JOURNAL OF APPLIED BEHAVIOR ANALYSIS 1988, 219 65-71 NUMBER 1 (SPRING 1988)
TASK CLARIFICATION, PERFORMANCE FEEDBACK, AND SOCIAL PRAISE:
PROCEDURES FOR IMPROVING THE CUSTOMER SERVICE
OF BANK TELLERS
CHARLES R. CROWELL, D. CHRIS ANDERSON, DAwN M. ABEL, AND
JOSEPH P. SERGIO
UNIVERSITY OF NOTRE DAME
Customer service for bank tellers was defined in terms of 11 verbal behavior categories. An audio-
recording system was used to track the occurrence of behaviors in these categories for six retail
banking tellers. Three behavior management interventions (task clarification, performance feedback,
and social praise), applied in sequence, were designed to improve overall teller performance with
regard to the behavioral categories targeted. Clarification was accomplished by providing dear
delineation of the various target categories, with specific examples of the behaviors in each. Feedback
entailed presentation of ongoing verbal and visual information regarding teller performance. Praise
consisted of verbal recognition of teller performance by branch managers. Results showed that
clarification effects emerged quickly, producing an overall increase in desired behaviors of 12% over
baseline. Feedback and praise effects occurred more gradually, resulting in overall increases of 6%
and 7%6, respectively. A suspension of all procedures led to a decline in overall performance, whereas
reinstatement of feedback and praise was again accompanied by performance improvement. These
findings extend the generality of behavior management applications and help to distinguish between
possible antecedent and consequent effects of performance feedback.
DESCRIPTORS: organizational behavior management, bank tellers, customer service, feedback,
darification
Few applications of organizational behavior the efficacy of similar procedures with bank tellers,
management (OBM) in the banking industry have who are considered by many in their industry to
been reported, despite evidence showing that work be the primary agents of customer relations.
performance in many areas of a financial institution, A second purpose of this study was to explore
and even bank profitability, can be improved with further the possibility that typical OBM feedback
these methods (Abernathy, Duffy, & O'Brien, 1982; interventions may exert an influence on behavior
Roberts, 1983). However, given today's highly both as antecedents and as consequences (Prue &
competitive financial services marketplace, it is im- Fairbank, 1981). One possible antecedent effect of
perative that banks optimize the performance of feedback is what may be called "task clarification,"
their human resources, especially in the area of or precise specification of the behavioral compo-
customer service. Several studies of retail selling nents of a job. In this capacity, feedback interven-
have demonstrated that customer-service-oriented tions may inform employees, perhaps for the first
employee behaviors can be increased through the time, of the particular actions that are desired for
application of behavior management techniques such successful completion of a required task. Such an
as feedback and praise (e.g., Brown, Malott, Dillon, effect of feedback may be primarily directional in
& Keeps, 1980; Komaki, Blood, & Holder, 1980). nature and may be especially useful for those tasks,
One purpose of the present study was to evaluate like customer service, that so often have been char-
acterized by ambiguity with regard to performance
criteria.
The authors thank W. D. Jones, J. Seitz, B. Nichols, T. Feedback may also have response-strengthening
Wyczkowski, L. Siroky,J. Behrens,J. DuBoyce, L. Hannum, effects apart from task clarification. Such effects may
L. Hau, and T. Schrek for their assistance with various phases be more reinforcement-like in nature and thus may
of this research. Request for reprints should be addressed to
Charles R. Crowell, Department of Psychology, University be dependent upon repeated application of the in-
of Notre Dame, Notre Dame, Indiana 46556. tervention. In this way, feedback may serve more
65
, 66 CHARLES R. CROWELL et al.
as a consequence than as an antecedent in the con- point" values were assigned to each category based
trol of behavior. on an arbitrary priority determined by bank man-
In the present study, separate task clarification agement. The sum total of all quality points for
and feedback interventions were used. The clarifi- any transaction was 100.
cation procedures used here were similar to those The specific categories targeted, together with
used in previous research (e.g., Brown et al., 1980; their quality point assignments, were as follows:
Geller, Eason, Phillips, & Pierson, 1980). We ex- 1. Time to service: latency of first verbal ac-
pected that the sequential introduction of clarifi- knowledgment of customer by teller, measured from
cation and feedback would first reveal any long- moment customer was within one yard of the teller
term directional effects of task specification fol- window. A latency of 0 to 5 s was assigned 12
lowed, perhaps, by further response strengthening points; 6 to 11 s was assigned 7 points; longer than
caused by explicit knowledge of performance. 11 s was assigned 0 points.
2. Greeting: opening teller verbalization, such
as "hello" or "good morning." A phrase emitted
METHOD by the teller at the start of a transaction from among
Subjects the list of acceptable greetings was assigned 10
The participants in this study included a male points.
branch manager, a female operations manager, five 3. Expression ofconcern: one of several possible
full-time female tellers, and one part-time female teller statements to the effect of "how are you
teller. The tellers ranged in age from late twenties today?" This category was assigned a value of 5
to late fifties. Five of the tellers had been employed points.
by the bank for 1 to 6 years, whereas one joined 4. Using customer's name: including mention
the bank shortly after the project started. One teller of the customer's first and/or last name somewhere
left the employ of the bank before the project was in the transaction. Using the name, which is nor-
completed for reasons unrelated to the study. mally available from papers, deposit slips, and so
on, was worth 10 quality points.
Setting 5. Ta/king only to customer: refraining from
This study was conducted in one branch of a conversation with someone other than the current
medium-sized Midwestern bank. At the time of customer during a transaction, /unless the teller is
this study, the bank had 12 urban branch locations specifically addressed by someone else or it is nec-
and reported total assets of approximately 500 mil- essary to obtain information. This category was
lion dollars. The particular branch selected for the assigned a value of 10 points.
project was one of the bank's oldest and had ex- 6. Additional assistance: one of several possible
hibited only a modest rate of growth in the pre- queries by the teller to the effect "can I be of further
ceding 2 years. assistance?" Also, this category, worth 6 points,
could be satisfied through referral to another bank-
Behavioral Targets ing product or service.
The specific behavioral categories targeted in this 7. Minimizing small talk: keeping unnecessary
study were selected on the basis of extensive pre- self-initiated conversation to a minimum so as not
liminary observations of teller-customer interac- to prolong the transaction unduly. Points were
tions, combined with input from bank personnel, awarded in this category based upon the percentage
including the tellers themselves. The final list of of a transaction during which the teller refrained
behavioral targets was largely a reflection of bank from small talk. Each transaction was divided into
management's decision that each teller-customer consecutive 10-s intervals. The percentage of in-
interaction should be characterized by certain key tervals without small talk was computed. For short
features, including a specific set of teller verbaliza- transactions (less than 90 s in duration), 6 points
tions. Once the target list was formulated, "quality were awarded if the percentage of intervals without
