[caseysoftware]

I worked for a dating website a long time ago.. and it's key to understand their business model:

- If you find good matches but not great matches, you stick around.

- If you get frustrated and give up, they lose one customer.

- If you find love and get married, they lose two customers.

Which one will they optimize for?

My writeup: https://caseysoftware.com/blog/working-for-a-dating-website

[solatic]

People who have difficulty on dating apps want to find a scapegoat, so they scapegoat the app.

The truth is that dating markets are lemon markets. People who are "dateable" tend to find success quickly, and people who are "not dateable" tend to stay on the market. Hence over time, the market will be dominated by "not dateable" people. No dating app on the planet will magically make you a "dateable" person.

To find success on dating apps, you have to work on yourself first, and only afterwards make sure that work shows through both in your profile and in your texting.

Source: was on the apps, undateable for eight years (depression and low self esteem), went to therapy, after making huge changes to my life and getting to a point where I felt like things were going well in everything but being single, a month later I found my girlfriend (now two years together).

[anal_reactor]

> you have to work on yourself first

I hate this phrase because it's a generic catch-all that says nothing but shuts down any discussion. If I'm friendly, responsible, honest, not poor, do sports, learn new things, keep the house clean, then the fuck more you want. Can we admit that social dynamics have completely changed and the value of "a relationship" dropped through the floor? 200 years ago bad relationship was better than no relationship because have fun trying to farm land on your own, but nowadays it's literally more convenient to live single than to deal with the inconvenience of living with another person.

Also, personally, I'm a minority within a minority, and I'm not going to cheat the statistics even if I shower twenty times a day.

[roenxi]

> If you find good matches but not great matches, you stick around.

I dunno, I have difficulty seeing how the dating sites could singlehandedly pull that off in the average case without the site users really leaning in to help. It would seem to run into the basic reality that men and women historically pick the best match from a fairly small pool of people. A dating sites can't do worse than that even if they're trying. If people are willing to use the same standards as all their ancestors then they'd pair off quickly.

It seems more likely that there is just a natural dead-sea effect because of that where the people on the sites over the long term are not the sort of people you'd settle down with, and there is also this subtle idea that the dating site is there to find someone a perfect match (probably doesn't exist to start with). Those are design issues that go a lot deeper than any algorithm the sites might be using.

[chongli]

That's because "matches" are the wrong criterion to look at. In aggregate, matches don't matter. What matters is the population of marriageable (or otherwise amenable to long-term relationships) people. And that's what the dating app calculus works against. Every time 2 marriageable people get together, they remove themselves from the pool. If there is not a significant influx of new marriageable people then over time the marriageability of the pool will decline. As it drops, the concentration of "serial daters" goes up.

In a high concentration of serial daters, no one wants to pair off because there isn't anyone worth pairing off with around.

[crazygringo]

This assumes that dating sites are able to give everyone great matches, but are somehow holding them back.

That's not the case. They don't have much idea at all who you're going to hit it off with. And most in person first dates don't lead to second dates, much less leaving the site.

So no. The reality is that dating sites really are trying to give you the best matches, but it's just a numbers game. So they make money on the numbers -- to see more profiles or send more messages you need to pay more.

That's all it is.

Because if they really could reliably make high-quality matches all the time, they could charge $$$$$ for that and make much more money in the end. But they don't, because the algorithm just doesn't exist.

[Dr_Birdbrain]

I don’t work at a dating company, but I do work in machine learning applications.

My best guess is this: they are not optimizing for good vs great matches, and they are probably not even building a model of what that would even mean, not even trying to represent the concept in their algorithms.

Most likely they are optimizing for one or more metrics that are easy to measure and hence optimize, and these metrics have the side effect of producing an excitement for the user without actually pairing them up.

Example metrics: - time spent on the site

- times they “swipe right” or whatever

- messages sent

- money spent

[smelendez]

I’ve wondered about this. Presumably they have some idea of who you will initially match with?

