Chapter 10 - Perils of Incumbency

Choosing the moment to make the daring leap from a successful technology to something new and more competitive requires the keen timing of the trapeze artist choosing when to leave his trapeze. Except there’s no nimble-handed catcher for the business leader. So the more stable the old business model, the flimsier seems the new—making it all the easier to just “wait and see.” Success can be a killer once the wheels of change begin to turn. Two competitive pairings—Tandem Computer versus Sun Microsystems and Cullinet versus Oracle— illustrate the perils of incumbency.

Tandem was formed in 1974, eight years before Sun, as the last significant minicomputer company built on a proprietary operating system. And what a system it was. Designed never to fail or lose even a single transaction, Tandem’s NonStop computer met almost instant acceptance from an intensely demanding group of banking and brokerage clients, including the Securities Industry Automation Corporation (SIAC), which provided IT to both the New York and American Stock exchanges. Over several years, NonStop became integral to the operation of the financial markets.

Sun appeared in 1982, one of the first and certainly most successful companies to start with Unix as its only operating system. For cofounder and CEO Scott McNealy, Unix’s “openness” provided at least three strategic advantages: (1) It drew engineering and scientific applications from third-party software suppliers; (2) it supported Sun’s sales proposition to customers by delivering lower switching costs than competitors’ “closed” proprietary systems offered; and (3) Unix smoothed the transition for customers each time Sun upgraded its microprocessor, the SPARC (for Scalable Processor ARChitecture). In the early years, Sun doubled its price/performance ratio every year or two, versus every five years for DEC or IBM.

Meanwhile, Cullinet Software, which entered the database market in the early 1970s (seven years before Oracle), grew into the world’s second-largest software vendor (after Lotus) in just ten years. Central to Cullinet’s growth was the rampant success of the IBM 4300 minicomputer, which underpinned half the company’s revenues. Oracle was started around the new relational database but at first had difficulty gaining traction against well-entrenched competition. Up against Cullinet, Applied Data Research (ADR), and Cincom in the software segment, Oracle also faced hardware majors IBM, DEC, and HP that fielded their own databases.

But Oracle was not without its charms, the first being ease of use. Its relational structure allowed a user to retrieve information quickly and easily. The second advantage was portability across most mainstream minicomputers. Three years after Oracle reached volume shipments, Cullinet was crushed.

In Tandem with Treybig

I first met Jimmy Treybig, Tandem’s founder and CEO, in late 1977, just after the company floated its initial public stock offering. Tandem was then still cozy enough that each box on the assembly line carried a small cardboard plaque commemorating the customer’s name and national flag. The Research Board was studying microprocessors, and since Tandem was relatively new, I assumed it was microprocessor- based. The slightly beer-bellied Jimmy was slouched back in his chair, cowboy boots propped on his desk. But when I mentioned microprocessors, he jumped up to take me on a rambling tour of the NonStops on the assembly floor. “You must know microprocessor part numbers. You see any of ‘em here?”

Our meeting was short, and that year’s RB report about the computer industry contained nothing about Tandem.