ArticlePDF Available

Effects of employees' opportunities to influence in-store music on sales: Evidence from a field experiment


Abstract and Figures

The effects of in-store music on consumer behavior have attracted much attention in the marketing literature, but surprisingly few studies have investigated in-store music in relation to employees. Conducting a large-scale field experiment in eight Filippa K fashion stores in Stockholm, Sweden, we investigate whether it is beneficial for store owners to give employees more opportunities to influence the in-store music. The experiment lasted 56 weeks, and the stores were randomly assigned into a treatment group and a control group, with the employees in the treatment stores having the opportunity to influence the in-store music through an app developed by Soundtrack Your Brand (SYB). The results from the experiment show that sales decreased by, on average, 6% in treatment stores when employees had the opportunity to influence the music played in the store. Interviews revealed that employees frequently changed songs, preferred to play high-intensity songs, and had diverse music preferences that were not congruent with the brand values of the company. Our results thus imply that employees choose music that suits their preferences rather than based on what is optimal for the store, suggesting that store owners might want to limit their opportunities to influence the background music in stores.
Content may be subject to copyright.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
Please cite this article as: Sven-Olov Daunfeldt, Journal of Retailing and Consumer Services,
0969-6989/© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (
Effects of employeesopportunities to inuence in-store music on sales:
Evidence from a eld experiment
Sven-Olov Daunfeldt
, Jasmine Moradi
, Niklas Rudholm
, Christina ¨
Institute of Retail Economics, Stockholm, Sweden
Soundtrack Your Brand, Stockholm, Sweden
Orebro University, ¨
Orebro, Sweden
JEL classication:
C93, D22, L81, M54, M31
Background music
Brand-t music
Music tempo
Consumer behavior
Job satisfaction
Atmospheric cues
Work environment
Field experiment
The effects of in-store music on consumer behavior have attracted much attention in the marketing literature, but
surprisingly few studies have investigated in-store music in relation to employees. By conducting a eld
experiment in eight Filippa K fashion stores in Stockholm, Sweden, we investigate whether it is benecial for
store owners to give employees more opportunities to inuence the in-store music. We randomly assigned the
stores into a treatment group and a control group, with the employees in the treatment stores having the op-
portunity to inuence the in-store music through an app developed by Soundtrack Your Brand (SYB). The
experiment lasted 56 weeks and sales data were also gathered 22 weeks before the experiment, resulting in a
total of 4626 observations. Our results show that sales decreased by 6% when the employees had the opportunity
to inuence the music played in the store, and the effect is driven by a reduction in sales of womens clothing.
Interviews with the employees revealed that they had diverse music preferences, frequently changed songs, and
preferred to play high-intensity songs. Employees thus seem to make choices regarding the in-store music that
reduce sales, implying that store owners might want to limit their opportunities to inuence the background
1. Introduction
Already in 1915, Thomas Edison tested whether music could be used
to increase the productivity of his factory workers. He reportedly found
no effects, perhaps because of the noisy work environment and the poor
sound quality at that time (Kellaris, 2008). Since then, the effects of
in-store music have attracted much attention in the marketing literature
(for overviews, see, e.g., Hargreaves and North, 1997; Kellaris, 2008;
Yorkston, 2010; Michel et al., 2017; Roschk et al., 2017). These studies
have almost exclusively focused on how in-store music affects con-
sumers, for example, by studying the effects of music presence (Garlin
and Owen, 2006), music choice (Areni and Kim, 1993; North et al.,
1999, 2016; Wilson, 2003), music tempo (Milliman, 1982, 1986; Oakes,
2003; Knoeferle et al., 2017), music volume (Biswas et al., 2019),
brand-t music (Beverland et al., 2006; Daunfeldt et al., 2017), and the
interaction of music with other sensory cues in the store environment
(Mattila and Wirtz, 2001).
Surprisingly, few studies have investigated in-store music in relation
to employees. This lack of research is puzzling considering that music
can be an important factor for well-being and the working environment
(Shih et al., 2012). On the one hand, an increased opportunity for em-
ployees to inuence in-store music could make them more satised and
service oriented and therefore increase sales, lead to more satised
customers and improve the image of the brand. On the other hand,
businesses may be afraid that their staff will make suboptimal choices
that will result in lower sales and less satised customers.
We contribute to the literature by conducting a eld experiment in
eight Filippa K stores to test whether sales are affected by the oppor-
tunities for employees to inuence in-store music. Filippa K is a Swedish
fashion retail chain with 50 Filippa K stores and more than 700 selected
retailers in 20 countries that sells classic clothes with a modern and
minimalist design. We randomly selected stores into a treatment group,
where the staff had the opportunity to inuence the in-store music, and a
control group, where the employees had no such opportunities. To
reduce geographical heterogeneity and to facilitate identifying the ef-
fects of the experiment, all stores in the experiment are located in
Stockholm, Sweden.
The in-store soundtrack that we use in the experiment was developed
* Corresponding author.
E-mail address: (S.-O. Daunfeldt).
Contents lists available at ScienceDirect
Journal of Retailing and Consumer Services
journal homepage:
Received 11 January 2020; Received in revised form 9 November 2020; Accepted 30 November 2020
Journal of Retailing and Consumer Services xxx (xxxx) xxx
by music curators at Soundtrack Your Brand (SYB), and the songs
included in the playlists were selected to reect the chains brand
values. The music chosen in both the control- and the treatment stores
are thus in congruence with Filippa Ks brand values, which is of
importance because brand-t music can align consumers with the
marketplace and thereby inuence their consumption and perceptions
(see e.g., Bruner, 1990; Beverland et al., 2006; Daunfeldt et al., 2017).
The difference between the stores in the treatment group and the control
group is that SYB developed an app that could be used by the employees
in the treatment stores, while no such app was introduced in the control
stores. The app made it possible for employees to change the volume of
songs, skip songs, share songs with consumers, and choose between a
brand-t playlist with high-intensity or medium-intensity songs. In all
other ways, the playlists were controlled from SYBs headquarters in
Stockholm, limiting the risk of noncompliance problems.
The experiment lasted 56 weeks, and data were also gathered during
the 22 weeks before the experimental period. The scale of our study
differs from most previous studies on in-store music, which tend to be
based on eld experiments that cover one store (Areni and Kim, 1993;
North et al., 1999), restaurant (Wilson, 2003), coffee shop (North and
Hargreaves, 1996), or shopping mall (Yalch and Spangenberg, 1990)
during a very short experimental period (usually a week). Due to the use
of a few stores and the short study period, the results might have been
driven by store-specic or time-specic conditions, thus limiting the
ability to draw causal inferences based on these experiments.
We investigate the effect of the eld experiment on sales by esti-
mating a difference-in-differences regression model that controls for
time-specic effects and time-invariant differences between the treat-
ment and control stores. To our knowledge, this is the rst attempt to
investigate the effects on sales of letting employees inuence in-store
music. We also conduct thirteen semi-structured interviews with em-
ployees at the Filippa K stores, providing us with more information on
employees attitudes toward the background music and their possibil-
ities to inuence in-store music.
The remainder of this paper is structured as follows: A literature
review on the effects of in-store music is presented in the next section.
Our experiment and research design are explained in Section 3, while
our empirical model and descriptive statistics are presented in Section 4.
The results from the eld experiment, concerning the effects of em-
ployeesopportunities to inuence in-store music on sales, are presented
in Section 5, while results from the interviews are summarized in Section
6. The results are then discussed in Section 7. Finally, Section 8 con-
cludes with theoretical contributions, managerial implications, and
suggestions for further research.
2. The effect of in-store music: a literature review
To position the present paper towards previous studies, this section
summarizes research on the role in-store music plays for employees and
its effect on customers buying, respectively. The research gap is pre-
sented in the intersection between these two areas of research, empha-
sizing how employees inuence in-store music and its effects on sales.
