Customer Satisfaction

The SIMALTO benefits outlined elsewhere in these notes are particularly relevant to service products and customer interface services in general.  Such studies usually involve more than the 6 or 7-attribute limit of conjoint approaches.  In fact, keeping the number of attributes below 25-30 can sometimes be difficult.  And often many of the attributes are correlated to some degree, invalidating a major requirement of regression analysis, if the ßeta weights are to be used in their own right (as conjoint utilities?)

Recognition of the inappropriateness of conjoint has led many researchers to fall back on simple rating and ranking methods - provide a score out of 5 or 10, say, for how well your product performs on each attribute, and also some attribute importance measure, shortfall from expectation, or ranking of attributes.  These methods are sufficient to show the relatively good and poor performing attributes, but crucially, do not tell the client the perceived actual level of performance (benchmark, and distribution of perceptions around an average), and, which changes are the priority ones for the customer to improve satisfaction - should the supplier further improve already good performance on some attributes, or does he need to remedy perceived faults on others?

 SIMALTO begins with a grid showing all attributes and written alternative levels of performance and then directly, unambiguously records expected performance (from a "blue chip" supplier - not necessarily the "ideal" performance which may be cost prohibitive), current perceived performance, and that of a rival supplier (if required), and unacceptable performance levels - i.e. those that if delivered would cause a customer to seek an alternative supplier. 

 It then records each customer's priorities to improve from known levels of an example service - this may be his perception of his current service, or some pre-specified service specification provided by the client.  This latter is particularly recommended when respondents may not know for sure the levels of service delivered on some attributes.

 The sequence of questionnaire completion can be illustrated as follows:

Expectations

Current Perception

Unacceptable

First Priority

Second Priority - Grid Fully Completed

Importantly, these improvement priorities can be ascertained with relative costs attached to the various levels of service.  This recognises the real life situation to suppliers that some things cost more to provide than others (not all the best things in life are free).  Service provision is a business, and suppliers want the best Return for their Investment - return in terms of more loyal customers who will use their service more if this is possible, and attracting rivals' customers.

The benchmarking aspect of SIMALTO data collection is particularly useful when combined with a segmentation of customers into "degrees of loyalty" categories.  Are potentially disloyal customers receiving the same service as very loyal customers (and if not, what is the difference and why) and do they have the same priority of improvement benefits?  And if the disloyal customers received the benefits they prioritise - or the same level of service as loyal customers now receive, would they become more loyal, or are they naturally fickle?