When you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind: it may be the beginning of knowledge, but you have scarcely, in your thoughts, advanced to the stage of science.
-- William Thompson, Lord Kelvin, (1824-1907)
The Electronic Design Engineering Hall of Fame in December 2010 honored inventor Melville Eastham, founder of General Radio, for his forward-thinking leadership in fostering outstanding employee policies and contributions to the electronics industry.
Hewlett-Packard today is world-renowned as an electronics and computing giant; General Radio, however, was acquired by Teledyne in 2001, and is known to but a few history buffs. What happened?
General Radio (GR) was a Boston firm which created the entire field of electronic instrumentation, beginning with the American advent of radio. When World War II broke out, General Radio was five times larger than any worldwide competitor. It had a reputation not only for developing terrific instrumentation for electrical engineers, but also for establishing leading business policies. GR revenue was twenty-five times larger than that of the fledgling three year old Hewlett-Packard at the start of 1942.
Eastham and GR had crafted a close alliance with the Massachusetts Institute of Technology (MIT) for many years, and when MIT spearheaded the academic /government alliance for electronics research and development during World War II, GR supplied key leadership and technology. MIT was without question the cradle of American electronics innovation at the time, underscored by winning an astounding $117 Million in government awards for military electronics work during the war. The spin-out Harvard Radio Research Labs garnered another $30M. Although Stanford had pioneered much radio technology, their pre-eminent radio engineering leader, Fred Terman, moved east in early 1942 to manage the Harvard Radio Lab. The net result: Stanford was awarded a paltry $50K during the entire war.
Eastham, 53, took a half-time leave from General Radio to manage a radio navigation system project at MIT's Radiation Labs as well as serving in Washington as a co-director for the prestigious Office of Science Research and Development (OSRD) that guided the university funding. Donald Sinclair, 32, who would succeed Eastham as GR's Chief Engineer, then President and finally Chairman of the Board, also divided his time and led the Search Receiver section at the Harvard Radio Research Labs for Terman.
General Radio management, faced with huge wartime demand, decided to subcontract half of their orders, mostly for the less complex equipments, to other manufacturers instead of developing a much larger manufacturing presence. Terman and Eastham steered several contracts to HP. "HP's contributions were mechanical only - machine lapping and honing surfaces, which improved electrical performance."
While Hewlett-Packard grew substantially during the war, the post-war wind-down led to a seventy percent reduction in HP staff, to sixty-five people, a fact little known about HP; General Radio, by contrast, maintained its four hundred and forty person workforce. Entering 1948, General Radio revenues and employees were six times larger than HP. Moreover, at a time that television was becoming the new electronics fad for America, General Radio had products ready for television production, broadcast, and reception, while Hewlett-Packard had nothing of consequence for television engineers, operators, or users.
Television did indeed 'take off' - from a base of 3.6 million television sets in 1950, America added another twenty million sets in the next three years, overlapping the Korean War. And yet, when the Korean War ended, HP was thirty percent larger than General Radio. What happened?
Eastham grasped in 1915 the idea that the nascent radio industry lacked high quality measuring instruments. Precise tools existed to measure electrical parameters (voltage, resistance, and direct or alternating current) but Marconi's radio waves and the signals possible with de Forest's audion tube were not yet able to be quantified. The electronic instrument business evolved slowly, GR leading the way, first with oscillators and the first commercial vacuum-tube voltmeter. Shortly thereafter, they built the first Standard Signal Generator, soon followed by the first Frequency Standard.
As the Great Depression engulfed America, things became dire. Shipments, $934,000 in 1929, slipped modestly to $843,000 in 1930. But in 1931, they plummeted to $593,000 with a loss (first since 1919) of $17,000. In 1932, billings were $515,000, and orders even less at $381,000.
