Mad Dog 21/21: The Show-Me State
June 9, 2014 Hesh Wiener
Between 1981 and 1993, IBM liquidated an extraordinary asset. Its value had been building since Thomas J. Watson became head of Computing-Tabulating-Recording Company, IBM’s birth name, in 1914. That asset was a vast base of rental equipment. Rentals had fueled and stabilized IBM and fostered its binding relationship with customers. Once gone, that rental base could never be rebuilt.
At the helm when this sea change began was John Opel, who hailed from Missouri, the Show-Me State. He probably didn’t understand what IBM showed him. His successor, John Akers, caught on, but by then it was too late.
In 1993, IBM brought in Lou Gerstner to run things. Gerstner came up with a strategy that produced results similar to the outcome IBM had previously achieved using its rental base. Basically, IBM figured out how to take over some or all of the IT department technical functions at customers. Today, IBM’s services operation dwarfs its hardware manufacturing business, which, combined with IBM Global Finance, account for pretty much all IBM’s intake directly connected to equipment. Gerstner also increased the cost-cutting efforts begun by John Akers. The IBM downturn slowed. Then the company’s fortunes began to improve. After that, IBM was able to grow again.
Today, IBM’s revenue is just shy of $100 billion, up considerably from the $64 billion of 1993. Its payroll is 431,000, about the same as it was at IBM’s prior peak size when John Akers began his term as CEO. Along the way, IBM’s headcount dipped so something in the vicinity of 250,000. There is a lot more to this than a period of redundancies and attritions followed by one of hiring. In Akers’ time, Americans were the most numerous IBMers, today that distinction belongs to Indians.
It is very hard to explain IBM’s behavior during the Opel and Akers years. The company’s formidable legal team from Cravath, Swaine and Moore broke the government’s grip on IBM in 1982. The lead lawyer for Cravath was Tom Barr, from Missouri. He showed ’em. And Barr, unlike the fellow Missourian heading IBM, understood very well what he did.
As a result of Cravath’s court success, IBM had more freedom to operate its business as it pleased than at any time since 1956, when the company signed a consent decree that considerably restricted its behavior to end a very effective government antitrust suit. If IBM had chosen to continue basing its equipment business on a rental base, it could have done so. Moreover, some policies that had been constrained or curtailed by the 1956 consent decree would no longer be restricted.
IBM might have been able to not merely continue its rental scheme but instead to reshape it so it worked more effectively in the business environment of the 1980s. That environment made leasing very attractive to corporations, particularly during the first few years of the Reagan Administration, when tax policies favoring lessors–including banks, leasing specialists, and finance subsidiaries of manufacturers and conglomerates–attracted considerable interest. But there was a lot more to IBM’s rental plan than the part bean counters saw. The rental relationship involved trust and mutual dependency on the part of IBM and its customers. IBM trusted customers to keep using its equipment and paying rent on a monthly basis until they purchased the gear or traded it for something newer. Customers trusted IBM to provide impeccable support and service along with practical pricing (even if it was on the high side), making the executives who ordered from IBM look competent to their colleagues and superiors.
Rental with accruals financing plans are still offered by some leasing companies, but it is a scheme that only works for capital equipment that depreciates gracefully. The asset at the heart of the deal must have a long economic life for the technique to work out well for the customer and leasing company alike.
Here is a sketch of the plan used by IBM: A customer would rent equipment for what seemed to be a reasonable price, perhaps 2 to 3 percent of the full list price. The rental fee might have included maintenance. So, equipment that sold for $100,000 might rent for $3,000 a month. But the rent got the customer more than use of the machine along with basic support. It entitled the renter to eventual purchase at a discounted price. The discount increased each month until it reached a specified maximum, after which the discounted amount would no longer accrue. A specific case based on our hypothetical $100,000 machine might have a 40 percent accrual and a 50 percent maximum total accrual. Each month 40 percent of the rental, $1,200 in this example, would be added to the customer’s accrual fund until it reached $50,000, at which time accruals would cease. At the end of the forty-second month of rental, the customer could buy the machine for $50,000. If the customer bought the hardware, the only continuing cost would be for maintenance, which was most often provided by IBM but was sometimes available from third-party maintenance companies.
