Monthly Archives: September 2009

Internet Industry

The Network Effect

Of all the things I learned in during my MBA, the strategic importance of the network effect is the one I see most misunderstood in the tech industry.  The network effect is present when the value of the a product or service is in large part determined by the number of other users of that product or service.  EBay is perfect example of the network effect.  Creating an Internet auction site is not very difficult and is an extremely lucrative business.  EBay had a gross margin of 74% in 2008 because the business never handles physical goods and the cost of running any auction marketplace is low.

So why does EBay have no serious competitors in the US?  Amazon tried to break in to the auction space without success.  EBay’s competitive advantage is the network effect.  If I want to sell something, all the buyers are on EBay.  If I want to buy something, all the sellers are on EBay.  Moving to another site as a buyer or seller means I’d lose value since all the other participants are not there.  Thus an auction site could be better than EBay and cheaper or even free and still not succeed because the value for either party is driven by the presence of the other party.

The principle applies to Craigslist as well.  It would not be difficult for someone to build a classifieds site with better features than Craigslist which still looks circa 1999 .  Web technology has advanced considerably in the last decade but Craigslist is in a time warp.  When I look for furniture on Craigslist, it’s incredibly annoying that I can not see my search results with pictures without having to click through to every listing.  However, I do not use a more functional classifieds sites since the number of listings is too low to be useful even with better features.

Facebook also benefits enormously from the network effect.  Social networks are useless unless your friends are there.  Since Facebook is the largest and almost ubiquitous in some circles, the only general purpose social network it makes sense for people to join is Facebook.  Thus, social networking is very likely to have an industry structure similar to Internet auctions where one player dominates and the 2nd and 3rd place players are dramatically smaller.  To counter the network effect, the smaller networks like My Space have to carve out a defensible space where Facebook can not effectively play for some reason.

Standard
Internet Industry

Content Economics

A few months back there were a series of analyst reports and subsequent blog posts about how much money YouTube was losing for Google.  What struck me about this coverage was that many people seemed surprised that YouTube would be hemorrhaging money since they had a lot of users.  Since then I’ve realized much of the industry does not think about the economics of serving content.   The smaller content is in terms of the bytes the better off the provider is.  Large content like video consumes tons of bandwidth and storage both which cost real money.  User content like YouTube uploads is even worse since there is a lot of it store and much of it is rarely if ever used.  Even photo storage and serving is a significant expense.

To offset these costs, heavy content needs superior monetization and since consumers rarely pay for content on internet regardless of what the old media types want to believe, this means advertising.  Video has the advantage that video ads have much higher CPMs than standard display.  However, any user generated content suffers from the problem of much lower CPMs than editorially produced content since the brand advertisers are wary of associating their brand with such unpolished and potentially offensive content.

Profitability data on individual properties is hard to come by since most properties are part of the big providers like Google, Yahoo, AOL and MSN but if it were available, I’m confident that the ratio of revenue per view to bytes served per view would be a strong predictor of profitability. As an approximation, ad CPM / bytes served per view would work and ad CPMs are easier to estimate from the outside since the CPMs of most ad types does not have that much variance across similar properties.

At the other end of spectrum of the content, Facebook recently announced they are cash flow positive.  Most of Facebook’s content is the news feed which is very small in terms of bytes.  Facebook does have the challenge of keeping track and assembling small pieces of content which has its own expenses but from what I’ve read, Facebook engineers have been able to make that process reasonably efficient.  Facebook also had to invent a new type of advertising to get to positive cash flow.  Traditional display advertising has never been high CPM on user generated content.

I haven’t done enough analysis to know whether YouTube can be profitable on its current model.  Google has advantages of both scale and scope.  From their other businesses, Google knows how to run massive data centers cheaply.  Being a huge purchaser of bandwidth and infrastructure hardware surely gives access to rock bottom prices.  If Google does mange to make YouTube profitable, these other advantages will protect them against new entrants.  My guess is that without a new advertising model that raises their revenue per view, YouTube can not over come the costs of serving video for free.

Standard