, which is what a lot of these projects do, we actually made a user interface out of a sensor network? Let’s take these little smart things and make them something that you actually put your hands on and move them around. That was exciting so we started running with it, and it meshed with the beliefs we had about interfaces needing to be more humane.
My email inbox has NEVER been the same! That was really a turning point. My TED talk was the catalyzing moment. We had thought about commercializing our project beforehand, but we weren’t there yet. Some of the early videos we had posted had gotten a good reception on tech blogs, and journalists had come to MIT to interview me and do news pieces. That was a tiny sliver of the attention that we got when the TED Talk went online. When the talk went out it had more than a million views really fast! That was our kickstarter moment – before there was such a thing as Kickstarter. We knew then that there was something there, something that we needed to get out to the world.
“When the talk went out it had more than a million views really fast! That was our kickstarter moment - before there was such a thing as Kickstarter.”
The short story is that my adviser got invited to give a TED Talk and we managed to split one invitation into two. The reason we managed to do it was she had given a talk at TED 10 years earlier and Chris Anderson, who curates TED, kept asking her to come back and give an update about what her group was doing at MIT.
When Chris heard about Siftables—this prototype that I had been working on with Jeevan—he got really excited and wanted that to be the focus of the talk. I thought, this is great, why don’t we do that? My adviser Pattie Maes initially pushed back and said well, let’s definitely mention your project but I also want to talk about all the other things that the group is doing these days.
I saw this opportunity and I asked, “Why don’t we pitch Chris that I’ll give a talk about Siftables, and you can give a broader talk about the group’s other work?” Pattie didn’t think that would work! She said, “They are so picky about who gives the talk – I think he’ll say no.” I pushed a bit more, saying, “Well, it doesn’t hurt to ask, right?” so I convinced her to ask. Chris wrote me and said, “I heard you want to give a talk about Siftables – convince me! What’s it going to be?” Jeevan and I spent 24 hours writing a script and then sent it to Chris. He saw it and said, “That looks great, let’s do it!”
We’ve completely re-engineered the product twice now, since we spun out of MIT. We left MIT with pretty good working prototypes, the ones you saw on TED Talk. To make that into a commercial product, we completely re-made it: new hardware, new firmware, new software, new applications. Everything. We re-built because there are such different priorities when you’re making a prototype for research compared to a product that’s robust and that people are going to spend their money to buy.
We re-built. We raised money. I think the TED talk helped with that. We got some National Science Foundation grant funding and I think that also helped. It showed that we had some momentum. Then we raised money from True Ventures. They have been great partners and investors. With the funding from True, we built a productization-ready version of the prototype that we built at MIT. We raised more money from Foundry Group. Foundry is based in Boulder, Colorado and one of their investment theses is that human-computer interaction is one of the next big things in tech. We met Jason and Ryan from Foundry and had a total meeting of the minds.
Yes. We first launched our Gen1 product in a small-batch way that we called early access. We launched Gen1 at CES — the Consumer Electronics Show —in 2011. We made about a thousand units available and sold them all out in 13 hours. We were amazed! We thought: We’ve been planning to come to CES to promote our early access program, and it just sold out — what should we do now? In retrospect, we should have just kept that pre-order open. Why limit it? That’s what the recent successes in crowdfunding have shown everyone. We shipped the early access units in March. But we didn’t really launch the system for real to the public until September that year.
We got a lot of great press. David Pogue from the New York Times wrote a nice article about Sifteo Cubes. We found that it was pretty easy for us to get press because the product was totally different than anything else on the market. As the product got out into the hands of customers, we started to hear certain feedback consistently. If you look at David Pogue’s article for the New York Times, it sums it up really nicely. He was very positive about the product — my favorite quote was, “Sifteo cubes are a kick in the imagination.” He really fell in love with the idea. He also had some great feedback, commenting that if Sifteo could bring the price down, eliminate the reliance on a computer to run the games and make the game library more varied, we’d have something really compelling.
We thought about it, and decided: You know what? He’s right. We have to do that. So right as we were launching Gen1, we had a company brainstorm where we asked the team, “OK, how can we satisfy these requirements that have been laid down so clearly for us?” It was an after-hours brainstorm where we broke into small groups, came up with ideas, came back together, and pitched them to the group. The exercise resulted in a document that had a bunch of variations on a theme — different takes on how to build a better version of the core idea.
