Hacker News new | past | comments | ask | show | jobs | submit login
Why We Think We Can Prevail in the Notoriously Crowded Online Dating Space (date.fm)
28 points by achalkley on Dec 6, 2011 | hide | past | favorite | 40 comments

This doesn't solve anything. The core problem is that women get tons of spam messages and men get very few responses. This means the matching algorithm is broken. This site is just a tweak on hot or not.

There are a few core problems here.

The first is that dating sites typically don't care about people actually finding a good partner, because there's a negative economic incentive for it. If you find your perfect man/woman, you won't come back to the site and keep paying.

The reason why women typically get loads of spam is because many sites are expressly designed to encourage this. A common pattern is where it is not made clear whether or not a profile is active. Since only a small percentage of profiles on the site are active, it's hopeless for a guy to try to send detailed, personalized messages -- odds are, they'll all fall on deaf ears, inactive accounts.

This is intentional -- because it lets the site appear as if they have far more profiles than they actually do. Once they've suckered someone into paying to send messages, they find that they get almost no responses, and end up finding themselves forced to send hundreds just to reach a few live people.

Which of course leaves the recipients flooded by spam... so they leave. Adding to the glut of inactive accounts.

Combined with the general sleaziness of the industry and the type of people who use general-audience dating sites (read: the type who tolerate all these problems), it's no wonder the entire industry is a disaster.

There are definitely niches available for matchmaking sites not built around the typical model, that try to pair up people for long-term relationships based on what actually matters instead of providing sex-starved 20-somethings with people of the opposite sex to ogle. But it's hard to market such a thing, even in a specific niche, because the entire industry tarred themselves with their own brush, leaving a huge number of possible customers permanently turned off from the idea of ever using such a service.

"The first is that dating sites typically don't care about people actually finding a good partner, because there's a negative economic incentive for it. If you find your perfect man/woman, you won't come back to the site and keep paying."

Dating sites that think like this are in no actual danger of having their users fall in love. It's hard enough even for well intentioned sites to set up enjoyable dates, let alone ones that lead to marriage.

If anyone actually had a dating service that worked so well they were losing users due to excessive happy relationships, that would just mean they weren't charging enough to get in: Imagine what people would pay for a service where 30% of all users end up with the love of their life within a year. Hell, imagine the marketing value of having large numbers of satisfied couples going around telling people they met on a particular site.

>The core problem is that women get tons of spam messages and men get very few responses. This means the matching algorithm is broken.

This matching algorithm has carried us, humans, through millions years. Various opinions and statements can be made about this algorithm, yet it is definitely not "broken" :)

Most women I know who use dating sites don't pay much attention to the unsolicited messages they get. Instead, they browse the available men and send the ones they like PMs.

To be honest, its a lot like recruiting - the people sending their resumes to every company on monster rarely get hired for the best jobs. Those jobs instead go to the "good looking" people on LinkedIn, who are approached directly.

No reason this app can't work the same way.

And don't forget, Hot Or Not was an insanely successful dating site.

But this app only lets users chat--not send out messages. You can copy+paste a generic message to 3 dozen women, but about the only thing you can use generically in conversation is "hello"--plus you have to be actively involved in a conversation. Also, the app insulates against spam because you can only chat with people who have mutually liked each other.

> "I don’t think conversations start with “I don’t smoke, want 3 children, and am Catholic”. But this is basically what online dating sites do: they put a ton of information in a person’s profile, and ask other users to evaluate this data. There are certainly busy people who can’t be bothered with speaking to people, and want a dating service to run algorithms and find them someone. Maybe these people have played the dating game, and have had enough. They actively want something unnatural, and don’t mind if the computer sifts out people who aren’t their carbon copy."

I don't necessarily disagree with most of what else you say in the blog post, but if this is seriously your conception of how people use dating sites, then you are doomed to fail.

Speaking from the experience of living with various roommates who have done e-dating with the popular services, none of them have ever approached them in this mechanical way. The data helps refine the search, but they're still looking for attractive people to chat with.

And that data part is important. If you're a 20-year-old looking to have fun, you're going to want to weed out anyone who is in an age demographic that wants to have a life-long relationship. And vice versa. So not having that data is a disadvantage that you have to make up for.

