Better, Faster, Cheaper—Pick Any Two

It used to happen at the kickoff meeting for every new program. The engineer managing the development team would come in with an armload of overhead-projector transparencies for the business-unit managers and marketing people, who were there to learn about features, design challenges, schedules, resources—the usual stuff. Inevitably, about the third foil down, would be one of those triangle graphics with a bubble at each vertex (Fig. 1). One bubble would say "faster," another would say "cheaper," and the third would say "better." Then, in boldfaced capital letters was the exclamation: "PICK ANY TWO!" From the scratches on the foil, you could tell that the project manager had used the same foil in every kickoff meeting since he had finished that night-school MBA.

The foil was a threat to those designers who couldn't—or wouldn't—look at a program-planning tool such as a PERT (Program Evaluation Review Technique) or even a Gantt (after Henry Gantt) chart. Its subliminal message was that if you come back to the team with "feature creep" (adding new bells and whistles) or a budget crunch, there will be an impact on the delivery schedule or performance numbers, or both.

That warning never stopped feature creep or budget cuts or delivery slips from happening, though. But the culture of "pick any two" placed some sort of restraint on those engineering annoyances.

When I was a young engineer, working with other designers who were dealing with Apollo program subsystems (I was involved with military transports), NASA and the U.S. Air Force were pretty comfortable with the concept.

If an astronaut complained that lithium hydroxide (LiOH) from the oxygen-recovery system was getting into his helmet and making him weep, a special task force would work on the problem day and night for six weeks, and NASA would fund the effort. (It wasn't the LiOH. It was Tang orange-drink powder, the task force concluded.)

In 1992, NASA administrator Daniel Goldin decided that it was possible to have all three. Under his "faster, better, cheaper" management philosophy, NASA launched 146 payloads valued at a total of $18 billion. All but 10 were successful.

Then came the debacle of the $125 million Mars Climate Orbiter (Fig. 2, left). This satellite was lost because Lockheed Martin failed to convert English units to metric units when coding critical software. It was followed by the loss of the $165 million Mars Polar Lander, which also was attributed to software errors.

An independent review said that there was a lack of guidelines and policies to implement the "faster, better, cheaper" concept. But more significantly, the Mars program had been underfunded by at least 30%. It seemed like "faster, better, cheaper" was a smokescreen for "cheaper."

Yet the pre-Goldin, $1.4 billion Galileo spacecraft kept going like the Energizer Bunny (Fig. 2, right). Designed for a two-year survey and launched in 1989, it deliberately plunged into Jupiter's atmosphere in 2003 after a journey of more than 2.5 billion miles. Along the way, it took detailed images of asteroid Gaspra as well as a tiny moon orbiting asteroid Ida. It then observed comet ShoemakerLevy as it crashed into Jupiter.

When Galileo reached Jupiter, it released a probe that parachuted into Jupiter's atmosphere and sent back nearly an hour's worth of Jovian weather data. Galileo then orbited the gas giant 11 times, observing its moons as they passed. During 14 more orbits, it concentrated on Europa. Next, it went deep into Jupiter's radiation environment to study Io. Galileo's suicide dive yielded data on the mass of Amalthea, one of Jupiter's inner satellites.

About the only thing Galileo couldn't do was bring back a monolith. And at $1.4 billion, or $100 million/year for its 14-year lifespan, each year of Galileo's life was on par with the lost Mars missions. But that was before "faster, cheaper, better."

WHAT ABOUT ELECTRONIC DESIGN?
That's not to call Daniel Goldin a fool. But some managers who picked up the "faster, cheaper, better" ball and ran with it blindly could fall into that category. In unmanned space exploration, it makes sense to push managers to take risks, within limits, to achieve innovation and economies.

But some form of "faster, cheaper, better" is happening in more terrestrial pursuits without a mantra and without anybody taking much notice of the process. Semiconductor chipmakers have assumed some of the subsystem-development activities of their original-equipment-manufacturer (OEM) customers, shortening the OEMs' development timelines and amortizing the costs among multiple (they hope) OEMs.

As a result, OEMs create faster development cycles using fewer (if more expensive) components, trading off marginally higher bill-of-material (BOM) costs for lower assembly and inventory costs. That covers "faster" and "cheaper." "Better" arises because the chipmakers know how to avoid the design traps that go with mating basic high-performance components. Today, there's less chance of making layout mistakes.

That's the big, simple view. But it's really a lot more complicated because of the evolving changes between suppliers of all sorts (not just chipmakers) and OEMs. Remarkably, it still seems to deliver "faster, cheaper, better" results.

Here's how the relationships are evolving. A few years ago, particularly in the consumer-product arena, chipmakers dealt directly with their overseas OEM customers. Customers focused on manufacturing. Chipmakers supplied the circuit knowledge, tied up in silicon intellectual property (IP). Along with that, they got to define the performance specs.

More recently, emerging independent design houses (IDHs) in China and India have intermediated between the chipmakers and the OEMs. The chipmakers still controlled the silicon IP, but the IDHs dictated the specs. This shifted the balance of power in the development process. Lately, chipmakers have been buying the IDHs and turning them into remote design centers, shifting the balance once more.

BREAKING COMMUNICATIONS BARRIERS
Overseas design centers sound like an answer to the "cheaper" component of the triad, and if they're staffed with competent engineers, that supplies the potential for "better." But how about "faster"? At last fall's audio engineering conference, Kevin Leary of Analog Devices discussed the company's recent acquisition of AudioAsics A/S, a Danish hearing-aid company with its own design center in Bratislava, Slovakia.

Leary, who is responsible for ADI's remote design centers, said that when he takes on an overseas design organization, he first finds some engineers stateside who are young, unattached, and looking for adventure and sends them over there. That's not to groom junior managers, he said. It's more a case of opening up unorthodox lines of communication.

Leary looks for engineers who'd go out and drink with the locals, get to know them, and then come back to the States or go on to another remote design center. Ultimately, there would be an alternative global communications channel that lets engineers get things done. This may offer an answer to "faster," considering that some companies' so-so remote design centers suffer from too much hierarchy.

HOW DO THEY DO IT?
According to Bill Murphy, product line manager for ADI's integrated amplifier product line, ADI drives developments into leading applications and then migrates the basic technology into other applications with minimal changes. With this approach, a second set of customers (fast technology followers) can benefit from the development time of the technology-driving application.

For example, automotive manufacturers concerned with diesel-injection control and engine management were the lead customers for ADI's AD8205, a 42V difference amplifier often used for high-side current sensing. "Those applications drove a higher and more cost-effective level of performance than was previously available," says Murphy.

"When inquiries came in from a home medical electronics manufacturer, it was a quick task to adapt the feature set and provide a suitable solution for the medical application," he continues. "Because the high performance with low-cost structure had already been built into the basic product, the time it took to service the newer medical application was very short."

ADI also has a family of variable gain amplifiers that were originally developed for cart-based ultrasound products. Now, they've been rapidly configured to the needs of adaptive cruise control. Essentially, the company uses a platform strategy and executes it well.

International Rectifier CEO Alex Lidow has put a lot of his company's assets behind a targeted platform called iMotion, an adaptable set of building blocks for appliance motor control. He says that iMotion's focus isn't on "faster, better, cheaper" versions of previous chips, but rather to use "faster, better, cheaper" to displace previous electromechanical approaches.

"The first application was in a washing machine, where it replaced an antiquated and inefficient subsystem consisting of a mechanical gear, belt, pulley, and complex mechanical switch driving the motor," Lidow said (see "Washing-Machine Controller Wrings Multiple Design Wins From One Platform"). "This old mechanical system's $30 BOM cost target became IR's budget for an electronic replacement that could extend the washer's feature set and cut its energy consumption in half."

The next targets were air conditioners that had a price point of $55 and refrigerators with a $20 price point for the existing mechanical solution. In both cases, IR tailored or is tailoring the solution to fit those price points while providing energy efficiency, essentially for free, and with faster development cycles.