2.1. In-store music and the employees
In research on work environments, scholars have pointed out how
the physical setting affects job satisfaction, well-being, productivity, and
motivation (Sundstrom and Sundstrom, 1986; Bernstein and Turban,
2018; Otterbring et al., 2018, 2020). Skandrani et al. (2011) investigate
atmospheric cues in stores to determine their effect on employee atti-
tudes and behaviors. Related to music specically, they nd that em-
ployees prefer music over silence, that music decreases employees
feelings of boredom, and that familiar music or music employees like
increases their performance and motivation. A lack of variation (in songs
and rhythms) and incongruence with the time of day negatively affect
In a study by Knifn et al. (2017), the positive inuence of music is
shown in terms of willingness to cooperate, while Korczynski (2003),
who provides a historical overview, attempts to explain how managers
may have a negative idea about music in the workplace. Despite a
broadly positive relation between music in the workplace and em-
ployees, there are studies acknowledging variations in preferences
among individuals (Bruner, 1990; North and Hargreaves, 1996). A
number of studies have indicated that, for example, the effect of music
depends on whether the individual has an extroverted or introverted
personality (e.g., Furnham and Bradley, 1997; Cassidy and MacDonald,
2007). Extroverts generally listen to music more often than introverts,
and they are less affected by music than introverts when completing
tasks (Furnham and Bradley, 1997). While there are a number of studies
indeed focusing on how music affects employees, the specic circum-
stance of store employees (see Skandrani et al., 2011 as an exception of
research on store employees) and the further link to customers seems to
be absent in present research.
2.2. Music and the customer
According to previous research, music is one of the environmental
factors or atmospheric cues that affect not only image, but also behav-
iors (Bruner, 1990; Bitner, 1992; Turley and Milliman, 2000; Mari and
Poggesi, 2013; Helmefalk, 2017). The marketing literature has
addressed not only how music, in terms of tempo and volume, may affect
how long a customer stays in a store and the amount spent but also
whether the genre ts with customers (and customer preferences) or
alienates them in (Milliman, 1982, 1986; Areni and Kim, 1993; Hui
et al., 1997; Mattila and Wirtz, 2001; Garlin and Owen, 2006) and
whether the music creates a t with other values and cues exposed to
customers (Biswas et al., 2014). It has also been acknowledged that
music affects customer perceptions, such as the amount of time that
customers believe that they have spent in a store and the amount of time
waiting (cf. MacInnis and Park, 1991; Baker et al., 2002), while the
mode and intention of shopping also impact how customers experience
In the right combination, expressed as pleasantness in the literature,
music positively affects behavior and increases customer purchases
(Morrison et al., 2011). In a meta-analysis on research on the effect of
in-store music on customers, Garlin and Owen (2006) summarize
research by pointing out how music (rather than silence) creates a
feeling of pleasure and benevolence (cf. Andersson et al., 2012); how
music that is familiar and preferred positively affects benevolence; how
a slower tempo, a lower volume and familiar music lead customers to
stay in stores for longer periods of time, while a higher volume and
tempo and music that does not meet the taste of customers cause cus-
tomersperception of time to be longer than the actual time passed.
Garlin and Owen further conclude how tempo has the highest impact on
the customers behavior.
The positive effects of in-store music on consumers are conrmed in
a more recent meta-analysis by Roschk et al. (2017), showing that the
existence of background music was associated with higher pleasure,
satisfaction, and behavioral intention ratings compared to a situation
with no in-store music. However, Michel et al. (2017) question the
positive effect of in-store music in another literature review, nding that
many studies present negative or non-signicant effects of in-store music
on sales in comparison to a situation without in-store music. They
concluded that the design of in-store music contributes more to the
explanation of the musical effects than does the existence of in-store
music(p. 30). In particular, they emphasize that tempo seem to have
a positive effect on consumers emotions, but a negative effect on sales
volumes and time spent in the store. Their results are supported by
Hynes and Manson, 2016, who found that the presence of background
music in supermarkets mainly acted as a distraction when consumers
were aware of it.
However, the effects of in-store music on consumer behavior seem to
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
be context dependent. Knoeferle et al. (2017), for example, found that
high-tempo music moderated the negative effects of retail crowding,
suggesting that fast music might be preferable to slow music under such
conditions. Biswas et al. (2019) showed that music volume might affect
consumersproduct choices, with low volumes promoting sales of
healthy foods, suggesting that managers might want to manipulate
music volumes depending on what they want to sell. Adding to this is the
idea on brand-t (Beverland et al., 2006; Daunfeldt et al., 2017), where
the in-store music can help to establish a brand and keep customers
associations with it. Daunfeldt et al. (2017) showed that such t could
raise sales, and thereby how music is not only about meeting customers
taste but to create an entirety with the store and its offerings.
2.3. Synthesis
Bitner (1992) made an early attempt to link together research on
what she refers to as the impact of physical surroundings on customers
and employees, thus adopting a broader set of stimuli than music alone.
Later studies have also conrmed a linkage between physical working
environments and aspects such as job satisfaction, well-being, and pro-
ductivity (Bernstein and Turban, 2018; Otterbring et al., 2018, 2020).
In studies further developing the ideas of Bitner (1992), which are
normally linked to service encounters, the employee and music are
mostly treated as independent variables to explain customer attitudes
and behaviors. Levy and Grewal (1993), for instance, point out the
interaction between ambient factors of music and light and how the
number of and friendliness of employees affect customerspurchase
likelihood. There are also suggestions about how music affects how
employees are perceived by customers (Morin et al., 2007), and while
the effect on employee behavior and the link to customer attitudes and
behaviors are rarely present, there is one exception suggesting that
dance music causes employees to start dancing, thus scaring off
Through largely treating music and employees as unrelated stimuli
to customers in-store behavior or buying, there is a missing link be-
tween variables that only and partly comes forth in the rare study by
Morin et al. (2007): how employees may have an interactive effect be-
tween the music and the customers behavior. Exploring this path
further while specically linking it to the employees ability to inuence
the music in a store, there are some potential and previously unexplored
options at hand: If the employee is allowed to inuence the music,
he/she could assume to be further motivated in the work (cf. Sundstrom
and Sundstrom, 1986; Kumar and Pansari, 2014) with positive effects on
the employee-customer interaction. There are though the risk that the
employees would select music that does not t with the brand, or music
that has an advert effect on customersbuying (e.g., too loud, too fast, or
not tting the customers preferences or is unfamiliar for the customer,
cf. Garlin and Owen, 2006). These risks may be amplied as various
employees would represent different music tastes (cf. Furnham and
Bradley, 1997; Cassidy and MacDonald, 2007). Furthermore, the
employee may based on the music selected act in ways that nega-
tively affect the customers experience (such as the dancing). Whether it
is benecial for store owners to give employees more opportunities to
inuence the in-store music is what this paper sets to address through an
in-store experiment and interviews with store employees.
3. The experiment and research design
3.1. Field experiment
To test the effects of in-store music on employees, we conducted a
eld experiment in eight Filippa K fashion stores in Stockholm, Sweden.
The background music was supplied by SYB, which created a soundtrack
that was supposed to be congruent with the brand values of the com-
pany. More specically, the in-store music at Filippa K was selected to
signal the following brand values: (i) exclusive, (ii) elegant, (iii)
innovative and forward thinking, (iv) expressive, and (v) non-explicit.
Regarding elegant music, examples may include songs that are light,
airy and shimmering; the opposite of elegant music is rugged music,
which has a rougher and more robust sound. Innovative music repre-
sents the sound of tomorrow, where styles and genres are mixed; in this
regard, conventional and classic songs are the opposite, i.e., no sur-
prises, and thus, such songs should be avoided in Filippa K stores.
