RESEARCH DESIGN AND DATA COLLECTION
The data collection process began with a series of in-depth exploratory interviews conducted
with bank managers and bank customers, followed by a survey in which a questionnaire was
distributed to a stratified random sample of bank customers. A total of seventeen in-depth
interviews took place, nine interviews with bank managers and eight interviews with bank
customers. These in-depth interviews allowed an initial grasp of the important issues for both
sides in the provision of a satisfying banking service, provided clarification of the research
problem and assisted the formulation of a questionnaire for the quantitative research stage.
The sample comprised four strata including users of the main bank delivery channels:
the Internet-based delivery channel, the telephone-based access to the bank, the debit-card
access (through the ATM network) and a stratum of users of the branch network of the bank.
The strata construction followed the definition of a user of a delivery channel as a client who
had used that particular delivery channel in two specified months of that year (namely,
January and February, considered to be typical months for the banking business).
One major bank1 agreed to collaborate in the research by distributing the
questionnaire to its customers. In order to have the strata as mutually exclusive groups of
users, it was decided to have a ‘pyramidal definition’ of the strata. Following the bank rules
on access to the delivery channels, the strata were successively extracted from the global
customer base: the first stratum to be identified was users of the Internet-based access,
followed by users of the telephone-based channel, debit card users and branch users. This
sequential process avoided duplication of customers among strata. The branch users received
a shorter questionnaire, as questions about the technology-based channels did not make sense
for these customers.
The questionnaire was mailed to 6,000 customers of each stratum, leading to a total of
24,000 questionnaires sent to users of the various delivery channels. Although the response
rate differed among the strata, the overall response rate was 20.2%. A total of 4,852
questionnaires were received, of which 3,250 were considered usable (1,467 from the on-line
channel users, 682 from the telephone-based access users, 731 debit card users and 370
branch users). The difference between the response rates was considered to result from
differences in education and in understanding the benefits of the bank’s effort: the
questionnaire was presented as coming from the bank with the intent of increasing customer
satisfaction. The survey responses were not re-weighted to correct for differences in the
1 The identity of the organization that provided the data is kept confidential to honor the wishes of the sponsor.
13