Tuesday, September 28, 2010

Early Indications September 2010: The Power and Paradoxes of Usability

Usability is among the most difficult of topics to define and analyze.
At one level, it is much like the famous Supreme Court justice who
noted of potentially criminal extreme sexual images, "you know it when
you see it." At another level, the number of daily moments that
reinforce the presence of poor design can be overwhelming. Examples
are everywhere: building entrance doors with a grab handle you're
supposed to push but that you instinctively (and unsuccessfully) pull,
all manner of software (in Outlook, does hitting "cancel" stop the
transaction or clear a meeting from the calendar?), and pinched
fingers and scraped knuckles. Usability may be easy to spot, but it
is clearly very difficult to engineer in.

Systems

Why is this so? As Don Norman, one of the heroic figures in modern
usability studies, puts it in a recent ACM piece, complex products are
not merely things; they provide services: "although a camera is
thought of as a product, its real value is the service it offers to
its owner: Cameras provide memories. Similarly, music players provide
a service: the enjoyment of listening." In this light, the product
must be considered as part of a system that supports experience, and
systems thinking is hard, complicated, and difficult to accomplish in
functionally-siloed organizations.

The ubiquitous iPod makes his point perfectly:

"The iPod is a story of systems thinking, so let me repeat the essence
for emphasis. It is not about the iPod; it is about the system. Apple
was the first company to license music for downloading. It provides a
simple, easy to understand pricing scheme. It has a first-class
website that is not only easy to use but fun as well. The purchase,
downloading the song to the computer and thence to the iPod are all
handled well and effortlessly. And the iPod is indeed well designed,
well thought out, a pleasure to look at, to touch and hold, and to
use. Then there is the Digital Rights Management system, invisible to
the user, but that both satisfies legal issues and locks the customer
into lifelong servitude to Apple (this part of the system is
undergoing debate and change). There is also the huge number of
third-party add-ons that help increase the power and pleasure of the
unit while bringing a very large, high-margin income to Apple for
licensing and royalties. Finally, the 'Genius Bar' of experts offering
service advice freely to Apple customers who visit the Apple stores
transforms the usual unpleasant service experience into a pleasant
exploration and learning experience. There are other excellent music
players. No one seems to understand the systems thinking that has made
Apple so successful."

One of the designers of the iPod interface, Paul Mercer of Pixo,
affirms that systems thinking shaped the design process: "The iPod is
very simple-minded, in terms of at least what the device does. It's
very smooth in what it does, but the screen is low-resolution, and it
really doesn't do much other than let you navigate your music. That
tells you two things. It tells you first that the simplification that
went into the design was very well thought through, and second that
the capability to build it is not commoditized." Thus more complex
management and design vision are prerequisites for user
simplification. (Mercer quoted in Bill Moggridge, Designing
Interactions
(Cambridge: MIT Press, 2007))

Because it requires systems thinking and complex organizational
behavior to achieve, usability is often last on the list of design
criteria, behind such considerations as manufacturability or modular
assembly, materials costs, packaging, skill levels of the factory
employees, and so on. The hall of shame for usability issues is far
longer than the list of successes. For every garage door opener, LEGO
brick, or Amazon Kindle, there are multiple BMW iDrives, Windows
ribbons, European faucets, or inconsistent anesthesia machines:
doctors on a machine from company A turned the upper right knob
clockwise to increase the flow rate, but had to go counter-clockwise
on company B's machine in the next operating room over. Fortunately,
the industry has standardized the control interface, with a resulting
decline in human endangerment. (See Atul Gawande, Complications: A
surgeon's notes on an imperfect science
(New York: Macmillan, 2003))

Paradoxes

As Ronald Rust and his colleagues have shown, usability presents
manufacturers of consumer electronics with a paradox. In purchase
mode, buyers overemphasize option value in their purchase
consideration: if multifunction device from company D does 13 things
and a competitor from company H performs 18 actions, the potential
utility is overemphasized even if the known need is only for, say, six
tasks. Watching the evolution of the Swiss Army knife testifies to
this phenomenon: very few of us, I suspect, have precisely the tools
we a) want or b) use on our knife.

