Wednesday, February 10, 2016

In Video Era, Mobile Zero Rating is Essential

Some people do not like zero rating, the policy of allowing consumers to watch video without incurring any usage charges on their data plans.

As with many other issues in life, this a matter of competing values, philosophies, benefits and policies.

Allowing consumers to watch entertainment video on their devices is a “good thing,” as it avoids the huge data charges those consumers might otherwise incur.

On the other hand, some argue, such policies do not “treat all apps or bits the same.”

T-Mobile’s Binge On, Verizon’s zero rating of go90 or Walmart’s Data Saver are examples of such practices.

Policy, in this case, is bumping up against business economics, specifically, the bandwidth intensiveness of video. Put simply, if people are to be able to consumer entertainment video the way they want, at prices they can afford to pay, then zero rating, or at least “very low costs,” are essential.

If policymakers really want robust competition between platforms, industries, networks and competitors, they must recognize the video is quite different, in terms of network load, supplier cost and end user willingness to pay.

It is easy enough to explain why video entertainment consumption poses a huge--some would say nearly fatal--challenge to mobile operators: there if a fundamental mismatch between revenue and bandwidth required to deliver narrowband services (voice, messaging) and that required to support full-motion video.

Simply, revenue per bit for messaging and voice can be as much as two or more orders of magnitude higher than for full-motion video or Internet apps.

The revenue per bit problem is easy to describe, for fixed, mobile or untethered networks.

Assume a fixed network ISP sells a triple-play package for a $130 a month retail price, where each component--voice, Internet access and entertainment video--is priced equally (an implied price of $43 for each component).


How much bandwidth is required to earn those $43 revenue components? Almost too little to measure in the case of voice; gigabytes for Internet content consumption and possibly scores of gigabytes for video.

By some estimates, where voice might earn 35 cents per megabyte, revenue per Internet app might generate a few cents per megabyte. At one level, a network engineer might argue that such fine distinctions do not matter. The network has to be sized to handle the expected load.

Business strategists do not have that luxury. Entertainment video simply generates too little revenue per bit to support gigabit networks, especially in all wireless domains.

McKinsey analysts have argued in the past that a 3G network costs about one U.S. cent per megabyte. The problem, in many developing markets, is that revenue could drop to as little as 0.2 cents to 0.4 cents per megabyte, for any mobile Internet usage, before we even load on extensive video consumption.

That implies a strategic need to reduce mobile Internet costs to as little as 0.1 cent per megabyte, or an order of magnitude. Tellabs similarly has warned about revenues per bit dipping below cost per megabyte, leading to an "end of profit" for the mobile business.

Of course, all of that analysis occurred under conditions where it was web browsing that largely represented Internet bandwidth demand.

Streaming video is another order of magnitude or two orders of magnitude sort of problem, though, in part because it is so hugely bandwidth intensive and because it will represent as much as 70 percent of all Internet bandwidth consumption, in a few short years.

Consider the wide variance in revenue per bit represented by a few different potential mobile Internet use cases.

One use case is a $20 a month smartphone data plan and 2GB of usage, representing retail revenue of $10 per gigabyte.

A Netflix subscription generates no direct revenue for a mobile operator but could represent network consumption of between a few gigabytes and  30GB of traffic, if usage approaches fixed network levels. Revenue arguably is zero dollars per gigabyte.

A work environment might represent $100 a month revenue and consumption of between 10 GB and 50 GB. So revenue might range between $2 to $10 per gigabyte.

And that’s the problem with video: much of it does not actually represent revenue for the ISP. But even if it does, what is the revenue and cost per gigabyte?

Even if one assumes consumption of just one hour of standard definition video, and that product is owned by the ISP, revenue might be $1 to $2 per gigabyte. But most video content is not owned by the ISP. It simply has to be delivered, at prices the consumer will pay.

Some would argue the cost per gigabyte for a mobile ISP is higher than $1 to $2 per gigabyte.

And it is almost nonsensical to think that as video moves from linear to on-demand and streaming, mobile consumers are going to pay for bandwidth representing multiple gigabytes per day, per device, at $5 to $10 a gigabyte.

If today’s linear video consumption were instead viewed on mobile devices, consumers might really expec to pay $200 to $400 a month--or more--in mobile Internet access charges, to say nothing of the actual retail price of the content service.

That is unworkable, and explains why zero rating, or very low cost rating, will be necessary.



Marketers might argue that revenue per account is what matters, for a multi-product business.

That is true, up to a point. An ISP might fare okay if providing a mix of products with disparate revenue per bit values.

The revenue earned from text messaging is almost arbitrarily high, as SMS is a byproduct of using the signaling network. Voice revenue might be moderately high, if users can be coaxed or compelled into paying for access to the feature, rather than for usage.

Ericsson hs calculated the cost per bit for a mobile network at about one Euro per gigabyte. So total revenue per bit has to exceed that cost.

Heavy video consumption--especially of third party content-- is likely to exceed cost per bit under almost any scenario, unless zero rating, or very low cost rating, is allowed.

No comments:

"Tokens" are the New "FLOPS," "MIPS" or "Gbps"

Modern computing has some virtually-universal reference metrics. For Gemini 1.5 and other large language models, tokens are a basic measure...