Rethinking the 4G business model

Wireless Asia: Some investors are reportedly skittish on 4G's short-term prospects, even though they traditionally see telecoms as a long-term play. What do they have to be nervous about?

Herns Pierre-Jerome: Data is far less predictable. They see that slide of traffic growing faster than revenues and run from the room screaming.

How can cellcos get into the value added side of things and not just be a fat LTE pipe?

A significant challenge there is getting the data to the RAN. We have the issue of backhaul and how much you can cost-effectively provide to support these large channels. You can get away with T1/E1s when you're using a 1.25 MHz or 5 MHz channel. Once you go up to 20 MHz, you need a big chunk of backhaul right off the bat, and you need to overprovision and anticipate the potential for growth. So the question then becomes what kind of operator are you - a classic wireless operator from the last 20 years, or one that's better positioned for this - a multifaceted one that has a wireline business and a wireless business that actually talk to each other and like each other that have access to Internet interconnect, in-house resources for transoceanic connections?

Beyond that, the business models fundamentally are rooted in 3G six or seven years ago when it was in its infancy in terms of conceptualizing how we're going to use this platform. It was like, "It's a fairly fast connection, you can download information with it, here, have at it, and here's a walled garden for you to play in, but you can't go outside of it." So the mindset has changed - for example, look at Verizon's Open Device Initiative.

But how much innovation are we really seeing with wireless data business models?

It depends what you compare it to. If you look at innovations in voice, the voice market really took off, especially in emerging markets, when they started prepaid as an option, because it allowed people to quantify what they were doing in a different way. So I know that I need to spend $20 this month, but maybe next month I'll need to spend less, or more. So I can spend at whatever rate I want and top off when I'm ready, and I don't get penalized or have to deal with these crazy plans.

Compared to that, data services are still in the Stone Age. How do you quantify something like wanting to have certain content streamed to your phone, which means I don't need a super hi-res stream, so maybe I should pay less for that than I would if I got the same content streamed to my plasma TV in my living room. But we're not set up to even determine or auto-detect certain devices - the RAN knows who you are through the SIM, but doesn't know what device your SIM is in.

So we don't have a good way to effectively charge for data. It's not that we don't know how people consume data now, but we don't have a model that makes sense for different forms of media, the different usage patterns and behaviors, and the different bearers that can deliver that. And there are so many angles to this that we have to get our heads around - the delivery mechanism, security, billing, regulatory definitions of services.

And get everyone on the same page for all of those.

Exactly. And that's just in one market, let alone the other 200. So we have a very long road ahead to make that possible. And I don't see it happening in a very planned, architected fashion. A lot of people are working on all of these things and have proposed various frameworks for doing it. But look at how the Internet evolved. Years ago, the notion was that we can interconnect a bunch of these networks to send files around, share documents between universities and governments. But then the Web came along, and all of the uses that proliferated after that were never planned. They were never part of the original vision. The same thing will happen with the wireless Internet.

Is that why outsiders like Google and Apple have shaken things up so much in the mobile space?

That's in part because they come from the more mature portion of the ecosystem. Mobile is starting 20 years later than them, and we're evolving our understanding of our business models in the same way the IP guys did during the dot-com bubble. And many of those never panned out either. But that experimentation has already happened in the fixed-line world. It hasn't happened in the wireless world because in a way it's a much different world. You do need stable operators. In the fixed world, you can have ISPs that aren't that sophisticated, of all sorts of various sizes. In wireless, the business of running and managing a wireless network and wireless delivery platform is very complex in terms of technology, customer support and customer relationships.

What's going to be challenging going forward, is that this structure doesn't lend itself to the types of things we're seeing in the future. You still need someone to play the role of reliable network provider, but it's no longer just of case of delivering my Internet to your device and I'm managing that for you. It's much broader than that. For the services we're talking about where you want to download this track and share it across these seven devices without paying for it seven times or mash it up, there are capabilities in the software to make this more context-aware, or adapt the content to the platform. The software can do that, but it's tough to implement because there's all these requirements like DRM and SLAs that don't allow us to do that.

As we look to this massive ramp-up of data consumption, the answer isn't necessarily to figure out how I need to charge to deliver data cost-effectively, rather than how can I price it in a different way that aligns with what people want to do with that data and are willing to pay for it so that I can monetize it. Everyone talks about growing ARPU, but what is ARPU when I have an iPhone, a laptop, a MID and a wireless toaster and a wireless picture frame and telematics in my car? What am I using those for and am I sharing it with others? All of those things don't fit our business models today. No one's going to pay separate subscriptions for each of these devices from different providers. I want it all integrated and consolidated in one place and one bill so I can manage it better and reduce my package.

Meanwhile, customers will expect more for the same price as time goes on.
I agree. And also, when you think about LTE, you hear about how that will force an evolution of the business model because how do you monetize a 10-20 MHz pipe? Slice it up for tiered services? Or fill it up and let someone use BitTorrent as long as no one else is on the cell at the time and charge them for it? This notion of reflecting the cost of delivery is too rigid, because it changes depending on congestion, time of day, what your upstream provider charges you per GB of content flowing through your interconnect. There's so many variables, but operators have to figure out whether it costs them more to have an empty pipe, have ten users at a flat-rate ARPU, or a way to stack up different services and pricing schemes so the pipe stays full, and maybe you won't make as much money as you want, but the cumulative effect gives you a greater return. The whole notion of ARPU is archaic when we talk about data, because it doesn't mean much. What's Amazon's ARPU with Kindle? Can you even call it ARPU? It's not a good measure of optimal loading of the network to keep costs and revenues in balance so you're maximizing ROI.

How can cellcos mitigate that from a bandwidth management standpoint?

Bandwidth management makes sense when it allows you to continue to leverage the entire pipe, so you can have those power users, but when others need the service, you can use QoS to throttle them down, and depending how you charge for data, you're still maximizing your revenue stream. You're still allowing everyone to use your service and pay you for it without artificially constraining usage with data caps or throttling them all the way down to 120 kbps like I've seen some operators do.

Qualcomm is clearly interested in the embedded market with SnapDragon and Gobi. Once wireless consumer electronics go mass-market, will cellcos be ready for that business model?

Once we see more wireless device categories, like MIDs and e-book readers and consumer electronics, it comes back to what I said before - we need business models that reflect consumption in a more effective manner.
Another aspect of that is multimode wireless, where these devices will be able to choose either Wi-Fi or Bluetooth or HSPA or something else to relay information. Maybe it's based on the cheapest connection or the fastest. It will be helpful to have that choice, and it's another example of how operator business models aren't equipped to address that. Suppose one of those services is supplied by another provider - do you block that if you subsidized that device? Do you collaborate with those providers to make it context-aware and less painful to the consumer? There are all kinds of things to consider there. Billing and privacy are other issues, too.

How about M2M? That's been around for ages but is only just now getting some traction in the cellular space.

One of the challenges is that if I'm a utility company with 10,000 meters and I download the data from them on a daily or a monthly basis, how does the operator charge me for that? A set monthly amount for all the devices, or for the amount of data? We haven't come up with the models to facilitate these different vertical apps.