Maybe they have enough data to say things like “when someone like user x matches someone like user y, they are relatively likely to both stop using the app within a month?” But that has to be so noisy.

[xigoi]

> This assumes that dating sites are able to give everyone great matches, but are somehow holding them back.

From what I’ve heard, OkCupid used to be really good at finding compatible people, then it got deliberately nerfed when sold to Match Group.

[Xeago]

What about an alternative business model, pay-per-date? There's an application in NL that charges €7.50 to arrange a dinner, disallowing chat until 2 hours before the arranged date for practicalities. They partner with restaurants and you each get a 'free' (you paid for it..) drink; but with the commitment that your date also paid for it and will therefore show up.

This removes a lot of the meat-grading and endless swiping; with the platform prompting you why you're not working to scheduling your existing matches. Whilst I have no experience with the absence of any scheduled matches, this gives the platform insight into whether you're a worthy date (remember, each date is profit!).

One date on tinder/hinge/bumble in a 5+ years to a finding my partner in a few months. Paying for the actual date experience was so much less and so much more fun than the footing the subscription on the other platforms - even accounting for the cost of food.

[bradlys]

This exists outside NL now too. However, it’s possibly the most shallow app possible. You know literally nothing about the individual except what they look like. For a country like NL where there’s high homogeneity, probably works out. For the US, this is a disaster.

In the US, you’d only get matched if you were a hot guy. It was more brutal than tinder, hinge, etc. Women in the US aren’t gonna spend a single cent on a guy unless he’s mega hot.

[samrus]

This is really interesting. The platform's incentive in engineered more towards finding you a match than to keeping you looking

I guess this is still corruptable. The paltform could make more money by getting you matches thay look good but dont work out. But id imagine thatll only be a problem once they scale and ROI becomes a larger priority

[anal_reactor]

Can you tell me more about it? I live in NL and I'd be curious to try it

[quantummagic]

That doesn't account for the good-will and word-of-mouth generated from any successful matches, which presumably could lead to many more customers than those lost due to marriage.

[raincole]

Very anecdotal, but in my experience people have no attachment to or enthusiasm for dating apps. I've heard (acquainted) couples say the met on dating apps. No one ever said which ones.

[dpe82]

That's a very difficult metric to measure whereas "did this user return and continue paying" is easier. The tyranny of metrics in action.

[bawolff]

I feel like that kind of word of mouth is not enough to compensate. Like how many customers is word of one sucessful match expected to attract?

[bitshiftfaced]

Pie in the sky idea: users sign up and deposit to an escrow. If after x years, the user has been married to another user on the site for 1 year without being divorced, the matching site receives the funds. Otherwise, the user receives their funds back.

Might also work with using the users' registered home addresses instead of marriage. There are ways to game it and ways to make it less game able, but you get the idea.

[dang]

There was a thread about that one:

Working for a Dating Website (2015) - https://news.ycombinator.com/item?id=34368601 - Jan 2023 (196 comments)

[t-writescode]

It’s really disappointing because, a human matchmaker, on the other hand, *does* optimize for “losing 2 customers”. Wouldn’t it be way better for the company’s long-term-health if they charged an appropriate price for making, actually, great connections?

“I found my wife on FindLove” is one hell of a marketing campaign for *future* sales. It’s not like people never break up, and it’s not like people don’t continually enter the dating market or move or whatever.

[ergocoder]

Number 3.

Imagine if you can advertise that 50% of the matches on your app leads to marriage.

[hikkerl]

You might found out the hard way that a lot of people say they want that, but very much want to avoid it.

The main allure of these apps to young women is all the attention from far more attractive men (relatively). Take that away - show her men who might be her "equal" in terms of marriageability, men who might be willing to commit to her - and your service will soon be dismissed and abandoned for only showing ugly men.

You need to sell the fantasy, sell the delusion. Sell hope. The reality hits too hard.

[pwdisswordfishy]

Like, lying? You can already do that.