Wireless Wide Area Network Client Attach Forecast
By Roger L. Kay
Not since doing end user research in 2000 have I seen an emotional response of this
magnitude.  At that time, flat panel displays were being introduced as desktop computer monitors
(notebooks had already been based on panels for a long time, but were rare in the client mix
because of their $1,000+ premium).  The desktop displays were still high priced, and the
response by all segments was notable: everyone loved the panels when the survey didn't
mention price.  They warmed right up to panels of any size, but when conjoint analysis matched
panels with realistic sales prices (e.g., $1,250 for a 17" LCD), they cooled right off.  I made the
judgment call that when panel prices came down, adoption would be widespread.

Having gauged that response on a purely emotional level, I would say that notebooks enabled
with Wireless Wide Area Networks (WWANs) invoke a level of affinity and enthusiasm closer to
that gold-medal flat-panel response than any capability that I've seen introduced since.

One anecdote comes to mind indicative of this market sentiment.  I happened to get an upgrade
on a flight this past year, and I was seated on the aisle across from a gentleman who had his
notebook open as the plane was loading.  He was happily typing away, and I could see Web
pages from where I sat.  Others could see, too.  Pretty soon a guy in a pressed shirt and tie
leaned over to him and said, "Are you connected to the Internet?"  And the guy said, "Yeah," and
proceeded to show the other guy the antenna from his PC WWAN card.  Everybody in the cabin
was either listening or frankly getting up to see better.  The scene could have been ad copy for a
WWAN service provider.

Analogously to flat panel adoption, WWAN penetration will also be inhibited by price, but based
on what I've seen, I'm ready to forecast a fairly important attach rate to notebooks over the next five
years as prices come down (and capabilities increase).  Although an argument can be made for
WWAN attach to desktops, the market will be almost entirely a notebook phenomenon.

The value proposition differs slightly between commercial customers and consumers.  
Commercial customers want convenience and reliability: to get online anytime while traveling
without having to ask their host's permission or pay extortionist fees to a local provider.   For
these customers, an all-you-can-eat plan makes the most sense.  Consumers, on the other
hand, are not quite as mobile as business users, and may use a wireless wide area connection
only occasionally.  This segment is more likely to be responsive to a pay-as-you-go scheme.  The
"killer app" for both these segments is essentially ubiquitous connectivity, unrestricted access at
a decent data rate.

2005 Baseline

The first year when WWANs had any impact at all was 2005, when offerings for the most part
came directly from carriers like Cingular, Sprint, and Verizon, and radios were mostly PC card
types.  The volume of these cards has grown slowly since 2002, when they first began shipping,
but they were still numbered in the low hundreds of thousands in 2005, primarily because of high
service pricing, low data rates, and difficulty of use.  The difficulty of use issue is one of the key
factors driving vendors to embed their WWAN offerings.  The vendor can preconfigure much of the
setup, and the user has only to interact with a simple interface to make a few service and setting
choices.

The first company out of the embedded WWAN gate was Sony, which in May 2005 introduced an
ultraportable notebook with an embedded second-generation wide area radio, the EDGE-based
VAIO T-Series, offered in partnership with Cingular for $80 per month with unlimited usage or $50
per month with a two-year contract.  Although the T and follow-on TX are sold primarily through
retail channels, Sony indicates that most of the traffic buying the WWAN-enabled notebooks
would be classified as business users.  Late in the year, Sony brought out similar capabilities on
its explicitly-business BX series notebooks.  The downlink speed of these EDGE-based systems
averages about 100Kbps, certainly better than a 56Kbps modem, but really nothing to write home
about.  Because of the expense, the fact that the notebook was tied to a particular carrier and
plan, and the low data rate, Sony WWAN-enabled notebook penetration was somewhat limited.  
But because of its early lead in embedded WWAN attach, Sony constituted the bulk of WWAN
notebooks in 2005.        

The next company out of the WWAN chute was Lenovo, which in November brought out its wide-
screen Z Series with embedded EVDO WWAN capability.  Lenovo partnered initially with Verizon
and was the first to bring out a true 3G capability with downlink data rates of 400-700Kbps.

Both Sony and Lenovo intend to expand their offerings in 2006, moving embedded WWAN into
more of their notebook lines, bringing on more carriers as service partners, and upgrading their
technologies.

But in some sense 2005 was just a warm-up for things to come.  Dell, HP, and Panasonic
announced WWAN business notebooks at CES in January, with shipments expected to begin in
earnest in 2Q06.  All vendors have indicated that they intend to spread the capability to other lines
later in 2006 and into 2007.

