From Hate Mail to 1 Million Likes

Wow. Webcam Toy now has more than ONE MILLION LIKES on Facebook!

I am stunned! Only 5 months ago, Webcam Toy didn’t even have a Facebook page. 10 months ago I was being sent hate mail for making such a poor app!

So why the turnaround? How did it go from zero to one million likes so quickly?

At first glance the app is very simple: take photos with your webcam, add some funny effects or nice filters and share on a social network. It’s not even a new idea. Instagram, Photo Booth and loads of other apps have done this before, yet despite the competition Webcam Toy is a huge success and continues to grow in popularity.

I think this is a good milestone to reflect on Webcam Toy’s progress, to share my story of ups and downs, and to hopefully encourage fellow developers who are thinking of making their own app.

Zero likes

I created Webcam Toy in July 2011 after redesigning the rest of my website, Neave.com. It was adapted from a similar app I had already created called “Neave Webcam” which allowed people to record 3 second video clips of themselves and add them onto a wall of other user’s clips. The interaction between people on a large, shared grid of video clips was great fun, however some people uploaded rude videos and I eventually had to remove the wall.

People hated this new “Webcam Toy”. It only had 30 effects, you couldn’t take photos and you could only view yourself. I added the app to the newly launched Google Chrome Web Store, following a reasonable success with another web app called Planetarium, but it was getting panned with 1 star reviews. So with damage limitation in mind, I pulled it down after a week.

I was inundated with complaints, some exasperated, some hateful, incredulous that I dare remove functionality from the app they loved. I never knew! It was a small audience but a vocal one. People are loathe to unexpected change.

Take a photo

In an attempt to assuage the angry emails (and some more polite requests) I hastily added the ability to “take photos” — you know, pausing the live video and presenting a still image in a white border to make it look more photo-like.

A little while later I added a download button so people could save photos to their computer. Then I set to work studying the Facebook Graph API and soon managed to add a button that allowed people to upload photos to directly to Facebook from within the app. I quickly put up a Facebook fan page too. Why not, I thought. It can’t hurt.

Without any respite, Twitter announced their own photo sharing service so I quickly added the option to upload to Twitter too. I had to learn about OAuth, cross-domain resource sharing, bandwidth issues and loads of other little details, but I got there in the end. Each tweet also contained the #webcamtoy hashtag, which turned out to be a great way to see other people’s photos.

Since then, Webcam Toy has grown in popularity purely by word-of-mouth. I’ve made no effort to promote the app aside from the odd tweet or Facebook update. I added more effects over the next few months, from its initial 30 or so effects to where it is now at over 60. My only focus was to make the app simple, elegant, quick and easy to use.

Too popular

All was going well until March when my server started to strain under the ever increasing traffic. At weekends when there were over 6,000 visitors at any one time, my server frequently crashed. After a lost week of panic and digging around to find what I could do to fix things, I found the problem was all down to the way in which photos were uploaded to Twitter.

Facebook’s API allows apps to upload images directly to its servers, but unfortunately Twitter’s API doesn’t. To post an image to Twitter you need to upload the image to your own server first, then post it across to Twitter’s servers. This uploading activity was crippling my site, so I quickly rented a new server and started using it exclusively for uploading images to Twitter. Webcam Toy and I both breathed a sigh of relief.

The stats

Despite the setback with Twitter, things really took off with Facebook. Without Facebook I doubt my app would be anywhere near as popular as it is today.

So for the curious, here are some stats:

• The Facebook page went from zero to 1 million likes in 5 months. I only created the page in late December 2011.
• Webcam Toy has around 500,000 page views per day, which goes up to around 800,000 page views per day on weekends.
• Around a further 80,000 people use the Chrome Web Store app every day. According to Google that’s around 750,000 active users. My advice: don’t ignore the Chrome Web Store. Web apps are as valid as native apps.
• 100,000 people post photos to Facebook every day. According to Facebook’s stats, that’s around 1 million monthly active users.
• 600,000 photos are uploaded to Facebook every day. On weekends that figure almost doubles to over 1.1 million photos per day — I mean, wow, that’s about 12 photos per second! Kids are the main users, so usage goes up after school and at weekends.
• Since its launch, 105 million photos have been downloaded (saved to the computer), 75 million photos posted to Facebook, and over 1 million photos posted to Twitter. Proof, if any were needed, that Facebook is the dominant photo-sharing social network.
• 20,000 new Facebook “installations” (new users) every day.
• Around 5,000 new Facebook likes per day, but it varies a lot.
• The main demographic is South American teenage girls! Well, I never intended to target this audience, but I guess the simple identity, the sociability of Facebook, and the fact that my site has been localised into Spanish helps explain this.

There are plenty more stats, but you get the point. It’s doing rather well, and doing better all the time. Here are a few graphs from my Google Analytics, tracking usage from October 2011 to May 2012:

As you can see the traffic yo-yos every week, peaking on Saturdays. The Christmas and Easter breaks meant that kids could use the app every day, so there was more sustained usage. And in retrospect, you can now see where the number of photos posted to Twitter faltered as my server struggled.

To the next million!

So there you go. One million likes on Facebook. Not bad for an app that’s made in Flash — yes, Adobe Flash — and is only available online, in the browser. It doesn’t work on mobile. At all. It flies directly in the face of the perceived wisdom that native mobile and HTML5 are the only ways to make a successful app these days.

Having said that, recently I have been busy making an HTML5 WebGL version of Webcam Toy which should replace the Flash version in a few months once the new HTML5 device API is available in modern browsers (Internet Explorer users will still get the Flash version as IE doesn’t support WebGL). I also hope to make a native mobile app eventually.

My lesson to developers is simply this: make a great app and people will use it. Make it in Flash. Make it in HTML5. Make it in native code. Make it work on mobile only, or desktop only. However and wherever you do it, find your niche and make a neat, simple app that does one job very well. Listen to your users and improve and iterate. And don’t be put off by competition. There’s plenty of room out there for your app too.