Over the next week we boiled the collection of ideas down into a plan. The engineering team jumped into action and had a working prototype of the core of our Gen2 system pretty quickly that the games team began building against. We got this second gen system to market with new games over the course of a year — a completely new system.
“That's the thing that is exciting but nerve wracking about a company like Sifteo: you're essentially betting the company every year, on the holiday season.”
PR helps, both for our system lanunches and when new games are coming out. Game launches give us more regular newsworthy items to talk about. We also do a lot of events. We were at Maker Faire, we go to Girl Scout events, we do Game Jams here at the office where we invite game developers to code for one or two days and build a new game that they invent on the spot. Any product that’s for families, or that can be a gift, is seasonal to some degree — so the holidays make a difference. Almost any consumer electronic product has this feature. We have advisers from the original Flip video team who told us, “Every year, you have to just invest in building that inventory and getting marketing ready for the holidays. Then the next year you do it again — placing a bigger bet every year.”
That’s the thing that is exciting but nerve wracking about a company like Sifteo: you’re essentially betting the company every year, on the holiday season. I wish were different — I’d sleep better at night!
The good thing is that venture gives you runway to do research and product development. The company doesn’t have to become profitable for a while. That’s necessary if you need time to get the product to a point where you can sell it.
I think crowdfunding is really interesting. Kickstarter, Indiegogo, and the other options out there are providing an alternative to venture, to some extent. What I’m realizing about crowdfunding though is that you can get burned if you don’t manage it carefully. If you launch a campaign before the product is very far along, you might realize that you need to change the product because you didn’t understand some flaw in the original idea, and then you’ve got a bunch of people expecting one thing, meanwhile you realize you should actually build something else.
Or you might realize that it’s going to be more expensive than you thought to build the product — you find out that you’ll wind up breaking even, or come up short. Or you could just realize that it’s going to take longer than you thought, and you need a certain sized team to build that product — and it’s not in the budget.
If a successful pre-order campaign has validated the idea and collected the money you need to actually build the product, you probably also need money to fund the team! That’s where venture can come in and work hand-in-hand with crowdfunding. This is what OUYA and Pebble are doing — two of the most successful Kickstarter campaigns to date. The model is: raise a giant Kickstarter campaign and THEN ALSO raise venture capital to fund the team and operations. So venture is still necessary, unless you have managed to get to the point of having a shippable product that you can just turn the crank on (with a very small team) by the time you run your Kickstarter campaign. People are starting to think of Kickstarter as primarily useful as a marketing tool — it’s most important property is getting the word out to early customers.
We had the good fortune to incubate the product inside the university as a research project. That helped because we had a couple years of having our basic expenses covered. Having a budget to buy materials and prototype things and a lot of smart people around to bounce ideas with was awesome. The community of a place like the MIT Media Lab helps you push an idea forward to the point where it’s a working prototype, then you can learn from it. That was how we got to create our first functional prototypes.
Then as a company, raising money to build a small team was the next thing. We raised money from True Ventures and I think we grew to five people on that first fundraise. It was two of us founders and the first engineer who is a super-talented full stack engineer who did all kinds of stuff — from desktop software to graphic design to web architecture. Everything. Then we hired a great firmware developer and a dedicated game developer/designer, and the product really began to take shape.
Back then we had this miniature version of our entire current team in the form of five people. At the earliest stage it is critical to have the best, the right five people to set the course. Jeevan and I can’t really take credit as total masterminds to have concocted that perfect team in the beginning, but I think good people were attracted to us and to what we were doing and we managed to bring the right people together.
We do a lot of play-testing for new games. We bring kids and families in to try them out along the way to make sure that they’re not confusing, to make sure that they’re fun, to make sure that they’ll sustain interest. Essentially, to make sure they’re good! There’s no substitute for repeated testing during that process. Then on the platform itself — although we have a strong design vision for what this experience is, we couple that with a lot of testing.
