Keeper

1.9K posts

Keeper banner
Keeper

Keeper

@KeeperAI

Engineering love at first match. 1 in 10 today. Dating apps: 1 in 10,000.

NYC Katılım Nisan 2022
331 Takip Edilen6.6K Takipçiler
Sabitlenmiş Tweet
Keeper
Keeper@KeeperAI·
Finding someone to marry is harder than ever. We used to marry for practical reasons and grow into love. Now we hold out for a soulmate. That’s good! Marriages today are much happier. But soulmates are extraordinarily rare, and too many people are struggling to find theirs.
Keeper tweet media
English
18
52
457
103.7K
Keeper retweetledi
Keeper retweetledi
Wes Myers (itl)
Wes Myers (itl)@realwesmyers·
Women want to close the "gender pay gap" and also refuse to date men who earn less than them. Pick one.
English
1
2
23
817
Pisan Milk Hotel
Pisan Milk Hotel@neutralcarseat·
@KeeperAI 6 quintillion dollar ai startup with engineered proprietary algorithm be like: try smiling
English
1
0
10
1.2K
Keeper
Keeper@KeeperAI·
You don't need a new jawline. One of your photos is losing you matches. Find out which one:
English
15
10
252
3.3M
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
The more physically attractive an older man is, the younger the woman he wants.
Jake Kozloski tweet media
English
80
63
995
244.7K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
% of Keeper users with at least one tattoo: Women: 2.8% Men: 3.3% US baseline (Pew 2023): ~32% of adults have a tattoo. Keeper users are 10x less likely to have a tattoo than the general population.
English
21
3
286
55.1K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
Divorce clusters in social networks. Framingham research found that having a divorced friend or relative substantially increases your own divorce risk, detectable up to two degrees of separation. Marriage is a social institution that needs other marriages around it. Communities of intact marriages reinforce each other. Communities that start losing them lose them in cascades. If you want your marriage to last, surround yourself with people whose marriages are lasting.
English
8
14
197
20.8K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
If you had sex two nights ago, you're probably still feeling closer to your partner than you would otherwise. A Psychological Science study found that sexual satisfaction stays elevated for about 48 hours after sex, and the strength of this afterglow predicts marital satisfaction years later. Sex resets the emotional baseline of a relationship for about two days at a time. When that rhythm goes away, couples often drift apart even when nothing specific has gone wrong.
English
13
39
886
85.9K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
"Would you marry someone with substantial debt?" Outright "no": Women: 69% Men: 44% Women are 56% more likely to rule it out.
English
50
7
212
1.5M
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
Women are 2x more likely than men to refuse to change their style for a partner, per Keeper data. Hard no — Women: 42% | Men: 19% A few possible reads: – Style is more identity-laden for women – Men's looks carry less weight in attraction, so less to defend – Women have invested more in their wardrobe, harder to abandon Not sure which is right. Maybe a mix. What would you add?
English
33
8
151
25.1K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
Everyone has a take on dating-app height dynamics. Here's what 219k Keeper users actually said they'd accept: 5'0" man → 5% of women 6'0" man → 93% 5'5" woman → 92% of men 6'6" woman → 16% The curves cross at 5'9". Luckily, all you need is one.
Jake Kozloski tweet media
English
38
27
434
68.3K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
The average Keeper man is a 5'10" 27-year-old with an IQ of 117. The average Keeper woman is a 5'4" 24-year-old with an IQ of 112.
English
24
7
311
33K
Keeper retweetledi
Wes Myers (itl)
Wes Myers (itl)@realwesmyers·
Top 10% and bottom 25% tend to be hardest. At the bottom, "ugly" often reduces to fat, which points to a stack of fixable things left unfixed. At the top, you get people who never had to develop much character. In either case, looks are mostly a tell for the real problem.
Larry Fink Jr@larryfink_jr

@jakozloski Do you have a harder time matching ugly people vs attractive? Apologies for the politically incorrect words but you know what I mean

English
3
2
111
14.1K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
95% of Bumble's users never paid them a dollar. The 5% who paid generated about $28 a month. The math problem the dating industry has spent fifteen years trying to solve is how to extract more revenue from those users. The obvious solution is to reliably match them to someone they'll fall in love with and charge for the value delivered. This has never been tried.
English
32
7
486
53.3K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
If you signed up for Keeper, thank you! We’ve got a huge backlog of demand — more than our system can handle right now. We don’t make a match unless we’re extremely confident in it, and that takes time. We’re working on increasing throughput to get to you soon!
English
0
2
52
2.7K
Keeper
Keeper@KeeperAI·
Not necessarily. The point is that the bias of the rater is taken into account in both directions. So if a typically harsh rater rates a photo well, that will get up-adjusted, and if a typically permissive rater rates a photo poorly, that will get down-adjusted. Both are intended to get the score closer to the latent true global consensus. So in your case, you might have had a harsher than normal set of raters across the board. I'm happy to look up your test though just to confirm everything looks right - I sent you a DM.
English
0
0
1
63
FlowMan
FlowMan@FlowMan92·
@KeeperAI @jakozloski Thanks. The point on "harsh raters" is a bit counterintuitive as effectively this might mean that a photo in question got less "yes" from not so "harsh raters" or even got "no" from them which basically is sign of low quality and probably should be penalized
English
1
0
0
173
Keeper
Keeper@KeeperAI·
It can be surprising how competitive the swiping apps are. Getting a decent amount of yeses is often enough to put you in the 70th+ percentile, although that doesn't necessarily guarantee success. It's also worth noting that the tendencies of the raters is factored in. So if the people who rated your photo 'Yes' are typically very harsh raters, that will result in your score being higher.
English
1
1
1
2.1K
FlowMan
FlowMan@FlowMan92·
@KeeperAI @jakozloski I let it test one photo and the rating is actually great (as I think a person on photo is like 5-6/10), but there are lots of "no" and "somewhat", which make me wonder how come the rating can be 8.4/10. And then also, this photo actually did not help get matches on dating apps.
FlowMan tweet media
English
1
0
0
3.9K
Keeper retweetledi
Jake Kozloski
Jake Kozloski@jakozloski·
Men’s top don’t-want in a spouse is “overweight” by a 2.5x margin over the next item. Women’s top don’t-want is “rude” by a 1.5x margin. Men’s rejection filter is dominantly physical. Women’s is dominantly character. Both sexes’ top wants are identical though: kind and funny. (Keeper data on millions of stated preference clauses.)
Jake Kozloski tweet media
English
215
168
1.9K
450.5K