Friday, November 21, 2014

Zero Rating is a Normal Part of Content Services, on Internet, or Not

Ironically, zero rating of Internet apps--opposed by some because it favors some apps over others--is a standard practice in other parts of the media and content universe. Consider newspapers, magazines, broadcast television and broadcast radio, which are zero rated.

Some argue zero rating is unfair because it favors some apps over others. One could make exactly the same argument about TV and radio stations, newspapers, books or magazines. Ownership of broadcasting licenses, or simply editorial discretion, favors some content providers over others.

How regulation of Internet access fosters or hinders application and software innovation is a legitimate policy concern. But there are huge private financial interests intrinsic to the policy concerns.

Zero rating is about revenue models, for example. Many app suppliers subsidize content consumption by selling advertising, and not charging end users directly. Zero rating is another form of doing that, especially where an app provider and a specific Internet access provider have a business arrangement where the app supplier pays the access provider for data consumption.

The point is that “non-neutral” pricing and availability are a fundamental part of the content delivery business. And much of the Internet’s top activities are about content products.

Analyst John Strand points out that network neutrality rules in the Netherlands have, in one important instance, not lead to a flowering of new apps from new providers. Instead, Netflix has grown from zero consumption of Internet capacity to 20 percent of all downstream network capacity almost immediately.

“Neutrality” rules do not, in other words, prevent highly-popular or large content providers from gaining dominance in a market. Neither will an absence of zero rating rules prevent markets from operating, either.

No comments:

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...