For example: one of the things that we changed between the Gen1 and Gen2 systems was that we had a clickable button on the top of each cube in the Gen1 days. It turned out that users found it confusing. We made it clickable in the beginning because the tactile sensation of a click matched the tactile design sensibility of the system. But every time a game would present multiple options and would say “press to continue” people would apply a very light touch, then they would be confused about why it didn’t do anything. So we realized that for Gen2 it had to be touch.
In our second generation system, we wondered: should it be capacitive touch, could it be resistive touch? Or no touch? That is, could we eliminate touch from the interaction vocabulary altogether?
We had a game in the works called Sandwich Kingdom that helped us make the touch decision. The developer + designer of that game—Max Kaufman—was insistent that the system needed touch because without it, the game just wouldn’t be playable. Max’s voice was in our minds while we struggled to make touch work at a reasonable price point. Since we believed in Sandwich Kingdom and other games that would be enabled by touch, we kept the R&D intensity up and made sure that we made the tradeoffs that made touch possible on the system.
“What Nintendo accomplished in the last decade with the Wii for TV-based game systems, we're going after with a portable system.”
At the highest level, Sifteo is about animating everyday things with the magic of interactivity. We’re currently animating the humble block — the most basic, iconic, platonic, plaything.
We think that there’s this opportunity where play hadn’t changed for a really long time — until video games came along. In the old days, play meant blocks, dominoes, Jenga, Mancala, activities with classic play objects.
When the Atari 2600 and Commodore 64 became popular, they brought electronic, interactive play into people’s living rooms for the first time. That type of play was so engaging — like a drug for people’s brains. These experiences were pretty virtual because they used TV screen and a joystick or later on a gamepad. This was the first wave of democratized interactive play. Then with Guitar Hero, interactive play started to bend back in this direction of traditional play experiences with real everyday things like guitars. The play interface started to look like a real thing again — an object that people know from their pre-electronic lives. Guitar Hero and the Wii are based on the idea that the form of interactive play can be linked to play from the real world. Players swung the Wii-mote the same way they swung a tennis racket, allowing a more natural experience that’s linked to existing forms of play.
For Sifteo, we’re placing a similar bet. What Nintendo accomplished in the last decade with the Wii for TV-based game systems, we’re going after with a portable system. There are a lot of great, timeless, natural play patterns that are based around little physical objects. And that if we can animate those activities with the magic interactivity, we can transform them in a way that’s amazing and fun.
Make your product awesome. Don’t sacrifice things about it that you know in your heart are important. Really listen to what people say when they interact with your prototype — whether features price point or whatever else, take that feedback very seriously. Try your darndest to build a product that fully instantiates your customer-informed product vision without making any compromises. If you hear something more than once, it’s probably worth paying attention to.
What’s your current state of mind?
When and where were you happiest?
When I was learning to build embedded systems for computer music, at the CCRMA.
What’s your favourite quote?
"It is not the critic who counts; not the man who points out how the strong man stumbles, or where the doer of deeds could have done them better. The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood; who strives valiantly; who errs, who comes short again and again, because there is no effort without error and shortcoming" -Teddy Roosevelt
What is your favourite book?
Hackers: Heroes of the Computer Revolution, by Steven Levy
What is your greatest fear?
Not making a difference to the world.
Which words or phrases do you most overuse?
"through the lens of.."
Which talent would you most like to have?
Playing kickass, expert-level rockabilly electric guitar on a Fender Telecaster.
What is your idea of misery?
Being stuck in a conversation with someone who wants to talk but not listen, who wants to lecture but not interact.
When did you have the idea that inspired the Sifteo Cubes?
7 years ago at MIT Media Lab.
What's your background?
As an undergrad I studied computer science and psychology.
I do it now for very personal reasons — I have a son who is two years old. And even though our product is not for kids his age, I know that he’s growing up in a world that’s going to have a lot of interactive things around him and he’s going to have a lot of choices about what kind of technology he uses.
I know there’s something special for human beings about interactions with real, physical things. If I can just steer the world of interactivity back in the direction of the physical world, that’s a real contribution that I can feel proud of.
Sifteo Cubes are the kind of device that I want my son to have—and that I want to be in the world generally—and that’s why I’m excited to be part of a company that’s making the kind of products we are.
What question would you pose to our next interviewee?
In what key way have you have grown as an entrepreneur in your current company?