Getting Banned From The App Store Was The Best Thing That Happened To Us

“Brace yourself, Marco…we’ve just been removed from the App Store.” With those words, I visualized everything we’d slaved over for months crumple itself into a ball of trash and — ironically — flinging itself into the bin beside me. But it hardly came as a surprise.

Despite stellar traction at over 4 million users and 15 million flings opened per day after launching less than a year ago, Fling had morphed into something entirely different from what I initially conceived.

Fling’s Conception

It was October 2013, and I was on a plane from Hong Kong to London. It’s a 13-hour journey, so I had plenty of time to kill. But instead of tuning out to in-flight movies, I found myself oddly drawn to watching the plane’s flight path.

It was the same dull, slow-moving animation I’d seen countless times before, but this time was different. I’d spent a lot of time thinking about what the next big app in social messaging was going to be, and as I flipped through British Airway’s in-flight magazine that showed its hundreds of routes around the world, a vision started to crystallize.

“I need to make an emergency call,” I said.

There was apprehension, and possibly a faked medical emergency involved, but finally I managed to reach our COO.

“Emerson, I’ve got an idea, and it’s either gonna be worth zero or a billion.”

Despite the skepticism, I got to work anyway, pulling up Photoshop and completing Fling’s designs by the end of the flight.

The vision was clear: Fling was going to be a platform that allowed you to send any real-time message to 50 random strangers in the world. We built the app in a matter of weeks, and within a month we had nearly half a million downloads and incredibly active users. They were sharing snippets of their lives all over the globe, from America to Zambia.

Fling’s vision was coming to life without any of the roadblocks I’d expected. It seemed too good to be true…and it was.

Fling’s Content Minefield

As more and more users began flocking to Fling, we were fortunate enough to receive increasing PR and coverage from various publications. Unfortunately, this didn’t come without its drawbacks.

Vanity metrics can be toxic if you’re not careful with them.

As with any platform that allows any degree of anonymity, we had to deal with another very different type of PR — something we jokingly (kind of, but not really) dubbed ”the penis rate.” Hopefully that illuminates the problem sufficiently and I don’t need to go into more detail.

Unless you’re working on something that’s prone to this issue, I don’t think it’s possible to appreciate the frustration. It was like a perverse game of Whack-A-Mole — we would ban users and delete inappropriate flings, only to have them reappear more creatively named and obscenely photographed than ever.

We eventually had a full-time team devoted to keeping dicks at bay. But despite our efforts, we couldn’t decrease the rate of reported flings to a number lower than 10 percent.

Knowing we had to fundamentally change the app in order to fix the problem, we began devising the next iteration. Fling 2.0 would be the most viral platform in existence for discovering authentic, unfiltered content.

But making this move wasn’t easy. The sexual nature of Fling had become the fuel of our vanity metrics, and with billions of flings delivered per month it was difficult to bring ourselves to tamper with this incredible traction.

One of the most humbling learnings I’ve ever faced as a CEO is that vanity metrics are as addictive — and dangerous — as any hard drug. We knew we had to change the culture of our user base; we even completed the designs of Fling 2.0. But like any self-described addict will tell you, change isn’t easy. We needed an intervention.

Why We Were Removed From The App Store

To be clear, Apple officially removed us because they implemented a policy against apps with randomized messaging. But there’s no doubt in my mind that the reason they pulled the trigger was because much of Fling had become a playground for men to harass women. Take a look at these two graphs:

The more flings that women sent on their first day, the more unlikely they were to come back. Compare this to the analogous graph for men, which is quite normal:

To give you some context about the ridiculousness: Imagine a restaurant where those who eat the most food are the least likely to return.

Sadly, we knew about this gender dichotomy earlier, but we didn’t move fast enough. Apple did the right thing by removing us, and was kind enough to work closely with us around the clock in order to get back in; I’ll always be grateful for that.

How We Handled The Removal

In the short term, we had to relaunch as quickly as possible. This meant identifying the absolute minimum set of features required for relaunch. We expected user backlash, but also considered it an opportunity to rid the app of inappropriate users. We whittled the required core features down to two: following users and sharing flings (reflinging).

Startups are an art, not a science.

The next priority was removing all roadblocks until relaunch. This meant things like ensuring someone from each team was always on-call so communication wouldn’t be bottlenecked. As far as my role went, it meant things like booking Airbnb stays and late-night Uber rides to reduce travel hassle. It also meant smaller things, like volunteering for coffee runs when I had downtime.

As it turns out, there’s nothing that will motivate and mobilize a team like fighting to get your app approved again, and the vibe in the room was not unlike going to war with the clock. Fling 2.0 was ready for release within a fortnight.

The longer-term solution required a bit more introspection. Truth be told, our surging vanity metrics made everyone, including (especially) me, a bit complacent. But even with billions of flings sent per month, something seemed amiss.

That “something” was that our original vision had been lost along the way. We created Fling to create connections powerful enough to disrupt one’s social graph. How many of these connections could truly be meaningful when the norm was to troll for sexual conversation?

We knew we had to ensure two things:

  • Never run into a content issue again.

  • Make sure that the metrics we tracked were aligned with success.

We tightened our rules around moderation and tweaked our design anywhere users had the opportunity to create inappropriate content. For example, we de-emphasized the usernames of strangers, which sometimes contained terms that slipped through our filters, and instead focused on their country of origin. We also started tracking report incidents and implemented tools like Periscope to display these rates company-wide.

But this was also a wake-up call to improve good content, and we put a content team in place to surface the best flings. We also designed a quality score based on characteristics of amazing flings and tracked their related metrics. One of them happened to be average fling engagement from female users.

Screen Shot 2015-08-03 at 4.13.41 AM.png

Content quality changed almost overnight. Selfies were quickly replaced by glimpses into people’s lives that would normally go unseen. The best flings started traveling at an incredible speed in the form of reflings, which today make up 50 percent of all content (for comparison, retweets made up only 2 percent of tweets in Twitter’s early days).

Lessons Learned

Change is difficult. Sometimes it requires a cataclysmic event. Getting banned from the app store was the scare that we needed, and it could have been much worse. As CEO, I never should have let it get to that point, but I’m glad things turned out the way they did.

Here’s what I learned:

  • Vanity metrics can be toxic if you’re not careful with them. They don’t necessarily reflect the health of your product or company. If you only focus on vanity metrics, you’ll hit a local maxima sooner or later, and it’s worth sacrificing them to raise your app’s true ceiling.

  • Startups are an art, not a science. Metrics (and quantitative data in general) should be taken with context. You need to continually assess the qualitative side, as well. For example, vanity metrics state we have lost some users, but that needed to happen. Imagine what would have happened if we simply added new features without purging original harassers. (“Great, I’ll refling and follow a bunch of people…now back to harassing people over chat.”)

  • Everyone at your company cares even if they don’t tell you something like this happens. The first thing that our head of design told me was, “I love what we do here; I love this place, and I love my job…I’m willing to do whatever it takes to keep it, so let’s get to work.” If you’re a startup, it’s unlikely people would be there unless they’re passionate.

  • Your product needs to be aligned with your vision at all times. If there’s a moment where you find it’s not, then move quickly to fix it. Don’t wait until it’s out of your control.

I wish I could offer up a Cinderella ending here. It would be nice to end this by saying that Fling has twenty million active users and we are now one of the world’s most successful startups.

We’re not there yet, but I know 100 percent that we will be. The same gut that first told me that we needed to change now says we’re on the right path, as long as we stay true to what we’ve learned. And this time, I am listening to it.