NCR: Grand National

Over its 112 years in business, NCR has seen many competitors fall by the wayside. How has it survived for so long and how is it planning to face the future? asks Sean Hallahan

If any company can be said to have created the culture that made IBM the most powerful force in the computer industry it is NCR. It was in 1895 that IBM?s founding father Thomas Watson joined NCR, then National Cash Register, as a salesman. The style of management that he learned at NCR was to be superimposed on IBM when it was founded in 1924.

Watson was sacked by NCR in 1913 and joined the Computing Tabulating Recording Company, but he took with him the aggressive sales-led culture that had been imbued in him at NCR.

The two companies continued to compete in both the hardware arena and in NCR?s core business, automated telling machines (ATM) and point of sale (POS) systems, throughout the succeeding decades. In 1991, NCR merged with one of IBM?s major rivals, the US telecoms giant, AT&T, which had been seeking to build a presence in the computer industry just as IBM had tried to force its way into the telecoms world. The merger of AT&T and NCR, called AT&T Global Information Systems, was short-lived. In 1995 the companies agreed to go their separate ways.

For all the talk of the convergence of telecommunications and computers, partnerships between the two sectors do not seem to work out. As AT&T was acquiring NCR, IBM was passing the remaining half share of the telecommunications company, Rolm, which it had acquired in 1984, on to Siemens. The merger between AT&T and NCR made more sense than the IBM-Rolm alliance. NCR was an early convert to Unix, which for many years was owned by AT&T, and there was a natural synergy between the two companies.

The separation from AT&T meant that NCR had to restructure its business. The company?s chairman and CEO Lars Nyberg outlined to employees the shape this restructuring would take in a speech to employees last year. The restructuring was to be based around five ?turnaround principles?, none of them overly radical but all containing more than a grain of common sense.

The principles were encapsulated under five headings: focus, accountability, expense-level discipline, process improvement and maintaining a sense of urgency. The focus initiative was basically a plan to concentrate the company?s efforts in the areas where it felt it was strongest and to abandon others where it was weak.

Like all the major players in the early 1980s NCR also sold PCs, but in 1995 it decided that the market was too crowded and the rewards too small and pulled the plug on the business. It will still supply PCs but only as a part of a larger hardware package based mainly around its Unix servers.

?Basically we were not making any money out of PCs. We manufactured our own PCs and had a $2 billion business but it was not profitable. Now we have a three-quarters of a billion dollar business and we are making a profit. There are really only two companies making money out of the PC business and they are Intel and Compaq,? said Peter Yarwood, UK channel director for NCR. The company now badges machines that are manufactured by SCI.

Yarwood admits that the break with AT&T worried some people at the time. AT&T is much larger and richer than NCR and there were fears that floating off AT&T GIS to emerge as a born-again NCR would hamper sales. In fact, according to Yarwood, NCR emerged from AT&T with a billion dollars in the bank and a firm product base.

NCR?s main strength lies in the Unix market and in the supply of banking automated telling machines and retail POS terminals. When the demerger with AT&T took place, NCR took with it Teradata which NCR itself had acquired in 1991. Teradata was one of the first relational database companies to recognise the importance of symmetrical multiprocessing (SMP) and massively parallel processing (MPP) and could be said to have led the field.

The importance of Teradata in the world of parallel databases is acknowledged in the Bloor Research report, Parallel Database Technology: An Evaluation and Comparison of Scalable Systems. ?There is a sense in which the existence of this report and all the evaluations within it, is a result of a database machine known as Teradata,? the report said.

Originally, Teradata was both a supplier of hardware and software but the situation changed with its takeover by NCR. NCR produced its own range of machines but to ensure brand continuity, retained the Teradata name for the database software. Apparently it was felt that the name Teradata, while being recognised in its own narrow field, did not fit well with NCR?s global ambitions. The Teradata DBC1012 was renamed the 3600 to give it the appearance of continuity with the earlier range of NCR machines.

?An anonymous name was required to reassure customers that this was simply a natural upgrade path from the SMP systems such as the 3500,? the Bloor Research report said.

Teradata specialised in symmetrical multiprocessing and massively parallel processing systems and was one of the first companies into the data warehousing field. NCR committed itself to focusing on the key technical areas of scalable data warehousing and high availability transaction processing (HATP). Although it is offering to provide data warehousing and HATP for any company that requires it, NCR is offering industry specific solutions to the financial, retail and communications sector where the company?s main strengths lie.

Until the start of this decade, NCR machines were based around the Motorola chipset. It had been highly successful in the 1980s, selling its Tower Unix-based medium to high-end servers. But around 1990, the company decided to move its chipset over to Intel.

