The Performance Effect of Feedback Frequency and Detail: Evidence from a Field Experiment in Customer Satisfaction

• DOI: http://doi.org/10.1111/1475-679X.12184
• Published date: 01 December 2017
• Date: 01 December 2017

DOI: 10.1111/1475-679X.12184

Journal of Accounting Research

Vol. 55 No. 5 December 2017

Printed in U.S.A.

The Performance Effect of

Feedback Frequency and Detail:

Evidence from a Field Experiment

in Customer Satisfaction

PABLO CASAS-ARCE,

∗SOFIA M. LOURENC¸O,

†

AND F. AS´

IS MART´

INEZ-JEREZ

‡

Received 17 June 2015; accepted 11 July 2017

ABSTRACT

This paper presents the results from a field experiment that examines the ef-

fects of nonfinancial performance feedback on the behavior of professionals

working for an insurance repair company. Wevary the frequency (weekly and

monthly) and the level of detail of the feedback that the 800 professionals

receive. Contrary to what we would expect if these professionals conformed

to the model of the Bayesian decision maker, more (and more frequent) in-

∗Arizona State University; †ISEG, Universidade de Lisboa and Advance, CSG Research Cen-

ter; ‡University of Notre Dame.

Accepted by Philip Berger. We thank Alberto Abadie, Larbi Alaoui, Ramji Balakrishnan,

Javier Bartolom´

e, Seth Berry,Margaret Christ, Juanjo Ganuza, Gar y Hecht, Chris Ittner,Michal

Matejka, Katherine Miller, Andrei Shleifer, Sriram Somanchi, two anonymous referees, and

seminar participants at the AAA Annual Meeting, Arizona State University-University of Ari-

zona Workshop, Carnegie Mellon University, IAP Cambridge Research Symposium, GMARS,

IE Business School, Midwest Summer Research Conference, MAS Midyear Meeting, North-

western University, Temple University Accounting Conference, Universidad Complutense de

Madrid, Universitat Pompeu Fabra, University of Notre Dame, and University of Toledo for

helpful comments. Sofia M. Lourenc¸o gratefully acknowledges financial support from FCT—

Fundac¸˜

ao para a Ciˆ

encia e Tecnologia (Portugal), research grant PTDC/EGE-GES/119607/2010

(national funding), and Pablo Casas-Arce from the Spanish Ministry of Science and Innova-

tion research grants MINECO ECO2011-28965 and MINECO ECO2014-59225-P.All errors are

our own.

1051

Copyright C,University of Chicago on behalf of the Accounting Research Center, 2017

1052 P.CASAS-ARCE,S.M.LOURENC¸O,AND F.A.MART´

INEZ-JEREZ

formation does not always help improve performance. In fact, we find that

professionals achieve the best outcomes when they receive detailed but infre-

quent (monthly) feedback. The treatment groups with frequent feedback, re-

gardless of how detailed it is, perform no better than the control group (with

monthly and aggregate information). The results are consistent with the in-

formation in the latest feedback report being most salient and professionals

in the weekly treatments overweighting their most recent performance, ham-

pering their ability to learn.

JEL codes: C93; D81; D91; J01; M40; M41; M52; M54; M55

Keywords: feedback; incentives; salience; performance; customer satisfac-

tion; field experiment

1. Introduction

One of the main roles of accounting information is to facilitate decision

making. Timeliness and detail are usually regarded as desirable character-

istics of information because they enable prompt and adequate responses

to business threats and opportunities. However, the intensity of these at-

tributes must be weighed against the decision maker’s ability to process the

relevant information. Too frequent information may result, for instance,

in an overreaction to short-term factors, whereas too detailed informa-

tion may cloud a decision maker’s ability to identify general trends or is-

sues. The progress of technology, dramatically expanding the amount of

data collected, and the speed with which this information can be made

available to managers for decision making, has exacerbated these ten-

sions. In this paper, we use a natural field experiment to analyze how

the frequency and detail of performance feedback influences employee

behavior.

In collaboration with Multiasistencia—the leading Spanish business pro-

cess outsourcer of repairs for insurance companies—we design and imple-

ment a field experiment in which we manipulate the nonfinancial per-

formance feedback received by 800 home repair professionals (such as

plumbers, masons, or painters) who work with the firm. We intervene in

the reward and feedback system by introducing a bonus that rewards the

achievement of a target for customer satisfaction—a metric that has been

found to be a significant leading indicator of the financial health of a firm

(Ittner and Larcker [1998])—as well as two process indicators. We vary the

frequency of feedback information (weekly vs. monthly) and the level of

detail included in the report (the average score of all jobs performed dur-

ing the period by a professional vs. the scores for each of the individual

jobs).

If the professionals in our setting conformed to the ideal of the Bayesian

decision maker, they would efficiently use all the information available to

them to improve customer satisfaction. Therefore, more detailed and more

THE PERFORMANCE EFFECT OF FEEDBACK FREQUENCY AND DETAIL 1053

frequent feedback should lead to better performance. However, more (and

more frequent) customer satisfaction feedback does not always result in

improved customer satisfaction scores. In fact, we find that professionals

achieve higher scores when they receive detailed but infrequent (monthly)

feedback. These results are consistent with the latest feedback report being

most salient, and professionals overweighting the information it contains.

As a result, although detailed customer satisfaction feedback supplies infor-

mation that helps professionals to improve the service they provide, feed-

back that is more frequent (and that consequently focuses on a shorter time

horizon) ends up harming service, as previous information is disregarded

in the face of new information.

Notably, we also show that the deterioration in performance for the pro-

fessionals in the weekly treatments cannot be explained by their rational

abandonment of the monthly bonus. It is true that professionals in weekly

treatments may review their decision to exert effort every time they receive

a feedback update. It is also true that, if they learn in a weekly update that

they are unlikely to qualify for the bonus, they may (rationally) decide to

abandon the pursuit of the monthly target, resulting in lower performance

than participants in monthly treatments, who do not receive weekly up-

dates. However, we find that such dynamic incentives cannot be the main

driver of our results, because weekly treatments do worse than monthly

treatments in the first week of the month, particularly when they follow a

negative report in the last week of the prior month. This result cannot be

explained by the rational abandonment story, as bad performance in the

last week of a month has no impact on the chances of achieving the bonus

in the next month.1

These differences among treatments do not exist with respect to the pro-

cess indicators included in the bonus system (e.g., the use of the Internet to

schedule a service or finishing a repair on time). This is because the profes-

sionals receive immediate feedback simply by executing these tasks. Thus,

the differences in the features of the formal feedback system do not result

in any additional information, and do not affect professionals’ knowledge

about their performance or the way they process that information (Annett

[1969]). Consistent with this, we show that operational performance is in-

deed the same in all four treatments.

This paper contributes to several streams of literature. First, we con-

tribute to the literature in information economics. In this literature, the

contracting stream of research on performance measurement mainly fo-

cuses on how properties of information affect their inclusion in contracts

(Feltham and Xie [1994], Prendergast [2002], Moers [2006]). The ulti-

mate objective of this literature is to judge the strength of the performance

metric in providing information to the firm about employees’ choices (the

1Moreover, there is no reason for professionals in the weekly treatments to be more likely

to infer that the bonus is not achievable. Both weekly and monthly treatments have the same

information about past performance during the first week of each month.