A novel pay scheme allowed GR to maintain its R&D innovation pace. The rapid rise of European exports helped, surging from 18% in 1932 to 39% of revenues in 1937. New orders in 1934 rose to $744,000, with shipments at $676,000. Europe was enjoying the rise of radio popularity that the United States experienced a decade earlier.
But in fact, the creative R&D labs were the real savior. An unusual product, the "Variac ® adjustable auto transformer, which ... provided for the first time a means to adjust 115-volt and 230-volt power-line voltages smoothly from zero to something above line voltage..." It quickly became a top profit generator. It and two other 1933 releases showed the versatility of the company's engineering staff - the first was "the Type 687-A Electron Oscillograph, the first complete oscilloscope ever marketed. The second was the wave analyzer, a very advanced instrument for its time which allowed making accurate, harmonic analyses of waveforms by direct measurement."
GR history is remarkable for the business elements unique to the instrumentation discipline, e.g. the very wide number of products at very low volume, that necessitated a quite different set of manufacturing, quality, and inventory management skills. In 1928, the catalog listed 168 'major' products, with average sales per product about $500 per month. This is much more akin to the sales of individual parts from the back shelves of Radio Shack than the sales of cell phones or personal computers. The HP of 1950 wasn't much different.
GR, 25 years old in 1940, had 215 employees and $1.25 million revenue. Twenty percent of the employees had college degrees, most in electrical engineering, and fully ten percent of revenues were re-invested in R&D. Eastham continued to add benefits for his employees, putting in Blue Cross-Blue Shield in 1938 as a fully-paid company benefit for employees. Vacations were extended to three weeks for all employees in 1940, and a pension plan for retiring employees was launched in December 1941.
Pearl Harbor Day - December 7, 1941 - changed irrevocably the fortunes of both General Radio and HP. Nearly every company in almost every industry became part of a massive re-allocation of effort. Even though much was subcontracted out, General Radio's revenues leapt from $1.25 million in 1940 to $4.5 million in 1944; full-time employees grew from 220 to 440. HP, by comparison, went from $106 thousand in 1941 to $1.54 million in 1944 in revenue, and 6 to 210 employees. At war end, sales fell for each - GR in 1946 sold $2.98 million, down 34%, and HP sold $790 thousand, down 49%. GR kept all of its full-time employees, while HP cut back severely, to sixty-five full-time folk by year-end.
Bill Hewlett and Dave Packard, both 1934 Stanford graduates, were 25 and 26 when they began their company in 1938 at Fred Terman's urging. Hewlett's thesis described his invention of a novel RC-oscillator that even after licensing a GR patent was able to be priced nearly 80% lower. It proved a solid foundation for the company. Packard soon contributed a market winning voltmeter, and Hewlett completed a sophisticated wave analyzer just before he was called up fulltime for Army duty in December 1941.
Packard remained at HP's helm. His old mentor, Fred Terman, "helped steer several engineering jobs to HP; they produced finished products from the working prototypes developed at MIT or Harvard Labs. Thus, HP was able to move from the Audio range - 20 cycles to 20 Kilocycles - into the UHF (Ultra High Frequency) Radio Frequency range with the Model A, designed by the U.S. Navy labs..."
Coming off World War II, commercial television caused enormous public interest. GR introduced new TV monitors in 1947 and 1948 that were the first in the field. Also, in 1948, after field tests and demonstrations lasting two years, the unique Type 874 coaxial connectors (for signals to 8 GHz) were announced. GR seemed poised to profit extremely well from this new broadcast medium; HP was absent in this arena.
But strangely, General Radio passed up microwave signal generators. Not only did GR invent the first wholly integrated oscilloscope, they had pioneered wideband signal generators and frequency sweepers. These tools were fundamental for both the MIT and Harvard Research Labs, as well as for the Navy and Army Air Force during the war. At war end, the Navy tried to sell their unfinished program back to industry - notably to both GR and to Varian, which had produced the essential microwave tubes. Each declined, believing that the commercial markets were too small.