IBM used its rental with accrual scheme to control the behavior or its user base. As equipment became old or obsolete, IBM might raise rentals to get more customers to buy the old gear. Alternatively, in cases where it was clear that the market value of the capital asset was falling rapidly, IBM might reduce the sale price while keeping the accrual amounts in place, in effect making the maximum accrual contribution to the purchase price more than 50 percent. Changes in IBM’s pricing details altered the propensity of customers to buy gear they had on rent.
Basically, IBM had considerable influence over customers’ behavior. When it liquidated its rental base, moving to a mix of sales and more conventional leasing, it lost control.
Most observers of IBM during the early 1980s didn’t foresee the consequences of IBM’s change in strategy. Instead, as IBM sold off its rental portfolio, they watched Big Blue pick up a phenomenal amount of cash. Later, towards the end of that decade, people inside and outside the company realized that IBM’s rising cash wealth came because it stopped selling corn and instead sold off the whole farm.
IBM added to its own misery by never finding a way to get a good grip on the PC business it entered in 1981. Makers of PC clones, notably Compaq, proved to customers that they were faster on their feet than IBM. Big Blue might have had a chance to reboot its PC business by moving from its open product architecture to a proprietary one. Flush with cash, IBM could have, for instance, bought Apple and, if it didn’t ruin the company, end up with a product line as distinctive as it had in mainframes or midrange business computers. (The PS/2 was an attempt at such proprietary differentiation.)
After Opel and Akers were gone, when Gerstner was in charge and IBM was on the mend, the company tried to compensate for losing its grip on PCs. Unfortunately, it tried to do this by buying into the PC software business via Lotus, missing the Apple application market the way it missed the Apple hardware market. Lotus had already been overtaken by Microsoft in its key product categories–productivity software and messaging–so the best IBM could do was sloppy seconds. As was the case with PC hardware, IBM’s PC software offerings were good, but not good enough. IBM’s competitors were excellent, better, in many cases, than IBM.
Fortunately for IBM, the information technology business remains vibrant and inventive. Fresh opportunities arise all the time. Unfortunately for IBM, the competition in emerging segments is as smart, swift, aggressive, creative and determined as it was in the PC business, storage, printers, high volume servers, and other areas where IBM has not come out on top or, in the most extreme cases, where IBM has actually been driven out of the market.
IBM’s financial performance has been weak for the past couple years. Quarter after quarter, the company has frustrated trusting investors who yearn for the kind of success IBM had during the 1990s, and which it had before that during most of the years between 1914 until approximately the time it announced the PC.
IBM’s revenue has been going nowhere. Software and some services activities are on the rise but equipment sales and portions of its services business are shrinking at about the same rate. Half the time the company has reported profits during the past eight quarters, hardware has gone negative. Software has been fabulously profitable for IBM, as has enough of the services business to put the company in a good position. So, notwithstanding its apparent struggle, IBM hasn’t been showing awful bottom line results. Moreover, by buying back billions of dollars of its own stock, IBM has been able to deliver improved earnings per share and to promise even greater EPS improvements during the next year or two.
But, as it did during the 1980s, it might be liquidating irreplaceable assets to hit its quarterly targets. In particular, IBM might be trading customer-specific services arrangements for more generic cloud deals. Many of IBM’s key services contracts are tied to applications suite hosted on proprietary IBM systems. These systems can be moved to cloud data centers, but the nature of cloud computing makes it more economically effective when workloads can slide around server farms. That’s difficult enough to do well when the servers are more or less standardized X86 processors. It is probably impossible, or at least impractical, to get similar results using IBM mainframes or Power hardware, because the immense pool of relatively portable applications and data objects doesn’t exist in mainframe country, or i land, or AIXville. As for Linux on non-X86, well, it is not the same as Linux on X86. This is not a technology issue so much as a customer count scale issue.
I am not saying IBM will fail. But right now, as I think about getting this material into our servers, machines we use to show you what we’ve observed from, where else, a hardware hotel in the Show-Me State, I regret that I had to leave IBM in order to survive as observers of its unfolding fate. Our broadcasting, well, maybe narrowcasting technology has roots in IBM’s creation, the PC, but comes from other vendors’ factories or maybe other vendors ‘subcontractors’ factories. Our web software is de facto standard Windows with a few adjustments by our server hotshots to give us a bit more resiliency. It is definitely not in the same league as z/OS or IBM i Internet technology. But it is the best we and sponsors can afford, and it’s not living on a smarter planet but rather on an ordinary one. If IBM has a more practical way for us to do this, we’d love for them to show us. And to show you, too.