Collaborative Filtering/Personalization

Collaborative filtering and personalization are two processes by which websites are able to gain users’ attention. By offering recommendations that are relevant to the user either based on the suggestions of others (collaborative filtering) or based on the previous buying/viewing patterns of the active user (personalization), sites are able to focus users’ attention to products that they might enjoy, which could lead to an increase in purchases made by the users.
Collaborative filtering is the process of generating recommendations for products for one user by collecting information of many previous users regarding their taste choices. This process exemplifies the idea of Web 2.0; users generate reviews and ratings of products that are then used to aid other users in finding products that they might enjoy as well. The more users rate products, the more successful the collaborative filtering practice becomes. By using information from other users, the program can generate a list of products that the active user may like as well; the idea is that if people agreed on one product in the past, they will tend to agree about similar products in the future. With the myriad of choices available on the long tail, it is hard to sift through everything that is offered on a site; collaborative filtering makes the process of finding and purchasing items easier. Information that is relevant to the user is sorted through the system of collaborative filtering, so the user’s attention is focused on items that they will most probably like. This saves time for the user and is also beneficial for the business, which can gain more revenue by drawing the user’s attention to similar products that they will likely purchase.
There are two types of collaborative filtering: active filtering and passive filtering. An example of active filtering is amazon.com; the site generates recommendations for the present user based on the ratings of past users. iTunes presents both active and passive filtering. When one searches for a certain band or musician, there is a window of recommendations that displays similar bands that other users liked and purchased; this is an example of active filtering. An example of passive filtering on iTunes is the list of recommendations iTunes generates based on your own previous purchases. It does not rely on information of other users, but it analyses your own buying patterns to create suggestions for future purchases.
Passive filtering is closely related to the idea of personalization; sites track the buying or viewing patterns of a user and change their account accordingly. The difference between collaborative filtering and personalization is that collaborative filtering gleans from information provided by numerous other users in order to generate recommendations for the active user. Personalization utilizes the information given solely by the active user; this gives the website power to command the user’s attention because it is presenting information that is significant to that particular user and not anyone else.
Although there are differences between the practices of collaborative filtering and personalization, the goal of both procedures are the same. They both are concerned with commanding and focusing the user’s attention for the benefit of the site. It gives the user a sense of individuality because it seems that the site is tailored to their needs and desires. As a result, the user will frequently return to the site because it is easier to find the products they may like; increased traffic to the site will probably result in an increase in purchasing from the site. The user and the site both benefit from collaborative filtering and personalization.