My friend wrote greasemonkey scripts for POF and OKCupid, all it does is repeatedly search for his "type", builds a list of users and then messages them just after they come online. Same thing if someone views his profile. There's a 15 minute delay for that one so he doesn't come off as creepy.

Yeah it's spammy, but surprisingly effective, I'm encouraging him to write an A/B testing module for it so he can test the effectiveness of different messages.

Contrary to what one might think he's just using this for LTRs.

Are these posted anywhere or could I send an email asking for these? :)

The 6 most important metrics in online dating is gender, location, age, children, active site user, and attractiveness. After that many guys just start spamming accounts.

The only spam they can do is "Like" someone. Sending chat messages can only happen when someone likes you back.

Why not just "Like" every user of the opposite (or target) sex, then just decide if you are actually interested when you see a positive response from one? It seems like that would be the best strategy with this app, and it means you get to see who likes you without actually disclosing any useful information yourself.

If you're not fussy you can do that. You could widen your range to anywhere in the world and ages 18-99. But you may not be within the person you like's age range or geography criteria, so that strategy would only take you so far.

Thanks for your feedback. I've rewritten this bit to hopefully make the point better. I agree this is NOT what users actively do on these sites. But I think this is what the service is essentially doing under the hood in many cases.

I agree that there are people who prefer the time-saving methods used with many of the major online dating services, and they will certainly continue to use those, but we feel they may be missing out on people that they didn't expect.

Why not bring back messages but put a crowd sourced spam filter in?

Instead of matching profiles, match messages, people looking for long term populate one set of bayesian filters, and people looking for short term populate another set of bayesian filters. The spam cut and paste message go away pretty quickly.

The mechanism could be as simple as after a message is read they are prompted as to whether they liked the message or didn't like it, combine with a feature sorting messages by their bayes score?

All of that data is also crucial if you want to sell advertising against your user base.

I don't think we have to deal with that. iAd and Admob provide the ads so we don't have to.

Oh, I had assumed you were selling the app and not going to show ads.

Displaying your own ads might be something to think about - collecting user info over time and delivering highly targeted ads might be more profitable.


I have no idea on how targeted iAds or Admob are. I'd assume there would be some tracking of sorts.

Good idea, I'll check it out.

one comment: "We thought about how most people meet. Generally, a person will see someone, be attracted to them..."

That's not the first step. The first step is context, e.g. you're at a party thrown by a mutual friend.

That first step is important, it's context and that's what needs to be fixed. In the real world "dating contexts" are fairly limited, e.g. there are only so many parties you can go to in a week, only so much time to participate in a hobby with someone else, etc, etc.

I look forward to seeing this app evolve. If you address the "context" issue it could really take off.

I think they underestimate how much of a role context plays in meeting people in the real world. You may not have an online dating style bio sheet on your t shirt, but that you and your crush are in the same place, likely doing the same thing, provides a lot of information that you might be compatible.

More than that, there is social proof, with dating sites you have to establish social proof by yourself, where as in person you're usually introduced, or know others at the event.

This sounds oh so very much like Loopt Mix. While it may seem to be working at first, as the app gains popularity the quality of the average match will fall, the attractive/desirable people will leave, and you'll end up with a sketchy hookup app where people pay for sex with iTunes gift cards[1].

I wish them luck, and would love to be wrong, but I'm skeptical.

[1] Yes, this actually happens in the real world. Nothing surprises me anymore.

You know, we discussed many ideas that have such public elements, like "Where you are" public statements, but we ultimately decided against it because we felt it might be susceptible to stalker behavior. And I don't think a lot of people would be comfortable putting that kind of information out into the world publicly. We kept coming back to the realization that a private chat is where information has to be exchanged.

We also feel that the simplicity will help encourage new users as the months and years go by. We think most people stay away from online dating because of the unnecessary complexity, when really they just want to be able to start chatting with people who like them, that they also like.

We'll keep you posted!

This is essentially a mobile advertising-supported implementation of HotOrNot's MeetMe service.

