Six Sigma is usually viewed with a jaundiced eye in Services as "those quality guys from Manufacturing" and many companies I know, actively discourage widespread usage of the phrase within their company. They may be using Six Sigma and Lean Six Sigma techniques for Process Optimization widely, though.
There is nothing wrong with Six Sigma techniques per se but what may be erroneous may be inappropriate application of techniques from Manufacturing without thinking about the differences that the Services side may entail.
I have seen many a call center adhering strictly to reducing variation in Average Handle Time. It is a worthwhile goal to achieve reductions in variation but a different picture may emerge when the agents whose AHT metric may be poor have very high Customer Satusfaction Scores! Now the AHT violators may be the ones winning new customers for the company!
In Manufacturing, uniform variation reduction may be a good thing across all metrics. However, services are, for the most part, a person-person-machine business while Manufacturing may be a person-machine-machine business. Manufacturing may be more subject to variation reduction than services in this context if it can be proven that one performance constraint that is violated is compensated by another one that is done well.
Which brings us to the concept of what is meant by Optimal Performance in Services? I would postulate that if you consider all Performance Measures holistically and if one measure compensates for another, the agent should be commended for performing well!
More processing time on a data process may be compensated by accuracy and precision measures being high. Again, only if accuracy and precision are NEEDED and add value to the company!
I think it may be better to think in terms of Optimal Combinations of Performance Constraints rather than consideration of each one in turn.
If you are a young company, competing against a larger competitor, you may want to go easy on the AHT metric and achieve high Customer Satisfaction (CSAT) scores. It may bring in valuable additional business and cross sell and upsell efforts on the phone may be more successful.
If you are an established company providing Customer Support over the phone for a $400 computer where your margin is about $50, you may want to achieve low AHT numbers even if the CSAT scores are not stellar. Customers should not expect Nordstrom services on inexpensive products.
If you are in the Nordstrom Call Center, throw out the AHT measure al together and make sure that the CSAT score is good!
It may be better to view Process Metrics, especially in Services as a set of Performance Contraints that you balance against each other and determine what works for you, given the stage of your company, your strategies and tactics in the marketplace.
Blindly following what worked in manufacturing may not work at best; at worst you may be turning off a lot of people to some very useful techniques by misapplication!
A cow is a very good animal in the field; but we turn her out of a garden – Samuel Johnson