The strategic choices at this point would prove crucial. Both MIT and Harvard Labs had heavily invested in microwave radio, radar, and sonar during the war with great success. Each had dabbled with computers as well, but the key contribution in that arena was at the University of Pennsylvania with ENIAC. Radar, sonar, and computing were all 'digital' technologies, not amenable to test instrumentation that either GR or HP built, but tailor-made for oscilloscopes that GR had pioneered a decade earlier.
HP stepped into the void for high frequency communication test instrumentation, and a new company, Tektronix, jumped on the oscilloscope opportunity. Both would soon dwarf the longtime leader, General Radio, in instrumentation revenues.
MIT at war's end voted heavily for computing as the next wave to ride, led by both Ernst Guillemin and Vannevar Bush. Analogous to Terman's bold work in radio design two decades earlier, Guillemin wrote a textbook that provided unifying mathematics to link the analog and digital realms via Fourier transforms, giving analytical backbone to the value of oscilloscopes as the primary analysis tool. Vannevar Bush provided an even more exciting contribution with his seminal paper, "As We May Think" which caught the public imagination about how computers might impact the 'new world'.
Terman, arriving back at Stanford in 1946, was indignant about the vast sums of money that MIT and Harvard garnered from OSRD during the war while Stanford was ignored. He felt strongly about a bright microwave communications future, especially with his ex-students who had founded companies near Stanford working with these technologies. And Stanford had practically no capability in digital design technologies.
Terman tried, unsuccessfully, to persuade Packard to buy the rights to the Navy signal generator program. When Hewlett returned, though, the duo not only bought the rights, but hired four key microwave designers from the East Coast labs. Packard meanwhile had interviewed Howard Vollum at Hewlett's request; Vollum demurred, opting to set up Tektronix in Portland, Oregon company in 1947 to build oscilloscopes.
Curious twists - the structure of irony. Eastham, born and raised in Oregon, would watch Tektronix emerge as a major instrument competitor, with a product concept that GR pioneered, which was vital for the computer technology that neighboring MIT was espousing. Meanwhile, HP and Stanford were chasing commercial microwave communications for which GR had created the tools, but squandered the rights.
Strategic shifts take time to evolve. In the interim period right after the war, GR was able to re-establish a strong export business, fueling sixty-six percent growth to $4.00 million over the next two years. HP grew by 175% in the same two years, to $2.18 million in 1948, selling higher frequency tools to domestic companies and radio stations, groups that had hitherto only bought instrumentation from General Radio. Tektronix, just underway, managed to sell fifty 'scopes in 1948, for $257,000.
In 1948, HP launched its first microwave signal generator, followed by six more in a decade. This was a commercial version of the Navy's microwave signal generator, invented by General Radio engineers for the Navy but oddly ignored at GR after the war.
Key designers Bruce Wholey and Art Fong had both become good friends with Bill Hewlett. With their skills, HP changed the rules, focusing on the microwave spectrum, several orders of magnitude faster than any broadcast medium - whether AM or FM radio, or VHF or UHF television. Instead of relying on the local radio or TV station, HP's bet was on AT&T building up the national grid of microwave repeater stations.
An ancillary bet at HP focused on frequency stability of the local broadcast station. Anyone who experienced the 'fading in and out' of radio signals in a local area during the 1930's to 1950's, followed by a mad dash to retune the station in order to hear the football or baseball play, knew how frustrating this could be. For this problem, the new HP Frequency Counter was far more effective than GR's monitors. In 1950, HP launched the Frequency Counter. It quickly became HP's best selling product.
General Radio had no counter in its lineup. Beckman, via Berkeley Labs, was the only early competitor. Al Bagley, the inventor of HP's counter, told a small audience in July 2010 that he got a call from a GR designer after two years of HP sales, who sought to know the counter's manufacturing cost (GR management believed that HP was losing money on this product). Bagley related that things were so cordial between the designers that he shared the data. Two weeks later, the same engineer called back, ruefully saying that the project had been canceled because GR management didn't believe the supplied data. Thus, GR missed another critical opportunity.