Once they get that 18-way gadget home, however, option value recedes
and usability comes to the fore, and the extra controls, interfaces,
and other factors that drive complexity can make using the more
"capable" device frustrating at best and impossible at worst. At
consumer electronics retailers, most returned items function
perfectly, but are often returned because they are too hard to
integrate into everyday life. (They may also be returned because
consumers routinely seek better deals, get tired of a color or finish,
or use the purchase essentially as a free rental, performing a task
then returning the device.)

Hence the paradox: does the designer back off on features and
capabilities, and thus lose the head-to-head battle of shelf-side
calculus in order to win on usability, or do purchase rather than use
considerations win out? There are some ways out of this apparent
paradox: modular add-ons, better point-of-sale information, and
tutorials and other documentation (knowing that the vast majority of
people will never read a manual). The involvement of user groups is
growing, for both feedback on products in development and support
communities for stumped users. (Roland T. Rust, Debora Viana Thompson,
and Rebecca W. Hamilton, "Defeating Feature Fatigue," Harvard Business
Review
, February 2006)

At its worst, overwhelming complexity and other forms of poor
usability can kill, as the anesthesia example makes clear. Nuclear
power plants, military hardware, and automobiles provide ready
examples. Especially with software-driven interfaces becoming the
norm (even for refrigerators and other devices with little status to
report and few user-driven options to adjust), the potential for
either bugs or unforeseen situations to escalate is becoming more
common.

Beyond Gadgets

This essay will not become a tribute to Apple or Southwest Airlines,
however, if only to escape the cliche. Instead, I'd like to discuss a
recent video
by TED producer Chris Anderson. In it he looks at the
proliferation of online videos as tools for mass learning and
improvement. Starting with the example of self-taught street dancers
in Brazil, Japan, LA, and elsewhere, he argues that the broad
availability of video as shared show-and-tell mechanism spurs, first,
one-upmanship through imitation and then innovation. The level of TED
talks themselves, Anderson argues, provides home-grown evidence that
cheap, vivid multimedia can raise the bar for many kinds of tasks:
futurist presentations, basketball dunks, surgical techniques, and so
on.

Five things are important here.

1) The low barrier to entry for imitator/innovator #2 to post her
contribution to the discussion may inspire, inform, or infuriate
imitator/innovator #3. Mass media did some of these things (in
athletic moves, for example: watch a playground the week after the
Super Bowl or a halfpipe after the X games). The lack of a feedback
loop, however, limited the power of broadcast to propagate secondary
and tertiary contributions.

2) Web video moves incredibly fast. The speed of new ideas entering
the flow can be staggering once a video goes "viral," as its
epidemiological metaphor would suggest.

3) The incredible diversity of the online world is increasing every
year, so the sources of new ideas, fresh thinking, and knowledge of
existing solutions multiply as well. Credentials are self-generated
rather than externally conferred: my dance video gets views not
because I went to Julliard but because people find it compelling, and
tell their friends, followers, or colleagues.

4) Web video is itself embedded in a host of other tools, both social
and technical, that are also incredibly easy to use. Do you want to
tell someone across the country about an article in today's paper
newspaper? Get out the scissors, find an envelope, dig up his current
address, figure out correct postage (pop quiz: how much is a
first-class stamp today?), get to a mailbox, and wait a few days.
Want to recommend a YouTube or other web video? There are literally
hundreds of tools for doing so, essentially all of which are free and
have short learning curves.

5) Feedback is immediate, in the form of both comments and views
counters. The reputational currency that attaches to a "Charlie bit
my finger" or "Evolution of dance" is often (but not always)
non-monetary, to be sure, but emotionally extremely affecting
nonetheless.

With such powerful motivators, low barriers to participation, vast and
diverse populations, rapidity of both generation and diffusion, and a
rich ancillary toolset relating to online video, Anderson makes a
compelling case for the medium as a vast untapped resource for
problem-solving on multiple fronts. In addition, because it involves
multiple senses, the odds that a given person will grasp my ideas
increases as the viewer can hear, watch, or read text relating to the
topic.

