The Nexus One debacle: How does one beta test a phone?

The first absolute confirmation of the existence of the "Google Phone," as no one calls the Nexus One now, came when the company admitted it had distributed a few hundred models to its employees, now exactly one month ago. It could be the single shortest beta test period in the history of Google, especially compared to the number of years Gmail bore the "beta" tag before it was quietly removed last spring.

Clearly after the first week of Nexus One's retail availability, Google has proven itself not ready for instant launches. In one sense, the problem is actually systemic: For a new product to build market momentum leading up to its launch, its manufacturer needs not only to maintain secrecy but to nurture that secrecy like a cash crop -- Apple is easily the best company at nurturing secrecy as a virtue of any company, in any industry, in history. By contrast, the whole beta test process, and the basis of Google's software development model up to this point, has been transparency -- a type of openness which, when applied to the realm of hardware, feels more like nakedness.

Among the hundreds of messages pouring onto Google's product support board this week (including just one from Google itself, which amounts to a slightly longer variation of "Please stand by") are a few from defenders of the company, some who are veterans of Android phones since the first T-Mobile G1. Chill out, they tell complainers. We're all the earliest adopters of a new and basically untried product. Did we really expect everything to go off perfectly?

Perhaps not, but in light of that fact, argue disgruntled responders, shouldn't Google and T-Mobile (the carrier of choice) and HTC (the manufacturer) have prepared for the eventuality that nothing may go right, at least for a sizable plurality of customers?

The driver of the largest body of customer complaints thus far has been an apparent tendency for the Nexus One to switch off of 3G back to 2G EDGE service after just seconds of use, and often to switch back and forth between 3G and 2G -- thus wasting several seconds of performance in the interim. Some customers report no 3G service at all, in areas where other phones -- including the T-Mobile G1 -- report 3 bars of coverage. Other customers report reduced 3G capability on both the Nexus One and their other T-Mobile 3G phones in the same region.

Summing up the experience for those who have already lost hope is this message on Google's forum from a customer in Indiana who named himself Fiasco: "I am really disappointed in the way this has all panned out and am afraid that any future issues (defective updates, patches, etc.) will be met with the same silence. It is not worth it for the cost of this device, and the size of the company involved (Google) which should have an entire group dedicated to this phone since they pushed it with such fanfare. T-Mobile is caught in the middle, with no tech docs for the device and apparently their own smartphone department has received zero training. This has turned into Nexus Beta, and I cannot justify the expense for a device that does not work for so many people."

In another era -- one which appears to be rapidly fading into history -- when product cycles for both software and hardware were measured in years and not months (or even weeks), it was possible for a small team to adequately test the efficacy and stability of a product in relative secrecy, so that a vendor could release that product relatively assured of its reliability and even, in the legal sense, merchantability. But too many factors have conspired to render the circumstances for developing any such well-funded product under a known brand unattainable:

• Almost everything requires connectivity, whether it's a device or an application. The level of cooperation necessary for a large enough group to test the connectivity of a product over the Internet -- in the modern era, the only data network there is -- defies the secrecy needed to keep its existence under wraps, and to generate market momentum.
• Newfound processing efficiencies are driving platforms like afterburners attached to tricycles. Just last week at CES, we saw the unveiling of portable connectivity platforms that were at least three times more powerful than those shown the previous year, including from Intel and Qualcomm. With smartphone platforms slated for production this year that already not just match, but exceed, the processing power of the average notebook computer in everyday use, the software platforms behind some apps projects begun just last spring are already obsolete, forcing apps developers to upgrade and restart their efforts before ever having launched version 1.0.
• The network can't sustain the traffic. There's a genuine concern among customers now that even a rumored software fix from HTC won't be able to overcome the issues T-Mobile is facing with an overburdened network. As fast as processors are evolving to make ultra-high-bandwidth connectivity theoretically feasible, the capacity of current wireless networks to marshal all that traffic may not exist yet. And after AT&T sounded its warnings to that effect last month (though in a manner which appeared to blame customers, not architects, for the problem), the FCC -- busy conceiving its broadband plan for "7G" and "8G" for Congress in March -- reacted to those warnings as if it was only now hearing them.
• Consumers' expectations for product renewal are too high. Some of the response we saw to last week's news of Palm's announcement of the forthcoming Pré Plus and Pixi Plus for Verizon Wireless, was that Palm was merely substantiating an old platform -- webOS, which was the cutting-edge story exactly one year ago. Perhaps testament to its continued marketing genius, Apple's forthcoming announcement of a likely tablet PC as its "next big thing" -- something the market may not really need -- has already successfully diverted the attention of its faithful supporters away from the subject of "the next iPhone," which they would probably have already expected by this time otherwise.
• "Old" is too effective a marketing tool. It's on a very rare occasion that I can go try out smartphones at some retailer, and not have a salesperson notice my BlackBerry 8830 and compare it to an extinct species -- a phone whose contract has only just now expired. In order to generate a desire among consumers to upgrade, vendors instill the notion that their current units' viability has expired, long before the contract date. As the development cycles of apps synchronize themselves with those of their platforms, the window of opportunity for testing becomes narrower and narrower.

Though few observers of the gadget business will consider the US/Mexico border surveillance system a "platform," a comment made by a Boeing official on 60 Minutes last Sunday stated as much about the changing nature of the technology industry as anything I've heard in years. Boeing had been awarded a government contract to install a security monitoring network from one end of the border to the other, but well after the completion deadline, it had only managed to cover a small fraction of the border -- and even then with a platform that couldn't tell the difference between an immigrant and a raindrop. Now, the government finds itself working with Boeing to rip and replace the portion of the network it installed.

The Boeing official admitted to CBS' Steve Kroft that the first try was a failure. But in writing it off, he applied this term to it: He called it just the "beta."

The word that once was synonymous with "cutting edge" is being leveraged to refer to the incomplete, failed first attempt.