Stop talking algorithm when it comes to relationships
… how can online dating platforms help you find your better half
Be it in precursory science-fiction books like This Perfect Day by Ira Levin, or in more recent movies like The Island produced by Michael Bay, futuristic scenarios often depict in their own way a world run by advanced technologies that end up being misused by ill-intentioned humans. Interestingly the main scenario usually contrasts with an authentic love relationship between two characters.
Reflecting on our fears and ideals, it seems that human relationships are still perceived as a small island that technology will never reach and control, even in the most tragic scenario.
Since their inception, most dating apps or websites have tried to sell their services showcasing all the benefits of their “algorithm”, with the promise that they will maximize your chances to find your better half.
This promise raises several questions. What is the objective of an algorithm in general? How can that apply to relationships? Does an algorithm try to reproduce an existing human pattern? How can technology better foster great relationships?
Why do we develop algorithms?
Algorithms are built to achieve a clearly defined objective that can be reached thanks to a definite number of rules. The algorithm will process data according to the rules established by a human (with a personal bias) and produce an output. Through machine learning you can test the rules and assess whether your output meets your objective in better way and hence improve your algorithm.
How does that apply to relationships?
To apply an algorithm to relationships, we first need to define the objective of the algorithm, and then the data and rules needed to meet the objective.
- Data and rules: any kind of data about a person could be included to increase the chances to find a correlation (the rule) between some of the characteristics and the chances to match with another person. This includes a personality test, your studies and work, your values, your “ideal partner” description, etc.
- Objective: defined by Robert J Stenberg, a member of the Psychology Department at Yale University, “the triangle theory of love” covers 7 types or stages of a relationship. The ultimate stage is defined as the consummate love that combines, “intimacy”, “commitment” and “passion”
- Intimacy is the feeling of closeness and attachment to somebody. This feeling can arouse from similar or complementary affinities. It can also satisfy a need to be put on a pedestal or on the opposite to find the Freudian notion of your “ego ideal”
- Passion is related to the strong feelings that can arouse from an encounter mostly related to sexual attraction and goes beyond the physical aspect
- Commitment is the conscious decision to stick with each other. The commitment can be shown through promises, loyalty to each other, or strongly supporting something the other one is involved in
So far online dating platforms have tried to foster relationships, mostly targeting the notions of “intimacy” through affinities and social clusters and “passion” through the physical appearance:
- Physical appearance with the profile picture at the center of the experience. Depending on how “desirable” you are, the algorithm allows you to match with somebody else that is equally “desirable”
- Social clusters that ensure some kind of selection at the entry based on criteria like a high level of education, social class, circles with friends of friends, etc.
- Affinities with algorithms that select people you might get along with based on what you and other people like
Based on the data gathered (pictures, your interests, your work, etc.), online dating platforms have built algorithms to supposedly increase chances to find your better half.
Does that algorithm reproduce an existing human pattern?
A research lead by the department of psychology of the University of Utah “Is Romantic Desire Predictable? Machine Learning Applied to Initial Romantic Attraction”, tried to find out whether at least one of the three following assumptions could be met:
(1) For a given person, is there a pattern to predict which people the person would be attracted to?
(2) For a given person, is there a pattern to predict who could be attracted to him/her?
(3) And finally, is there a pattern to predict a unique desire between two people?
The primary purpose of the present research was to test whether it is indeed possible to predict unique romantic desire using measures collected before the two individuals have met.
To test these assumptions, they gathered on two occasions two samples of 163 and 187 undergraduates students, with an equal number of women and men, aged 19.6 years-old in average.
All students had to fill a 30-min questionnaire covering multiple topics including personality test, mating strategy, values, self-portrait and an “ideal partner” description, etc.
After that they were all invited to attend 4-minutes speed dates at the end of which they had to comment on their dates, e.g. “I really liked my interaction partner”, “I am sexually attracted”, etc.
These results suggest that relationship science has uncovered many traits and preferences that can meaningfully predict people’s tendencies to desire others, and be desired by others. However, models drawing from these background measures were consistently unable to predict how much one person could especially desire another person.
In other words it was impossible to predict that two people could match during a 4-min speed dating, based on the information they shared prior to meeting each other. This questions all the different methods dating apps and websites have used so far to try to match people with one another.
How can technology foster great relationships?
How many times has your perception of a person changed in seconds as you were getting to know each other? This man, you were judging quite attractive two minutes ago, suddenly became ordinary, or this woman who you thought was ugly but then something charming illuminated her face.
It is fair to say that we all have a certain level of “acceptability” defined by our own basic set of (a priori) criteria based on physical appearance, social clusters or even some affinities. We tend to think that by narrowing down the search to the maximum we will find the ultimate match, but the previously mentioned study revealed that no criteria could predict it.
Online dating platforms certainly have benefits as they offer you the opportunity to meet people outside your network and sometimes allow you to filter some of your “minimum requirements”. But none has managed to reproduce the “chemistry” that you can feel in real life.
It is all about content and time
The online dating experience could go further. It should put more focus on content and give people the opportunity to really take the time to discover each other, before they even meet.
- Content is king: on the existing most famous dating apps, you can create a profile with only one selfie, your name and age and that’s it.
However the more you know about a person, the more chances you have to have some kinds of feelings. A lot can already be done online, whether by revealing personality traits through textual elements, or by using the power of images to show something different than just a selfie. On top of that the usage of video, almost inexistent today, can bring more to the experience, as movements create feelings through a smile, a glance, an attitude, a voice.
Content also creates natural conversations because it gives a proper context to engage a discussion and avoid the unfortunately famous “Hey, how r u?” - Give it more time: it goes hand in hand with content. Endless swiping is counterproductive as the choice overload reduces our own capacity of making a choice. You never really get the time to enter into somebody else’s world. It’s like swiping through hundreds of CVs and never looking at the experience of the candidate in detail, and then organizing an interview based on who answers the phone first.
Finding the right partner takes time and people should acknowledge that. Online dating platforms should limit the number of profiles presented per day and encourage people to take a better look at one another.
In conclusion: the experience of online dating is going to evolve from a “supermarket-like” experience to something more real and, immersive that will leave room for people to reveal their personality and bring online more of that chemistry that can happen in real life.
So far the usage of smartphones, the proposed user experience and the maturity of the dating market in general have surely limited the appearance of new breakthrough products. A lot of companies have tried to introduce products with more engaging content like videos but they have always failed so far, expecting too much effort from users while they were not ready to go that far in the experience yet.
It’s a matter of time…
This story is published in The Startup, Medium’s largest entrepreneurship publication followed by +443,678 people.
Subscribe to receive our top stories here.