Thus the power of extreme usability transcends gadgets, frustration,
and false-failure returns. When done right, giving people easy access
to tools for creation, distribution, interpretation, and
classification/organization can help address problems and
opportunities far beyond the sphere of electromechanical devices.
Apart from reducing frustration, improving safety, or increasing
sales, lowering barriers to true engagement (as in the web browser,
for example) may in fact help change the world.

Wednesday, September 01, 2010

Early Indications August 2010: Rethinking Location and Identity

Even though they're sometimes overlooked in relation to spectacular
growth rates (50x increases in wireless data carriage), successful
consumer applications (half a billion Facebook users), and technical
achievement (at Google, Amazon, Apple, and elsewhere), location-based
technologies deserve more attention than they typically receive. The
many possible combinations of wired Internet, wireless data, vivid
displays, well-tuned algorithms running on powerful hardware, vast
quantities of data, and new monetization models, when combined with
location awareness, have yet to be well understood.

Digital location-based services arose roughly in chronological
parallel with the commercial Internet. In 1996, GM introduced the
OnStar navigation and assistance service in high-end automobiles.
Uses of Global Positioning System (GPS, which, like the Internet, was
a U.S. military invention) and related technologies have exploded in
the intervening years, in the automotive sector and, more recently, on
smartphones. The widespread use of Google Earth in television is
another indicator of the underlying trend.

Handheld GPS units continue to double in sales every year or two in
the North American market. As the technology is integrated into
mobile phones, the social networking market is expected to drive far
wider adoption. Foursquare, Gowalla, numerous other startups, and the
telecom carriers are expected to deliver more and more applications
linking "who," "where," and "when." Powerful indications of this
tendency came when Nokia bought Navteq (the "Intel inside" of many
online mapping applications) for $8.1 billion in 2007, when Facebook
integrated location services in 2010, and when the rapid adoption of
the iPhone and other smartphones amplified the market opportunity
dramatically. Location-based services (whether Skyhook geolocation,
Google Maps and Earth, GPS, and others) have evolved to become a
series of platforms on which specific applications can build, tapping
the market's creativity and vast quantities of data.

In the process, the evolution of location taps into significant questions:

-Who am I in relation to where I am? That is, what are the
implications of mapping for identity management?

-Who knows where I am, when I'm there, and where I've been? How much
do I control the "information exhaust" related to my movements? Who
is liable for any harm that may come to me based on the release of my
identity and location?

-Who are we relative to where we are? In other words, how do social
networks change as they migrate back and forth between virtual space
(Facebook) and real space (Mo's Bar)? What happens as the two worlds
converge?

Variations on a Theme

While location often seems to be synonymous with GPS, location-based
data services actually come in a variety of packages. Some examples
follow:

-Indoor Positioning Systems
For all of the utility of GPS, there are numerous scenarios where it
doesn't work: mobile x-ray machines or patient gurneys in hospitals,
people in burning buildings, work-in-process inventory, and
specialized measurement or other tools in a lab or factory all need to
be located in sometimes vast and often challenging landscapes,
sometimes within minutes. GPS signals may not penetrate the building,
and even if they can, the object of interest must "report back" to
those responsible for it. A variety of wired and wireless
technologies can be used to create what is in essence a scaled-down
version of the GPS environment.

-Optical
Such well known firms as Leica and Nikon have professional products to
track minute movements in often massive structures or bodies: dams,
glaciers, bridges. Any discussion of location awareness that neglects
the powerful role of precision optics, beginning with the essential
surveyor's transit, would be incomplete.

-WiFi mapping
As we have seen, the worldwide rise of wi-fi networking is very much a
bottom-up phenomenon. Two consequences of that mode of installation
are, first, often lax network security and second, considerable
coverage overspill. Driving down any suburban or metropolitan street
with even a basic wireless device reveals dozens of residential or
commercial networks. Such firms as Google have systematically mapped
those networks, resulting in yet another overlay onto a growing number
of triangulation points. The privacy implications of such mapping
have yet to be resolved.

-Cellular
Wireless carriers can determine the position of an active (powered-up)
device through triangulation with the customer's nearby towers. Such
an approach lacks precision when compared to approaches (most notably
GPS) that reside on the handset rather than in the network. In either
case, the carrier can establish historical location for law
enforcement and potentially other purposes.

