Wednesday, July 6, 2011

Network effects and self-perpetuating incumbents

As part of my travels to and from the latest user innovation conference, I’ve been traveling through a lot of airports.

Airport hubs are an example of both network effects and more general IO economics (i.e. Michael Porter) barriers to entry and supplier power. With enough market power — such as by reducing rivalry through mergers — large airlines can offer mediocre service at a high price and people will be stuck paying it. (Or, they won’t fly.)

This is particularly true since airlines have accepted commoditization of their service, with few offering differentiation. (Frequent flyer plans were once intended to offer differentiation, but once everyone copied them, all they do is create switching costs which discourage price shopping by business customers.)

Flying through European capitals this week, I realized how my view of airline competition was distorted by the US experience, where a few Midwest (or non-coastal) cities got nabbed by airlines to build fortress hubs: Chicago (United, American), Dallas (American, Delta), Houston (Continental), Atlanta (Delta), Pittsburgh (US Air), Minneapolis (Northwest), St. Louis (TWA), Salt Lake City (Western) — or secondary hubs like Denver (United), Kansas City (TWA), Cincinnati (Delta), Detroit (Northwest) and so on. A few international airlines had the bulk of their international departures from key “gateway” cities like New York, Miami, San Francisco and Los Angeles, but only later did these become international hubs.

In Europe, the pattern is very different. Most countries had one airline which offered non-stop flights from their capital (or largest city) to key international destinations. And of course for minor obscure languages, the national airline meant customer service in a language you understood.

Only later did they convert their capital to a through hub to attract through passengers. Judging from my efforts to book connecting tickets to small EU cities online, the most successful seem to be British AIrways (London), KLM (Amsterdam) and Lufthansa (Frankfurt).

Airline hub theory states that to make a hub work, you need both connecting passengers and terminating passengers (whether local residents or visitors). So big cities like NY, LA, London etc. have a lot of built-in demand to support the hub.

A few things surprised me about this airline competition during trip. One is the reminder that 40 years ago, every country (even Belgium) wanted its own airline no matter how sub-critical mass it was. Exhibit A: Sabena Airlines.

Meanwhile the Scandinavians (except Icelanders) swallowed national pride and combined the local demand of three countries. Even so, sitting in the SAS hub in Copenhagen today, it seems surprisingly weak compared to my first visit 15 years ago — in terms of size of planes and number of destinations.

I guess the problem is that Copenhagen is the connecting hub but Stockholm has twice the local traffic, so neither is particularly viable. Also, the rise of the Star Alliance and code-sharing means that it’s easier for SAS to code-share with longhaul partners than to offer its own tiny fleet of long-distance jets. (It was interesting to note that nonstop traffic between Japan and Germany is carried by both Lufthansa and Star Alliance partner ANA, with every flight by one code-shared on the other.)

The final observation is the reminder that the natural reaction to rent-seeking and extortionist monopolies (the aspirational goal of most airlines) is to encourage new entrants to do a better job. In the US, we hear about Ryanair (more penny-pinching than Southwest) and EasyJet but not the dozens of others that have sprouted up.

For this trip, to save $500 I bought a separate ticket from Copenhagen to the Vienna conference on AirBerlin. However, I ended up on Niki, the smaller Vienna-based Austrian discount airline that it acquired in 2010. Both seem to be German-speaking versions of Southwest, catering to German-speaking travelers, but (unlike Ryanair or EasyJet) offering slightly more amenities than Southwest (i.e. a free cold meal). I guess AirBerlin (which has longhaul flights to the US, the Caribbean, Africa and Southeast Asia) is a German version of Virgin Atlantic, but without Sir Richard and his out-of-this-world ego.

Niki is strictly a European short-haul carrier. Even though it is tiny (21 planes), it was clear at Vienna airport it was cutting into the business of Austrian Airlines (the former national carrier acquired in 2009 by Lufthansa). In Copenhagen we had Sterling and Norwegian, and of course European authorities showing (a limited) interest in competition has transformed the airline industry to make these carriers cheaper and quicker (if not quite as convenient) alternatives to the monopoly national train systems.

But then, that’s not that different than the US, either. American’s attempts to control Dallas passengers helped fuel the success of Southwest. Delta’s control of Atlanta begat ValuJet (rebranded to AirTran after that terrible crash) which is now a division of Southwest.

Interestingly, it seems that big destination cities are impossible to dominate. No single airline has dominated NYC or LA the way the middle-American hubs have been dominated. (Perhaps PanAm did so once upon a time, but that was well before my time). And if you look at London, while BA is by far the largest carrier, it has lots of competition.

In the end, it comes back to a first-year strategy principle: businesses hate competition and choice (except with suppliers) and customers love it. The entire field of IO economics is about how incumbent firms can develop monopoly (or oligopoly or monopsony) strategies and how new entrants and customers conspire to destroy them.

No comments: