When Douglas Ivester took over the chief executive officer job at Coke, following in the footsteps of the celebrated Robert Goizueta, he was pretty much universally praised as the perfect guy for the job. Perhaps bearing in mind how poorly that worked out, observers of yesterday’s announcement that Jeffrey Immelt will take over General Electric’s top job from the even-more-celebrated Jack Welch are falling all over themselves to deliver cautionary assessments. One business professor quoted by the Times suggested that the successor to a heroic top executive “is destined, if not doomed, to a life of insignificance and insecurity.”
Wow. That’s pessimism. So is there a positive case to be made for Immelt’s future prospects? Of course there is, since the lesson of the Ivester hype really ought to be that there is no way of predicting how a new CEO is going to perform and what factors may affect that performance. Given that, I’m not going to hazard a guess as to how things will work out for Immelt. But it’s worth taking a look at the most intriguing component of the Immelt bulls’ brief, which is the power of something called Six Sigma.
So what is it? Like a lot of management-related ideas, the goals of Six Sigma are pretty unsurprising: better quality, more efficiency, etc. In this case, you could say the key underlying idea is something like better quality through extreme measurement. At its heart, the Six Sigma theory is all about collecting lots and lots of data and analyzing it in clever ways that yield otherwise elusive efficiency insights.
This is probably not the place to go on about statistical science (and I’m not the person to go on about that, anyway), but Six Sigma programs are rooted in statistics, and if you’re curious, sigma, the 18th letter of the Greek alphabet, is used by statisticians to denote deviation from a norm. To cut the chase a bit: Achieving Six Sigma efficiency or quality means producing just 3.4 defective widgets out of every 1 million. The Sigma literature also addresses lower sigmas, which are not as good, but may be steps on the road to greatness: Sigma three, for instance, yields 66,800 bad widgets per million. The further refinement of all this is to look at every single step in the widget-making process: Bringing each step as close to perfection as possible should result in hitherto unimagined efficiencies to the product and process as a whole.
According to Six Sigma lore, the theory was invented around 1985 by a (now deceased) manager at Motorola to improve that company’s manufacturing processes. What GE is perhaps most famous for among people who care about this sort of stuff is its success in applying the same ideas to services–and thus improving, for example, response time to customers of the company’s financial division, and so on. GE now reportedly employs hundreds of Six Sigma experts–preposterously referred to as “black belts”–who ferret out inefficiency, all on the larger theory that if you are good enough at doing this then the results will more than pay for the efforts. GE says its Six Sigma program saved the company $2 billion in 1999.
I’m oversimplifying, of course, but Welch’s enthusiasm has helped create a cottage industry on Six Sigma significance, so you can read lots and lots more in recently published books such as The Six Sigma Way, Six Sigma: The Breakthrough Management Strategy Revolutionizing The World’s Top Corporations, Rath & Strong’s Six Sigma Pocket Guide, Implementing Six Sigma, Managing Six Sigma, The Six Sigma Handbook, the eight-volume Vision of Six Sigma, and (inevitably) Six Sigma Simplified.
Is any of this really meaningful? Who knows? Six Sigma skeptics complain that this is simply a repackaging of previous ideas with names like statistical process control and multivariate analysis. That said, I suspect that rigorous attention to quality is, on the whole, a good thing, and the key is getting managers and employees to buy in, so repackaging an old idea isn’t necessarily bad if it’s a good idea. Six Sigma seems to have produced good results at Jack Welch’s company and the medical systems division (run by Jeffrey Immelt), which was a Six Sigma pioneer there.
But that doesn’t necessarily mean Six Sigma would work as well elsewhere or even that it will continue to have as great an impact on GE when it becomes Immelt’s company. The dilemma facing Immelt will remain that heroic CEOs generally earn their reputations by building or improving companies, not by maintaining them, and GE is going to a be a lot harder to improve than it was when Welch took over back in 1981. I don’t think that his mastery of Six Sigma notions is gong to make difference on that score. In fact, it may easily turn out that GE has done a lot more for Six Sigma than Six Sigma, going forward, will do for GE.