The world of technology has been buzzing with discussions about the merits of one recommendation technology over another. And in an effort to showcase their “perfect” recommendation algorithm, companies keep trying to out-shout each other.
However, even the most intelligent algorithms do not perform like a silver bullet.
1. Algorithms are like consultants -- with inaccurate input, they provide faulty output.
2. Algorithms are not without bias. They work based on their past experience rather than your past experience.
3. Just because an algorithm spits out a salient recommendation does not mean that someone is going to actually follow through on the advice.
For example, the semantic recommendation algorithm works great from day one (what’s called Cold Start) as it is not dependent on usage; however, it doesn’t reflect user preferences, and is language dependent. Collaborative filtering, on the other hand, answers these two shortcomings, but requires extensive usage data - a challenge in Cold Start scenarios for both new content and new users.
Last, but definitely not least, a recommendation engine, as sophisticated as it may be, is useless without strong and relevant user experience.
As recommendations and discovery are highly dependent on the user’s ability to easily find relevant content, a strong recommendation engine must have a strong link to the way its recommendations are presented to the end user. The user interface and experience have a profound effect not only on the chances that the user will consume the recommended content, but also on the overall trust between the user and the recommendation engine. This is because, at the end of the day, the UX is a content store front. If specific content is being pushed too aggressively, users will notice and won’t trust the recommendation that the system provides them.
So what does this mean then for recommendation technology? It means that while algorithms are important, it’s more important to provide your customers with a reliable presence that transcends a variety of situations. In other words, make sure you establish trust so that the recommendation is actually taken.
In the following posts, stay tuned for exciting news about "COMPASS," Viaccess-Orca's innovative content discovery platform that easily enables both subscribers and content service providers to fully benefit from the content-rich world of interactive television.