Dear community,

i conducted a SERVIMPERF customer satisfaction study and retrieved 21 customer ratings for perceived importance and satisfaction with 15 items.

In order to prioritize the items for improvement, various books advise to build a quotient out of these two values, dividing the importance values by the satisfaction values for each item (KPI=Weight/Satisfaction). By doing so the items with the highest weight and lowest satisfaction will have high rankings for future improvements.

Considering the stated satisfaction it is logical to use the mean or average value for satisfaction (for x surveyed customers).

My question is about the importance values. Should i use for the calculation of the KPI:

- average importance for x surveyed customers

- relative importance (as used in the conjoint analysis) for the impact of each item?

On the one hand conjoint analysis uses relative values for importance. On the other hand i am not sure if it is statistically wrong to divide relative values by average values and can be claimed as "bad math"?

If i compare the outcomes of the two calculations, they also result in different rankings.

I appreciate any help on this statistical issue.

Thank you a lot in advance.

Best Regards

Dimitri