As an old econ professor of mine used to say, it's not true that anything worth doing is worth doing well. It is true that anything worth doing is worth doing optimally.

BTW, in the cases Bob cited in his original article, bad processes were directly costing the company money, not just inconveniencing a customer.

I currently work as a sales rep for a very large on-line computer company in one of their mall kiosks. This past Thanksgiving weekend (Black Friday 06), the company's web site couldn't handle the traffic, regardless of what time of day or night it was. It took IT 5 days to correct the problems, but by then it was too late. Many orders were lost.

When we complained to our management, their response was, but we did a lot of business. Way more than plan.

My response was, "You should be outraged that we had these problems on the busiest weekend of the year. You shouldn't just be pleased with doing better than plan. Think about all the orders you could have had, but lost. How about all the customers that won't get their orders? Where's the customer service?" It fell on deaf ears.

It's also a cultural thing. In my group if 1 account was wrong once we fixed the program so it would be correct in the future, even if the chance of those circumstances coming up again were more unlikely than remote. In the system we connected to if there were 50,000 accounts wrong there weren't enough to consider 'wasting time' a program change. If there were 100K they'd think about it. If there were 250K they'd give it serious consideration. And if there were 500K or more accounts incorrect then they'd eventually get around to finding a solution to the problem.

OK, I worked on a current conditions system that involved financial calculations. The other system was a historical system that displayed information to our offices. Yeah, they could give the customer a wrong number, because that was what was on their screen and they office users had no idea it wasn't correct. Even if they were incorrect the calculations would be OK because we used our own data, not what they stored. What we got from them was passed through from the user interface, supposedly unchanged.

Finding the optimum answer can be more complicated than it looks.

Perhaps you've run the numbers and concluded that it's okay to "screw over" 3% of your customers.

Do you rely on repeat business? How many turns of the wheel does it take before you've alienated enough customers to make your results sub-optimal?

Maybe you don't rely on repeat business. Do your screwed-over customers tell others about their experience with your company? Did your calculations factor in future business lost due to bad word-of-mouth?

Perhaps it's possible to come up with a complex equation that will factor business model, costs, productivity, etc., weighing all the different forms of negative blow-back, and will spit out a precise number to tell you how many customers you can optimally screw over. Frankly, I've never been good enough at math to figure out that equation. I prefer the imprecise metric known as "karma." It might not fit in a spreadsheet, but I think it's a decent predictor of the future consequences of screwing over your customers.

Whether customers get screwed over does not necessarily enter into the equation. In fact, many companies are doing quite nicely while screwing over 100% of their customers. This works best if you have their money before doing so, but it's not necessary. It also helps to have friends in Congress (DMCA), have a virtual monopoly (cable, dominant software firms) or if your "competitors" all do the same (cell phone, airlines).

At least one airline made recent news with a press release trumpeting even more ways they plan to screw over their customers. I guess the encouraging thing is that it did make news.

Sorry, MikeG, it's not that simple.

Take a look here for the exhaustive reason. The short version is to imagine a scenario where your best estimate is that you'll make $400,000 in profits on the holiday weekend.

Imagine two equally likely scenarios--above-average demand and below-average demand. There's a 50% chance that demand will be less than the average, in which case profit will be less than $400,000. That's the downside. But how about the 50% chance that demand outstrips the average of 2 million? Unfortunately, this would exceed the capacity of your server, so there's no upside to balance the downside: $400,000 is the most you can make, not the average.