Editor’s Note: Welcome to our weekly Reality Check column. We’ve gathered a group of visionaries and veterans in the mobile industry to give their insights into the marketplace.
This week’s article deals with a hot topic – bandwidth buckets, or, as Craig Moffett from Sanford Bernstein terms, “data rationing.” The concept of rationing is very hard to stomach in today’s telecom world because it’s viewed as a step back, a “takeaway” from the days of unlimited data, but “bucket of minutes” pricing has been around since the early days of wireless. Minutes, however, in a traditional circuit switched environment are subject to the design of telephony switches: because there is limited capacity to connect any two end points, market prices keep usage to a minimum during peak hours. Anytime minute pricing (e.g., 900 anytime minutes for $60) is a form of rationing – go over the limit, and you’ll pay an extra fee.
This capacity restriction does not exist with data, however, in the same manner. Our ability to connect to Pandora should be limited only by spectrum capacity, and, should there be an excessive surge of traffic, we might experience a slow-down at a very busy spot (e.g., a sporting venue). After we get past the wireless tower, the engineering is exactly the same as it would be from a large enterprise or small data center. Finding the most efficient way to connect to large content providers is no easy task, but, as AT&T Inc.’s announcement with Zynga last week showed, it could become a differentiator.
The economics of wireless data traffic are as follows:
1.) Spectrum
2.) Electronics
3.) Tower (allocated)
4.) Backhaul and access (allocated)
5. IP routing costs (allocated)
As bandwidth increases from any particular tower, the cost per megabyte/gigabyte plummets. Fiber economics tend to be better for each carrier once bandwidth levels justify a second high speed circuit. Combine the effects of several carriers from a particular tower, and you can see why cable, telco and independent fiber companies are smiling. Also, upgrading to fiber at a particular tower dramatically increases the quality of the circuit (vs. older copper), reducing maintenance expenses and increasing customer satisfaction.
AT&T, Sprint Nextel Corp. and Verizon Communications Inc. all operate tier-one IP backbones, meaning that their connection costs are lower (sometimes only slightly) than the tier-two providers. T-Mobile USA Inc. routes through a couple of well-known tier-one backbone providers, and, given their scale, their IP costs aren’t significantly higher than the tier-one group. But, as we have seen with cable company gross margins, the IP routing costs are low so long as the IP network has optimized their ingress and egress points.
As we discussed when Verizon Wireless introduced the mandatory $10 data rate on all but their most basic feature phones last year, the comparable yield per gigabyte on these plans is significantly higher than the tablet product (and, outside of the minor difference in electronics cost structures between 3G and 2G data, the “from the tower” IP costs are exactly the same). Better yet, name me one person who is going to frequently browse from one of these devices (at least for long). The result: a median yield on the 75 MB plan is likely somewhere in the neighborhood of 60 cents to 70 cents per MB(!). To use an old telco term, this is the equivalent to the “casual call” product in the old telephone world, and a source of disproportionate profit growth over the past 18 months for those customers who didn’t make the switch to smartphones.
AT&T Mobility has a similarly yielding product, offering 200 MB of data usage for $15. Nice price point, but the customer service transition over the past year has been very difficult as AT&T Mobility has seen an increase in customer service calls. Interestingly, in my last store experience they completely skipped the $15 offer after a few questions. “Do you have a lot of applications?” the store clerk asked me, thumbing through the 35 or so we have on our Mobile Symmetry test device. “Sure do,” I replied. “Do you spend a lot of time within a Wi-Fi zone?” “Not really, I use this around town a lot,” I replied. She paused, and, with a serious look, replied “Because one applications upgrade can blow your entire data limit.”
Frankly, the thought of using AT&T Mobility’s network didn’t even occur to me because I use Wi-Fi as much as possible with the iPhone. Not because I want to, but because so many applications require me to be in a Wi-Fi zone (FaceTime and Skype video being the most prevalent Wi-Fi only products). Apple Inc. has trained me to think about my network and its tie to applications availability – that’s not good for the carriers. But they (who??) economically benefit from it.
At the high end, Verizon Wireless makes another bet, this time on a telecom term called “breakage.” At $80 for 10 GB per month, you have to have a lot of video and streaming music usage to even get close. Even with a tethered product (or a hotspot, which carries an extra charge), the carriers bet that you won’t use exactly 10 gigabytes of usage. To use 10GB, you would need to have about 1 hour of streaming video each day as well as 8 hours or so of Pandora – not impossible with a long commute, but definitely not the norm.
So Verizon Wireless bets that you’ll use 8 gigabytes some months (which yields a 50% gross margin), and others you’ll use even less. The problem is that as applications become optimized for 4G devices, they will likely attempt to use as much bandwidth as possible to ensure the highest quality experience for the application. Give Skype and Facebook time, and that LG Electronics Co. Ltd. Revolution, HTC Corp. Thunderbolt, or Samsung Electronics Co. Ltd. Charge will display the quality that a 12 to 15 Mbps download can deliver. Think of it like YouTube, but always on the 480p (or higher) setting.
Enough about the profit margins – unlimited had to end, because more than a few customers were challenging the profit thresholds of the two largest carriers. Or did it have to end this way – a rationed world? I posed this question to several of you in preparation for this article, and was surprised to see how many of you said “no.” As one Sunday Brief regular who works for a wireless carrier so briefly said, “It’s a few applications driving the larger volumes. Charge the applications providers for using the wireless data.”
Would a surcharge per application be a better solution? Specifically, would a surcharge on Micro-Skype-Qik -Facebook, Netflix, Google or others be able to stand both a regulatory and a market test? Isn’t rationing easier than a “guzzler tax?” While the billing infrastructure is in place, I cannot imagine the customer service nightmare this would present for the carriers. My Netflix bill went from $8 per month to $80 because of bandwidth charges – who’s neck do I choke?
While a “data access” charge sounds good, anyone who has been through the nuances of switched access charges knows that the disputes and settlements are very complex. It would likely lead to more costs for carriers and providers like Netflix, not less.
So we are left with rations, a byproduct of a regulatory system that would not allow the carriers to think about a “premium lane” for “premium apps.” It’s time to re-think net neutrality. It’s time to re-think voice plans as a smaller percentage of wireless users approach their 450 or 900 minute anytime plans. It’s time to re-think texting vs. in-application messaging. It’s time to re-think separate hotspot and smartphone data plans (as AT&T Mobility
did with the iPhone). And it’s really time to re-think th
e effects of an increasingly Wi-Fi dependent society on the carriers’ brands. Until then, it’s a capped world with limited possibilities.
Jim Patterson is CEO and co-founder of Mobile Symmetry, a start-up created for carriers to solve the problems of an increasingly mobile-only society. Patterson was most recently President – Wholesale Services for Sprint and has a career that spans over eighteen years in telecom and technology. Patterson welcomes your comments at:jim@mobilesymmetry.com.