Expressive music is music that stands out, i.e., music that is catchy.
Every fth song is something that is slightly different, making the
soundtrack more expressive and interesting. Non-explicit music is low-
key background music.
The eight stores in the experiment were randomly selected into four
treatment stores and four control stores. The difference between the
stores in the experiment is that SYB created an app so that the employees
in the treatment stores could inuence the in-store music during the
experimental period, while the staff at the control stores did not have
this opportunity. More specically, the staff at the treatment stores had
the opportunity to inuence the in-store music by (i) adjusting the
volume, (ii) skipping songs, (iii) sharing songs with customers, and (iv)
choosing a soundtrack with high-intensity songs or medium-intensity
songs. These choices were made in collaboration with Filippa K and
used because previous studies have indicated that choice of volume (e.
g., Garlin and Owen, 2006), song choice (e.g., Jacob, 2006) and music
tempo (e.g., Milliman, 1986) might be of importance in inuencing
consumer behavior.
The experimental period started on 5 July 2016, and ended on 1
August 2017, which means that the experiment lasted over 56 weeks.
We also gathered data during the 22 weeks before the experimental
period, starting on 1 February 2016. Our experimental design means
that we can compare sales in the treatment stored during the experi-
mental period with sales in the pretreatment period and in the control
stores during the experimental period. Thus, we have a much larger
dataset than that in most previous studies, making causal inferences
more convincing. In all other ways, apart from the differences noted
above that were part of the experiment, the soundtracks were controlled
from SYBs headquarters. In other words, employees could not inuence
the choice of songs in the soundtrack during the experimental and
control periods, nor could they make any changes other than those that
the app allowed. This aspect of the experimental design limits the pos-
sibility of noncompliance problems, which is otherwise a common
feature in eld experiments (Krueger, 1999; Ortmann, 2005; Duo et al.,
3.2. Interviewees
To gain a deeper understanding of the results, semi-structured in-
terviews (cf. Sarantakos, 1998) were conducted with 13 employees (3
men and 10 women) in the four treatment stores.
The aim of the in-
terviews was to provide the staff with an opportunity to reect on the
soundtrack, their own personal experiences of the app and the increased
opportunity to inuence the music played in the store. All interviews
were conducted in Swedish, and the average length of the interviews
was approximately 20 min. The interview guide has been translated into
English and is presented in Appendix 3.
In the analysis of the interviews, interview transcripts were rst
empirically coded using a thematic coding feeding back to the interview
guide and previous research on music and employees. More specically,
the interview data was coded based on: the employees preferences and
use of the app, the employees way of changing the music and rationales
therefore, the employees interpretation of the brand image and about
how the changing of music affected customers. The interviews were
The proportion of male and female employees interviewed are based on the
proportion of male and female employees working in the stores when the
experiment was implemented.
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
rstly categorized separately and in a second step compared, meaning
that variations among staff but also them as an entirety were captured.
The former was linked to variations in music preferences (e.g. Cassidy
and MacDonald, 2007) and the latter to help understanding the inu-
ence and rationales of the employeesinuence on the in-store music. In
the step comparing the individual interviews, codes were theorized
(Pratt, 2009) in steps of reducing empirical codes in the comparisons
among the interviews and thereafter in relation to previous research.
The integration of empirical ndings from the interviews with previous
research was made to ensure the theoretical gap and how the interviews
helped to address that gap.
4. Empirical model and descriptive analysis
We investigate whether the opportunity for employees to inuence
in-store music affected sales by estimating four different models. First,
we estimate the following basic difference-in-differences model:
ln Sales
) +
where ln Sales
is daily sales in store i on day t, expressed in natural
logarithms. There are two reasons for expressing the outcome variable in
logarithmic form. First, the variable becomes approximately normally
distributed, which is good for drawing statistical inferences. Second, it
yields a semi-elastic model in which the estimated treatment effects can
be interpreted as percentage changes.
is an indicator variable equal
to one during the treatment period and zero otherwise; and tg
is an
indicator variable equal to one for stores belonging to the treatment
group and zero otherwise. The primary variable of interest in equation
(1) is the interaction term (tp
). The parameter β
measures the
percentage change in sales in the treatment stores when the experiment
is introduced compared to sales in the same stores in the pretreatment
period and stores in the control group over the whole study period. In
applied research, this type of difference-in-differences estimator is one
of the most frequently used tools for evaluating the effects of in-
terventions on the relevant outcome variables (Abadie, 2005).
We also estimate more general models that include rm-level xed
effects, time-level xed effects, and both. In other words, we now also
control for possible store- or time-level heterogeneity within the treat-
ment and control groups and/or within the control or intervention
period. The most general model that we estimate is as follows:
ln Sales
) +
are store-specic xed effects, and
are date-specic xed
effects. The store-specic xed effects control for possible heterogeneity
among stores in both the treatment and control groups that could affect
sales. Such variables include store-specic management skills, store
location, and opening hours, given that they are (at least roughly)
constant during the study period. The date-specic xed effects control
for all day-to-day heterogeneity, which could affect the level of sales.
Examples include not only payday effects, weather, and chain-specic
marketing campaigns or sales events but also long-term changes in
business climate, such as recessions.
An unbiased identication of the effects of the experiment using
difference-in-differences models requires that the treatment and control
group stores have parallel trends in the outcome variable in the absence
of treatment. Of course, such trends are impossible to empirically
observe since a store that receives the intervention cannot then also be
observed in the counterfactual state of not having received that inter-
vention. Instead, most researchers plot pre- and post-treatment trends in
the outcome variable to gain an idea of how plausible the identication
assumption is in the case at hand. Such trends (based on a sales index
where the average sales over all included stores during the experiment
period is used as the base) are presented for the pretreatment period
from February 2016 to June 2016 in Fig. 1..
The trends are nearly parallel throughout the rst four months of the
pretreatment period, but sales are at a higher level in the treatment
group stores. In the regressions, this difference is captured by the
treatment group indicator variable or, in our more general models, the
store-specic xed effects, and therefore does not cause any bias in the
estimate of the treatment effect. The last pretreatment month, June
2016, coincides with the summer sale in Swedish clothing stores, and the
sale seems to have had somewhat more of an impact on the treatment
stores than on the control stores, and this pattern is also found for the
winter and summer sales periods during the experiment period.
the sales are chain-wide events, the average impact of these events will
be controlled for by the inclusion of the date-specic xed effects in the
most general regression described in equation (2), while this is not
controlled for in our most basic model presented in equation (1). Note,
however, that our results will be biased upwards if there are remaining
differences in trends between the treatment and control stores that are
not accounted for by the date-specic xed effects in coming periods
had the experiment not been implemented, i.e., we will then be more
likely to observe a positive sales effect that is due to these differences in
trends rather than due to the experiment.
Descriptive statistics regarding sales and the specic variables
included in the estimation of equations (1) and (2) are presented in
Table 1. Because the retail chain wants to keep its sales information
condential, we are not allowed to present average revenues. Therefore,
in the descriptive statistics, the means have been converted into indexes
with the mean of all stores over the whole study period as a base being
equal to 100. All means are thus presented in comparison to this base
value, and it is easy to see the differences between the intervention and
Fig. 1. Trends of the clothing sales index.
The exact effect in percentage terms of a parameter estimate β
can be
calculated using formula 100 ×[exp(β
) 1]. However, since the parameter
estimates in our setting are small, the differences are negligible.
Since we also perform the estimations when the full dataset is divided into
sales of womens clothing and mens clothing, the graphs for these datasets are
presented in Figures A1:1 and A1:2 in Appendix 1. Once again, the graphs show
that for most of the pre-treatment period, the trends are clearly parallel; once
again, however, the impact of the summer sale in June seems to be more
pronounced for the treatment group stores than for the control group stores.
