Standing up to the customer

Quality is, by definition, a good thing, but it’s the definition of the word “quality” that I think needs a little rework. I was once involved (in a small way) with the redesign of an elastomeric gasket used for an automotive OEM who shall remain nameless. This was in the pre-metric era, and the print’s “block” specified plus or minus 1/64th on all fractional measurements unless otherwise noted. That was pretty standard, except in the case of this drawing, there were no fractional dimensions. Tolerances were, in fact, on the order of plus or minus five thousandth of an inch. That’s probably meaningless to the younger technical types, but it was a totally unreasonable tolerance on this high-volume, low-cost molded elastomer. Even more ominous was that the tolerance applied to every dimension, including the non-critical “lightening holes” added to reduce the resin mass.

Holding unrealistic tolerances on non-critical dimensions is a common request, but how many processors out there stand up to the customer and request a redesign? “Pushing back” is one of those soft skills that good companies master, and if you promise not to tell the customer, I’ll divulge my rules for getting realistic expectations for your product.

One is, check the “block” on the drawing and see how many redesigns it’s had. If it has gone through the alphabet, this may be one of those “bastard” jobs that nobody wants, and tolerances may be a reason why.

Another is to establish a design review team (or designated person) to offer “free” advice to the design engineer on ways to reduce costs by opening up non-critical tolerances. Rank individual dimensions by importance and scrutinize the bottom 20 percent. This strategy works very well where processors can put experienced engineers up against the customer’s younger designers, since they in turn can feed the redesign ideas up their chain of command and look like geniuses in the process. Do it the other way around, i.e. bright young processor’s engineer against older wiser, (read, more stubborn) customer designer, and it just won’t work. Grey hairs count here.

Another useful strategy is to use a tiered pricing strategy, showing how costs can be reduced by relaxing specific tolerances, i.e. “you can save 2.5 cents per piece by opening up that slot”, etc. If you’re going to complain about the requirement, do it loudly, and do it early, especially if you’re dealing with a large OEM, who will take considerable time to approve a change. And if you “win one” with a verbal O.K. to relax a tolerance, do it by E-mail or fax, and make sure that the communication refers to both the specific dimension, and the drawing number. An unofficial paper trail can save lots of grief.

Here’s one last idea I have on this subject, and it runs against the stream of current thinking about productivity in dimensioning and tolerancing: don’t rely too heavily on teleconferencing, E-mail and telephones. If you have a big problem with a spec, consider a face-to-face meeting over the drawing, on paper, not on the designer’s monitor. If possible, do it in a conference room, away from telephones and staff. In the old days, I understand the process involved modest quantities of Scotch, but that’s not recommended these days. Standing up to the customer is a delicate art, but occasionally it’s necessary to preserve everyone’s profitability, and sanity.