If you’re the type of single person that’s downloading an entire season of a recommended series to binge-watch in one night, or the type that’s swiping through people’s faces, trying to find a date – you have a lot of choices to make for this Valentine’s Day.
Regardless of which part of the spectrum you find yourself this Valentine’s, you have one thing in common: your choices are being controlled by algorithms. While finding a potential partner isn’t necessarily the same thing as finding a series to pass your time with, the user experiences of both these apps are based on similar algorithms. Algorithms such as “Collaborative Filtering”.
What is it?
Collaborative Filtering is sort of what happens in a clothing store: clothes and accessories that are bought together by people are displayed together, so as to influence people into buying more of them. Using data of what users watch, streaming apps like Netflix curates a list of shows or movies from the same genre- or shows and movies that other users with similar watch patterns had watched- and recommends this to its users. A majority of times, the question“What do I watch next?” is answered by the Recommendations section.
In essence, Netflix ends up have a direct bearing on the type of content that a user consumes. By reviewing the watched material and providing more data about your experience, the algorithm gets more and more refined. Netflix even uses different thumbnails for the same series or movie selected from myriads of frames in it that cater to the personal preferences of the users- based on the type of genre that they most frequently watch, or search for- so as to make the thumbnails visually more engaging to the users.
You’re Not My Type:
A similar clustering and filtering technique is used by dating apps. By swiping right or left, the algorithm refines its results, and shows you relevant results of the same “type” of people. In addition, the interests mentioned in your profile also affect your results. For example, if you’ve mentioned that you love animals, the apps try to match you with someone with a pet in their profile picture. However, human beings are a little more complex than TV shows or movies, and this type of clustering or filtering might not always provide required, efficient results.
Given that a lot of people lie on their profiles, (#wanderlust in a bio isn’t a verification mark for the love for travel- what a shocker), and tend to choose people who might not always align with specified interests, the algorithm faces the risk of errors, or of choosing people who might not match the user’s expectations at all.
Moreover, similar to how media streaming apps have review options that can refine results, dating apps rely on feedback to better suit their users’ preferences. It might not always be feasible to collect this feedback, as sharing more details usually requires the user to reveal personal information, and many users may hesitate in doing so.
Hope for Progress:
However, these shortcomings are not permanent. In the near future, the companies might be able to refine results based on the users’ posts on social media, and only the information that they choose to make public would be used for the above-mentioned.
In conclusion:
In a world of countless choices, Collaborative Feedback serves as a necessary evil, that, on one hand- significantly decreases the amount of time required to find relevant results that cater to personal preference of the user- and on the other, require the user to reveal more information about themselves, even at the cost of their privacy. Knowing how to navigate your way through these data-seeking websites, thus, becomes of immense importance. Moreover, companies should also devise methods of sifting results that don’t intrude upon their users’ privacy.