Note also that when we analyse demand for female and male clothing sepa-
rately, the number of observations will have the following pattern when
compared to estimations of the full sample: Store-days with non-zero total
demand (4,626) >store days with non-zero sales for males (3,511) or females
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
control stores and before and after the introduction of the experiment.
We do not present any information that would make it possible to infer
average sales levels in the tables or graphs. For the same reason, the
constants are not presented in the tables that report our regression
The average sales in the stores included in the dataset are fairly
constant over time. During the pretreatment period, the average sales
index was 99.86, while in the posttreatment period, the index was
101.95. The differences are more pronounced in regard to average sales
in the treatment and control group stores, with an average sales index
for the treatment group stores of 132.57 and an average sales index for
the control group stores of 75.29. Thus, these results are similar to the
numbers presented in Fig. 1, which also clearly indicate a higher level of
sales in the treatment group stores.
5. Effects of letting employees inuence in-store music
The estimation results regarding the effects on sales of letting em-
ployees inuence in-store music are presented in Table 2. Model 1
corresponds to equation (1), while model 4 corresponds to equation (2).
The main difference between these models is what type of heterogeneity
in the data that they control for. Model 1 only control for average dif-
ferences in sales between treatment and control group rms, and
average differences in sales in the treatment and control periods. Model
4 also controls for average differences in sales between stores within
each group using store specic xed effects, as well as average day-to-
day differences in sales using date specic xed effects. To select be-
tween these models, we use both R-squared and the Akaike information
criteria. Both the goodness of t measure and the Akaike information
criteria indicate that model 4 is preferred irrespectively of if the esti-
mation is for the full sample, or for female or male clothing separately.
As such, the results from the estimations of model 4 are those discussed
in the text unless otherwise mentioned.
Studying total sales, regardless of whether the goods sold are
womens or mens clothing, we nd that the experiment reduced sales in
the treatment stores by 6%. This effect is driven by a reduction in sales of
womens clothing of 11%; we nd no signicant effects of the experi-
ment on sales of mens clothing. Thus, giving employees the opportunity
to affect the music in the store reduces sales.
We also perform placebo testing by randomly assigning treatment to
what are in fact nontreatment time periods. These tests are performed in
the following way. First, we randomly assigned two of the ve pre-
treatment months as having been months where treatment was given. By
means of a random draw without replacement, months 3 and 5, i.e.,
April and June, were assigned to be fake treatment months, while
February, March and May remained controls. In the next step, we re-
estimated equation (2), i.e., our most general model. The results are
presented in Table 3, showing that none of the treatment effects in our
placebo estimation are statistically signicant, providing further evi-
dence that our estimated effects are due to the treatment given and not
some random artifact.
Finally, we investigate whether the effect of the music treatment on
sales is roughly constant over time, which we would expect because the
change in the store environment was a one-time shift and not imple-
mented over time. Finding that there is a trend in the effect of the change
Table 1
Descriptive statistics; the data include both female and male clothing sales.
Variable Mean SD Denition
Sales 100.00 94.33 Average sales index per day.
Sales, TP
99.86 97.70 Average sales index per day before the
experimental period.
Sales, TP
101.95 93.04 Average sales index per day during the
experimental period.
Sales, TG
75.53 71.29 Average sales index per day for stores belonging
to the control group.
Sales, TG
132.57 108.38 Average sales index per day for stores belonging
to the intervention group.
0.32 0.46 Interaction between tp and tg. Measures the
effect of the treatment on the treated stores.
0.71 0.45 An indicator variable equal to one during the
treatment period and zero otherwise.
0.45 0.50 An indicator variable equal to one if the store
belongs to the treatment group and zero
Note: The descriptive statistics are based on the nal sample used in the main
regression analysis.
Table 2
Estimation results, dependent variable ln Sales
Total sales Womens clothing sales Mens clothing sales
Model 1 Model 2 Model 3 Model 4 Model 1 Model 2 Model 3 Model 4 Model 1 Model 2 Model 3 Model 4
0.05 0.05 0.05 0.06* 0.10 0.10* 0.11* 0.11** 0.03 0.05 0.04 0.06
(0.05) (0.05) (0.05) (0.04) (0.06) (0.06) (0.06) (0.05) (0.07) (0.06) (0.07) (0.05)
0.05 0.05 0.12*** 0.12*** 0.02 0.04
(0.04) (0.03) (0.05) (0.04) (0.06) (0.04)
0.70*** 0.70*** 0.39*** 0.40*** 0.30*** 0.30***
(0.04) (0.04) (0.05) (0.05) (0.06) (0.06)
TG indicator. Yes No Yes No Yes No Yes No Yes No Yes No
TP indicator. Yes Yes No No Yes Yes No No Yes Yes No No
Store f.e. No Yes No Yes No Yes No Yes No Yes No Yes
Date f.e. No No Yes Yes No No Yes Yes No No Yes Yes
Observations 4626 4626 4626 4626 3665 3665 3665 3665 3511 3511 3511 3511
R-squared 0.13 0.41 0.37 0.65 0.03 0.30 0.32 0.59 0.02 0.34 0.30 0.66
AIC 11532 9803 11112 8449 9272 8098 9015 7202 9878 8483 9738 7502
Note: ** statistically signicant at the 5% level; * statistically signicant at the 10% level. Heteroskedasticity robust standard errors are in parentheses. The TG in-
dicator is an indicator variable equal to one for treatment group stores, while the TP indicator is an indicator variable equal to one for observations during the
treatment period. Store f.e. Indicates that the regression is performed using store-specic xed effects, while Date f.e. Indicates that date-specic xed effects are used.
AIC =Akaike information criterion.
Table 3
Estimation results, placebo tests, dependent variable ln Sales
Variables All Female Male
0.02 0.03 0.002
(0.07) (0.09) (0.08)
Store f.e. Yes Yes Yes
Day f.e. Yes Yes Yes
Observations 1288 1019 969
R-squared 0.58 0.58 0.66
AIC 2443 2132 2102
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
in store environment on sales would also make us question whether the
estimated effects are due to the treatment given. The results when we
estimate the treatment effect after 3, 5 and 8 months, as well for the
whole treatment period, are presented in Table A2 in appendi. The re-
sults conrm that the effects of the treatment are constant over time,
lending further credibility to the notion that the observed effects are due
to the changes made in the store environment.
6. Employeesattitudes toward in-store music
The interviews showed that the employees had very different music
preferences and ideas about what type of music should be played in the
stores. This is based on statements regarding not only the music
preferred in the store but also who should select the music and whether
the music played was related to personal preferences or what would t
best with the store or its customers. Some employees preferred to in-
uence the type of music to be played while still realizing the need for
the music to conceptually t with the store. These employees also stated
how the choice of music affected them in their work, relating the music
to intensity (music that is too slow) and repetition (the same songs being
played several times). Meanwhile, other employees preferred central
(not store-level) decisions on music but were equally opinionated about
the music played in the stores. They were inclined to mentally block the
music and appreciated silence more. Both groups of employees seemed
to prefer instrumental music to music with lyrics, with the group
preferring music with lyrics being more concerned about inappropriate
lyrics in the ears of customers.
In a sense, Filippa Ks music strategy is quite broad, captured by the
keywords elegant, innovative and forward thinking, expressive and
subdued. Once again, this breadth caused opinions about the t of the
music and what the music really stands for. When asked about what ts
with the store, the interviewees seemed to suggest quite a variety of
music, and they criticized some of the current music for not tting with
the store. The interviews seemed to imply that in regard to what ts, the
interviewees tended to express personal preferences: For instance,
someone with a taste in music for jazz would prefer jazz music. Once
again, there was an overall tendency to prefer upbeat, happy music
rather than soft music. Employees with strong opinions on the in-store
music talked about how it affected not only their work mood but also
their treatment of customers: It affects how I act towards the customers,
the energy level. (Interviewee VII).
