AT&T Makes A Second Stab At The Computer Market

Information Processing

AT&T Makes A Second Stab At The Computer Market

With A Revamped Sales Force, It Will Emphasize Its Forte: Communications

Geoffrey C. Lewis and Mark Maremont in Morristown, N. J.
Business Week

April 1, 1985

Need a primer on the mistakes to avoid when entering the computer business? The folks at American Telephone & Telegraph Co. have written the book. Last March, AT&T jumped feet-first into the market for data processing equipment. But its products proved inadequate and its marketing strategy ill-defined. The result was disappointment. ''We really hadn't sorted out precisely what customers needed and how our product line fit in,'' says a rueful Robert E. Allen, the new chairman of AT&T Information Systems Inc., which markets the computers.

Now, AT&T is ready to write a different story. The company has revised its computer strategy to take better advantage of its strengths as a communications company. On Mar. 26 it will trot out a series of new products, including a powerful new personal computer. An updated line of minicomputers will follow by early summer, calculated to show critics that the company's vaunted technical prowess can be translated into competitive computer products.

SPLASHY ENTRY

''We intend to make it clear to our customers and potential customers that we have staying power, a good strategy, and good products,'' the soft-spoken Allen told BUSINESS WEEK in his first interview since he took charge of the Information Systems group seven weeks ago. The lessons of the first year, he vows, ''haven't been wasted.''

Once the flush of enthusiasm over its splashy entry into the market faded, AT&T quickly realized it lacked a coherent marketing strategy. It positioned its 3B series of minicomputers as general-purpose machines, for example. But their aging technology failed to measure up to more powerful minis from market leaders such as Digital Equipment Corp. Moreover, the AT&T models lacked the ability to communicate with International Business Machines Corp. mainframe computers, the mainstays of most corporate computing operations.

It was soon clear that few computer users were about to rewrite billions of dollars' worth of software just to accommodate AT&T's Unix Operating Systems, the basic command structure that controls its computers. Developed in the late 1960s by AT&T's Bell Laboratories for internal use, Unix has become widely available under license from AT&T. Recently, many small microcomputer vendors have tried to make Unix an industry standard to run on different brands of computers, enabling them all to use the same applications software.

That hasn't happened. Unix still lacks the wealth of business software that exists for other operating systems. Says Equitable Life Assurance Society Vice-President James W. Johnson: ''We had a detailed presentation from AT&T a year ago, and I did not come away with any reason to consider Unix.'' Shipments of all Unix-based computers last year lagged 57% behind projections (chart, page 92). Dataquest Inc. estimates that AT&T sold only $205 million worth of 3B minis to outside companies, while almost four times that amount wound up in-house and at local Bell phone companies.

Nor did customers flock to AT&T's personal computer line when it was introduced last June. The AT&T PC 6300, built by Italy's Olivetti, received high marks for speed and graphics clarity. But it was seen as simply another clone of IBM's top-selling Personal Computer. Even $40 million in advertising failed to give the 6300 more than a 1% share of the retail personal computer market in 1984. This year, however, the computer has gathered momentum. Retailers MicroAge Inc. and CompuShop Inc. now sell one 6300 for every five IBM PCs. At New York area Computer Factory Inc., IBM outsells AT&T by nine to one.

Blunders also beset AT&T's efforts to sell directly to customers through its 6,000-strong sales force. While the salespeople were qualified to hawk telephone systems, few were schooled in tailoring computer systems to customers' needs. Even Allen now concedes the sales force was ''not trained adequately.'' Lack of competitive hardware and software further hobbled their efforts. ''We just were not in a position to solve the business customer's data processing problems,'' says Allen L. Prestegard, an AT&T regional sales vice-president until last summer and now president of United Technologies Communications Co.

STREAMLINING

By last fall, the problems had become alarming to AT&T management. Its solution: Reorganize AT&T Information Systems by product lines to coordinate product development and marketing better. Separate marketing and sales fiefdoms previously handled both telephone and computer products, while development and manufacturing took place at another division. That made it impossible to translate customer needs quickly into products.

To streamline its computer effort, AT&T consolidated all computer-oriented operations in a division under James D. Edwards, 44, a veteran computer marketer who worked at IBM, Xerox, and Bausch & Lomb before joining AT&T in 1983. In another shift that indicates the level of concern at AT&T headquarters, Chairman Charles L. Brown early this year pulled Allen from a top corporate job to head the troubled Information Systems group in Morristown, N. J.

