Research On The Comparability About The Customer Satisfaction Index

With the widespread of customer satisfaction index (CSI) in the world, the demand ofcomparing among different CSI is on the increase. The incomparability phenomenon thathas been captured needs to be further researched. Among the various influence factors onthe incomparability in CSI, we focus on three of them, including customer heterogeneity,generic measurement model and negative skewed satisfaction score.It is a pioneering research as there are few references about these three issues. It isfeasible that we take Monte Carlo simulation study as an exploratory analysis and takeempirical study as a further confirmatory factor analysis. One standardized service (CCTVNetwork News Broadcast) and another standardized product (Coca Cola) were put to CSIinvestigation by CATI (computer assisted telephone interview).The main findings of thispaper can be summarized as follows:（1）Customer diversity has no effect on the CSI under other conditions beingidentical. The nominal CSI complies with the real CSI when we measure CSI for onebrand among different populations.（2）The core of comparability is the consistency of the meanings of customersatisfaction, which means the manifest variables specification about the core latentvariable customer satisfaction in one CSI system should be consistent with that in anotherCSI system. Meanwhile, it’s found that the latent variables expect customer satisfactionand the relationships among all latent variables have no significant effect on thecomparability. （3）The negative skewed satisfaction scores tend to exert influence on thecomparability of CSI through the standardizing procedure. Then, we presented anapproach to solve the comparability problem in CSI by exerting an operator on the currentCSI estimating results.This research is exploratory in nature; major improvements and innovations areevident in the following aspects:（1）For the first time, this paper has taken direct research on customer heterogeneity,CSI generic measurement model and negative skewed satisfaction scores, which isexploratory and fills the gaps in CSI to certain extent. Furthermore, the main findingsabout these issues can be the guide and benchmark for national CSI in China as well as inthe world.（2）Combining mean theory and hypothesis test together, the paper has given anapplicable testing method. It tests the hypothesis of the influences of customerheterogeneity by the method of comparing the conditional mean with the unconditionalmean.（3）Based on statistical hypotheses test, it has proposed that the hypothesis test wascarried out by means of transforming the generic measurement model used by differentregions to different generic measurement model used by one region, which is a usefultesting method for comparability testing.（4）The paper has proved the influence by the skewed satisfaction scores on thecomparability and has presented an approach to solve the comparability problem in CSI byexerting an operator on the current CSI estimating results. It is referential experience forsimilar research.