Inside the Decisions of Real-Life Matchmakers

Over the past few decades, online dating has begun using algorithms to automate the work of matchmakers. Algorithms are more cost-effective than human labor, which has allowed the industry to expand by offering matching to users at scale. However, with more people growing disillusioned with online dating, The Atlantic reports that matchmaking is experiencing something of a renaissance. But how is matchmaking different from online dating? And what do matchmakers look for in a partner that an algorithm might miss?

In a new two-part study in the Journal of Social and Personal Relationships, I explore these questions and more using data from a professional matchmaking organization.

What Matchmakers Want

Study 1 surveyed 29 of the organization’s matchmakers about their preferences when selecting partners for their clients. All of the matchmakers were women and college educated, and they had a mean age of 38. They had been employed as matchmakers for five years, on average.

When asked what makes their approach different from online dating, they referenced the qualitative data they use to make decisions about partners. One matchmaker said:

On an app, the choices may be wants kids/does not want kids. But there are so many shades of grey in between those black-and-white options – what if someone has children from a previous marriage, shared custody, wants to adopt, is open to fostering, etc. etc., and they are a perfect match in all other ways? It takes a human conversation to avoid swiping past that opportunity.

Another matchmaker explained that their work continues where online dating’s ends, noting:

We go on everyone’s first date for them. There are things like seeing how a person treats the wait staff, their pitch of voice, or mannerisms that you just can’t tell from a static dating profile.

I also asked these matchmakers to rate their preferences for 50 qualities in a partner. The top qualities they prioritized were those involving financial and professional resources (financially secure, good job), substance use (does not smoke), individual traits (reliable, communicative, stable, honest), and physical appearance (nice body, attractive). This left me wanting to know how the preferences of matchmakers compare with those of their clients.

What Clients Want (and What Gives Them the Ick)

Study 2 used data about the preferences and dealbreakers of 31 couples they had successfully matched. Partner preferences are the qualities people desire in a mate, while dealbreakers are the “unideal” qualities they seek to avoid. The couples were comprised of men and women with a mean age of 40, many of whom reported higher levels of income and education than I am used to seeing in my samples.

The data were collected by matchmakers during their intake interviews with clients and coded for the qualities they rated in Study 1, as well as for dealbreakers. The top preferences were financial and professional resources (intelligent, ambitious), individual traits (communicative, good sense of humor, kind), physical appearance (attractive, sporty and athletic), and values and beliefs (likes children, religious beliefs).

The top dealbreakers were substance use (smoking, drug or alcohol use), individual traits (insecure, pessimistic, inflexible), physical appearance (tattoos or piercings), values and beliefs (political beliefs, religious beliefs, desire for children), lifestyle factors (animal ownership), and relationship history (separated or divorced).

In short, algorithmic matching has advantages, including access to vast computing power and massive databases of information that can provide unique insight into what people want in a partner. Yet unlike algorithms, matchmakers do more than act on these preferences. In the era of online dating, one of the most valuable functions of matchmakers may be in helping people discover and reflect on what they are looking for. As one matchmaker observed:

People believe they can check boxes and have a partner hand-delivered to their doorstep without much work or effort on their part. Algorithms create robotic and unrealistic expectations around what partnership is and can be.

Want to learn more? Listen to Talia Goldstein and me talk about dating apps versus matchmakers on The Harvard Data Science Review and You Seem Great podcasts.