Price – The Cost of Complexity
Next couple of posts’ll be about pricing. Simple is better than complex; and free, as the simplest of all, is sometimes the best of all. All of this relates to the mystery of why cooperative services like file sharing networks, wikipedia, and, potentially, America’s Antiterrorist Network work.
Premise is the academic one that price is a mechanism for exchanging goods and labor. Best price is the price that facilitates the most exchange. Arguably, that will always be a price within a range where both parties to the transaction are better off for having done the transaction. So much for being academic.
A long time ago I blogged about the success of flat rate, all-you-can-eat pricing in promoting subscriptions to AT&T WorldNet Service, a dialup ISP I founded within AT&T and which, for a brief time, had more new signups than well-established AOL at much lower cost of subscriber acquisition and which forced AOL to go to flat rate pricing. We didn’t invent flat rate pricing but we popularized it and had the strength of the AT&T brand. Also wrote about the success of NetFlix using the flat rate or subscription pricing model.
There are still people who think that Internet access pricing should be more complex. Some say that the American expectation for flat rate pricing makes Internet access unprofitable for our telcos and cablecos and so that’s why we don’t have better broadband access available at a better price. This argument ignores the fact that many developing countries have significantly faster access available at significantly lower monthly prices; but still, the argument is being made.
The telcos, at least, would like to complicate broadband pricing by charging differently for different levels of service somehow billed at the content provider level. Of course, they used to charge different rates per minute for calls between any two city pairs in the US and then multiply that with the complexity of time-of-day pricing. But they were a telco monopoly then. Competition did away with that complexity. But competition is diminishing.
From an economics point of view, what these arguments ignore is that pricing complexity has a cost. Often this cost overwhelms any possible efficiency that comes from more pricing complexity.
Back to AT&T WorldNet. We offered both hourly and unlimited access plans. Customers switched from hourly to the $19.95 flat rate unlimited on the average when they were paying about $11/month. Moreover, their usage didn’t typically explode after they changed plans. There were exceptions, of course, but people had only so many hours a day to go online (then) and they were tying up a phone line as well.
So why did they switch and voluntarily pay more money? They did it because they were willing to pay for simplicity and predictability. They didn’t want to pay the cost of complexity themselves so they were willing to pay a premium to outsource the complexity to us. They also didn’t want to take the risk that they’d leave their connection on accidentally and run up a huge bill. We were implicitly offering them “insurance” against that, which made sense, like all insurance, because we could spread the risk across a huge customer base.
Equally important and much less understood is how much cheaper it was for us to offer flat rate pricing than metered pricing. Of course, we didn’t need to accumulate and distribute call detail records (CDRs) for the online sessions of unmetered customers. We didn’t need to provide the customers with access to these records because they had no need to check their predicable bills.
The huge cost difference is that there were almost no billing disputes. These are obviously handled by human customer care reps and are very expensive to deal with. But the whole issue of whether a customer was actually logged on at a specific time goes away with flat rate pricing. Estimates if I remember right was that flat rate customers were about $2/month less expensive to handle than customers using $19.95 worth of metered service.
Churn (customers leaving the service) was lower among flat rate customers although this is not a completely reliable statistic because we couldn’t correct for the fact that the flat rate customers were self-selected. My theory is that billing disputes lead to unhappy customers which leads to customer attrition. Eliminate this point of friction and you keep your customers longer.
Also, in those days, Internet access was not as reliable as it is today. Connections did drop. Things went bump in the night. If someone was doing a long download – remember how long downloads could be in dialup days – and it got interrupted near the end and had to restart, she wasn’t very happy. If she was paying by the hour, she was really unhappy to have to restart from the beginning. The interrupt may not even have been our fault. No matter, we had an unhappy customer and one who might even imagine we had an incentive to hiccup occasionally to make her online bills higher.
Flat rate customers didn’t like having their downloads fail, either. But, from their point of view, we were bearing the cost of the extra online time. They knew we had an incentive to make things work well and quickly.
Next: but what about those who take advantage of simplicity to arbitrage or overuse?