As I posted earlier, some iPhone users are real bandwith hogs. Now Farhad Manjoo at Slate has proposed solution that does not necessarily require (or more accurately, doesn’t rely on) AT&T building out a lot more capacity. His simple solution is to introduce tiered pricing (The iPhone Is Not an All-You-Can-Eat Buffet). As he notes, adding more capacity may only exacerbate the problem:
Right now there’s only one thing stopping me from using my iPhone more—the network’s too slow. If I get a big e-mail attachment when I’m away from a Wi-Fi network, I normally don’t download it, because it would take several minutes to come through. … But if the network worked perfectly all over San Francisco, I’d probably change my behavior radically—I’d download every attachment, browse the Web like mad, and start streaming music every time I left the house. A lot of other iPhone owners would do so, too—and people who’ve held back on getting the iPhone because of network problems would sign up. And just like that, the network would crawl to a halt.
In some circles this is known as Braess’s paradox. Tiered pricing based on usage doesn’t suffer from this problem because, like congestion pricing on roads, it gives people an incentive to reduce usage.
There seem to be two stumbling blocks to making this happen (three if you count push back from Steve Jobs). One is simply whether AT&T’s user agreements allow it to impose an aggressive enough pricing plan quickly enough. The iPhone 3G launched less than two years ago and users were locked into a two year contract. For this to work, they need to move the installed base and not just new customers to the scheme. Second, the article reports that the average iPhone user consumes 400MB per month and proposes a charge of $10 per MB. The real question is how the distribution of usage is skewed and how it varies over the time of day. It may well be that there are only a handful of real bandwith hogs (say north of 700MB per month) but given all the problem with the network they must be doing a lot of network activity at off peak times. The may reduce their usage but if they do so off peak it doesn’t solve the problem. The real problem may be that too many “lite” users near the average usage level all try to use their iPhone at the same time. Without a time based charge, it is not clear that the problem goes away.
UPDATE: So iPhone usage is heavily skewed toward a few heavy users. According to PC World(AT&T Wireless CEO Hints at ‘Managing’ iPhone Data Usage, Oct 7, 2009):
But all that data usage is not evenly spread across AT&T’s wireless customer base, [AT&T Wireless CEO Ralph] De la Vega says–far from it. He cited AT&T research showing that just 3 percent of AT&T’s smartphone customers [read iPhone users] use 40 percent of all smartphone data, that they consume 13 times the data of “the average smartphone customer,” yet represent less than 1 percent of AT&T’s total postpaid customer base.
So it is not immediately clear that curbing that small slice of users will free up enough capacity for everyone else who is trying to use their iPhone on the train during their morning commute.
Update: The Wall Street Journal had an opinion piece about the woes of mobile web browsing (The Coming Mobile Meltdown, Oct 14). Here are three fun facts from the article:
A single YouTube viewing consumes nearly 100 times as much cellular bandwidth as a voice call. In Asia, some 200 million people already watch video on their smartphones.
Data collector AdMob reports that mobile Web page requests grew 9% from July to August—a 180% annual growth rate. And Motorola recently went public with worries that a handful of mobile Slingbox users (a video streaming device) could wipe out cell service in a whole neighborhood.
It too discusses usage-based pricing and concludes:
Mobile browsing could quickly become a very different experience for users, and require an entirely new approach by Web businesses. After all, flat-rate pricing inspired customers to sample the Web and its services a lot more freely than they otherwise would. Still, the resistance of Google, Microsoft, Amazon and others to usage pricing may be futile unless they’re willing themselves to subsidize delivery of their services to mobile consumers—which would turn net neut precisely on its head.