As I recall, OKCupid offers a way to only search for new friends, etc. I'd be interested in how many people on OKC have that marked in their profile, and how many successfully find new friends that way (that statistic may be harder to obtain).

OkCupid is pretty popular with a lot of people living "alternative lifestyles" although it's probably only a few percent of their main customer base.

We never worry about this. Helping people find happy relationships is hard, and there's no risk of us succeeding so completely that there's nobody left to pair up. People in happy relationships also tell their friends that they met on OkCupid, and we rely on that fact almost exclusively to bring new members. We spend no money whatsoever on marketing.

There isn't only a customer retention issue with people who do end up in relationships, but also those who don't. I'll admit that I am now on my third OKC profile as I usually try for a while and then give up. A few years later I try again. From the comments on other people's profiles, it is obvious that I am far from the only one who has experienced this.

> surely they stop using the site altogether? An interesting challenge.

Nope, a very simple one. Just keep the experience mediocre. Not bad enough for them to leave dissatisfied, but not good enough for them to leave satisfied. This will maximize their time on the site. This is largely the result of the monthly subscription model, which creates a fairly twisted set of incentives.

Interestingly, the old school matchmaker model doesn't suffer from this problem. You pay a fee for a successful "match". This creates an incentive for the matchmaker to do her best in finding the right match. This model is out of fashion now, because people prefer to have more control over the process and dating is no longer always about marriage.

Also, most relationships have a shorter duration than most leases, providing a very lucrative repeat business stream.

This assumes people are faithful in their relationships right? I wonder if any dating sites have quantified the churn rating for their usage and does it correlate with wider trends in dating such as average length of relationships % of people who cheat on a regular basis. Or are the subset of people who use dating apps those who tend to be in and out of relationships more?

Starting a dating site with fake or scraped profiles just makes it a ghost town. New members would leave quickly before you built up a real user base.