Competing Technologies

Another important aspect of WWAN adoption will be its ability to replace 802.11 and supplant
WiMAX.  Although 802.11 is widely available today, it has major limitations.  Among them are that
lighting up a large corporate campus with access points is costly.  A full infrastructure can run into
six figures.  On a national level, vendors and partners have tried to assemble a "warm fabric" out
of hot spots, but coverage is spotty at best and nonexistent in remote areas.  

In one ambitious case, Greene County, North Carolina, one of the poorest rural districts in the
United States and formerly one of the most economically dependent on tobacco, received a huge
subsidy as part of the tobacco industry's settlement with the Justice Department.  The county
decided to invest its windfall in, among other things, lighting up the entire county with 802.11
access points.  The plan called for putting antennas on the water tanks found on many farms in
the area.  Equipment alone was budgeted at $300,000.  Given that the county is about 265
square miles, it is possible to estimate the theoretical hardware cost of putting up 802.11
networks to cover the entire United States, which has an area of 3,537,441 square miles.  Using
the Greene County ratio, the cost of illumination would be about $4 billion.  By contrast, the
cellular carriers already have an infrastructure in place for transmitting data and can upgrade it to
3G for considerably less money than the cost of a wholesale project to put 802.11 all over the
place.

In addition, 802.11 frequencies can interfere with other electronics, such as medical equipment
and microwave ovens.  Finally, 802.11 does not allow for roaming.  In one classic case, I was in a
hotel and had signed up for the WLAN in the room.  I had to go downstairs to wait for someone in
the lounge and signed on to the wireless network there.  I was asked to register and pay again.  
After notifying the hotel management, I was credited for one of the sessions, but even within the
same establishment, roaming was impossible.  Of course, this shortcoming is even more true
for moving between establishments and cities, from indoors to outdoors, and across the nation.

WWANs give the user something unobtainable via 802.11: freedom.  The carriers already have
the back-end infrastructure, both technical and financial, to handle roaming, which is a standard
feature of cell phone coverage.  Once a customer has a data plan in place, he or she can connect
anywhere in the coverage area, which will expand for 3G networks in the United States throughout
the forecast period.  Although 3G networks currently cover about half the U.S. population, and this
coverage is growing slowly but steadily, utilization is still low.  Carriers will only expand network
coverage aggressively when utilization rates pick up, which should occur toward the back half of
the forecast period.    

Now, 802.11 does have some compensating benefits, and it is important to recognize them.  For
example, 802.11 is an established standard, and while cellular networks differ in different parts of
the world (and even, in the case of the United States, within the same country), 802.11 is the
same everywhere.  Many notebooks today come equipped with embedded 802.11 radios, and a
notebook bought in the United States works perfectly well on wireless LANs in Japan, China, Italy,
or the United Kingdom.  Also, some governments and public-private partnerships are investing in
creating free public access, and some partnerships are working on technology to share access,
either for free or at low cost.  These developments will tend to sustain 802.11 over time.  
However, the footprints of major 3G networks will be sufficiently large to create a vast roaming
area good enough for many users, and 3G's other advantages will tend to take share from 802.11
throughout the forecast period.

Finally, a note about WiMAX or 802.16, a much-touted technology for an 802.11-like service that
can transmit wireless data at high speed over a metropolitan area.  Since it has a much wider
range than 802.11, WiMAX infrastructure is supposed to be much cheaper.  WiMAX can achieve a
theoretical maximum data rate of 75 Mbps over a 30 mile range.  However, true average
download throughput is likely to be much lower, and WiMAX has yet to be tested in the real world.  
It's still a technology of the future.  Also, users don't really need speeds that high.  UMTS-HSDPA,
the 3G scheme used by Lenovo, gets from 550Kbps to 1.1Mbps downloading, which, if you can
really get it, is plenty for most of the tasks that people actually do.  And although WiMax is
sometimes positioned as a low-cost wide area network solution, infrastructure and endpoint
costs are likely to be similar to 3G.  Also, WiMax providers will need to purchase spectrum to
ensure quality of service.  

Thus, it seems likely that, once embedded WWAN technology catches on, it will tend to take over
from other wireless schemes because it will have a much wider coverage area, a better story on
financial and technical handoff between networks, and — with embedded hardware and software
— a simpler usage model.  When service and hardware costs come down, which they will as
scale increases, WWAN-enabled notebooks will become the norm for commercial users and
even many consumers.

