Peter Kafka

Recent Posts by Peter Kafka

Mixel, Take Two: After a High-Profile App Fails, Its Founders Try Again (Q&A)

Khoi Vinh had a lot going for him when he launched his iPad app last November: A fancy pedigree from the New York Times, high-profile backers and lots of positive press.

None of that mattered.

Mixel, a social app that was supposed to inspire people to create their own art collages using images they found on the Web, never took off. By this spring, Vinh and his co-founder, Scott Ostler, came to terms with their failure. Now they’re trying again.

Their new app uses the same name, but is entirely different: Mixel 2.0 is an iPhone app that lets users do something they’re already doing with their phones: Manipulate and share their photos. Much less ambitious and, hopefully, much more popular.

It’s definitely pretty easy to use, though you’ll need a Facebook or Twitter account if you want to try it out. Here’s my first effort (apologies for the salty language, but that’s life on the Jersey shore):


Start-up “pivots” are now a cliche, but you seldom get to hear the story behind them, or a candid self-assessment of entrepreneurs’ mistakes. Via email, Vinh explains why he thinks he failed the first time — for starters, he thinks he picked the wrong Apple device — and how he’s applying some hard-won lessons this time around.

Kafka: Let’s start with the basics. For people who didn’t try the old Mixel and have yet to try the new one, what’s the difference?

Vinh: It’s much, much simpler. Mixel for iPad tended to be pretty involved and, to our regret, required you to really work at it. It also became much more about creating art than we intended, and that resulted in a somewhat self-selecting user base.

Mixel for iPhone, on the other hand, is so simple anyone can make a collage as you walk down the street. Anyone. The app does the hard work of making you look great.

It’s also much less about making artistic statements and much more about sharing the pictures that people are constantly shooting, so that you’re telling the story of your life to the people that you care about.

So now let’s back up. When you launched Mixel 1.0 last November, what was the plan? Was that supposed to be the only app you produced, or did you always think that you were going to be launching other apps like Mixel 2.0 as the company progressed?

No, our plan was to focus on one app, one network, and to grow it to be a bajillion users strong. Obviously that didn’t work out. So we had a very protracted, emotional, difficult period of self-examination around March or April of this year.

We knew things weren’t working for many reasons that were becoming more and more apparent with each week. The one reason that stuck out at the very beginning was that we needed to be on iPhone — it’s just not possible to bootstrap a new social network entirely on iPad, at least not yet.

So the question became: What do we build for iPhone? A smaller version of the original app, or something new? Eventually, we realized we’d have to retire the iPad app and focus exclusively on something entirely new for iPhone. Again, there were many reasons for this, but this time, the simplest was we just didn’t have the resources to maintain two separate networks.

Love to hear more about the iPhone versus iPad discussion. Why do you think you can’t launch a new social net with a tablet but no phone? And when did you realize this?

It took us a few months to realize this, but in retrospect, we should have been looking more closely. We were distracted by our great-looking engagement numbers. Users of Mixel for iPad spent a lot of time in the app, and those numbers were consistently outperforming our benchmark category in the App Store.

But over time it became apparent that the usage patterns for the iPad are very different than for iPhone, at least at this stage of the device’s evolution. By and large, people leave their iPads at home — they don’t take them to work — so it’s really difficult to get someone to engage with the network during the course of the day. You lose tons of opportunities to get him or her to interact with their graph, and that’s basically like starvation for a new social network.

When the user gets home, they pick up their iPads and spend a few hours with it. That’s part of the beauty of the device; it’s very, very immersive. The underside of that quality is that, given a few hours after dinner and before bed with such an immersive device, most people will open just a few apps and spend lots of time in each (versus the dozens of sites they might visit with a laptop and a browser). If you’re one of those apps, that’s great, but if you’re not, you’re in a much more difficult competition for users’ attention.

When you launched in November, you got a lot of publicity, and shout-outs from influencers, some of whom were your investors. And you say you had good engagement numbers at the start, too. When did you realize you had a problem, and what was the metric that tipped you off?

The number that tells you more than anything is pure downloads. It’s not a perfect indicator, but it’s a reliable proxy for everything else. So, watching the downloads top out and then start to decline was the big red light for us. We knew we had to come up with a plan B. That was around March; we had a few features to finish off with the iPad app, and then we started working in earnest on the iPhone app in April.

Did you have an impulse to try to save the first app by introducing new features, or overhauling design, or doing a new publicity campaign? And how much input did your backers have in the decision?

Yes, definitely. We were all very attached to the iPad app. Personally, I poured myself into building that thing. And there was a wonderful community of people using it, all making truly amazing things. So we looked for ways to save it, to juice it up. I went “on the road” trying to talk about Mixel wherever I could, in front of whomever would have me.

Eventually, I came to the conclusion that the app was just not viral enough. Our core users were not attracting new core users, at least not in sufficient volume, and all our efforts in publicity just weren’t producing significant upticks in the numbers. Deep down, I probably recognized this reality a few months before we formally decided to build something different for the iPhone, but I still tried to chase down every possible lifeline I could before giving it up.

Our investors were great, and very supportive. They all believed in the original iPad app, but when we came to them with the prototype for what would become Mixel for iPhone, they got on board quickly. I really credit them for their faith in our team; I was initially a little nervous that when I showed them the prototype they’d say something like, “This isn’t the app we invested in.” Instead, they got right behind us.

Now you’ve got a new app, with the same name but a very different function. Mixel 1.0 asked people to do something most of them weren’t doing already — create digital art projects using other people’s images. And Mixel 2.0 seems to be a much more conventional photo app. I assume you made a point of trying to reach an existing market rather than create a new one this time around. But what else guided you to this version? What other ideas did you consider?

Well, we knew we wanted to create something for iPhone, because we wanted to increase our potential reach, and we knew we wanted to create something in the general vicinity of collage-making, because that’s what we were all passionate about.

A lot of ideas were kicked back and forth: Make a version of our iPad app reduced down to iPhone size, make an app that would let you pose questions and answer them in collage form, make an app that would let you make collages from preexisting parts.

There were other ideas, too, but we tried to vet every concept through a few basic principles: It would have to be much, much easier to make something than it was in our iPad app, and it would have to make the user look really, really great. We really wanted whatever we built to produce results that were delightful or even magical, so that new or potential users would say, “This looks amazing! I’d like to do this, and the app looks like it practically does it for me.”

That makes sense. So now that you’ve relaunched, any lessons you’re applying from the first go-round?

There’s so much. We purposefully went for a “soft launch,” meaning we wanted to see what the product does on its own, and not seek out press coverage, at least at first. Of course, when people like yourselves reach out to us, we’re smart enough not to pass it up.

That was important to us, so we could see as quickly as possible how people would use the app, what works for them and what doesn’t, without the pressure of press scrutiny. There are a few key features we held back just so we could get the product out there more quickly.

Meanwhile, we’ve been watching the early feedback very closely. And even then, how we interpret that data is colored by what happened last time. I like to think we’re better at distinguishing the meaningful feedback — the kind of stuff that actually makes a difference to the growth of the product — from the complaints from folks who just like to complain — stuff that just wastes time if you overreact to it. I can honestly say that, for me, anyway, knowing this the difference can only come from having done it wrong the first time.

Latest Video

View all videos »

Search »

Nobody was excited about paying top dollar for a movie about WikiLeaks. A film about the origins of would have done better.

— Gitesh Pandya of comments on the dreadful opening weekend box office numbers for “The Fifth Estate.”