Eastham retired, aged sixty-five, in June 1950 from GR, apparently oblivious to both the Hewlett-Packard and Tektronix competitive threats, done "under his nose". Bear in mind that all three companies were privately owned, so GR couldn't "get the numbers". At year end in 1950, GR booked $4.9 Million in revenues, while HP sold $2.3 million, and Tektronix $1.2 million. Three years later, GR managed only $9.8 million, 30% of it internationally. HP and Tektronix (with only U.S. sales), did $12.9 million and $5.9 million in revenues respectively.
Business historians usually focus either on one company at a time, or on broad general trends. Each approach is inadequate to understand the issues that surround the evolution of these three companies, companies that together provided much of the design capability in world electronics for the second half of the 20th century.
General Radio, the venerated world leader, grew revenues and profits by a factor of five in the post-war decade; they were both satisfied and a bit smug about it.
You might think that your company is 'doing fine' if you managed GR and didn't know the HP numbers. After all, the growth rate from 1946 to 1950, coming off the lows after World War II, was a compounded 15% per year, as fast as any peacetime growth in GR history. The Korean war buildup created an immodest three year growth rate of 36% per year. The inevitable aftermath still had a 6% growth rate per year.
Meanwhile, an upstart HP, enabled by product designs from the leader's designers during a wartime crisis, assumed the leadership mantle, growing revenues by twenty-five times. Starting in 1946, HP was one-third the size of GR; a decade later, they were 75 percent larger, a monumental shift in the instrumentation power structure.
Moreover, Tektronix had emerged in the same period, and in fact passed GR in 1956, with revenues of $16.1 million vs. GR's $11.2 million. In just six tumultuous years from the start of the Korean War until the second Eisenhower administration, the world's electronic test equipment leader - bigger than HP and Tektronix combined in 1950 - tripled in revenue, but astonishingly was now - gasp - a distant third, behind each of its rivals by a considerable amount.
Yet, for GR's 50 th anniversary biography (1965), CEO Arthur Thiessen glibly wrote: "The decade from 1950 to 1960 was marked by the almost explosive growth of competition and by the continuing trend in instrument design toward automatic or at least highly simplified operation. The Company shared well in this growth; its shipments grew from $4,450,000 in 1949 to $16,000,000 in 1960."
The irony had to be obvious for the assembled General Radio team. HP had "gone public" in late 1957, Tektronix did so in 1963. Now, the numbers were visible; they told a savage story of missed opportunity. In 1965, GR had sales of $20 million; Tektronix was four times that, and HP was more than eight times larger. Somehow, the industry re-structuring of instrumentation eluded GR.
|Figure 1. Annual Revenues in $ Millions|
It is key to understand the ‘early years’ where the strategies were shaped, in order to understand how the growth rate differences unfolded later
Figure 2. Revenue comparison and key events, GR and HP, 1942-1950
HP hiring of four key engineers in 1946 parallels the GR hiring of four key engineers and a chief engineer in 1928 just prior to the Great Depression.
Figure 3. HP vs GR skill-sets as a function of communication frequency circa 1945
Figure 4. Revenue comparison and key events, GR and HP, 1946-1952
Figure 5. Revenue comparison and key events, GR and HP, 1950-1956
General Radio decided against Microwave Test Equipment at the end of WW II, allowing HP to buy GR technology cheaply; Varian likewise sold its Microwave Tools Group to HP cheaply
Figure 6. Revenue and Profit comparison, GR, HP, and Tektronix, 1950-1956
Figure 7. Revenue comparison and key events for GR and HP, 1952-1960
Figure 8. Revenue and profit comparison – GR, HP, and Tektronix – 1956-1964
All Rights Reserved,
Charles H. House,
Charles H. House