Customer satisfaction and competencies: an econometric study of an Italian bank

Abstract

We empirically address how Customer Satisfaction (CS) and loyalty in the banking industry may affect profitability. This helps in identifying the strategy and competencies necessary to benefit from customer relationships, which are important sources for improved performance in the banking. We do this by analysing data collected on 2105 customers of 118 branches of one of the biggest banks of an Italian banking group. We find that CS impacts loyalty, which in turn has a direct effect on financial and nonfinancial customer value/total customer value/complex customer value. Moreover, loyalty is a mediator between financial and nonfinancial customer value and two sources of CS, namely, relationships with the front office and the branch, on the one hand, and the products offered, on the other.

Acknowledgements

We thank (without implicating) Riccardo Leoni, Torben Pedersen, Larissa Rabbiosi, Sabina Tasheva and Enrico Fabrizi for comments on earlier drafts. This research was carried out while Paola Gritti was a visiting scholar at the Center for Strategic Management and Globalization at the Copenhagen Business School.

Notes

¹In addition, some relationships between the branch level and the customer level will be considered.

²The retail customers of a bank include individuals and small businesses. Although 20 000 customers were asked to participate, the final sample was 2995 customers.

³The rating measures the profitability of customers for the branch, not only in terms of total revenue but also in terms of the number and value of the products they hold.

⁴AIR/BIR is a classification of customers on the basis of their income and age.

⁵For all of our results we calculated the marginal effects. They confirm the directions of the impacts and give their intensity. They are available on request.

⁶The moderation hypothesis is supported if the interaction, as measured by the product of the variables taken into consideration, is significant. There may also be significant main effects for the predictor (the independent variable) and the moderator, but, conceptually, these are not directly relevant to the test of the moderator hypothesis.

⁷The results can be shown on request.