The Yankee Group Inc. and other analysts are still projecting huge revenue from mobile commerce, or m-commerce for the hipsters. The latest estimate is that some 50 million cell phone users in the U.S. alone will buy US$15 billion worth of goods, services and "content" by 2006.
That's a little over 4 years from today. Enterprises may have better odds, and a bigger payoff, just by picking up a few lottery tickets.
A recent Yankee report, "Mobile payments: What are they worth?" like all such reports, is built on a whole nest of assumptions, some explicit, others implicit. And the authors leave lots of wiggle room: even at $15 billion, those purchases will only account for a paltry 2 percent to 3 percent of all noncash, electronic transactions.
Why is that, you might ask? Good question.
The reason is that most mobile applications today are bad, rotten, lousy applications. And there's little evidence they'll get better by 2006, when $15 billion is supposed to light up the lives of some happy investors and executives.
But there is one undeniably successful mobile application that can make use of wired and wireless connections: the ATM cash networks.
I went out the other night to buy a bottle of wine and some cold medicine. I didn't have dollar bills with me, but I did have my trusty ATM card, which the liquor store and the pharmacy accepted just as readily and almost as easily as cold cash. Then, I remembered we needed coffee for the next morning, so I stopped at the supermarket and bought some French Roast. Again, I used the ATM card. Again, there were no problems.
This is, by any definition, a viable, valuable mobile application: I can use the card almost anywhere there's a magstrip reader. The magnetic strip on my card and my PIN number are sent over the in-store net, through a telco net, to my bank, which confirms that I have funds in the account to cover the payments. Electronically, different banks debit my account and credit the retailer's account. I don't have to calculate in advance how much money I need or carry it around with me. Instead, I carry a "token" that authorizes me to an application, accessed by one or more networks in between.
This kind of simplicity is missing from almost all the mobile applications now being deployed, certainly in the consumer market. The ATM's simplicity is in two areas: an infrastructure that almost always works reliably and predictably, and a simple, clear application design.
The infrastructure for local and long-distance wireless is already in place: wireless LANs or wide-area cell phone nets used for wireless data calls, or a combination, have enough horsepower to move selected data pretty quickly and reliably. You don't have to wait for bigger, better and oh-by-the-way much more expensive future nets. Firewalls, Radius Servers and virtual private nets are needed security components.
But the key weakness remains application design.
Mobile applications are different from traditional PC applications. Recent tests of wireless application protocol (WAP) phones in London, for example, turned up the fact that it took a numbingly long time to train users how to navigate the various screens, make selections, and in short, get work done. The process remained numbingly slow and didn't improve significantly with practice. Most people in the test who tried to use the phone to check a television schedule for the evening, gave up when they realized it was faster, cheaper and easier to buy a newspaper and throw everything away except the TV listings.
This is a lesson for enterprise IT groups looking at mobile applications: What's the critical information that either needs to be collected from, or distributed to, the end-user via a wireless connection? What's the clearest and simplest way for the end-user to enter or view that information?
Enterprise IT has a huge advantage over the carriers and others who are trying to figure out the next "killer mobile application," which is probably a mirage anyway. The enterprise already has the data and applications that are essential to its success.
One vendor that seems to have figured this out is Kirkland, Wash., WaveLink Corp. The company makes tools, such as WaveLink Studio, to outfit enterprise applications for wireless deployment. WaveLink supplies sets of class libraries, each targeted to a different client operating system, for building the graphical user interface for a given device. An added plus: developers simply work with their existing application tool sets.
There's also a small piece of client code that's downloaded to each device in order to interact with the WaveLink Server. The client program can exploit a barcode scanner on the handheld or its serial port, and so on - something that's impossible if your client interface is simply a Web browser.
The WaveLink Server software runs a set of programs that act as services to the application. For example, the Discovery Manager detects a request from a device, identifies the device characteristics, authenticates the user, and displays the appropriate list of applications or data queries. The user doesn't have to enter a host ID or host TCP/IP port number for each application. The Connection Server program links the client device to a TCP/IP port to access a given application, which can run on Unix, Windows, Linux or AS/400 servers.
The new release of WaveLink Studio, Version 4.0, will be available at the end of this month, priced at $5,000 for the server software (which can run on any number of computers) and $295 per client, with the usual volume discounts.
In Version 4.0, WaveLink engineers added the following features:
* Load balancing, enabling connection requests to be spread over several servers.
* Better scaling, letting you add ports and servers as users and traffic grow.
* Streamlined auto-discovery, meaning there's no end-user fiddling needed to work with different applications.
* Improved session management, enabling users to start where they left off after a wireless link is restored.
According to WaveLink executives, most of their customers' applications are focused on the wireless LAN. Most of the applications consist of transaction processing that's moving out to mobile workers, replacing pen and paper with faster, more accurate electronic data updates to back-end databases. Cost savings, they say, can be substantial because this kind of data lets an enterprise avoid stock-outs and minimize excess inventory.
But increasingly, according to WaveLink, enterprises are using mobile applications and wireless connections to improve customer relations. One regional airline, JetBlue, based in New York's JFK Airport, uses Wavelink software on mobile devices as a so-called "line-buster," to speed up ticket processing during peak travel periods. Mobile technology saves the airline vast amounts compared to renting additional counter space and setting up fixed, wired workstations that lie unused except during peak periods.
All of these are applications with measurable returns on investment. For that reason, the only response to the worsening economy that WaveLink execs claim to see is that some wireless projects are being phased in, to spread the costs over a longer period. "These applications are deployed to reduce costs," says Eric Hermelee, WaveLink's vice president of marketing. "They tend to get deployed when times are tough."