In terms of brand image, there is also the discussion about whether
the music is primarily aimed at customer preferences or whether it aims
to reect the brand. That is, there may be discrepancies between the
two; customersmusic preferences may be different from the music
strategy of the store, and therefore, this difference may be intentional, as
the store may want to communicate specic values through the music.
The interviewees were concerned about such differences. Specically,
there seemed to be a concern with the broad age group that Filippa K
serves. Here, the individual stores, based on their locations, may well be
directed at or serve somewhat different customer clienteles, primarily
varying by age group. In other words, the interviewees may have had
somewhat different perceptions about what music tted the store and,
more importantly, what music did not.
The experiment above focused on employeesadjustment of music in
the store. The interviews highlight how employeesadjustment of music
was mostly explained in terms of adjusting to customers: Changes to
music followed from the number of customers in the store and the time
of day. As stated by one of the interviewees, On rainy Sundays, then
nonstress music ts better, music that creates that feeling that there is
room for everyone to shop Shopping does not have to be a party every
time (Interviewee IV). As stated by another interviewee in regard to
skipping songs, this happened to adjust the intensity of the music to the
number of customers; a store with many customers required up-tempo
music, which meant that slow songs were skipped: when the song is
too slow, while there are many customers in the store (Interviewee I).
The same interviewee continues, You can raise the volume when the
store is really full. Like today, during the event this morning, we played
really loud, there were lots of people, and it went well. Therefore, the
adjustments are basically about changing to more intense songs or
adjusting the volume up and down, with a higher volume and intensity
being linked to more customers in the store or more hectic times of the
week (see the quotation about Sundays above compared to busy week-
days, when people have less time to relax). Additionally, sales and the
season affected music choices, with sales as with the busy hours thus
being connected to a higher intensity and volume, while the season is
linked more to the type of music.
Adjustments were also made in terms of volume and in the selection
of songs for various parts of the stores. Specically, inappropriate
lyrics were considered critical in regard to the tting rooms, for which
there was a preference for a somewhat higher volume and feel-good
songs. In these cases, the music was adjusted when the employees saw
someone entering the tting room area, and attempts were also focused
on directing speakers so that the volume would be different there.
Furthermore, music was thought to attract people into the store,
leading to adjustments in terms of the selection of songs: When it feels
like the store is too empty, then it is nice to turn on a song to tempt
people to enter the store(Interviewee II). However, as stated by a
different employee, it is more common that the music is adjusted to
those who are in the store than to make people enter: I would say that
with a lot of customers in the store, we increase the volume and the
intensity. You realize that the pace changes, and the customers become
more alert. But, obviously, the ow is already there at that point. We
adjust the music to the ow, rather than choosing upbeat music and then
the customers come(Interviewee X).
In addition to arguments about songs being adjusted to the number
of customers in the store, the time of day and week, the season, sales and
attempts to draw customers into the store, the employees also changed
the volume or songs depending on their own mood and preferences,
which were often related to the intensity being too low or the rhythms
too repetitive for their taste. It affects our mood. In the afternoons,
when you are tired and youve been delivering the entire day and
something comes along that is too AW in style, then it happens that we
skip the song. Or if you feel like ‘I dont want to hear this song again,
then you also skip it. It does not happen frequently (Interviewee VII).
The preferences always seemed to be for a higher volume and a higher
tempo, as stated by one of the interviewees: It is really strange if you
enter a store with very soft music but that is crowded with hundreds of
people. Then, I change. As I said, I never change downwards, only up-
wards(Interviewee III). Similar stories are told by others: There may
be a song that does not t; something that doesnt make you thrive; then,
I skip the song(Interviewee V). Lastly, there also seemed to be a matter
of being liked through choosing songs that customers liked: Sometimes,
when there was a good song, I heard, ‘what a great taste in music, and
then, you feel as though you are credited(Interviewee IX).
While there were those changing songs or volume as a means to
adjust to customers or based on their own preferences, there were also
others who did not. There is thus a divide between employees engaging
in skipping songs and increasing volume and those not adjusting the
music under any circumstance, not for customers or for their own
preferences. This phenomenon links back to how some employees ten-
ded to block the music they did not like, which was their way of reacting
to music, rather than their actual opinions about the music. However,
this behavior also meant that they did not try to change the music to
adjust to customers, and as stated above, they also preferred playlists
that were constructed by someone else.
The interviewees surprisingly suggest how the employees are not
clear about the brand image of the stores and whether music should
ideally be tted to the store concept or the customers. This, in addition
to the inuence of employeesindividual preferences, likelihood of
inuencing the music and ideas on how any such inuence would be
good for the customer, create a much more complex setting than
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
revealed in previous research on in-store music.
7. Discussion
Studies in psychology have shown that individuals often convey who
they are with the music that they listen to and that they also use music
preferences to evaluate other individuals (Rentfrow and Gosling, 2006).
The results of such studies imply that business owners might also use
in-store music to communicate their brand values and therefore increase
sales and customer satisfaction (Beverland et al., 2006). Consequently,
the effect of in-store music on consumer behavior has attracted much
attention in the marketing literature (Hargreaves and North, 1997;
Kellaris, 2008; and Yorkston, 2010).
However, remarkably few studies have taken an interest in how in-
store music is related to employees and whether it is benecial to give
them more discretion regarding the choice of background music in
stores. This lack of research is puzzling considering that music is of such
central importance for many individuals, suggesting that in-store music
might have a major inuence on the well-being of employees which
would again expect to positively affect customers. Increased opportu-
nities for employees to inuence in-store music could, for example,
make them more satised and service oriented, thus increasing customer
satisfaction and sales (e.g., Kumar and Pansari, 2014). On the other
hand, employees might choose music that is based on their own pref-
erences rather than what is optimal for the store, suggesting that store
owners may want to limit the opportunities for employees to inuence
the music played in stores.
Our study is the rst large-scale experiment on whether increased
opportunities for employees to inuence in-store music affect sales.
Most previous studies on the effects of in-store music are based on very
limited samples; in contrast, we have studied eight Filippa K stores in
Stockholm, Sweden, for 78 weeks (56 experiment weeks and 22 pre-
experiment weeks). The stores were randomly selected into a treat-
ment group and a control group, with the employees in the treatment
stores having the opportunity to inuence the in-store music using an
app developed by SYB. The employees in the control stores had no such
The results from our eld experiment are disappointing for those
who believe that employees should have more discretion to choose in-
store music. Sales in the treatment stores decreased by, on average,
6% when employees had increased opportunities to inuence the in-
store music. We also performed placebo tests by randomly assigning
treatment to what were in fact nontreatment time periods. None of the
estimated treatment effects were statistically signicant in this case,
which further strengthens the proposition that the observed negative
effects on sales were due to the increased opportunities for employees to
choose the in-store music. This has many possible explanations. First,
the app for controlling the music is frequently used during the day,
which implies that employees might use the app rather than attend to
their customers. Another possible explanation is that employees use
their discretionary power in a way that is sub-optimal, for example by
playing music that is too loud or choosing high-intensity songs during
the wrong time of day.
We also found that the negative effect of increasing employeesop-
portunities to inuence in-store music was driven by a decrease in sales
of womens clothing, while the demand for mens clothing was unaf-
fected. One possible explanation is that women are more likely to make
impulse purchases than men (Dittmar et al., 1995; Coley and Burgess,
2003), and impulse buying tend to be inuenced by atmospheric cues
such as in-store music (Donovan et al., 1994; Grewal et al., 2003).
