After 2 years running product at Yahoo! Personals, it’s fun to see the ongoing interest in the category; in addition to Ignighter, my TechStars companions, I’ve probably talked with at least four other companies in the past year who are building online dating applications.
PCWorld has a piece today about the great revenue an online dating site can generate, the need to safety and security, how search works, converting users to paying subscribers, and whether they work or not (ie do you find a date and/or a partner). Robert Mitchell’s done one of the more articulate and through overview pieces I’ve seenand a great series overall,, but some additional points worth noting:
a) There are a couple of different ways to manage the algorithms: test based and attribute based.
- Services like eHarmony and OkCupid are test based; they ask a user to answer a series of question and give high weights to the scoring to assign and create a pool of prospective dates for someone.
- Other sites, like Match and Yahoo! are attribute based; users fill our a detailed profile and some–not all-of the responses are weighted and then matched to deliver the results.
- Test based sites are believed to deliver higher accuracy because they can aggregate a set of behavioral, values-driven and psychological results and use them to factor the match, users, however, prefer to higher picture value and ability to browse that non-test driven services offer.
- Some services, especially some of the free and niche ones, like Plenty of Fish, don’t invest much in search at all; once a fairly brief set of attributes are defined(like age, where you live, etc.), the user is presented with a set of profiles and photos to page through; some of the high time spent on site metrics from these free services have to do with the need to page through and review lots of people to find ones you like.
- Both kind of services compute results through weighing different elements: data you have contributed, behavioral data, results acted on by people like you (ie someone similar to you liked a specific profile), part of the dark art of the category is adjusting how much weight to assign to different elements to produce the best results.
- Unfortunately, while search, weighing algorithms, and computing results take significant investment from dating companies, they’re features users tend to take for granted; since a user only sees what he or she sees, there’s not much of an ability to objectively define better search.
- (Another interesting point: typically, men want a large selection with many photos, so they can gather lots of women of a specific type, say a 35 year old, athletic blond, and message multiple people who fit their criteria and have a look they like; women, on the other hand, wish the service could produce one or two men specific and perfect for them and just deliver these super matches up.)
The real bread and butter of the online dating experience though, isn’t the match, it’s the communication . Not only is it essential to NOT get emails from Russian brides, Nigerian spammers, sex offenders, and criminals, a service only works if people communicate with one another. Specifically, it only works if people communicate with you.
Sending missives out into cyberspace that no one responds to, or only getting contacts from people that don’t interest you is a quick recipe for dropping that particular service (and users generally have 2-4 they try out at any given time.)
So, the real mission of a dating site, once you’ve delivered up real people in a secure setting who are an appropriate match, is to trigger communication between them.
This is why games, online chat, anonymous calling, winks, emoticons,virtual gifts, email and so on are always prominently featured-they’re efforts to foster a light, yet meaningful communication. It’s getting communication from a potential date that is going to push someone who is not yet a subscriber into signing up.
Of course, the true business of online dating is to get people to sign up and pay for the service. while advertising may bring in some revenue; for many services either monthly fees or service-based payments (like LavaLife’s pay to communicate) are what makes it work.
For those services, the pricing, offers and conversion pipelines are everything.
As Rob Mitchell points out, this is a category with a huge upside–not only does it attempt to facilitate a core human need–to connect–it has a privacy, search, communications and filtering model that enables there to be a charge to make it happen,