Lowenstein's View: We need more data on data
Wireless data revenues are going to be over $50 billion in the United States this year and are clearly the wireless industry's growth engine. But reading a wireless operator's 10-Q filing is like a step back in time. They report service revenues, average revenue per user (ARPU), churn, and break out their prepaid, postpaid and wholesale businesses. However, the data on "data" is pretty thin: revenue, data ARPU, and percent of devices sold that are smartphones. An operator might open the kimono at an industry or investor conference and talk about data traffic having grown a zillion percent in the past three years--with the Y-axis left blank, naturally.
Contrast this with the metric deluge coming from the Internet and advertising side of things: what sites were visited, by whom, and for how long, and what apps were downloaded from what device and from where. But information on how much of mobile data traffic is Web vs. messaging? What is typical data consumption on phones vs. dongles? What does it cost to deliver a megabyte of data? This information might get discussed anecdotally but certainly not systematically.
This is not meant to be a "beat up the operator" column by any means. But there are some pretty important constituencies looking to make some pretty big bets on this industry. And they need some guidance. First is the investment community. They are trying to get their arms around an industry where annualized subscriber growth is about 7 percent, and Q2 2010 to Q2 2011 service revenue growth was about 3 percent, while EBITDA margins experienced modest declines. Other than knowing that voice revenues are declining and data revenues are growing, and that higher smartphone sales boost subsidy costs, they have to make a lot of assumptions and inferences to understand how things add up.
I can tell you, from the questions I get from the investment community, that they are having a very hard time understanding what long-term effect data is going to have on this industry and how to model this business. For example, what is the split in data revenues between highly profitable messaging business and more consumptive mobile Web? What effect is rich media having? What efficiency gains are being realized from the deployment of fiber backhaul, and HSPA+/LTE?
Substitution is another key question. For example, are services such as BBM, iMessage, and "private networks" such as Facebook substituting for messaging? Is the migration from messaging to data plans accretive or dilutive to the business? If there's not some transparency here, investors will get spooked like they did when out of the blue KPN in the Netherlands issued a profit warning because of text messaging substitution by WhatsApp rocketed to 85 percent penetration in a matter of months. Oh.
The other constituency is the content and developer community. They need better guidance on wireless industry economics in order to answer some pretty important questions about the future of content over mobile devices. Telling, for example, is the lack of cellular connectivity of the just launched Kindle Fire and Barnes & Noble Nook. Do wireless operators want to support video streaming services over services such as Amazon, Netflix, or Hulu? What are the costs of doing so and how might that influence prices? Right now, watching a movie on Netflix or downloading a movie off iTunes over cellular gets one into overage territory pretty quickly.
Then there's cloud services, which, as I wrote about a couple of months ago, require a different type of connectivity model. The development community is on the one hand super-excited about the coming wave of 4G networks and devices, yet uncertain as to how much effort to put into apps that consume a great deal of bandwidth. This is because they don't really know what the network capacity roadmap looks like or whether price structures will encourage use of rich media over cellular networks. In the same vein, should device OEMs and the connected device community be thinking more streaming model or download/cache model?
For operators, I realize this is highly sensitive stuff. If, hypothetically speaking, Verizon Wireless believes their cost to deliver a bit costs 30 percent less than AT&T Mobility, they're not going to announce this to the world. They'll either cut unique deals with content developers, bundle this cost advantage into prices or simply make more money.
But as data continues its inexorable growth, we're going to have to think as an industry about how to provide better guidance to some of the constituencies looking to make some pretty big bets on wireless. Some of this must start with the operators. Understanding they don't want to give away the farm, here's what I would suggest:
- Some measure of data profitability
- Some granularity on the source of data revenues, such as messaging/access/connected device
- ARPU from phones vs. connected devices (such as tablets and dongles)
- Some metric of data traffic growth, and impact of rich media
- An indication of cost of providing data, particularly with the move from 3G to 4G networks
- The extent to which there is substitution in their core voice and messaging services from over-the-top services such as BBM, iMessage, Skype, etc.
- Here's a wish list item: some level of reporting for enterprise, such as percentage of business from corporate liable. After all, they do it on the landline side.
Down the line, wireless operators might be pressured to break out their data business separately. In the meantime, some of the above measures would be a good start.
Mark Lowenstein, a leading industry analyst, consultant, and commentator, is Managing Director of Mobile Ecosystem. Click here to subscribe to his free Lens on Wireless monthly newsletter, or follow him on Twitter at @marklowenstein.