Dating: A mysterious constellation of facts
Are speed dating and network effects contradictory?
Here are a few things that seem to be true:
Dating apps are very popular.
Lots of people hate dating apps.
They hate them so much that there’s supposedly a resurgence in alternatives like speed dating.
None of those are too controversial, I think. (Let’s stress supposedly in #3.) But if you stare at them for a while, it’s hard to see how they can all be true at the same time.
Because, why do people hate dating apps? People complain that they’re bad in various ways, such as being ineffective, dehumanizing, or expensive. (And such small portions!) But if they’re bad, then why? Technologically speaking, a dating app is not difficult to make. If dating apps are so bad, why don’t new non-bad ones emerge and out-compete them?
The typical answer is network effects. A dating app’s value depends on how many other people are on it. So everyone gravitates to the popular ones and eventually most of the market is captured by a few winners. To displace them, you’d have to spend a huge amount of money on advertising. So—the theory goes—the winners are an oligopoly that gleefully focus on extracting money from their clients instead of making those clients happy.
That isn’t obviously wrong. Match Group (which owns Tinder, Match, Plenty of Fish, OK Cupid, Hinge, and many others) has recently had an operating margin of ~25%. That’s more like a crazy-profitable entrenched tech company (Apple manages ~30%) than a nervous business in a crowded market.
But wait a second. How many people go to a speed dating event? Maybe 30? I don’t know if the speed dating “resurgence” is real, but it doesn’t matter. Some people definitely do find love at real-life events with small numbers of people. If that’s possible, then shouldn’t it also be possible to create a dating app that’s useful even with only a small number of users? Meaning good apps should have emerged long ago and displaced the crappy incumbents? And so the surviving dating apps should be non-hated?
We’ve got ourselves a contradiction. So something is wrong with that argument. But what?
Theory 1: Selection
Perhaps speed dating attendees are more likely to be good matches than people on dating apps. This might be true because they tend to be similar in terms of income, education, etc., and people tend to mate assortatively. People who go to such events might also have some similarities in terms of personality or what they’re looking for in a relationship.
You could also theorize that people at speed dating events are higher “quality”. For example, maybe it’s easier to conceal negative traits on dating apps than it is in person. If so, this might lead to some kind of adverse selection where people without secret negative traits get frustrated and stop using the apps.
I’m not sure either of those are true. But even if they are, consider the magnitudes. While a speed dating event might have 30 people, a dating app in a large city could easily have 30,000 users. While the fraction of good matches might be lower on a dating app, the absolute number is still surely far higher.
Theory 2: Bandwidth
Perhaps even if you have fewer potential matches at a speed dating event, you have better odds of actually finding them, because in-person interactions reveals information that dating apps don’t.
People often complain that dating apps are superficial, that there’s too much focus on pictures. Personally, I don’t think pictures deserve so much criticism. Yes, they show how hot you are. But pictures also give lots of information about important non-superficial things, like your personality, values, social class, and lifestyle. I’m convinced people use pictures for all that stuff as much as hotness.
But you know what’s even better than pictures? Actually talking to someone!
Many people seem to think that a few minutes of small talk isn’t enough time to learn anything about someone. Personally, I think evolution spent millions of years training us to do exactly that. I’d even claim that this is why small talk exists.
(I have friends with varying levels of extroversion and agreeableness, but all of my friends seem to have high openness to experience. When I meet someone new, I’m convinced I can guess their openness to ±10% by the time they’ve completed five sentences.)
So maybe the information a dating app provides just isn’t all that useful compared to a few minutes of casual conversation. If so, then dating apps might be incredibly inefficient. You have to go through some silly texting courtship ritual, set up a time to meet, physically go there, and then pretend to smile for an hour even if you immediately hate them.
Under this theory, dating apps provide a tiny amount of information about a gigantic pool of people, while speed dating provides a ton of information about a small number of people. Maybe that’s a win, at least sometimes.
Theory 3: Behavior
Maybe the benefit of real-life events isn’t that they provide more information, but that they change how we behave.
For example, maybe people are nicer in person? Because only then can we sense that others are also sentient beings with internal lives and so on?
I’m pretty sure that’s true. But it’s not obvious it helps with our mystery, since people from dating apps eventually meet in person, too. If they’re still nice when they do, then this just resolves into “in-person interactions provide more information”, and is already covered by the previous theory. To help resolve our mystery, you’d need to claim that people at real-life events act differently than they do when meeting up as a result of a dating app.
That could happen as a result of a “behavioral equilibrium”. Some people take dating apps seriously and some take them casually. But it’s hard to tell what category someone else is in, so everyone proceeds with caution. But by showing up at an in-person event, everyone has demonstrated some level of seriousness. And maybe this makes everyone behave differently? Perhaps, but I don’t really see it.
Obscure theories
I can think of a few other possible explanations.