?It was a pretty difficult time for a couple of years. Tower was a well-known product and was badged by other manufacturers. We moved from being a supplier of a multiprocessing- based medium scale server to an Intel-based smaller scale server,? Yarwood says. All NCR?s systems are now Intel-based.

After the break with AT&T, NCR was forced to reorganise itself. The company formed itself into five business units, the computer systems group, financial systems group, retail systems group, worldwide services and Systmedia group, to ensure accountability for products and services. But in his speech Nyberg warned that these units should collaborate.

?Customers in the industries they serve should see NCR as a single integrated company. If customers think of and deal with the financial systems group almost like a separate company from the computer systems group, they have missed a chance to leverage off each other in winning more of the customer?s overall information technology business,? he said.

Nyberg also called for tighter control over expenses, much of which comprises employees? salaries. But he was at pains to reassure them that, with the exception of contract staff, the total of 37,900 full-time staff would remain much the same.

According to Nyberg, process improvement is the most important of the five turnaround principles and he told NCR to improve its internal operational processes. Finally, he said, the company should retain a sense of urgency if it is to achieve its goal of being a $10 billion company by 1999. The company?s present turnover is $8.16 billion.

In fact Nyberg?s statement is effectively a continuation of the tight financial control culture which NCR?s founder, John Patterson, imposed from the very beginning. The merger with AT&T provided an interesting clash of cultures which AT&T, as the larger partner, gradually came to dominate.

?We got pulled in all sorts of directions. You had two 100- year-old customers with very different cultures. AT&T was very marketing driven while NCR was financially driven,? Yarwood explains. Since the demerger, Yarwood hopes that some of the marketing culture of AT&T has rubbed off on NCR and that the telecoms giant will have learned some lessons about financial control.

NCR is often an ignored force in the hardware world, but in 1979 it was the sixth largest mainframe manufacturer behind IBM, Sperry Rand, Honeywell, Burroughs, and Control Data by value of installed base. With the exception of IBM and the now reborn NCR, those companies have merged, been taken over or pulled out of the market.

Sperry Rand and Burroughs eventually became Unisys, Honeywell was taken over by Bull, and Control Data pulled out of the mainframe market. But NCR is back in the game in its own right and in bullish mood. But there is no doubt that it has a hard furrow to plough if it is to regain its presence in the processor market.

Although, according to NCR, consultantcy IDC put it at number one in the data warehousing market and among the leaders in HATP and Unix and Windows NT servers, it is from ATM and POS systems that it makes most of its money. NCR itself admits ?these businesses, including professional service and customer service, are about 60 per cent of NCR?s total revenues, and growing?.

Although NCR has a policy of selling through the channel, it is more cautious in its approach than many of its competitors. Many companies have set themselves a target of moving towards at least an 80 per cent indirect, 20 per cent direct sales strategy, but not NCR.

?It is not our intention to move to the Digital model,? Yarwood says. The sales of the high-end machines for data warehousing effectively preclude NCR from selling through the channel. Because of the nature of the high-end machines, the complexity of the databases and the nature of the companies buying the 5100 data warehousing systems, there is a two-year sales cycle, Yarwood explains.

Even if a reseller had the in-house expertise and considerable support from NCR to sell and support data warehousing, it is doubtful if its cash flow could warrant such a move. NCR has a two-tier channel structure. Its partners in profit programme is aimed at the higher end of the reseller chain where companies can be guaranteed to deliver $75,000 of business a year.

NCR?s distributors, Ideal Hardware and Sphinx CST, serve another 150 resellers.

It is possible that with the falling costs of hardware, improved communications and a growing number of vertical market software packages, NCR could have drawn on its experience in the retail sector to offer services to smaller customers. In fact the company did attempt to trade on its reputation in the large retail systems world to sell products and services through the channel. But the Vars that were interested were generally very small and the experiment did not work.

?We tried it in the past and in the distribution channel and it did not work,? Yarwood says. NCR has now appointed a member of staff in the retail division to work with the specialised Vars.

The reborn NCR has 112 years of experience and culture to draw upon. That its culture, management style and methods of operating did not sit well with the culture of AT&T should come as no great surprise. Novell?s decision to buy Wordperfect to gain a presence in the applications market and provide it with a stick to beat Microsoft also failed, as did the IBM- Rolm alliance.

The new NCR may not be the force it was in the 1970s but in its specialised areas ? data warehousing, POS and ATM systems and mid-range Unix servers ? it still counts as a formidable player.

The bold decision to withdraw from the PC rat race can only serve to focus the company?s emphasis on its areas of strength.