Here at TechPolicyDaily.com, we have often discussed the potential benefits of usage-based broadband pricing (0). Last month, Netflix (a well-known critic of the practice) provided an unusually stark example of these benefits, illustrating how smart pricing encourages edge providers to become better corporate netizens.
Netflix has long reigned as one of America’s most significant Internet-traffic generators. Network equipment company Sandvine reports (1) that the video-streaming company is by far the leader in peak period traffic, responsible for more than 33 percent of all fixed Internet traffic during peak hours — more than twice the share of the next-biggest competitor, YouTube. This means that at times when the Internet is most susceptible to congestion, Netflix alone is responsible for one out of every three packets sent through the network.
Netflix’s plan to reduce its digital footprint
Last month, Netflix announced (2) a revolutionary initiative that should significantly reduce this digital footprint. Like most other video-streaming providers, Netflix codes a low, medium, and high-resolution version of each title and streams the appropriate selection to a customer based on his or her available bandwidth. But each movie is coded in substantially the same fashion. Netflix coders realized that this one-size-fits-all model is inefficient. Visually simple movies such as the animated feature My Little Pony need less bandwidth to transmit; the background is largely unchanging, and the characters primarily stand around talking. By comparison, blockbusters such as The Avengers involve a variety of changing visual effects against a complex urban backdrop, which requires more bandwidth. Armed with this insight, Netflix has announced that the company will embark on an expensive project to recode its entire collection, giving each title its own set of rules. If all goes according to plan, the project will result in improved video streams while using up to 20 percent less data.
What spurred Netflix’s investment in improved coding?
Netflix’s announcement was met with a chorus of accolades about how the white-knighted Netflix is riding to rescue consumers from the evil cable guy. For example, geek.com’s Ryan Whitwam reports (3) that the recoding reflects Netflix’s “plan” to “beat Comcast’s data caps,” which the cable company allegedly adopted “in hopes of ‘encouraging’ customers to continue using traditional cable TV.” Similarly, Brad Reed at BGR described (4) the initiative as Netflix’s response to “Comcast . . . trying to put the squeeze on them by implementing data caps for Internet users.”
But one could equally argue that broadband providers are encouraging Netflix to be a more responsible netizen. Netflix’s existing coding practices are inefficient, consuming more bandwidth than necessary to deliver the product to the consumer. Given Netflix’s scale and business model, this inefficiency translates into significant wasted network capacity, especially at peak times when congestion is most likely.
Before usage-based pricing, Netflix had little incentive to change its inefficient behavior. Recoding is expensive, and the company had no desire to pay the cost if it saw little gain. While Netflix pays volume-related transit fees on some of its non–Open Connect traffic, CFO David Wells noted (5) that these costs are insignificant compared to its content costs, and analysts estimate (6) that Netflix pays among the lowest transit fees in the industry. So reducing data consumption was unlikely to materially affect Netflix’s bottom line. In other words, Netflix’s inefficiency was a classic case of what economists call an externality: a cost the company imposes on society that is not completely reflected in the company’s own cost-benefit analysis. Consuming excessive bandwidth contributes to congestion. Netflix feels some of that congestion, in the form of downgraded or interrupted video streams. But the rest of that congestion is borne by other edge providers and consumers, who have less shared network capacity with which to do non-Netflix-related activities.
Usage-based pricing forced Netflix to be more mindful of the size of its digital footprint. Because they face potential overage charges, consumers are becoming more aware of the amount of bandwidth their online activities consume. This leads edge providers such as Netflix to develop more efficient methods of delivery, in response to increased consumer sensitivity. The result is a more efficient operation that benefits everyone by freeing up network capacity — which is like broadband providers improving speeds, but without having to install new network lines.
Lesson from the Netflix announcement: The importance of communicating prices
One lesson of the Netflix recoding vignette is that price signals are important. Economist Friedrich Hayek long ago explained how prices are an important way to communicate information throughout a market economy. By putting a price on consumer data consumption, broadband providers communicate that the size of a data footprint is important — which in turn leads edge providers to economize on their data footprints.
This insight would have been useful during the paid prioritization debate. Several commenters noted that in the event of broadband congestion, willingness to pay for prioritization signals to the broadband provider which applications would suffer from a delay and therefore which should be prioritized — just like priority mail helps the USPS decide which packages are the most important to get through quickly.
The paid prioritization ban has prevented price signals from helping broadband providers route network packets efficiently. Now, with its inquiry into usage-based pricing practices, the FCC may potentially make the same mistake again. The agency should appreciate the role that usage-based pricing played in encouraging Netflix to incur a cost that will improve the efficiency of its operations and that of the network as a whole.