What The FCC Left Out From Its Zero-Rating Analysis

As Congress packed it up for the holidays last year, the Federal Communications Commission (FCC) was working to push their agenda. The agency seems to have teed up action against both Verizon and AT&T. In two separate letters, the FCC decried zero-rating, a broad term for programs that exempt certain content from counting against consumers’ data caps. Both letters underscore the FCC’s end-run around Congress, which has warned against any controversial items before the new President, and the agency’s continued defiance of measured policy.

FCC Chair Tom Wheeler has presided over more straight-line party votes than any other Chair. Wheeler’s FCC will go down as one of the most activist in recent memory. He pushed the agency to adopt a 25 Mbps speed threshold for commercial broadband, creating a double standard within the agency since their own subsidies support broadband programs at a 10 Mbps level. Using media leaks and political pressure, Wheeler broke bipartisan compromise on a number of important topics. In a rare move, Wheeler changed course on the Open Internet Order caving to pressure from the White House. No new research has been produced, bucking a trend that is decades old. And his effort to overrule state legislation that manages government-run broadband networks was shot down by courts. Now in the waning days of the Wheeler Commission, the agency might make a final move, this time on zero-rating.

In the 2015 Open Internet Order, the FCC decided not to regulate zero-rating programs, choosing instead to handle concerns on a case by case basis. Among a litany of other issues, AAF criticized the Open Internet Order because the rules give the FCC broad discretion, which is on display in letters to AT&T and Verizon on their zero-rating programs.

AT&T’s Sponsored Data program and Verizon’s FreeBee Data 360 each allow edge providers, like Netflix, Google, and Facebook to purchase bulk data that is then exempted from subscribers’ data caps. AT&T plans to roll out DIRECTV Now, which will be zero-rated for all AT&T Mobility subscribers and include content from about 100 channels. The FCC worries that the terms and conditions of these offerings will make it difficult for other video competitors to get a foothold in the mobile market, since the affiliated content providers will have to pay the wireless data rates, which could be cost prohibitive.

The FCC thinks that the choice AT&T presents is especially unreasonable on its face since edge providers must “either pay a Sponsored Data rate (resulting in a $16-$47 per month – or higher – incremental cash cost not incurred by AT&T) that would make it very difficult, if not infeasible, to offer a competitively-priced service, or instead require its customers to pay significant amounts for their own usage of data while AT&T’s zero-rated DIRECTV Now service offers customers the same usage for free.”

Yet, the FCC notes “Sponsored Data rates are similar to the discounted wholesale rates paid by major wireless resellers,” so the cost of transmission is merely being shifted from consumers to producers. The cost of transmission has long existed in the form of data plans and is quite expensive for mobile, yet producers of content haven’t had to bear its consequences in the online mobile world. Since these externalities, as they are known to economists, aren’t included in the production of content, the mobile market might be considered allocatively inefficient. If however, edge providers were forced to bear these costs, then they would find themselves pressured to push for technological advances to economize on bandwidth.

The FCC’s comparison considerably narrows the market for online video and the competitive landscape. According to Neilsen’s most recent report, consumers watch just over 2 hours of video on a smartphone each month. This is far less than the 138 hours watched on live and time-shifted TV each month, and less than desktop video watching, which accounts for about 12.5 hours per month. Far and away, apps and general web surfing are the dominant players on mobile with consumers spending 46 hours there instead.

There are three other related problems with the framing of the FCC’s letters.

For one, zero-rating is simply an innovation on transmission delivery, yet both content and conduit are needed for consumption. As suggested by these letters, the FCC assumes that all content, whether it is zero-rated or not, conforms to an ideal of perfect competition. Clearly, not all content is the same. There are search costs, barriers to entry exist, content has market power and there are significant transaction costs. Would a new episode of Game of Thrones be completely overwhelmed by a zero-rated version of Marvel’s Agent Carter, which was cancelled? Hardly. People go to great lengths to see Game of Thrones. And for mobile, how might that Game of Thrones video be compared to scanning Facebook? In a converged and mobile world, the desire to consume specific types of content is highly varied, depending on both the content costs and the transmission costs.

Second, it is largely assumed that zero-rating provides an unassailable advantage in delivery, yet technological innovation hints at just the opposite. Zero-rating is just one way to compete in a world with quickly changing technological capabilities. Netflix serves as an example. Netflix now allows consumers to download movies and videos to be watched offline, including mobile users. Let’s call this zero-streaming. It is also worth noting that mobile users already offload a lot of downloads to WiFi. So, the zero-rated content would still be at a disadvantage to zero-streamed content since Netflix wouldn’t suffer from buffering issues or quickly changing video quality. Moreover, Netflix has upgraded its own delivery systems and is now using a codec that allows it to save 36 percent on bandwidth for Android phones and about 19 percent on iOS. Any advantage offered by zero-rating is nullified with zero-streaming and severely muted with new codecs.

Lastly, the FCC’s analysis fails to consider the expansion of choice and consumption. In a world before zero-rating, affiliated content that could be zero-rated and non-affiliated content are both constrained by the data cap. However, once affiliated content has been zeroed out, the cost to consume non-affiliated content is reduced. For example, let’s say your plan allows 100 hours of downloadable content. In a non zero-rated world, you consume 50 hours of affiliated content and 50 hours of unaffiliated content. Once affiliated content has been zero-rated, you effectively gain 50 hours of downloads, making it more likely that you’ll select unaffiliated content. To those who only see the marketplace in a static mindset, it is deeply paradoxical to learn that the Binge-On program nearly doubled video viewing for T-Mobile, which was able to sustain average billing rates even as it zeroed out the biggest sources of data usage. Reducing transmission costs is only one part of the larger story about content innovation.

But the FCC isn’t alone. Some of the loudest voices in the zero-rating debate fundamentally misunderstand the complex nature of content innovation.

One of the most widely cited research reports on zero-rating makes some serious blunders when it claims that, “Research shows that people strongly prefer zero-rated content over content that counts against their cap.” Two sources form the basis of this statement, a CTIA survey and an experiment from the publisher at Slate. Yet, neither really support the broad claims as emphatically as the author suggests. For one, the original CTIA survey cited only asked consumers if they would be more or less likely to consume content if it didn’t count against the monthly data allowance. The study didn’t ask if they would strongly prefer subsidized content over other content. This is similar to a survey finding broad support for lower priced oranges and then boldly claiming that oranges are preferred to apples. The second statement simply cannot be inferred from the first.

Slate’s experiments with their podcast marketing also gets misinterpreted. As Dan Check, the vice president of technology at Slate, explained to the New York Times, one method he used to get more users to their line of podcasts was to pay for the mobile data. For one group of potential listeners, it was found that they were 61 percent more likely to press play. But the next question that should be asked is, 61 percent increase from what level and compared to which kinds of users? Because little information is given about the base rate in the NYT piece, it could be that this zero-rating program bumped up a low performing podcast, putting it on par with other more desirable content. Nevertheless, it is tough to gauge the competitive effects without context.

Merely professing a love of innovation doesn’t get us to an innovative world. Content innovation is messy. While the FCC nominally encourages innovation, these two letters suggest that more analysis is needed from the FCC and for the sake of consumers, let’s hope the next FCC actually does that.