Question 01
How do AI matchmaking dating apps calculate compatibility? What data do they collect, and how transparent / legible are their policies to users?
Research in progress — background reading, methods, and findings will be published here.
Research
Setup investigates alternatives to the dominant matchmaking paradigm (swipe-based dating apps) in order to educate the public and improve our own peer-to-peer matchmaking app. We are especially interested in understanding consumers' appetite for AI matchmaking algorithms, and how these algorithms could impact society. Our research is based on quantitative and qualitative data from opt-in surveys and interviews, as well as policies and reports from matchmaking businesses.
Three main types of services are offering alternatives to the increasingly unpopular swipe-based dating apps: new dating apps that promise to be different, singles events that provide an in-person ways of meeting people, and professional matchmaking businesses that offer a bespoke experience. The following paragraphs explore the pros and cons of these services, and explain what a resurgence of peer-to-peer matchmaking could add to the mix.
The dating apps deliver scale like no human matchmaker or singles event ever could, but users often become fatigued sifting through a huge number of profiles, and any given match feels easily replaceable.
So the apps are currently pivoting their user experience from scale to curation. Algorithmic models promise to digest your data and send you your most compatible match, saving you hours of swiping and adding a sense of true-love-via-computation to your first date. But there's an issue with these dating apps: how is a team of entrepreneurs going to come up with a model for calculating romantic compatibility? Maybe they will analyze 10 million couples who got divorced and 10 million couples who stayed together. Maybe they will boil it down to income + education + attractiveness.
If dating apps crack the code and provide everyone in the world with a soul mate, all competitors will shut down and merrily pack their bags for their honeymoons. But if a user's first date isn't a success, they will lose trust in the algorithm's ability to calculate romantic compatibility. Will these apps make their algorithms open to public scrutiny, or will they market a "cutting edge AI model" without explaining what is under the hood? (And will they publish their revenue from data monetization?)
If these apps can't deliver on curation, users will be less likely to view their incoming matches as potential soul mates and revert to seeing each one as expendable. In this case, dating apps will need to improve their algorithm, or return to marketing their inherent strength, which is scale. Dating apps will remain great for finding casual hookups, but committed partners will have to find each other despite the apps' features rather than thanks to them.
Singles events are effective because they are incredibly human (sometimes painfully so), allowing people's first impressions to be based on an interaction rather than a profile. Many communities are doing a great job designing events that provide an environment for single people to meet and flirt.
Some singles events are starting to utilize real time algorithms to help singles find the "most compatible" person in the room, which in theory ensures that attendees don't waste their time speaking to unlikely matches. But the same question arises regarding how these algorithms calculate romantic compatibility. Is it the same way that the attendees would define compatibility or a likely match?
A new category of "dating apps" is coming out that actually fits more neatly into this singles events category than the dating app category. These new apps ping users when a compatible single is nearby and physically directs them toward each other, aiming to create a "meet-cute." This turns the whole world into a singles event by introducing people to each other outside of their homes; but users have to feel assured that this is more likely to lead to love than to stalking.
Modern professional matchmakers dedicate their career to developing a bespoke philosophy of romantic compatibility. They often coach clients on dating skills, getting to know them and helping them set their standards. A human matchmaker doesn't just make an introduction, they also tell you what to look forward to about a person that might not be apparent on the first date. This context is something neither a dating app nor a chance meeting can provide, causing many singles to ignore quality options and chase after incompatible ones.
The biggest drawbacks of professional matchmaking are cost and quantity of options. Both because they spend a lot of time considering matches, and because they need to make sure that their clients are serious, matchmakers charge rates that are inaccessible to most- often in the tens of thousands of dollars. Professional matchmakers' niche is curating matches based on spending time getting to know you, so they limit themselves to demographics they understand in order to give themselves bandwidth to think about interpersonal compatibility. So if a single exhausts options in their pool, the matchmaker has to refer them away.
Setup posits that the best matchmaking paradigm is distributed human matchmaking. What friends and family lack in matchmaking experience, they make up for in knowing the background, desires, and deeper psychological truths that a single-person might not even admit to themselves. Communities have an intrinsic motivation to see their friends happy. When friends set each other up, the new relationship is automatically embedded in a community, meaning social norms of good behavior and reliability are enforced.
The challenge is scale. In order to get more people involved, two barriers need to be overcome: make more people care about matchmaking, and make it easier to matchmake. Scale in this scenario isn't achieved by having a single matchmaker with thousands of singles, it is achieved by making it easy for every matchmaker to be connected with their 10 closest singles, and then reducing the friction when collaborating with friends.
There are some existing apps that attempt to scale in this way, but they fail at addressing the frictions that stop them from succeeding. These other apps haven't scaled because: singles and matchmakers have to pay to use the helpful features, singles download the app, single profiles are public which makes some not want to join.
Setup addresses this friction in a novel way compared to these existing apps: no user fees; singles can sign up quickly without having to download an app; the platform is private by design, only allowing singles' profiles to be viewed in real time during matchmaking conversations with their own matchmaker present.
Question 01
Research in progress — background reading, methods, and findings will be published here.
Question 02
Research in progress— background reading, methods, and findings will be published here.
Question 03
Research in progress— background reading, methods, and findings will be published here.
Question 04
Research in progress— background reading, methods, and findings will be published here.
Question 05
Research in progress— background reading, methods, and findings will be published here.