-Skyhook
A startup based in Boston, Skyhook has built a database of 100 million
wi-fi physical coordinates then added both GPS and cellular
components, making Skyhook most precise (inside or near buildings)
where GPS is weakest. A software solution combines all available
information to create location-tracking for any wi-fi enabled device,
indoors or out. Skyhook powers location awareness for devices from
Apple, Dell, Samsung, and other companies, and is now generating
secondary data based on those devices.

Landmarks

Noting a few historic transitions and innovations in the history of
location-based services reveals the scale, complexity, and wide
variety of applications that the core technologies are powering.

OnStar
With roughly 5.5 million subscribers in mid-2010, OnStar has become
the world's largest remote vehicle-assistance service. In addition to
receiving navigation and roadside assistance, subscribers can have
doors unlocked and gain access to certain diagnostic data related to
that particular vehicle. The service delivers important information
to emergency response personnel: when extricating occupants from a
damaged vehicle, knowing which airbags have deployed can assist in
keeping EMTs, police, and firefighters safe from the explosive force
of an undeployed device that might be inadvertently tripped. Knowing
the type and severity of the crash before arrival on the scene can
also help the teams prepare for the level of damage and injury they
are likely to encounter.

The service was launched as a joint venture. General Motors brought
the vehicle platform and associated engineering, Hughes Electronics
managed the satellite and communications aspects, and Electronic Data
Systems, itself being spun out from GM in OnStar's launch year,
performed systems integration and information management.

GPS
The history of GPS is even more compelling when considered alongside
its nearly contemporary stable mate, the Internet. GPS originated in
1973, ARPANET in 1969. Ronald Reagan allowed GPS to be used for
civilian purposes after a 1983 incident involving a Korean Air Lines
plane that strayed into Soviet airspace. The Internet was handed off
from the National Science Foundation to commercial use in 1995; Bill
Clinton ordered fully accurate GPS (20 meter resolution) to be made
available May 1, 2000. Previously, the military had access to the
most accurate signals while "Selective Availability" (300 meter
resolution) was delivered to civilian applications.

Since 1990, GPS has spread to a wide variety of uses: recreational
hiking and boating, commercial marine navigation, cell phone
geolocation, certain aircraft systems, and of course vehicle
navigation. Heavy mining and farming equipment can be steered to less
than 1" tolerances. Vehicles (particularly fleets) and even animals
can be "geofenced," with instant notification if the transmitter
leaves a designated area. In addition to latitude and longitude, GPS
delivers highly precise time services as well as altitude.

Trimble
Founded by Charles Trimble and two colleagues from Hewlett-Packard in
1978 (the first year a GPS satellite was launched), Trimble Navigation
has become an essential part of geolocation history. From its base in
Silicon Valley, the company has amassed a portfolio of more than 800
patents and offers more than 500 products. Much like Cisco, Trimble
has made acquisition of smaller companies a core competency, with many
M&A moves in the past ten years in particular. A measure of Trimble's
respect in the industry can be seen in the quality of its
joint-venture partners: both Caterpillar and Nikon have gone to market
jointly with Trimble.

The company has a long history of "firsts": the first commercial
scientific-research and geodectic-survey products based on GPS for
oil-drilling teams on offshore platforms, the first GPS unit taken
aboard the space shuttle, the first circuit board combining GPS and
cellular communications. The reach of GPS can be seen in the variety
of Trimble's product offerings: agriculture, engineering and
construction, federal government, field and mobile worker (including
both public safety and utilities applications), and advanced devices,
the latter indicating a significant commitment to R&D.

Location, Mobility, and Identity

Issues of electronic identity and mobility have been playing out in
quiet but important ways. Each of several instances is a classic case
of social or economic problems being tangled up with a technology
challenge. To see only one side of the question is to create the
possibility of unintended consequences, allow hidden agendas into
play, and generally confuse the allocation of sometimes-scarce
resources.

