Satan or saviour

Posted on 7 May 2008 by The Manufacturer

Six sigma creates fanatics – both for and against. John Dwyer goes to the battle lines and peers through the smoke to discover what makes it work

It’s loved and loathed in equal measure. Six sigma’s champions range from Motorola, which devised it, through US General Electric (GE), whose chief executive, Jack Welsh, was a six sigma fanatic, to bathroom equipment maker Kohler Mira and DuPont, whose English and Scottish Teijin Films businesses have been using six sigma since early 1999. Now the gospel of six sigma, a set of statistical and methodology tools for reducing variation and faults – defects – in products and processes, has spread to financial services and other ‘transaction-intensive’ businesses.
But the six sigma battleground is littered with corpses. One attendee The Manufacturer mentioned it to at a recent business improvement seminar said: “Yes, we started six sigma at a number of our sites, and it fizzled out in all of them.” Even Malcolm Holden, who calls his consultancy six sigma, says, “Six sigma is miles too far for 99 per cent of companies.” He cites three unnamed examples: “One was brought to its knees before they threw it out. The others had to throw it away and start again.”
Stephen Walsh, a partner in the Rugby, Warwickshire-based consultancy Burge Hughes Walsh (BHW) and a director of a six sigma joint venture with consultancy Bourton Group, agrees that manufacturing is disillusioned: “It has to be corporately supported. Therefore it’s hard to do and necessarily it’s top down. If it’s done bottom up it’s limited – it ends up contained in the quality department or something.”
Lean and six sigma are fundamentally different but, many would argue, complementary.
Walsh acknowledges that six sigma has a reputation as being technical and statistical – it is both. But the work is needed to get the results six sigma delivers.
Every process produces variations, says Walsh. If the variations are beyond what the customer will tolerate you need to understand where they are coming from. “After that,” says Walsh, “it’s about changing things, and you can’t make change until you understand.” Six sigma gives you that understanding.
His example is Network Rail (NR). NR has spent months using six sigma to force down its measure of ‘delay minutes’. “That’s complex,” he says. “It’s people, it’s organisation, and you have to break a problem like that down into hundreds of projects.”
Now ‘lean six sigma’ has developed to soften the technique’s rigour. Walsh can see the temptation for manufacturers and others to seek easier-seeming alternatives. The danger, he says, is that many lean six sigma programmes revert to waste elimination – the low hanging fruit – and that isn’t enough in the long-term.
Where six sigma has worked, says Walsh, it’s because of commitment at the top. GE saved $12 billion over five years and added $1 to its earnings per share partly because GE’s Jack Welsh famously insisted on seeing the results of six sigma programmes across the business.
Holden, however, has a list of six sigma reservations.
A particular problem is that six sigma’s yellow-, green- and black-belt practitioners begin to think of themselves as an elite: “They can be a real problem,” especially if they start telling mature staff what to do.
Andrew Smith doesn’t recall any such issues. But then the company where he is continuous improvement leader and a six sigma black belt, Perkin Elmer at Llantrisant, South Wales, is a stable company with a small workforce that has a lot of long service. “Everyone knows each other,” says Smith, and that kind of problem just hasn’t arisen.
Llantrisant makes instruments for the pharmaceutical, biology, semiconductor and other industries. Its Spectrum 100 (SP100) machine is an infra-red spectrophotometer which can be used in food safety for the non-destructive testing of drinking-water bottles and other food packaging.
Llantrisant assembles the instruments from base castings and other bought-in parts and subassemblies, some from suppliers and others from other parts of the group, then tests them. One or two sub-assemblies could be bought in but are put together here to protect Perkin Elmer’s intellectual property.
Every one of Llantrisant’s 86 employees has six sigma training, says Andrew Smith, even if it’s basic yellow-belt training. The site has 14 green and one black belt. Each green belt is involved in at least one improvement project. At the moment 39 projects are running and 49 more are proposed.
Perkin Elmer’s SP100 line is a classic example of a ‘lean sigma’ project. The usual progression is to validate the process steps and take out the ones that can be taken out. Suppliers now fit parts to the base casting instead of Perkin Elmer’s own operators. The next is to reduce the non-valueadded time in the remaining steps; and then, and only then, to use six sigma to reduce the variance in the refined process.
Some process variations can arise because the same work is done in different ways by different operators. If a process takes 15 minutes longer to do one way than another, there is a need to find the best way and stick to it.
Six sigma is proof against ‘tolerance matching’, the tedious, time consuming and hence expensive business of going through matching parts to see which fit. They should all fit, but if they don’t, assembly line operators try a succession of parts until they get a pair that match. But they don’t report the number of mismatches they get. Where possible, Perkin Elmer’s continuous improvement teams have eliminated the need for these activities.
The predictability you achieve with six sigma is valuable, says Smith. If a particular process takes, say, a consistent nine minutes, time after time, “you can load the line with the right amount of people for the output you need, and it factors into the supply chain.” Suppliers can be given exact quantities for each day’s production.
Perkin Elmer’s main six sigma success has been in making optical components. This process uses many steps to coat a pattern on curved or flat mirrors to fit a number of infra-red and ultraviolet (UV) light instruments. The consistency Llantrisant has achieved means it will be able to manufacture optics for use in three Perkin Elmer sites, work previously done by a US supplier.
Walsh accepts that six sigma evolved from the TQM and other initiatives of the 70s and 80s. But he doesn’t believe implementers have to do all the work before six sigma and install it as a last step: “I wouldn’t want people to think that they have to climb a hierarchical ladder of improvement techniques before they can get to six sigma.
Perkin Elmer’s experience was slightly different. “Ten years ago,” says Smith, “we started six sigma training but we couldn’t sustain it. We didn’t have the infrastructure in place and it wasn’t the right tool for some of the projects.”
Five years ago Perkin Elmer changed its approach from six sigma to continuous improvement – it would clean up its processes using 5S, pull systems and other lean tools. When you have done all that, says Smith, “you can see the issues and apply six sigma to those.”
Llantrisant’s 2007 savings target was $1.4 million. It achieved savings worth $2 million.
These included not just the cost savings made by improving processes – from the financial benefit of a change of setup, which increases capacity and reduces process costs to savings in warehousing costs or overtime – but also the cost of poor quality such as warranty, costs, scrap and rework.
This year’s target, originally $1.3 million, has been revised up to $1.54 million. But Smith is in little doubt that that too will be exceeded.
“It’s about process control,” he says. Llantrisant had to devise its own experiments to work out what samples and measures would bring it closer to the right result and then transfer its findings back into the production process to improve it. That, says Smith, is about, “using the tools, and translating the findings from the tools into meaningful actions.”
Smith agrees that, though this might be meat and drink to an organisation full of people steeped in scientific method, manufacturers in other disciplines would not find it quite as easy.
For six sigma to work, he says, “it needs to be part of a bigger continuous improvement setup.
You can only do what works for you in the end.” Though tools like 5S, lean and six sigma have been used both in manufacturing and in non-manufacturing activities, like call centres, “I wouldn’t advocate that our set-up can be used in all other businesses.”