The dynamic Edwards has begun to bring order to AT&T's computer thrust and to improve the skills of its sales force. ''I'll bet we could stage a Pepsi taste test now and you wouldn't be able to tell the difference between an AT&T and an IBM salesman,'' he boasts. Customers confirm the improvement. ''I am impressed with the quality of the people I'm seeing now. They understand they have to sell, and if they get some better products, they can do it,'' says J. C. Penney Co. Vice-President Robert Capone.

The sales force won't have long to wait for sexier new computers. First out of the gate in late March will be the AT&T Unix PC, a desktop computer built by Convergent Technologies Inc. Based on a more powerful version of the microprocessor that powers Apple Computer Inc.'s Macintosh, the Unix PC will start at $5,000 as a personal computer but can be expanded to handle 10 users.

Analysts give the new model kudos for ease of use. Like the Apple Mac, it has a ''mouse'' pointing device and a high-resolution screen that can be divided into sections to display several programs. Befitting AT&T's background, it also has a built-in modem and communications software.

POTENTIAL HANDICAP

Still, the machine is not likely to add much to AT&T's $33 billion in revenues. PaineWebber Group Inc. analyst Stephen K. Smith estimates that AT&T will buy only 40,000 Unix PCs from Convergent this year and 70,000 in 1986. And because it runs only Unix software, the machine faces an uphill battle to reach a corporate audience wary of non-IBM-compatible personal computers. Although the AT&T model will be introduced with a score of popular software packages, Lotus Development Corp. has not yet decided to adapt its best-selling Lotus 1-2-3 program for Unix, a potential handicap.

At the same time, to make its older personal computer more competitive with IBM's newer PC/AT, AT&T is doubling the 6300's available disk storage. Other new products will emphasize AT&T's communications skills. One, a computer terminal with a phone attachment, connects to business phone systems, so users can talk on the phone and scan data simultaneously. Another is a network that ties 6300s, Unix PCs, and other micros to a shared data storage system using ordinary phone wiring.

On the heels of these products will be an upgrade of AT&T's older, entry-level 3B2/300 microcomputer. But the big news will come in May or June. Edwards will field a superminicomputer that combines multiple AT&T processors to deliver power comparable to that of the largest superminis. It is expected to provide the missing links to IBM mainframes that large customers demand.

NEW GAME PLAN

When AT&T sends its forces into the field this time, they will be equipped with a new game plan. The strategy still revolves around Unix, but now AT&T won't try to persuade customers to adopt it for general-purpose computing. Instead, Edwards has narrowed the focus to highlight Unix' superior communications capabilities. ''Unix is really good at moving information around,'' he explains. That makes it particularly effective where IBM has always been weakest: distributed data processing. This technique, pioneered in the 1970s by minicomputer companies, takes computing tasks from mainframes and distributes them to smaller computers that can handle a department or division.

The tricky part is to get these smaller systems to communicate with each other and with the mainframe. So AT&T is emphasizing Unix' knack for tying together everything from personal computers to mainframes. Although it is only a slice of the data processing pie, Edwards estimates that distributed systems represent a $10 billion annual market.

AT&T seems to have convinced some customers already. ''I really think they can fill the gap when it comes to transporting data from one system to another,'' says Gary A. Schnorr, manager of office automation at General Foods Corp., who is ordering AT&T minis to tie together 6300s and other PCs.

The revised Unix strategy could still backfire if AT&T's new products fail to ignite more enthusiasm. With more than 100 other companies selling Unix products, there is no particular reason to choose AT&T. And most ominous, IBM has added Unix as an option on systems ranging from PCs to mainframes. So although IBM still emphasizes its own ways to tie different hardware together, it is ready to offer the Unix alternative if customers insist. ''AT&T still has to deliver price- and performance-competitive products that will convince people to buy their hardware instead of other Unix boxes,'' observes Peter R. Tierney, a vice-president at Relational Technology Inc., a Unix software company.

After a year of false starts, AT&T is in a hurry to prove its new products will measure up. ''There's a hell of a lot of urgency to this,'' says Allen. ''This is not a window that's going to be open forever.'' In an industry as fast-moving as computers, even a $33 billion company can't afford many more missteps.

Graph: UNIX HAS FALLEN FAR SHORT OF EXPECTATIONS Photograph: INFORMATION SYSTEMS' ALLEN: ''WE INTEND TO MAKE IT CLEAR WE HAVE STAYING POWER'' Photograph: AT&T'S PC 6300 WAS SEEN AS SIMPLY AN IBM CLONE, BUT NOW IT'S GATHERING MOMENTUM