Facebook is really pushing the limits to explore the boundaries of the privacy versus targeted advertising efficiency tradeoff. As I talked about before (see here and here), Facebook has a lot of data about its users that in theory allows them to target more appropriate ads to people, which makes Facebook’s platform valuable to advertisers. However, more aggressive use of the data probably requires a reduction in privacy, which people might care about. So Facebook’s tradeoff is to increase its efficiency to advertisers, versus making its users upset. This is a very interesting non-price aspect of its business and obviously no one really knows where the optimal solution lies, so Facebook’s been experimenting a bit lately.

At the bleeding edge of the privacy/matching efficiency frontier right now is Facebook’s ‘Beacon’ service, which links Facebook to online retailing partners. It seems to work like this: When you go to a partner retail site, the site checks for a Facebook cookie on your computer to see if you’ve got a Facebook profile. If so, data about what you purchase gets sent to Facebook and referenced to your Facebook profile. Thus Facebook gets more interesting data about the preferences of its members, which allows them to target ads ever more efficiently in future. In addition to this, things that you purchase may be displayed on your Facebook profile for your friends to see, unless you tell Facebook not to do that. This gives extra publicity to products and might generate extra sales if people tend to buy what their friends also buy.

As documented in this post on the NY Times Bits blog, the Beacon service has been evolving quite a lot lately. It seems that people aren’t too keen about having their purchases displayed on their profile (a number of potentially embarrassing products come to mind …). It’ll be interesting to see how this evolves in terms of how much revenue it generates for Facebook from advertisers or retailers versus how much it annoys their users. To my mind, it’s one thing for Facebook to know what I’m buying and use that data to show certain ads to me, but it’s quite another thing for Facebook to tell my friends about what I’ve been buying. I’m sure that people who are more paranoid about privacy than me can think of many more bad scenarios about what could happen to this data.

Relatedly, Tim Harford wonders if Facebook is doomed because it’s becoming too annoying. These latest developments have me wondering the same thing. It’s really not that hard for people to switch to an alternative social networking site. Sure it’s a pain to set up your profile again and add your friends, but if Facebook tramples its users’ privacy too much, I expect to see a competitor with a higher default level of privacy emerge and take users away from Facebook. Exponential growth doesn’t last forever. Let me finish with a graph:

It seems that everyone’s hiring research economists lately. Well, by ‘everyone’ I mean Google, Yahoo! and now Microsoft. Google’s econ team is headed by Hal Varian, who was doing research on Internet economics way before it was fashionable. Yahoo! has R. Preston McAfee as its econ guru, who is author of the superb open-source microeconomics textbook Introduction to Economic Analysis. Prof McAfee has also written many many interesting research papers about auctions, mechanism design, competition policy, and lots of other things. Now browsing through the latest JOE, I noticed that Microsoft is also advertising positions for ‘research scientists’ with econ PhDs.

So since it’s highly likely that Hal or Preston or Bill (Gates) are reading my blog, I thought I’d mention this paper and this paper that I wrote with my coauthor John Kennes, and which may be of interest.

A two-sided platform is a business that creates value by bringing together two distinct customer groups. I’ve blogged about these before — see here. Basically, the platform allows interactions or ‘matches’ between customers in the two groups. Two-sided platforms are common Internet businesses, because these are often information-based businesses, and the Internet reduces the costs of collecting and disseminating information. As a concrete example, think about Google’s search. It attracts web searchers by providing an efficient search engine. It then matches searchers with advertisements, and shows relevant ads to the searchers. The whole thing is a two-sided platform that creates value by matching searchers and advertisements. Facebook, Digg, Google’s mobile platform and iTunes are other examples.

The thing is, in many cases, not all matches between the two sides of the market are created equal. As I talked about before, the ‘quality’ of a match often matters for how valuable it is. In this case, we should not overlook the ability of prices to improve quality. One function of a price is to prevent people from buying something, thus you can use it to keep low quality customers out of the market and prevent them from making bad matches with your other customers. I can’t remember where I read it, but the following sums it up very nicely for me: Why do (heterosexual) nightclubs typically charge men for entry and not women? Because women prefer quality, while men prefer quantity. Setting a price for men keeps the low quality men out of the club, and increases the quality of potential matches for women. This increases the willingness of women to go to the club, because they know it will only be populated by men who found it worthwhile to pay the entry fee.

Now we can apply the same concept to online platforms. One of my pet annoyances at the moment is Google Gadgets. These are little software applications that you can add to your website or your personalised Google homepage. The gadgets do things like show weather or stock reports, and so on. My complaint is that the average quality of gadgets is low. There are some very good ones, but quite a few useless ones, and many that I’ve tried have bugs and so on. I think this is in part due to the fact that Google does not charge software developers to create gadgets. In fact, it offers funding to developers. While this will increase the quantity of gadgets available, it only serves to decrease quality, I think. Thus the value of matches between gadget users and gadget developers is low, and the value of the platform itself is probably less than it could be.

I admit that it’s difficult to know whether it’s more important to have a large quantity of applications available, compared to a smaller number with higher average quality. Personally, for software I think applications users have a high preference for quality. Apple also seems to think so, based on its strategy of a relatively controlled environment for its iPods and iPhones. It’ll be interesting to see whether Google takes steps to improve the quality of its gadgets over time.

Along similar lines as infidelity offsets, a new website called Music Donations allows you to appease your guilt about copying music by making donations directly to artists. As they say on their site:

You are still guilty of a crime whenever you copy, download or pirate music, but at least the artist receives money from your donation and you can feel better from making that donation to them.

A nice idea, maybe, but the way their system is set up at the moment suffers from a huge credibility problem — how do I know the money that I donate really gets passed on to the artist? It’s not like I can email Robbie Williams and ask him whether he received my five dollars. I guess an intermediary is useful, because it would be quite troublesome to find the necessary information to actually send five dollars to Robbie. So Music Donations can fill the gap by collecting this information and exploiting economies of scale in transaction costs. However, I’d be much happier if Robbie at least had a link to it from his site, or there was some way of auditing the donations, or something to make it more transparent. I don’t recommend using this service at all given the way it’s set up at the moment, because I’m highly suspicious that it’s actually a scam.

A new online journal in mathematics, Rejecta Mathematica, only publishes papers that have been rejected from traditional peer-reviewed journals. I’m not 100% sure, but it doesn’t seem to be a joke. I don’t know about mathematics, but in economics getting rejected by a journal is not necessarily a sign that your paper is bad. The refereeing process contains some signal, but also some noise. However, I don’t see how a rejection can be a sign of anything good per se, and surely the expected quality of a rejected paper is less than the expected quality of a not-rejected paper, unless the refereeing process is seriously biased or flawed. I also don’t really see why authors would want to publicise that they’ve been rejected by publishing in a journal that specialises in rejected papers. So it’s hard for me to understand how this journal is valuable and why authors would want to use it. The one possibly good argument the website gives is that by reading this journal you can find out what doesn’t work, so you don’t waste time trying it yourself. Maybe, but I think much better ways to improve traditional journals seem to be (a) speed up the refereeing process, and (b) reduce the time to publication.