Employees are thus more likely to affect the consumption of women
when they take the opportunity to inuence in-store music. Note also
that women tend to prefer slower and softer music, while men tend
prefer louder and faster music (Stipp, 1990), which has been conrmed
in eld experiments by Andersson et al. (2012). They found that the
presence of music had a positive effect on sales, but males responded
more positively to fast-tempo music than females. This is of interest
considering that the interviews revealed that the employees preferred to
play high-intensity songs when they got the opportunity to inuence the
in-store music, suggesting that female customers might have responded
more negatively to employeesmusic choice compared to males.
To gain a deeper understanding of our results, we also interviewed
thirteen employees in the treatment stores. The interviews showed that
the music preferences of the employees were diverse and, in most cases,
not congruent with the brand values of the company. We believe that
this implies an increased risk; that is, employees choose background
music that is not optimal for the store. The interviews also conveyed that
the employees used their opportunity to inuence the in-store music by,
for example, changing the intensity of the songs. Here, we nd that the
employees tended to prefer high-intensity songs at a high volume,
although previous research has shown that high-intensity songs at a high
volume might decrease both the time spent in stores and sales (e.g.,
Garlin and Owen, 2006), especially for women (Andersson et al., 2012).
The employees largely rationalized their choice to affect the music by
pointing out how the music was changed to t customers and was also
adjusted in different parts of the store (such as a higher volume near the
tting rooms). While such rationalization is shown and in many cases
follows the results of previous studies (such as a higher intensity if the
store is more crowded; e.g., Garlin and Owen, 2006; Knoeferle et al.,
2017), the interviews also point out a tendency to avoid low-intensity
music, although such music may have tted the time of day or the
number of customers in the store. Adjusting music to t customer
preferences rather than the store image was also shown in the interviews
and was manifested in statements about how the employees liked when
customers recognized their choice of music, with the risk that the music
choice (adjusting volume, skipping songs, sharing music, and choosing
intensity) may be self-fullling rather than based on what would be best
for the store. While previous research has indicated how familiarity and
preference positively affect benevolence (e.g., Andersson et al., 2012),
the interviews seem to imply how customers should be positively sur-
prised by the chosen music. The interviews indicate how the employees
preferences for up-tempo music and their beliefs regarding customer
preferences guided the music in store, which at times led to different
choices with regard to music that tted the store image or even the time
of day or the number of customers in the store. Based on how sales fell
and from the behaviors indicated by the interviews, it can be assumed
that brand image is more important for sales than employee satisfaction
(Sundstrom and Sundstrom, 1986).
8. Conclusions
Through uniquely focusing on how employees inuence the in-store
music and its effects on sales, we are able to conclude how such inu-
ence, while possibly positively being linked to employee engagement
and job satisfaction (Kumar and Pansari, 2014) and customers buying
(Baker et al., 2012; Garlin and Owen, 2006), has a negative effect on
sales. Through combing an experiment and interviews, we were also
able to shed light on the reasons for such a negative effect, where the
paper points at how the employees inuence on music is not only a
matter of individual preferences, but also of possible misconception
about the music brand image.
8.1. Theoretical contributions
Theoretical contributions are broadly made to the literature on
retailing and specically to research on in-store music. Through juxta-
posing that research with studies on musics effects on the workplace
and employee satisfaction, we are able to portray a complexity of
inuencing effects on customersbuying, which again relates to music
brand image (e.g., Daunfeldt et al., 2017). The added variable of
employee impact to understand in-store music helps to further contex-
tualize consumersin-store behavior related to atmospheric cues (eg.,
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
Bruner, 1990; Bitner, 1992; Helmefalk, 2017), while creating an inter-
action between employees and such cues, where previous studies have
separated atmospheric cues for customers and employees or seen music
and employees as unrelated variables (cf. Bitner, 1992).
8.2. Managerial implications
For managers, it would be important to educate employees about the
brand image, while also limiting employeesinuence on in-store music.
Given how the employees inuence indeed had a negative impact on
sales, advice would be to disable employeesadjustment of music, while
teaching the staff about the impact music volume and tempo would have
on customer buying. Related to how the experiment led to decreased
sales, and how the non-treatment group also was exposed to music
selected to t with the brand, investigating in music that ts the brand,
while creating positive buying signals in terms of volume and tempo,
would be important.
8.3. Limitations and further research
As always in eld experiments, we cannot conclude that the results
would hold under other conditions. The effects when employees have
increased opportunities to choose in-store music might thus be different
in another store environment and with other employees, and any
interpretation of our results outside the specic experiment should be
done with caution. We therefore believe that more research is needed to
validate whether our results also hold in other contexts. There is also a
lack of knowledge in general regarding the interaction effects between
in-store music and the well-being of employees, which we believe con-
stitutes an interesting area for further research. Repeated experiments in
other types of stores (low-cost, other types of items sold, service stores)
and workplaces would be interesting to see whether the relationship
among in-store music, employees and sales varies with the store envi-
ronment. In the experiment, we found differences in the results
depending on whether customers were shopping for mens or womens
clothing. It would be interesting to further study gender aspects,
including in relation to the gender of the employees.
Note nally that our main aim has been to test if increased oppor-
tunities for employees to change in-store music affect sales. The em-
ployees in the treatment stores were therefore given the opportunities to
change the volume of songs, skip songs, share songs with consumers, and
choose between a soundtrack with high-intensity or medium-intensity
songs, while no such opportunities were given to employees in the
control stores. However, this means that we from a strict causality
perspective cannot determine which of these decisions that had a
negative impact on sales. We believe that employees decisions
regarding specic choices of in-store music therefore constitute a fruitful
area for more research.
The authors would like to thank the participants at the 6th Nordic
Retail and Wholesale Conference for their valuable comments. Special
thanks go to Joel Brosj¨
o, Stefan Kragh, Andreas Liffgarden, Magnus
en, and Ola Sars at Soundtrack Your Brand for making this research
possible. We also want to thank Jenny Brostr¨
om and Anna Christidis at
Filippa K, and Teresa Garcia for help with conducting the interviews.
Financial support provided by the Swedish Research and Wholesale
Council (Handelsrådet), grant number 2016:692, is gratefully
Appendix A. Supplementary data
Supplementary data to this article can be found online at
Appendix 1. Pretreatment trends
As mentioned in the main text of our paper, identication in difference-in-differences regression models assumes that intervention and control
group stores would have had similar trends in the outcome variables in the absence of treatment. Of course, such trends are impossible to empirically
observe since a store that receives the intervention cannot then also be observed in the counterfactual state of not having received that intervention.
Instead, most researchers plot pre- and post-treatment trends in the outcome variable to gain an idea of how plausible the identication assumption is
in the case at hand. In the main part of the paper, we presented these trends for all sales, while we present the pre- and post-treatment trends in an
average daily sales index divided into womens and mens clothing below in Figure A1:1 and A1:2.
Fig. A1:1. Trends of the womens clothing sales index.
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
Fig. A1:2. Trends of the mens clothing sales index.
Appendix 2. Timing of the effects
To investigate whether the effects are immediate when the treatment is given and constant over time, we also estimated equation (1) but restricted
the length of the treatment period to continue to 1 October 2016, 1 December 2016, 1 February 2017, and 1 August 2017 (where this last one is the
whole period and where the results thus coincide with those presented in Table 2 in the main text of the paper). As shown by these additional es-
timations in Table A2, the effects appear within three months after the change in store environment and are fairly constant during the period under
Table A2
Estimation results. The treatment period ends at different points in time, the dependent variable ln Sales
Total sales Womens clothing Mens clothing
1 Oct
1 Dec
1 February,
1 August,
1 Oct
1 Dec
1 February,
1 August,
1 Oct
1 Dec
1 February,
1 August,
0.04 0.06 0.05 0.06* 0.12* 0.15*** 0.11** 0.11** 0.09 0.03 0.01 0.06
(0.05) (0.05) (0.04) (0.04) (0.06) (0.06) (0.05) (0.05) (0.07) (0.06) (0.06) (0.05)
Store f.e. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Date f.e. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 2092 2624 3065 4626 1655 2077 2427 3665 1586 1985 2322 3511
R-squared 0.64 0.64 0.65 0.65 0.58 0.57 0.61 0.59 0.64 0.64 0.65 0.63
AIC 3913 4937 5669 8449 3372 4240 4835 7202 3401 4319 4943 7502
Note: *** statistically signicant at the 1% level; ** statistically signicant at the 5% level; * statistically signicant at the 10% level. Heteroskedasticity robust
standard errors are in parentheses. Store f.e. Indicates that the regression is performed using store-specic xed effects, while Date f.e. Indicates that date-specic xed
effects are used.
