Lessons from the Skunk Works management style still resonate today.

Clarence L. “Kelly” Johnson (1910-1990) was a legendary aircraft designer and aerospace program manager. He and his Lockheed teams, known organizationally far and wide as the Skunk Works, were responsible for creating some of the most celebrated military aircraft in history, including the F-104 Starfighter and the SR-71 Blackbird Mach 3 Reconnaissance Aircraft. Lesser known but still renowned and enduring within the aerospace community are Johnson’s 14 Rules of Management, honed from more than 40 years in the pressure cooker environment of shepherding high-profile, yet top-secret, government contracts from conception to completion.

Johnson’s life and career offer valuable lessons that can be applied to any business, including test engineering.

A good example of this is the use of plainspoken clarity and stick-to-it-iveness as a day-to-day business practice – in written, spoken, and digital words.

Why? Because in our business, what often passes for plain speech and follow-through is more like jargon and lip service. Buzzwords are frequently employed to obscure, confuse and mislead, often in the service of diverting attention, rather than expanding the frontiers of knowledge.

Which is a pity, considering how many of us live and work in the technological center of the English-speaking universe (Silicon Valley). Home of allegedly well-spoken, smart people. In law-abiding, image-obsessed California. Where many of us still harbor barely suppressed yearnings to keep it the acknowledged center of digital supremacy it has always been. Somehow we lost our way, and we need to regain it.

Kelly Johnson’s lifelong motto was “Be quick, be quiet, be on time.” Short. Sweet. Indisputably clear. His 14 Rules of Management amplified that philosophy.

Johnson firmly believed that world-beating results were best achieved by small, highly-talented, firmly-focused, authoritative teams, holding oligarchic responsibility for delivering an extraordinarily complex working product (more often than not a supersonic aircraft) on time, within budget, and to customer (usually the US Air Force) specifications. Those teams managed documentation systems kept deliberately simple and extremely flexible. Procedures were subject to ruthless reinvention as circumstances demanded. Meetings were limited to small gatherings, often no more than a handful of people, to a maximum of 15, thus encouraging open participation and rapid feedback to design changes. Blowhard reports were discouraged; in fact, Johnson hated reports exceeding 15 pages in length. Reporting relationships were short. The customer was kept well – and regularly informed. Costs were carefully monitored. Surprises were minimal. Trust was all: between design chief and team; between prime and subcontractors; between customer and prime.

Johnson’s rules were the result of necessity. Most of the design, technology and materials for revolutionary aircraft like the SR-71 had to be developed from scratch.
The physics behind the specifications demanded it. National security inspired it. The Lockheed team was in uncharted waters, and had to apply its skills to making its own charts. A rigid system of design and contracting rules would never have enabled the crown jewel of American aeronautics, a plane conceived to outrun and outsoar everything shot at it or flown in pursuit after it, to see the light of day. Or night, where the SR-71 often operated. From this fearless willingness to invent anew when confronting technical obstacles, great things were accomplished. Skunk Works designs were significant contributors to the winning of the Cold War.

Back to Earth: What lessons can we test engineers draw from this history?

1. Keep It Simple, Stupid (KISS, another Johnson innovation). Follow a 20th Century variant of Occam’s Razor (i.e., given the choice between a simple path of thinking or action and a more complicated path, choose the former). Example: Avoid being hamstrung by overly-annotated Statements of Work (SOWs) that read like combat after-action reports. I recently received an 18-page SOW from a customer intent on defensively micromanaging every byte, bit and pogo pin of an in-circuit test fixture and program development because of the neglect of a previous supplier. It was a Dutch dyke-plugger’s manifesto in its attempt to catalog every engineering goof of the preceding 10 years. The sins of old are visited on the new.

2. Listen to the Customer. Really. Listen carefully to what the customer wants, however harebrained it may initially sound. Customers have legitimate needs, wants and biases, often borne of bad past experiences. (That’s why they are in your office, right?) Customers also have half-formed (some would say half-baked) ideas, occasionally needing the guiding nudge of hard-bitten expertise to make them real. Either way, respect them. Just take notes and suppress the snarky opinions. There will be ample time to pass judgment later.

3. State what you are prepared to do, then do it, and keep the customer informed while you’re doing it. Avoid the impression your work relies heavily on Black Magic (unless, of course, it does). Too often in the test business, the impression is given, with well-measured condescension, that a genius is at work and is not to be disturbed, by anyone, until the masterpiece is unveiled. And heaven preserve the poor Buyer who simply and reasonably wants a clear response to the question of when their product will be done. The bolder among them might even hazard a query about how it works. Such effrontery is often met today with the email equivalent of a malevolent stare, often in 140 words or fewer. Genius works that way. Rather than promoting hostilities, better to state your intentions in your quote, then back them up in the execution once the order is yours. Back them up again with accurate coverage reports. In all things make clear to the customer exactly what it is they are paying for. Look past the customer’s doltish behavior; they are Our Dolt. Patience, until further notice, is still a virtue. And their checks are still cashable.

Truth in advertising is the exception nowadays. Make it your strength.

4. No less important, declare unequivocally what you will NOT do. Kelly Johnson’s Rule #10 states this definitively: “The specifications applying to the hardware must be agreed to well in advance of contracting. The Skunk Works practice of having a specification section stating clearly which important military specification items will not knowingly be complied with and reasons, therefore, is highly recommended.” Don’t attempt to hide this uncomfortable truth. Better to address and enumerate the technical shortcomings head-on and early, before the recriminations start.

5. Above all else, eschew obfuscation. Say that 10 times fast. Worship clarity in all things, and put theory into practice. (Sounds insultingly basic, but so many of us don’t do it.) This is your lodestar. Everything flows from this. Observe this rule and you will separate yourself from the herd in a business where vagueness is deployed to competitive advantage.

Common sense yet again.

For some, anathema. For others, refreshing. Choose wisely.

Any takers? From somewhere on high, Kelly’s watching.

Robert Boguski is president of Datest Corp., (datest.com); rboguski@datest.com. His column runs bimonthly.

Boundary scan and embedded test will need to make up for ICT gaps.

Inline ICT is not as cumbersome as it used to be, and in the longer run, will help manufacturers save costs.

The idea of operating in-circuit test as part of the surface mount technology line is not new. Options available to volume manufacturers include modifying ICT systems with bridge handlers to receive printed circuit board assemblies from the input line to the tester. Others link test units to increase throughput.

Variables include the number of outputs tested at one time and spikes in output.

Conventional automotive electrical test methods only test each output from the electronics one by one and sequentially. This leaves a gap in the test coverage where faults, or outliers that occur simultaneously due to unexpected outputs from the electronics, are not tested or checked.