Maybe speed dating serves a niche. Just like Fitafy / Bristlr / High There! serve people who love fitness / beards / marijuana, maybe speed dating just serves some small-ish fraction of the population but not others.
Maybe the people who succeed at speed dating would also have succeeded no matter what. So they don’t offer any general lessons.
Maybe creating a dating app is in fact very technologically difficult. So while the dating apps are profit-extracting oligopolies, that’s because of technological moat, not network effects.
I don’t really buy any of these.
Drumroll
So what’s really happening? I am not confident, but here’s my best guess:
Selection is not a major factor.
The high bandwidth of in-person interactions is a major factor.
The fact that people are nicer or more open-minded in person is not a major factor, other than through making in-person interactions higher bandwidth.
None of the obscure theories are major factors.
Dating apps are an oligopoly, driven by network effects.
Basically, a key “filter” in finding love is finding someone where you both feel optimistic after talking for five minutes. Speed dating is (somewhat / sometimes) effective because it efficiently crams a lot of people into the top of that filter.
Meanwhile, because dating apps are low-bandwidth, they need a large pool to be viable. Thus, they’re subject to network effects, and the winners can turn the screws to extract maximum profits from their users.
Partly I’m not confident in that story just because it has so many moving parts. But something else worries me too. If it’s true, then why aren’t dating apps trying harder to provide that same information that in-person interactions do?
If anything, I understand they’re moving in the opposite direction. Maybe Match Group would have no interest in that, since they’re busy enjoying their precious network effects. But why not startups? Hell, why not philanthropies? (Think of all the utility you could create!) For the above story to hold together, you have to believe that it’s a very difficult problem.
Alternative theories:
1. Many, perhaps most, people are happy with dating apps but don’t publicly voice positive sentiment. I had a pleasant time on them and found someone, but I don’t go around saying how great the apps are. We mostly see grievances in the same way that we mostly see negative reviews of insurance providers (failed claimants post more often).
2. Back in the 2010s it was fashionable for dating sites to have data blogs. Hinge once revealed that the top 10% of men and women received 58% and 46% of all likes, respectively. The bottom 50% received 4% and 8% (no typos). Basically we collectively (1) roughly agree on who’s attractive and (2) disproportionately pursue these people rather than our ‘proper’ match. The result is disappointment for these optimistic overshooters. Maybe in real life people are more pragmatic, either by choice or constraint.
3. Desirable/functional people are only on there a short time because they find someone and drop off. Uggos and weirdos stay on there a longer time, and so you’re more likely to see them than in some hypothetical random encounter model. This leads to sour experiences because no one likes them, not even other uggos and weirdos.
4. Higher sociosexuality in men leads to lots of lying about dating intentions, i.e. they claim to be looking for a relationship but in fact just pursue casual sex (intention is an actual field you fill out in most apps). If 30% of men are there for casual sex but only 10% of women are, then the excess men have an incentive to lie. This leads to bad experiences for women seeking relationships and engenders general cynicism (‘where have all the good men gone’). In a similar dynamic to #3, these deceivers are on the apps for longer than the ones with a genuine intention to partner up, so you run into them more than you seemingly should.
I feel like the intersection between this question and the question "why is flirting hard?" is pretty large.
TL;DR: not only are dating apps an algorithm, they make us also into algorithms.
In other words, I don't think "high bandwidth" quite captures all of what's important about in-person. It's also "low latency". In other words, it's interactive in human time. Both "high bandwidth" and "low latency" are essential to natural human interactions.
To expand on that, lets say we're sitting across the table at a speed dating event.
- We do basic introductions
- I say something I think is a bit funny, wait nervously for a reaction
- You laugh a little bit, signaling it's not really that funny but you acknowledge the effort
- I appreciate your reaction and adjust accordingly
- You say "that reminds me of..." and tell a slightly personal story
- I respond emphatically
... and on we go.
Now let's try the same thing on a dating app.
- We do basic introductions
- I say something I think is a bit funny, wait nervously for a reaction
- 24 hours go by while I get even more nervous that I blew it with this person with a hot picture
- After waiting the internet-approved period of time designed to most heighten anticipation, the other person responds with mild encouragement
- I spend eight hours carefully considering what next step will maximize my chances of staying in the game
- ...on we go.
In brief: There are no authentic interactions online. Maybe that's what "high bandwidth" is meant to capture, but swapping videos could also be "high bandwidth"*. So I think that "low latency" should be called out explicitly.
By the way, people who work in systems design will tell you that "high bandwidth" and "low latency" are often in tension with one another. I think that the technical term you want here is really "wide spectrum". In person we are signaling with not only our words but our tone of voice, timing, body posture, facial expression, nodding, little flirtations like playing with our hair, maybe even little touches if it's going well. We are using lots of "channels", my network engineer friends would say.
----
*Business idea: Dating app x TikTok. All contact on the app is in the form of back-and-forth short-form videos. You're welcome.