Jacques Mattheij

Technology, Coding and Business

Turning around the Titanic, part 1, triple the signup rate

For a long long time I’ve been wondering what to do with ww.com, the site that has been my bread and butter for the last decade.

I’ve tired of the concept of webcams, that happened about 5 years ago or so, and since then I’ve found it hard to get motivated to work on the site. On top of that there was still an issue with the shares to be resolved, which essentially meant that if I worked on it I would have to probably pay more for it.

Since November last year though, I’ve decided to make a go of it. The first thing that happened was that we said ‘goodbye’ to the oldest component of the site, the streaming software that has been the engine of the site since its inception. The software was a windows only executable, written using Borland C++builder. It performed quite well but it was definitely showing its age, the streaming was jpeg-push based so the frame quality was good but the framerate quite low.

Kevin K. wrote a replacement using flex/flash and we implemented that during the early days of December last year. Just working together with someone else on the project definitely helped to get me motivated. We still allow ‘die-hard old-timers’ to use the old software, nobody gets forced to upgrade, the change is slow but unstoppable, now about 45% of the users have upgraded to the new stuff, and I expect the rest to follow in the next couple of months.

July this year, on the day of the transfer of the final block of shares the site hit the lowest number of subscribers ever, and it was time to start to turn around the ever decreasing trend.

The reasons the number of subscribers was decreasing are legion, the major ones are: no development for years, boring content, broken sign-up process and a very bad landing page and general design of the site.

After July the 10th I started to tackle these issues one by one. Little by little it started to have effect, first the drop in subscribers was arrested, and then ever so slowly it turned around and the site started growing again. It was like watching an ocean liner doing a 180 degree turn to avoid running in to a bunch of icebergs. In July the site still lost a tiny number of subscribers but the steadily dropping numbers seemed to be history. Along rolled August. In August I spent some time abroad on a holiday, but I did bring along a little laptop and bit by bit I tackled other issues that needed fixing, little things but still, overall it made a big enough difference that even though the number of signups stayed the same the number of cancellations dropped enough that we actually had some growth, about 0.5%. That’s not a whole lot, but it was quite a change after years of monotonous decline.

Then came September. September was to be spent on all kinds of stuff when I spotted a posting on HN by someone looking for a contract job. http://news.ycombinator.com/item?id=1629224

Enter Charles, the person behind http://www.hackernewsers.com/ and a couple of other sites that I had already spotted over the last year.

The first job I asked Charles to do was to do a write-up of ww.com from an SEO perspective, what we could do to increase organic search traffic and how we could improve the site.

The report was drawn up and indicated a very large number of places where the site could be improved. Several problems though, I’m no SEO guru by any stretch of the imagination and I absolutely suck at design (give me a back-end problem though and I’m your guy).

So I asked Charles if he would like to be the one to execute the plan. Charles agreed, and now, three weeks later we are partially underway with the changes to the site.

First of all, we’re replacing the foundation of the site, the ‘lack-of-a-framework’ that was driving the site before is being replaced by ‘YII’, a lightweight php framework.

The homepage and all the userpages are now created by YII in combination with memcached for speed. And the signup process has been streamlined, a better homepage was made and rolled out over the last week.

The effect? Triple the signup rate. Very impressive, and this is really only just the beginning. Ww.com has quite a bit of potential, and we hope to continue our improvements to make our users happier and to get a larger number of subscribers on board.

For the next couple of weeks the plan is to continue the migration of sections of the site to Yii and to further improve the various portions of the site in anticipation of a complete redesign once the whole site is brought under Yii. Charles is a pleasure to work with, he works quickly, doesn’t bitch too much about the mess that ww.com had become over the years and in general is all that you could wish for in a founder.

He also really knows his stuff, which is great because my knowledge (and the site) are pretty dated, and new brooms really do seem to sweep a lot cleaner.

Our work division is that I handle all the customer interaction and feedback, Charles concentrates on building stuff. We have a bunch of ‘trusted’ users that beta test for us and that are welcome in our virtual office (an IRC chatroom) where we hash out the details of what’s going on. The fact that we’re only one timezone apart really helps with this.

So, let’s see how much life we can breathe in to this old girl, youtube had better watch their six :)