-Social Networking Goes Local
Whether through Dodgeball, (a New York startup that was bought by
Google in 2005 then left unexploited), Foursquare, or Facebook Places,
the potential for the combination of virtual and real people in
virtual or real places is still being explored. Viewed in
retrospect, the course of the Dodgeball acquisition raises the revenue
questions familiar to watchers of Friendster et al: who will pay for
what, and who collects, by what mechanism? Who owns my location
information and what aspects of it do I control? Much like my medical
records, which are not mine but rather the doctor's or hospital's,
control appears to be defaulting to the collector rather than the
generator of digital bread crumbs.

-The Breakdown of 911
After a series of implementations beginning in 1968, Americans on
wireline voice connections could reliably dial the same three-digit
emergency number anywhere in the country. As the Bell System of the
twentieth century fades farther and farther from view, the presumption
of 911 reliability declines proportionately with the old business
model even as demand increases: the U.S. generates about 12 million
calls a day to 911. The problem comes in two variants.

First, a number of Voice over IP customers with life-threatening --
and as it turned out, life-ending -- emergencies could only reach a
recording at Vonage saying to call 911 from another phone. The Texas
Attorney General is raising the question after a 911 call failed
during a home invasion in Houston. A baby's death in Florida was
blamed on a Vonage 911 failure. According to the Wall Street Journal,
"In a letter to Florida's Attorney General, [the mother] said the
Vonage customer-service representative laughed when she told her that
Julia had died. 'She laughed and stated that they were unable to
revive a baby'. . . ."

For their part, Vonage includes bold-print instructions for manual 911
mapping during the sign-up process, but it's been estimated that up to
a quarter of the U.S. population is functionally illiterate. One
feature of VoIP is its portability: plug the phone into an RJ45 jack
anywhere and receive calls at a virtual area code of the customer's
choice. Navigating firewalls, dynamic IP addresses, wireless
connections, and frequent network outages taxes anyone but the most
technically adept Internet user. Children are also a key 911
constituency. Taken collectively, these overlapping populations raise
dozens of tricky questions. At the infrastructure level, the FCC and
other agencies face the substantial challenge of determining the
fairest, safest set of technical interconnection requirements
incumbent on the Regional Bells and VoIP carriers.

From the Bell perspective, 911 obviously costs money to implement and
maintain, and declining wireline revenues translate to declining 911
funds. Connecting 911 to the Internet in a reliable, secure manner is
nontrivial -- network attacks have used modems to target the service
in the past -- and until contractual arrangements are finalized there
is reluctance to subsidize the same firms that present themselves as
full wireline replacements.

911 isn't just a VoIP problem either: cellular users represent nearly
75% of emergency callers, but math and economics conspire to make
finding them difficult or impossible. In rural areas, cell towers
often follow roads, so attempting to triangulate from three points in
a straight line can limit precision. States have raided 911 tax
revenues for budget relief.

-Cell phone tracking
The wireless carriers offer a variety of services that give a relative
(often a parent, or an adult child of a potentially confused elder)
location information generated by a phone. the service has also been
used to help stalkers and abusive spouses find their wives in hiding.
Women's shelters routinely strip out the tracking component of cell
phones; according to the Wall Street Journal, a Justice Department
report in 2009 estimated that 25,000 adults in the U.S. were victims
of GPS stalking every year. In addition to the carriers, tracking
capability is being developed by sophisticated PC users that spoof the
behavior of a cell tower. Keystroke and location logging software is
also available; one package, called MobileSpy, costs under $100 per
year.

Conclusion

As the telephone system migrates from being dominated by fixed lines,
where identity resided in the phone, to mobile usage, where identity
typically relates to an individual, location is turning out to matter
a lot. Mobile number portability was an unexpectedly popular mandate
a few years ago, and the fastest technology adoption in history was a
phone feature: 55 million people signed up in a matter of months for a
service -- the Federal Do Not Call registry -- that didn't exist when
it was announced. (That's even faster than the previous champ,
Netscape Navigator's zooming to 38 million users in 18 months.) Given
the global nature of some of these questions, not to mention numerous
issues with ICANN and DNS, the discussions and solutions will only get
more complicated. As the examples illustrate, getting social
arrangements to keep pace with technology innovation is if anything
more difficult than the innovation itself.