“Big data” is a buzzword that has become increasing popular as the Internet continues to reach into our everyday lives. Suddenly everything can be recorded and calculated, from what we say on social media to where we are at any given moment. According to Steve Lohr, big data is information that measures every small detail about an environment or situation. It’s challenging to explain big data using a broad definition because big data itself involves micro-level information. So to simplify it, let’s look big data in the context of an emerging market: online dating.
What’s the Data?
Many popular online dating websites, such as eHarmony and Match.com, market themselves on having complex algorithms that sets up users with similar interests. The algorithm is fueled by the big data of online dating, mostly generated through long questionnaires. According to eHarmony executive Joseph Essas, qualitative data falls under three categories:
- Data on psychological compatibility such as traits, values, beliefs
- Data on interpersonal chemistry such as likes and dislikes and shared hobbies
- Data on physical attraction such as general preferences like hair color or height
Another group of big data in online dating algorithms includes attribute ranking. Match.com includes attribute ranking in its questionnaire, which expands or constricts the data pool depending on the importance of the attribute. So for example, if I identified that being taller than me was an attribute of high importance, the algorithm would only include men over 6 feet tall. Meanwhile, if I said that being taller than me was an attribute of low importance, the algorithm would include men of various heights.
Why Does It Matter?
Amy Webb’s TED Talk “How I hacked online dating” explains that while these algorithms are good, users need to be aware of the data they input. Big data on online dating sites is more dependent on the users being aware of the data they input than any other big data collector. For example, while customer feedback does help retailers determine marketing promotions, big data like sales, economic data, and even weather patterns can affect a marketing plan.
Back to online dating, Amy mentioned that you need to not only consider the result you want (aka, matches with desired characteristics), but also, on the flip side, consider what kind of data (characteristics) your desired matches are interested in. She found that the content of an online dating profile does matter, not only in terms of the characteristics, but also in terms of how you present those characteristics in word choice, profile length, and communication timing.
While now, the quality of the results relies on the quality of the data, online dating services are continuing to fine-tune their algorithms to learn from user communication. Maybe someday, you won’t have to answer that “Are you a cat person or a dog person?” question for the algorithm to find some deep, vague underlying personality. Instead, the algorithm will recognize a picture you post with your dog and note that you probably like dogs.