Dating apps are using data analytics to get better at matching people for dates.
I’ve been binge watching on this new Netflix show, The One. In it the protagonist discovers a method to find the perfect matches for people based on their DNA. It was a good show albeit with a less than satisfactory ending but it got me wondering, is there a way we could really do this?
I mean, if recent divorces involving billionaires Bill Gates and Jeff Bezos are anything to go by, we still haven’t found the answer to fulfilling and lasting relationships. In fact, divorce rates, especially in the West, are up, in addition to declining fertility rates.
Technology has changed the world in profound ways, how we work, communicate, transport and more relevantly, dating. Dating apps are fast becoming the mode of choice for singles worldwide. In addition to increasing computational power and internet penetration, the convenience and effectiveness of dating apps have made them gain wide popularity.
As a prospective dating single, I can tell you first-hand how hard it is to find a date in an ever-individualized society. Between my job and my studies, I can barely find the time to socialize. And when I do dip my foot into the dating pool the odds of successfully finding a date are slim to none. The old way of meeting people in a bar, a restaurant or even via a friend just isn’t working as it used to.
Dating apps are just more convenient and effective. From your phone you can look for someone you find attractive, engage with them a couple of times via chat and maybe set up a face to face date if you like them enough. If not, there is no harm done. There will be other options for consideration.
Dating apps are so easy to use they have attracted a lot of users and with them. The US is one of the biggest consumers of dating apps in the world and thus is a model of how powerful these apps can be if used well.
1 out of 10 Americans use dating apps and up to 25% of couples have found their soul mates on these platforms. Moreover, a Kelton study estimated that about 1/3 of Americans (about 80 million) have used an online dating site at one point of their lives.
Users are reported to spend 22 minutes daily on average on these apps. This adds up to a whopping 12 hours every week.
Despite the bad press that dating apps get they, are surprisingly effective at matching people and even get them to commit to each other. More than 60% of users successfully date on such platforms, with about 23% choosing to marry the spouses they found online.
The total revenue from these apps and websites was around $1 billion in 2019 and is expected to reach $1.1 billion in 2024. This market is estimated to be just over $2 billion in the US and around $153 million in Canada.
These numbers have attracted various players in this field, as each tries to cut their own piece of this lucrative market. Tinder is the undisputed leader. Other major players include OkCupid, Match.com, Hinge, eHarmony and Chemistry.com. Match.com, Tinder, Hinge, OkCupid and Plentyfish are all owned by dating conglomerate Match Group.
In addition to these, there are various other apps that have specialized in niche markets, these are such as Bumble (women initiate interaction), ChristianMingle.com (for Christians looking out for singles with similar values), JDate.com (for Jewish singles), OurTime.com (for serious daters over 50) and BlackPeopleMeet.com (for African Americans).
How they Work
As a techno-enthusiast, I was interested to understand how these apps/websites go about the business of matching people. The difficulty of connecting two people looking for love cannot be understated. I mean, if it were that easy then we wouldn’t have apps in the first place.
These apps use the best of data analytics and data science to first understand the user so that they can recommend the best matches for them. Granted, each app has its own ‘unique’ approach, they all heavily rely on data to make their platforms work.
What these apps propose is that love, or at least an element of it, can be brought down to figures. That you can completely understand a person; their personality and behaviour, express them in numbers and then use those numbers to find a similar/compatible individual in a pool of thousands or sometimes millions of unique individuals.
Sounds hard to believe? Well these apps don’t seem to think so. In fact, they’ve built their business models around this belief and they have the numbers to back it up.
Take Tinder for example. From an outside perspective, Tinder doesn’t look that complicated. If one user swipes right, meaning they like the person (at least physically), and another user also swipes right then they match! Hooray! They start chatting and before you know it you receive a wedding invitation.
Sounds simple, right?
Well as you’ve probably suspected, its not. Its actually very complicated. What if, for instance, one user, say a desperate guy who hasn’t had much luck in dating, decides that he will swipe right for every option that appears. I’m sure you’ve heard of such and hey, maybe you do it yourself (no judgement). While this strategy may work for the user, and sure enough it will, it lowers the value of the right swipe. The person on the other end of the swipe, usually a girl, will not value right swipes as much and Tinder will not be as fun or effective any more.
To solve this Tinder set a limit to the number of right swipes one is allowed each day. This proved surprisingly effective in curbing the behavior without having a significant impact on user membership.
Tinder also uses sophisticated data analytics software called Interna to collect data from their users. The software analyses client data when they input queries. These queries are then entered into a database after which appropriate responses can be generated.
eHarmony’s model is a little more complicated. They market themselves as a relationship focused app and so have go out of their way to ensure that they can get the ‘perfect’ matches for their users. They boast an astonishing 13 million matches, daily.
To make this happen, the app collects huge amounts of data about their clients. This adds up to almost 4 terabytes each day. With machine learning, data analytics and a number of complicated algorithms, they are able to sort through this data to generate such high numbers of matches.
Match.com makes use of questionnaires to better know their clients. The number of questions can vary from fifteen up to one hundred. Users are then assigned points based on how they answered the questionnaires. These points are then used for matching where people with similar points are paired together.
Contrary to eHarmony and Tinder, Match.com uses data analytics to find inconsistencies within their data. They will for instance compare your answers to your activity online (Yes, they track your online activity. With your permission of course).
The algorithm will then correct any inconsistency found and then maybe a more suitable match can be found. They even take the extra step of using facial recognition to determine what you really like.
Note here that the developers noticed that a lot of times clients lie about themselves on dating apps to look better. A guy may for instance lie about his height and income while a girl may lie about her age.
These apps utilize the best tools in data analytics and artificial intelligence at to get a complete understanding of their users. This way they eliminate errors and find the best matches for them.