Appendix 3. Interview guide
1. Gender
2. Age
3. What is your role, and how long have you worked here?
4. Describe how you listen to music as a private person. What type of music? When? How often? What is the role of music in your everyday life?
The store atmosphere.
1. Describe your thoughts on the music that has been played in this store since July last year to todays date.
2. How do you think the music matches the Filippa K brand? Why? Why not?
3. What type of music would you say matches the Filippa K brand? Why?
4. Do you consider that music at work affects your motivation and productivity? Describe how.
5. Have you heard any feedback from customers regarding the music? If yes, provide examples.
6. How important do you think the music is in affecting customers? In what way?
7. If you could inuence/change one thing about the music played, what would it be?
The music.
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
1. How would you describe your ability to affect your work place generally? Does the staff remote have a role in this? How?
2. Do you feel that you are/have been able to inuence the music in this store?
3. Would you like to be able to inuence the music in your work place? If yes, why? Additionally, how would you want to do it? If no, why not?
Spotify remote.
1. You have been able to inuence the music in this store by changing songs and adjusting the intensity and the volume by means of the Spotify
Remote on the iPad. Could you tell us your experiences with this?
2. In what situations did you use the staff remote to skip songs? Give a few examples. Why? What did you want to accomplish? (Did you dislike the
song or think the music was repetitive?) What exactly did you do? What was the effect?
3. In what situations did you use the staff remote to increase/decrease the music intensity? Give a few examples. Why? What did you want to
accomplish? What exactly did you do? What was the effect?
4. In what situations did you use the staff remote to increase/decrease the volume? Give a few examples. Why? What did you want to accomplish?
What exactly did you do? What was the effect?
5. Did you experience the music to be repetitive at any point? Please discuss. How did it affect you? How did it affect the customers? Did you do
something about the situation? What?
6. How did you feel about different music intensities? What ts best with the store/brand? What have you liked the best? What type of intensity do
you think best matches the store/Filippa K?
7. Did you nd the staff remote easy to use? If yes, why? If no, why not?
8. Do you think that the staff remote was important for your ability to affect the music in the store? If yes, how and why? If no, how and why not?
9. Do you consider the staff remote to be a useful tool for affecting the music in the store? If yes, how and why? If no, how and why not?
10. What do you think about the design of the staff remote? I like it, because (please tell why). I do not like it because (please tell why not).
11. What do you think about the existing features of the staff remote?
12. If you had been able to choose freely, would this service/product ease your work? Energy level, volume, music style, inuence favorites and
nonfavorites, play a song less/more
13. Do you have additional feedback on the staff remote?
Abadie, A., 2005. Semiparametric difference-in-differences estimators. Rev. Econ. Stud.
72 (1), 119.
Andersson, P.K., Kristensson, P., W¨
astlund, E., Gustafsson, A., 2012. Let the music play or
not: the inuence of background music on consumer behavior. J. Retailing Consum.
Serv. 19 (6), 553560.
Areni, C.S., Kim, D., 1993. The inuence of background music on shopping behavior:
classical versus top-forty music in a wine store. Adv. Consum. Res. 20 (1), 336340.
Baker, J., Parasuraman, A., Grewal, D., Voss, G.B., 2002. The inuence of multiple store
environment cues on perceived merchandise value and patronage intentions.
J. Market. 66, 120141.
Bernstein, E.S., Turban, S., 2018. The impact of the openworkspace on human
collaboration. Phil. Trans. Biol. Sci. 373 (1753), 20170239.
Beverland, M., Lim, E.A.C., Morrison, M., Terziovski, M., 2006. In-store music and
consumerbrand relationships: relational transformation following experiences of
(mis) t. J. Bus. Res. 59 (9), 982989.
Biswas, D., Labrecque, L.I., Lehmann, D.R., Markos, E., 2014. Making choices while
smelling, tasting, and listening: the role of sensory (dis) similarity when sequentially
sampling products. J. Market. 78 (1), 112126.
Biswas, D., Lund, K., Szocs, C., 2019. Sounds like a healthy retail atmospheric strategy:
effects of ambient music and background noise on food sales. J. Acad. Market. Sci. 47
(1), 3755.
Bitner, M.J., 1992. Servicescapes: the impact of physical surroundings on customers and
employees. J. Market. 56 (2), 5771.
Bruner, G.C., 1990. Music, mood, and marketing. J. Market. 54 (4), 94104.
Cassidy, G., MacDonald, R.A., 2007. The effect of background music and background
noise on the task performance of introverts and extraverts. Psychol. Music 35 (3),
Coley, A., Burgess, B., 2003. Gender differences in cognitive and affective impulse
buying. J. Fash. Mark. Manag. 7 (3), 282295.
Daunfeldt, S.-O., Rudholm, N., Poppius, H., 2017. Effects of Brand-Fit Music on
Consumer Behavior: A Field Experiment. HUI Working Paper No.121. HUI Research,
Dittmar, H., Beattie, J., Friese, S., 1995. Gender identity and material symbols: objects
and decision considerations in impulse purchases. J. Econ. Psychol. 16 (3), 491511.
Donovan, R.J., Rossiter, J.R., Marcoolyn, G., Nesdale, A., 1994. Store atmosphere and
purchasing behavior. J. Retailing 70 (3), 283294.
Duo, E., Glennerster, R., Kremer, M., 2007. Using randomization in development
economics research: a toolkit. Handb. Dev. Econ. 4, 38953962.
Furnham, A., Bradley, A., 1997. Music while you work: the differential distraction of
background music on the cognitive test performance of introverts and extraverts.
Appl. Cognit. Psychol.: The Ofcial Journal of the Society for Applied Research in
Memory and Cognition 11 (5), 445455.
Garlin, F.V., Owen, K., 2006. Setting the tone with the tune: a meta-analytic review of the
effects of background music in retail settings. J. Bus. Res. 59 (6), 755764.
Grewal, D., Baker, J., Levy, M., Voss, G.B., 2003. The effects of wait expectations and
store atmosphere evaluations on patronage intentions in service-intensive retail
stores. J. Retailing 79 (4), 259268.
Hargreaves, D.J., North, A.C., 1997. The social psychology of music. In: I Hargreaves, D.
J., North, A.C. (Eds.), The Social Psychology of Music. Oxford University Press,
pp. 124.
Helmefalk, M., 2017. Multi-sensory Cues in Interplay and Congruency in a Retail Store
Context: Consumer Emotions and Purchase Behaviors. Doctoral, p. V¨
Hui, M.K., Dub´
e, L., Chebat, J.C., 1997. The impact of music on consumersreactions to
waiting for services. J. Retailing 73 (1), 87104.
Hynes, N., Manson, S., 2016. The sound of silence: why music in supermarkets is just a
distraction. J. Retailing Consum. Serv. 28, 171178.
Jacob, C., 2006. Styles of background music and consumption in a bar: an empirical
evaluation. Int. J. Hospit. Manag. 25 (4), 716720.
Kellaris, J.J., 2008. Music and consumers. In: Haugtvedt, C.P., Herr, P.M., Kardes, F.R.
