by Mark Lowenstein
There once was a company that raised hundreds of millions of dollars from technology industry titans to build a data-centric wireless network using unlicensed spectrum in major cities in the United States, whose initial offerings were for laptops using wireless modems, but planned to support an array of devices using the network for a host of innovative applications. That company was Metricom, and it was one of the more visible and spectacular failures of the first wireless data bubble, circa late 1990s. I've been thinking about Metricom lately, given the commitments being made to data-centric 4G networks: to WiMAX and Clearwire by wireless industry challengers (cable, Cisco, Intel, Google) and to LTE by wireless industry incumbents (Verizon, AT&T). Before we cut the ribbon and begin building in earnest to LTE, let's consider what 4G is, whether we need it, and how it will not be Metricom redux.
First, let's look at the wireless data scorecard as of mid-2008. Yes, we have seen spectacular growth over the past several years, but more than 75 percent of data revenues in 2007 came from applications for which even 3G's capabilities are more than sufficient: messaging, email and "personalization" applications such as ringtones. Mobile video and mobile music, both of which need 3G in order to deliver a decent user experience, have not yet taken off. On the other hand, wireless access from laptops using 3G cards has been growing very steadily, offering an experience roughly equivalent to low-end DSL in the home There is also pent-up demand among smartphone users for a zippier browsing experience, which we see on some 3G devices. So, that's the data scorecard as of mid-2008.
Now, let's put 4G in perspective. 4G will offer important advancements in throughput and latency, and IP "elements" that look more like typical broadband networks. One aspect that has not been made that clear is whether 4G will support a more "constantly connected model" across a large number of users.
That said, mobile broadband networks will not approach most people's broadband connection at home or the office. The wireless-to-wireline delta will likely widen over the next several years. By the time 4G is widely available, the average landline broadband connection speed at home will be 10x today's 3G services, delivered at a fraction of the cost. Think about how your home Internet experience has changed over the past few years: the integration of video, media-rich experiences such as YouTube and Facebook, and the ability to stream or download TV shows and movies to your PC. Lowenstein's Law of mobile networks is that the present state of the art in wireless is the low end of the previous generation's wireline broadband network. Physics, economics, and spectrum make it so. This tells us that wireless displacement of data will not happen like it has with voice.
With that framework in mind, let's discuss whether the "build it and they will come" theory applies to 4G? In the case of the previous 3 "Gs", there was an identifiable need and a major opportunity driving the investment. With 1G, it was all about cellular voice, mainly for business people. 2G (digital cellular) was about the mass consumerization of mobile, affordable phones and prices, and on the data side, access to email from Blackberries and personalization applications such as ringtones. 3G finally made it possible to have a decent remote access experience from laptops, and a better experience on phones for browsing and some multimedia content.
What is the driving force for 4G? Is Citius, Altius, Fortius an adequate business case justification? Today's mobile broadband experience is adequate, for most users, most of the time. The main reason "addressable" users don't subscribe to this service is because it is too expensive, or the need isn't great enough given other options--free WiFi and so on. Few have said they don't subscribe because the speed is "too slow."
On the phone side, 3G provides, again, an adequate data experience for browsing, downloading music, viewing clips and uploading pictures and short videos. Some of the "slowness" on 3G phones is attributable to device-related limitations (i.e. processor) and poor user interface design. Leaving aside smartphones, what would you do differently on a 4G feature phone compared to what you do on a 3G feature phone?
The biggest catalyst for 4G seems to draft off the growth in the Internet: wireless fills a hole in the need be "constantly connected," and to provide a mobile extension to the sorts of applications growing over broadband. But is this more abstract notion enough, given the billions of dollars of incremental investment involved, and the data-centrism of 4G? I think we'll need some creative new ideas on top of some of the more intangible aspects.
Here are three broad categories that are ripe for innovation in my view:
- Portable Data Devices. This pertains to a whole new class of devices--often referred to as MIDs--that fall somewhere between a laptop and a cell phone. The best example of such a device today is the iPod Touch, which when you really think of it, is an evolution of the original Palm Pilot. PDDs will have fast processors, large amounts of external memory, significant multimedia functionality, graphics accelerators, GPS and great browsers. Voice is a secondary consideration on these devices. In fact, VoIP might be a reasonable alternative to a traditional cell phone service.
I believe PDDs will operate in a more "constantly connected" mode--sort of like your Blackberry does for email--with updating, refreshing happening in the background. The ability to support a constantly connected model, with appropriate economics, has to be a key attribute of 4G networks.
- Whole new class of 4G "embedded" devices. There's enormous opportunity to embed
wireless connectivity into other portable devices. The three that come immediately to mind are
digital cameras, digital video recorders and portable gaming devices. Think about instantly uploading a video from
your video camera, or playing a multi-player game over wireless.
We are also at a tipping point with machine-to-machine (M2M) is about to take off. Certainly not all M2M devices/applications require a 4G connection, but the right elements--capable devices, network capacity, chipsets, and capital--are beginning to coalesce around this market. We'll need "more network" to support that.
- More flexible pricing options. This is where the wireless industry is going to get outside its comfort zone. When we imagine people with multiple devices, or applications requiring varying types and levels of connectivity, we are going to have to become much more creative about pricing. With all the competition for a consumer's disposable income from the panoply of digital devices and services, we will need some options beyond today's $30-$60 monthly subscription plans for wireless data (which generally require a two-year commitment), especially for consumers with multiple mobile devices or with application-specific devices (such as a digital camera). Some alternative pricing models I envision:
- Time-based pricing (hourly, daily, weekly), taking a page from the WiFi model
- Application-centric pricing. For alternative types of devices, requiring occasional, "bursty" connectivity, the user gets charged for a specific action, such as uploading a video.
- Consumption-based pricing. Some applications and use cases have heavier network requirements. Should a user be charged more for streaming video than Web browsing? Or pay a premium for QOS on VoIP? Remember that wireless is fundamentally a capacity-constrained network. There's always an element of robbing Peter to pay Paul. This is a controversial topic--gets us into "net neutrality" territory. But I do think some of the app requirements will force the issue.
- Holistic data pricing. Most of the major stakeholders in 4G already provide fixed line broadband. They should be creative around bundling fixed/wireless services. Think $75 for a combined package that includes home and mobile broadband.
Success in 4G is not guaranteed by simply plotting the growth trajectory of today's services. It will require thinking outside the box in terms of services, devices and pricing structures.