A matching problem

by Eszter Hargittai on April 1, 2006

This year’s Google April Fool’s joke is Google Romance, a service that will help you find your romantic match. It’s sort of cute, although I think some of their past jokes have been better.

The site does bring up something I have been meaning to blog about so I’ll take this opportunity. It concerns the paradox of matching services such as dating Web sites or job search sites. I haven’t thought about this issue too much, but enough to blog about it. (What’s the threshold for blogability, by the way?:)

Services such as dating and job search sites promise the user to find a perfect match, whether in the realm of romance or the labor market. But deep down, is it really in the interest of these sites to work well? After all, if they do a good job then the seekers are no longer relevant customers and the sites lose their subscribers.

One way to deal with this is to offer additional services that go beyond the matching process. For example, the match-making site eHarmony now has a service for married couples. It is an interesting idea. It seems like a reasonable way to expand their user (subscription!) base so they are not dependent on keeping matchless those whom they promise to connect. Moreover, I can see that they may have quite a loyal user base in those whom they helped find their matches. Job sites can also offer services that go beyond the initial match. Nonetheless, I think there is an interesting tension in all this.

On a not completely unrelated note: Happy Birthday to GMail! Fortunately, that was not an April Fool’s two years ago. I came across the Google Romance notice on Google’s homepage, because I saw the GMail birthday icon and wanted to see if they had it in bigger on the Google homepage (a page I never visit otherwise, because why would I in the age of search toolbars). The birthday image is not reproduced there, but I did see the Romance link. (Yes, I’m obsessed with knowing how people end up on various sites and I’m projecting here by assuming that anyone else cares.)



Brett Bellmore 04.01.06 at 11:21 am

While it’s true that successful matching services probably don’t get a lot of return business, they DO get a good deal of new business by referals from happy customers. I certainly recommend the site that introduced me to my fiancee: http://www.filipinaheart.com

They seem to make most of their money by providing services like floral deliveries… It’s not like they’re charging much more their basic service!


Sean Carroll 04.01.06 at 12:04 pm

About the threshold for blogability: mathematicians have proven pretty rigorously that the “least bloggable unit” is, in fact, zero.


Sebastian Holsclaw 04.01.06 at 2:17 pm

It seems to me that the we have the same problem with dating services as we do with a company finding a cure for AIDS–and the same economic answer. While the industry as a whole would be hurt by someone finding the formula for the perfect match, developing the reputation of being the perfect match company would be fantastically lucrative for whichever company actually found it.

As far as relationships go, what is the perfect match anyway? I suspect it is something like a good complement of present personalities and present life situations. There are vast numbers of people who won’t (or perhaps can’t) maintain a good complement over long periods of time as the personalities and life situations change. These people are eligible as repeat customers on a long horizon (maybe ten years) without necessarily feeling that the service did a bad job.


FXKLM 04.01.06 at 5:55 pm

I got a laugh out of the Google Romance philosophy: Don’t be medieval. I’m surprised I hadn’t heard that before. It’s such an obvious parody of “don’t be evil” especially for an innovative technology company. Has that joke been around for a while and I’ve just managed to miss it?


soubzriquet 04.01.06 at 6:55 pm

4: `Don’t be evil’ is Googles (official) policy…


soubzriquet 04.01.06 at 6:56 pm

2: Sure, but we only have asymptotic bounds on that :)


FXKLM 04.01.06 at 7:51 pm

Soubzriquet: I know that “don’t be evil” is Google’s official policy. I get that “don’t be medieval” is a parody of that. I’m just surprised that I hadn’t heard it before since it seems so obvious in retrospect.


vivian 04.01.06 at 10:28 pm

I go to google’s home page periodically just in case they have a special logo. With a fast connection, it doesn’t waste time to click the search button without terms (which takes you to the main page) while thinking about the precise query phrasing.


Allan Friedman 04.02.06 at 2:33 pm

After all, if they do a good job then the seekers are no longer relevant customers and the sites lose their subscribers.
Doesn’t this apply to a huge swath of services? If I cure your specific disease / settle your lawsuit /integrate a specific tech into your company, then part of a successful job completion is obviating your need for future services. I am with Brett that the reputation gains from a successful match are better than the return business of a single customer (exponential vs. linear growth).

Even if having 20 somethings hooked on a single service for 10 years works, think about the slightly sub-optimal, profit-maximizing match–one ends in a predictable time frame but maintains user confidence in the underlying matching system. Is this possible? What are the factors that maximize this without jeopardizing either short-term gains (less buzz) or the long-term gains (less repeat business)?


Tom T. 04.02.06 at 11:59 pm

The right way for a match-making site to maximize profit would be to diversify into other stages of life: a bridal magazine, a day care, a for-profit college, a career service, a funeral home.


Tracy W 04.03.06 at 12:20 am

This analysis implies that such dating services can both:
a) get people together
b) having gotten people together, subtly stop them falling in love.

I can think of ways to unsubtly stop people who have met from falling in love, but methods such as shooting anyone whose first date went well would rather defeat the purpose of getting return custom.

I don’t think love is as controllable as this analysis, or the websites themselves, imply. There’s always enough chance in who winds up together that I think any policy of, say, assigning people who love biking to meet up with people who love knitting would result in as many marriages as assigning people who love biking to meet with people who love kayaking.


Scott 04.04.06 at 11:50 am

But deep down, is it really in the interest of these sites to work well? After all, if they do a good job then the seekers are no longer relevant customers and the sites lose their subscribers.

Same can be said about any govt program favored by either the left or the right.

Comments on this entry are closed.