People often talk about scaling technology (getting more traffic, doing more transactions etc). But how do you scale content?
Video content is very popular on 12WBT.com. It makes sense – people want to watch a workout and do it at the same time. Some of the really popular ones are 45 minute videos, which have several trainers doing a workout. Behind the scenes lots of people are involved, this kind of production needs lots of rehearsal, choreography, a dedicated set, and several cameras. Before it is ready, it needs to get edited down and uploaded to a video streaming server.
So they’re popular, but they take a while to make. They don’t scale particularly well.
For the new data driven videos, we took a different approach. A user still gets a 45 minute workout to watch, but we get there via a very different approach.
Instead of shooting everything as one continuous narrative, each exercise is filmed separately. Say the exercise was a push up, we start with someone doing a single pushup – carefully cut so that it can loop seamlessly. We do this for each exercise. Then the fitness content expert uses the CMS to build a workout (a list of exercises) and adds instructions.
When the user gets to the page, we can present a set of HTML5 videos. Each exercise is looped for the correct number of reps. We can even use data from a user’s profile to customise the workout. So if you’re on the strength program you might be asked to do 10 pushups, but a beginner weight loss might be asked to do 5 knee push-ups instead. The end result looks like a normal video, but is actually dynamically assembled.
Taking this approach, the video content becomes scalable. Rather than just a few long videos, we can create literally millions of different workouts.