DMAIC

Hello Kiruba: Before I am in a position to answer your questions I’ve some additional questions: - How do you know that the communication skill is running at 85-90%? Since you write that 68% of your customers are not satisfied with the communication skills of the agent, I don’t understand where the 85-90% come from and how you’ll measure 97%? - Why are you concentrating on communication skills? From a ustomer point of view, it seems to make more sense to focus on better resolving their problems (especially if 68% of your customers say that their issues remain unsolved)? - Do you want to improve the communication skills of your agents or the customer ratings of your agents’ skills? Both might have different root causes. Since you write that you will maybe concentrate on problem resolution to improve the communication skills, it looks more like that you want to improve the ratings rather than the skills. - Why do you want to normalize the data with Minitab? Normalizing data only makes sense, if you want to use a tool or statisitc procedure that requires normally distributed data? (But since “customer ratings (1-4)” and “the number of customers that give a particular comment on the agents’ soft skills” are discrete data normalizing data with Minitab will be impossible, anyway) Maybe you can answer my questions to help me better understand your stuation. My intial comment on your question is that drilling down negative communication skill ratings in order to identify root causes makes much ense. But since most of your customers did not specify their ratings you should ask (maybe a sample of) the customer again why they negatively rate your agents. On the other hand, I don’t think you have to consider all 5 issues since some of those reasons may also be part of the “no reason specified” group. If your’re interested in improving the skills (not only the ratings of the skills) you should think about additional methods to better identify the root causes (like videotaping agent-customer interactions or role-playing plus additonal ratings maybe by colleagues) 
I’m looking forward to hearing from you!   Arne      

Kiruba:
 
Thanks for clarifying my questions.
 
So, I think you’re on a good track with analyzing the data. Some ideas I have in order to improve your agents’ soft skill (ratings) are:
 
-         If possible, ask a sample of customers why they rated your agents soft skills down (as well as why they rated them up!): Why is accent a problem for the customer? What can you / your agents do to better satisfy your customer needs. I think, your customer data base (only the responses to your Csat survey) is not enough. Try to clarify the CTQs: When do you and your customer agree that your agents’ accent is good?  etc.).
-         Maybe use a fishbone diagram or the 5-Whys-technique to deeper analyze why accent is a problem for the customer (One hypothesis I have is that accent is not the ‘real’ cause but rather the most obvious and, thus, rated one.) You might get additional ideas how to solve accent problems, especially if you find out that it is difficult to improve somebody’s accent.
-         Conduct simulations (role-playing) of different agent-customer situations (especially those where the agent has difficulties to solve the customer’s problem). Have other agents rate and analyze soft skills of the role-playing agent. You should get a better picture of what goes wrong and how to improve.
-         Finally (as you already mentioned) improve the agents’ problem solving skills. This will actually not improve your agents’ skills themselves but maybe the ratings and – what is most important – customer satisfaction.
 
For any questions, don’t hesitate to ask me. Unfortunately, I can’t provide an ITES study.
 
Kind regards
Arne Buthmann