Forecast

Based on an assessment of 2005 shipments and assumptions about decreasing costs and
improving usage models, Endpoint believes that the U.S. embedded wide area wireless
notebook market will rise to nearly 20 million units in 2010 (Figure 1).  


Figure 1
The attach rate for U.S. commercial notebooks will hit better than 55%, and the consumer
notebook attach rate will reach nearly 16% (Figure 2).

Figure 2
Assumptions

The following assumptions were used as input to the forecasting model:

  • Attach for WWANs will be entirely notebook; zero desktops will be equipped

  • Forecast includes only notebooks with embedded radios; cards purchased separately or
    in the aftermarket are excluded

  • Although hundreds of thousands of  PCMCIA cards were sold in 2005, the embedded
    market was still tiny

  • Sony constituted the bulk of the market in 2005

  • Lenovo introduced 3G at end of 2005, accounting a small portion of the year's attach

  • 2006 will be the first year of wider adoption, with Dell, HP, and Panasonic joining Lenovo
    and Sony in the market with integrated units; other companies may announce later in the
    year

  • Major drivers include independent access while traveling and greater coverage than 802.11

  • Short-term inhibitors include slower data rates than wired, cost, onerous conditions on
    access plans, and a lack of standards; these factors will be mitigated over the forecast
    period

  • Service prices will decline from ~$80/month in 2005 to $25/month in 2010

  • Forecast is for hardware attach only, not activation rate

  • Although activation rates are not included in the forecast, they will likely rise from low
    double digits in 2005 to more than 50% in 2010

  • All major (and many tier 2) vendors will have integrated offerings in their lineups by 2008

  • Data rates above 200Kbps will be sufficient to stimulate the market

  • Carriers and PC hardware OEMs will make significant investments in demand generation

Conclusions and Recommendations

Although the WWAN attach rate for notebooks in the United States is currently quite low, the
market is only at its inception.  A happy convergence of propitious factors will lead to widespread
market adoption over the next five years.

For PC hardware OEMs, the advice is clearly: get this capability in your lineup if you haven't
already.  Also, there's a great marketing story to be told about WWANs: they confer freedom, a core
American value.  Tasteful advertising campaigns that play off this theme are likely to strum the
heartstrings and open the wallet.  Co-op relationships with carriers represent a good way to share
marketing costs.  

As far as technology is concerned, vendors should recognize that technology will improve over
time, but that waiting for higher speed links will delay market entry and the ability to build product
and brand awareness in the space.  A prudent strategy will involve bringing out first products
essentially now with the best technology available today, and to launch subsequent generations
with more advanced technology as it comes on the market.  

Activation rates are low now and will remain below hardware shipment levels throughout the
forecast period, although the gap will narrow, but one feature that would help stimulate usage of
the capability is Web activation, a simple method for users to sign up.  It is important for users to
get a sense of instant gratification during the signup experience.  

Finally, hardware vendors should give some thought to the effect universal connectivity will have
on form factors.  The high degree of mobility implied by WWAN access will place demands on
OEMs to come up with better mobility features, such as longer battery life (with an assumption that
the wide area radio will be on constantly), lighter weights (perhaps dispensing with the second
spindle, the optical drive, on more models), smaller sizes (for enhanced mobility), and daylight
screens (to improve outdoor readability).

For carriers, be aware that this will become a major market and an important source of new
revenue.  To stimulate this market, service pricing should be brought down as soon as is
practically possible.  Also, carriers should consider creating pay-as-you-go plans to reach into
lower demographics that do not like to be committed to a monthly payment and to appeal to
casual users who want to access the network only once in a while.  Such session-based pricing
will stimulate general demand as well.  Any doubts about the efficacy of pay-as-you-go can be
instantly dispelled by observing how widespread it is in the 801.11 world.  Of course, partnerships
with PC hardware OEMs will be important.  An important part of these partnerships will be joint
investments in awareness campaigns and demand generation, focusing on marketing elements
such as promotional offers, tie-ins, and advertising.

For those investing in alternative wireless technologies, keep an eye on this development.  You
may want to throttle back if this forecast appears to be coming true.  Particularly for backers of
802.11 and WiMAX, WWAN could be a disruptive development.

Finally, the technology needs a better name.  802.11 has greatly benefited in terms of public
awareness from the Wi-Fi moniker.  WWAN needs a similar handle.  Endpoint will be working with
the industry to try to establish such a name over the next weeks and months.  Suggestions are
welcome.
© 2006 Endpoint Technologies Associates, Inc. All rights reserved.
Freedom is Not Having to Ask
The Next Big Thing