My understanding is that MeetMe was hackable by 'liking' every single profile just to see who likes you back (although there is still a risk that people on the other side are also 'liking' every single profile which would make the match worthless); curious if this service handles this differently.

I encourage everyone to spend 10 minutes on hotornot.com, and then try date.fm for 10 minutes. Once you've done that, I think you'll see just how different it feels. While our monetization strategy is a simple ad banner, we think it's much better than a complex system of paid accounts, credits and other upgrades that most sites rely on for revenue.

Since your idea is to recreate the experience of meeting in real life (I'm not sure if this makes sense, because people could then just choose to meet IRL - people date on-line for a reason), sell the same benefits that would convince cool people to attend a party and meet up.

Your app is party without a headlining DJ or popular promoter.

Sell the benefits that would convince cool people to go to a party:

- Convince females that your app discourages loser males who will spam them for sex.

- Convince males that there are plenty of females who are not justing wasting men's time.

It seems like those are the benefits you are comfortable offering. There are other benefits to users of dating sites, but it seems like date.fm isn't on-board with the things people normally look for from on-line dating: pre-qualified partners and easy sex.

I'd rather go to a party than meet girls on a phone app any day, but I'll def try out date.fm and give some more feedback, if it gets any traction.

We'd love to hear how you got on.

Argh! If you want to improve dating using technology, stop focusing on how to arrange first dates. That problem is basically solved to 90%.

Focus on how to fix the real problems in dating, i.e. the things that happen between deciding to go on a date and agreeing you're in a relationship. No, I don't know how to solve this in a scalable way. Yes, you will create a lot of value if you can do this.

>the things that happen between deciding to go on a date and agreeing you're in a relationship

...are those which 99% of people don't want to outsource to technology. That's the fun part of dating: the fluttery-chested, "why can't I breathe right oh god does she notice I'm not breathing" excitement. Let's not make An App For That™.

Online dating is a huge potential market.

Whichever service eventually "solves" online dating wont call itself a dating service. People don't use online dating because it is just that - online dating. The social stigma isn't going away.

Facebook can disrupt online dating by introducing you to Facebook users near you with common interests. Unfortunately, I am not convinced a startup can do the same, or I would be doing it.

Would you want to expose people to your entire social world when you're still experimenting with whether or nor your even compatible with them? Dating experiments often go sour, or you might find out that someone is truly crazy. That's why we decided on a completely separate ecosystem that doesn't integrate with any other social platform. When things go bad, we think users would rather be able to quarantine the experience. In Date.fm, users can simply swipe and delete a match from their list, and it's removed on both sides. Neither can bother the other beyond that.

We totally agree about the online dating stigma, but we've got to be identifiable in some way.

Also friend of friends with similiar interests or letting the user match two of his/hers friends that he/she think might get along, etc. I quite surprised that they didn't get into that yet.

I'd be interested to see if the third party could be bothered matching two of their friends. It seems they'd have to go out of their way to use an app (on facebook) to do that.

That sort of stuff would be initiated in a conversation with a singleton.

I can only image they would. I often see people complaining that they don't have anything to do on facebook, what means that they get bored, they want more useless stuff to procrastinate.

We've posted some follow-up comments on how we're trying to improve the quality of Date.fm here: http://blog.date.fm/post/13881727508/how-quality-and-simplic...

Would love to hear your thoughts.

Is there some reason this is an exclusively iOS app? Seems like something generally desirable... if appropriately designed. Or is there some business reason for the restriction?

A number of reasons:

It's to guarantee a certain quality in the user base. Uses GPS to get general location and new iOS 5 face detection to guarantee 1 clear face in the profile pic. So we shouldn't get people just showing their eyeball or a group photo were you don't know who you're liking.

We're also native iOS developers. We're a two man band and have done everything in-house, iOS, Rails and Node development. We've done the promo video and composing of music in-house too. (Wife did the voiceover :)) All to minimise costs.

The revenue from our other iOS apps will be subsidising server costs.

If and when we launch on other platforms, we'll be native and use the APIs to guarantee a certain level of quality both in the App and in the user base.

Generally speaking, depending on how much income we generate will determine expansion on other platforms. We'll have to react to demand.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact