Online dating has become mainstream, but I still find the whole endeavor frustrating. I’m on OkCupid and though I’ve been in a couple of relationships because of it, I sometimes feel like the whole process just isn’t as efficient as it should be.

The Space

There is definitely not a lack of alternative dating sites out there. There are sites that are for finding a specific ethnicity or religion (e.g. J-Date, RedBean, etc). There are some novel approaches that aren’t exactly mainstream including:

• Thread: Now defunct but the premise was to have your friend (found via Facebook) be a match maker. Good concept, but I think as much as online dating has become acceptable, we still don’t want to expose it to our friend graph.

• DuoDater: The concept here is to have you date in pairs. I had a similar idea as a sort of speed dating/mixer concept awhile back, but I’m not entirely sure how viable this is.

• Some Facebook Integrated thing on Hacker News: I can’t remember the site because I don’t think it took off, but basically there was an essay arguing that integrating Facebook would solve some problems, for example it would hide you from your friends. Not sure the privacy messaging was correct on this because I don’t want to connect Facebook to a dating site.

• Nerve Dating: I’m not sure what the hook is here. Maybe a design it and they will come? I think they’re trying to emulate a Facebook/Twitter method of getting more user generated content that will somehow lead to better matches.

There are of course the old guard:

• Match: I’ve done this site and it’s basically a meat market. It takes everything about the offline world of dating and brings it online, along with everything that makes it hard.

• eHarmony: Extensive, and anecdotally exhausting, process driven. My mom is on this.

• OkCupid: Data driven. Still my favorite as they do some things differently, but it would seem that innovation has ceased since they got acquired.

And there are plenty more.

What to Do

Now here is a basic summary of what I see are problems in the current dating realm and some possible solutions.

Mind you, this is the perspective of a 32-year-old hetero male living in a large city. I’m somewhat shy but not anti-social. I have less than average self-esteem, but I think most people on dating sites might consider themselves that even if they wouldn’t say that in a questionnaire. I’m also looking for a relationship.

Mutual Attraction

I spend a lot of time browsing profiles, but it’s damn near impossible to figure out if it’s a waste of time to message someone. It’d be nice to know that someone I’m attracted to would be interested.

Profiles are often too generic to determine this. Everybody wants someone intelligent and funny. They often want someone outdoorsy or into activities.

OkCupid attempts to solve some of this with their questionnaires, their rating system, and with their visitor list. I think this is a great start, but more information is to be had.

For one, it’d be nice to see who the person has visited who is visiting me. That is, if I see all the photos of guys that a girl I like has visited and I would stick out like a sore thumb, then I’m probably not going to waste my time trying to message her. This is sort of self-defeating and the romantic in me thinks that I could break through, but let’s get real.

Playing on this idea, it’d be nice to pull out interests and characteristics from the set and highlight things that she is looking for. This saves me the trouble of messaging someone who isn’t interested in me and saves her from ignoring, rejecting, or otherwise being bothered with someone she isn’t interested in.

This also plays into the theme of using behavior (conscious or subconscious) to drive preferences rather than stating them. I think OkCupid did a great blog post on how bisexual people really are by looking at the gender of profiles they visited.

Further expanding on the idea above, I think there should be more detailed, if obtuse, ratings. That is, the current rating system in OkCupid doesn’t provide enough data to be useful. I rate people high if I like how they look and if their profile is interesting. This should be broken down into more ratings. Like, “I think this person is physically attractive”, “I think this person is funny”, “I think this person has an interesting profile” (more on this one later).

Rejection

One of the crappy things about meeting someone in real life that really hasn’t been solved by online dating is rejection, both the fear of giving and receiving it.

The norm seems to be that if you’re not interested then you just don’t reply. But at the end of the day it has to happen, either implicitly or explicitly. So why not force it?

I think an online dating site that made a response mandatory would make the process a lot more open and transparent. Sure you would lose that mystery and maybe quash the hopes of many, but in the end isn’t knowing concretely more helpful and possibly educational?

Here I think I’d change things so that you would have to respond in sequence to any messages you received. You can’t go and respond to the hottie without dumping the loser first.

Also, let’s try to make everyone better and more self aware including those that we don’t want to date. You don’t have to write a detailed message, which can be painful or daunting. You would have canned responses: “I’m not attracted to you”, “Your profile is strange”, “I’m looking for someone older”.

Let’s learn to close the door, properly.

Better More Authentic Profiles

I’ve already mentioned how most profiles seem almost too generic. On the one hand you may put stuff on there that is tangential to you just to make yourself be normative. On the other you strive to be zany, funny, or out there in order to stand out.

I’ve found myself “padding” my profile to project a persona that has broader appeal that is probably truly in my interest. It’s not lying, but is it really that important to me?

For one, let’s provide better profile feedback. Each section or interest item would have a sort of thumbs up or thumbs down annotation. Users will be able to improve their profile based on the feedback from other users who they may potentially date.

Also, I think utilizing the Facebook graph of stuff to like would be immensely better than the free form data that is currently out there. Ideally we would be able to import stuff from our Facebook profile.

This is where integration would be helpful. I’ve already like a whole bunch of crap, why don’t I get that stuff in my online dating site. Granted this can be touchy, but if the messaging is right, I think Facebook integration would be a veritable information gold mine in online dating.

Along with the interests, it’d be nice to import photos from Facebook and know that they were imported from Facebook. We could also examine EXIF data and the like to determine how recent the photo was taken. Photos would each have their own feedback widget so you’ll know you should take down that unflattering blurry photo of you.

The micro content strategy I think can definitely work here. Importing photos from Instagram or other likes as well as providing a place for witty commentary like Nerve can make for a richer, complex profile. What about videos? There’s an underutilized element.

Finally, suggested tags for profiles would be nice. The way I see it, it’d be cool to have people who visit my profile tag me with stuff. The more frequent tags are floated to the top. I wouldn’t be able to change that information. User generated meta data.

Search and Discovery

The OkCupid system of using the questionnaires to find similar people I think was a great step in using data to find good matches. I definitely subscribe to the alike are better for each other than opposites. However, it can be improved.

The mutual attraction data from above would be helpful in the search process. It’d be nice to have people to message that I know should like me as well.

Also, better granularity with respect to distance would be nice. That might just be a problem with someone like me living in San Francisco. Also it would probably show up more of a ghost town that would hinder adoption.

In terms of discovery, I find that surfacing people who aren’t talked to is a useful tactic in play at OkCupid. I would add to that the mutual attraction data as well as the common interest data. Show me people who might actually respond.

Again, using data from rejections, visits, messages, etc can go a long way here.

Misc

There are some other sort of miscellaneous ideas that I’ve had:

• Thread had mixers and I think that having an in-person presence to online dating is underutilized. I don’t know if a mixer is the right thing to do, but definitely something organized would be nice, like meetups for different interests. Or how about a better photo “trade-in”. You go to the event where there is a professional photographer to help take your profile photo.

• Update: It looks like Match is doing this now with their “Stir” events.

• The use a friend to aid in the process is also looking for the right use. Thread and DuoDater obviously come to mind. I would like to see a feature where you could recruit a friend. They could help you with your profile and browse (and rate) people you’ve visited. There’s something to the actual engagement with people, but using them to help with the process has until now been unexplored.

Today I released my first open source project, Hogan.groovy. It’s a simple port of Twitter’s Hogan.js to Groovy, my favorite JVM language, allowing you to render Mustache templates.

This isn’t major (there are several Mustache template engines on the JVM) and it wasn’t a lot of my own code (since it’s a port), but I have been wanting to do this for awhile. When they first released Hogan.js and I took a look at the modular code, I knew it was something I could kick out in a few days (which of course was a few months ago, I got distracted).

Here’s an example of how you’d use it:

One thing that I had liked about Hogan over other Mustache implementations when it came out was the fact that it could traverse up the context stack to find things. It could also use dot-notation to get at deeper items.

Handlebars.js offers some of these improvements as well as a general helper function concept which I liked, but it’s parser was generated and for awhile I contemplated trying to generate the parser using something like parboiled, however once I saw Hogan.js’ really small and simple hand-rolled scanner/parser, I knew I’d rather just port it instead.

Now that I’ve gotten my hands dirty with the scanner/parser (I even submitted a Pull Request on a bug) I might try to add some features to bring it to parity with the other implementations.

But the long term plan is to build a Grails plugin so you’ll be able to render Mustache templates in their views or easily include compiled JavaScript templates that use Hogan.js to render on the client-side.