When you think about websites with heavily personalized content, Netflix will probably be one of the first that comes to mind. And that is not a coincidence. More than 75 percent of the user activity on the site is driven by its recommendation system, meaning you are very likely to watch whatever Netflix suggests to you after login, and there is only a slim chance that you will need to use a navigational tool or the search bar to pick a movie.
If you ever went to a video rental store or spent countless hours trying to choose the right movie based on IMDB ratings, it’s not hard to see why creating a service like this was such a great achievement.
The California-based company has been using personalization tactics for over a decade now. In 2006, they offered 1 million dollars to any scientists, who could improve the accuracy of their recommendation system by 10 percent. This system was called Cinematch, and it was developed to predict what rating a user will give to a movie.
After a year, Netflix offered a progress prize to a group called Korbell for improving their formula by 8.43 percent, with 2000 hours of work and the use of 107 algorithms. The Grand Prize was taken by another group two years later.
At the time, the recommendation system’s goal was to help visitors find suitable DVDs to rent through the company’s website. Since Netflix started its streaming service, the amount of data the company can analyze to improve its service has grown exponentially. Besides the ratings, now it can also take into account your viewing patterns, browsing habits, location, and search history, and combine this knowledge with a very deep categorization of movies and artificial intelligence to find the perfect suggestions for you.
A bit of tagging
While many web services provide personalized experience based on location, the device used, or your friends’ recommendations, what makes Netflix unique is its deep understanding of movies. In order to narrow down the selection for the visitors, the company developed a very complex tagging system for their content.
They fit TV shows and movies into nearly 80,000 categories. Some of these contain information about the genre, the director, or the decade when the movie was made, and some others are used to describe who they think the movie is for, but most of them are created by blending these factors (for example: Comedies-About-Marriage-With-Happy-Endings-For-Young-Couples).
The movies are tagged manually by the analysts of the company, and algorithms decide which categories will pop up when you visit the site. The majority of them try to find similar viewing patterns between users, since they are very likely to represent similar tastes. The predicted ratings play a less important role nowadays because there could be a huge difference between what people value highly, and what makes them happy. Maybe you would give a high rating to a documentary about social injustice, but that doesn’t mean you would like to watch it on a Wednesday night after work.
Personalization and user engagement
The success and rapid growth of the company show that building a service around personalized recommendations can pay off big time. Discovering what satisfies its visitors most, and turning this knowledge into suggestions, makes it possible for Netflix to continuously grow its subscription base and increase the time spent on the site.
How can personalization help your business?
Netflix’s recommendation engine was developed by hundreds of engineers, and it analyzes the habits of millions of users. Most website owners will never have the resources and manpower to develop such a robust system, but that doesn’t mean that a smaller company can’t take advantage of personalization.
Many of the ecommerce platforms offer plugins that personalize the shopping experience. Every piece of advanced email marketing software allows you to tailor your messages based on what you know about your subscribers.
Our company’s solutions may also help you to better understand your customers and show only relevant articles or offers to them. If you are interested in the possibilities that content personalization can provide to your business, don’t hesitate to contact us.