(Eds.), Handbook of Consumer Psychology (S. 837856). Taylor & Francis Group/
Lawrence Erlbaum Associates, New York.
Knifn, K.M., Yan, J., Wansink, B., Schulze, W.D., 2017. The sound of cooperation:
musical inuences on cooperative behavior. J. Organ. Behav. 38 (3), 372390.
Knoeferle, K.M., Paus, V.C., Vossen, A., 2017. An upbeat crowd: fast in-store music
alleviates negative effects of high social density on customersspending. J. Retailing
93 (4), 541549.
Korczynski, M., 2003. Music at work: towards a historical overview. Folk Music J. 8,
Krueger, A.B., 1999. Experimental estimates of education production functions. Q. J.
Econ. 114 (2), 497532.
Kumar, V., Pansari, A., 2014. The construct, measurement, and impact of employee
engagement: a marketing perspective. Customer Needs & Solutions 1, 5267.
Levy, M., Grewal, D., 1993. An experimental approach to making retail store
environmental decisions. J. Retailing 68 (4), 445460.
MacInnis, D.J., Park, C.W., 1991. The differential role of characteristics of music on high-
and low-involvement consumersprocessing of ads. J. Consum. Res. 18 (2),
Mari, M., Poggesi, S., 2013. Servicescape cues and customer behavior: a systematic
literature review and research agenda. Serv. Ind. J. 33 (2), 171199.
Mattila, A., Wirtz, J., 2001. Congruency of scent and music as a driver of in-store
evaluations and behavior. J. Retailing 77 (2), 273289.
Michel, A., Baumann, C., Gayer, L., 2017. Thank you for the musicor not? The effects of
in-store music in service settings. J. Retailing Consum. Serv. 36, 2132.
Milliman, R., 1982. Using background music to affect the behavior of supermarket
shoppers. J. Market. 46 (3), 8691.
Milliman, R., 1986. The inuence of background music on the behavior of restaurant
patrons. J. Consum. Res. 13 (2), 286289.
Morin, S., Dube, L., Chebat, J.C., 2007. The role of pleasant music in servicescapes: a test
of the dual model of environmental perception. J. Retailing 83 (1), 115130.
Morrison, M., Gan, S., Dubelaar, C., Oppewal, H., 2011. In-store music and aroma
inuences on shopper behavior and satisfaction. J. Bus. Res. 64 (6), 558564.
S.-O. Daunfeldt et al.
Journal of Retailing and Consumer Services xxx (xxxx) xxx
North, A.C., Hargreaves, D.J., 1996. The effects of music on responses to a dining area.
J. Environ. Psychol. 16 (1), 5564.
North, A.C., Hargreaves, D.J., McKendrick, J., 1999. The inuence of in-store music on
wine selections. J. Appl. Psychol. 84 (2), 271.
North, A.C., Sheridan, L.P., Areni, C.S., 2016. Music congruity effects on product
memory, perception, and choice. J. Retailing 92 (1), 8395.
Oakes, S., 2003. Musical tempo and waiting perceptions. Psychol. Market. 20 (8),
Otterbring, T., Pareigis, J., W¨
astlund, E., Makrygiannis, A., Lindstr¨
om, A., 2018. The
relationship between ofce type and job satisfaction: testing a multiple mediation
model through ease of interaction and well-being. Scand. J. Work. Environ. Health
44 (3), 330334.
Otterbring, T., Bodin Danielsson, C., Pareigis, J., 2020. Ofce types and workers
cognitive vs. affective evaluations from a noise perspective. J. Manag. Psychol. (in
Pratt, M.G., 2009. From the editors: for the lack of boilerplate: tips on writing up (and
reviewing) qualitative research. Acad. Manag. J. 52 (5), 856862.
Rentfrow, P.J., Gosling, S.D., 2006. Message in a ballad: the role of music preferences in
interpersonal perception. Psychol. Sci. 17 (3), 236242.
Roschk, H., Loureiro, S.M.C., Breitsohl, J., 2017. Calibrating 30 years of experimental
research: a meta-analysis of the atmospheric effects of music, scent, and color.
J. Retailing 93 (2), 228240.
Sarantakos, S., 1998. Social Research, second ed. Palgrave Macmillan, New York.
Shih, Y.N., Huang, R.H., Chiang, H.Y., 2012. Background music: effects on attention
performance. Work 42 (4), 573578.
Skandrani, H., Dahmane Mouelhi, N.B., Malek, F., 2011. Effect of store atmospherics on
employeesreactions. Int. J. Retail Distrib. Manag. 39 (1), 5167.
Stipp, H., 1990. Musical demographics. Am. Demogr. 12 (8), 4849.
Sundstrom, E., Sundstrom, M.G., 1986. Work Places: the Psychology of the Physical
Environment in Ofces and Factories. Cambridge University Press, Cambridge.
Turley, L., Milliman, R., 2000. Atmospheric effects on shopping behavior: a review of the
experimental evidence. J. Bus. Res. 49 (2), 193211.
Wilson, S., 2003. The effect of music on perceived atmosphere and purchase intentions in
a restaurant. Psychol. Music 31 (1), 93112.
Yalch, R., Spangenberg, E., 1990. Effects of store music on shopping behavior.
J. Consum. Market. 7 (2), 5563.
Yorkston, E., 2010. Auxilary auditory ambitions: assessing ancillary and ambient sounds.
In: Khrisna, A. (Ed.), Sensory Marketing: Research in the Sensuality of Products.
Taylor & Francis, New York, pp. 157167.
S.-O. Daunfeldt et al.
... Third, we extended the effects of working conditions to brand performance measured by sales. Sales are typically time-variant and thus, for a detailed investigation, require panel data in order to control for non-treated stores and pre-treatment sales (Daunfeldt et al., 2021). Finally, we were able to observe the employees' in-store performance over time, thereby augmenting the confidence in the stability of the observed results (Gong et al., 2017). ...
... Brand performance: Our main hypothesis posits a positive effect of working conditions on sales, and we relied on the well-known difference-in-difference technique to assess the treatment effect (e.g., Bommaraju & Hohenberg, 2018;Daunfeldt et al., 2021). For this technique to be effective, it is necessary to observe a sample of treated and untreated units over a time prior to and during the treatment. ...
Through an experimental field study, we link improved working conditions for supportive jobs to enhanced brand sales. Results show that improved working conditions for in-store merchandisers based on monetary and non-monetary rewards yield a significant increase in sales, which might be facilitated through enhanced on-shelf availability. We underline the strategic importance of the job profile and identify explanatory mechanisms, enabling managers to optimize training and compensation in the context of supportive jobs in retailing. The study enhances the literature on the importance of supportive jobs in driving brand performance in accordance with the motivational framework based on rewards and recognition.
... As a result, in a corporation or organization, employee performance is critical to accomplishing organizational or company goals. (Daunfeldt et al., 2021;Fu et al., 2021;Wibowo et al., 2020) Business developments in the field of package delivery services and logistics services in Indonesia are currently experiencing a significant increase, this is evidenced by the increase in the ranking of the 2018 Indonesia Logistics Performance Index (LPI) Survey report, which rose 17 ranks, from the previous 63rd rank, to 46th rank in this year. This happens because more and more Indonesian people use package delivery services or delivery of documents to companies, it is proven that every year companies engaged in these services are increasing and increasing. ...
Full-text available
This study aims to analyze Job Satisfaction as a Mediator in Improving Employee Performance through Talent and Knowledge Management. The data analysis used is verification analysis. The population is focused on 16 logistics service companies in the city of Bandung. The analytical tool used is the Structural Equation Model (SEM)-PLS, data obtained through questionnaires to 160 employees. The findings reveal that talent management and knowledge management make a positive contribution to job satisfaction, either partially or simultaneously, which has an impact on employee performance. This is because logistics services will always innovate in accordance with good knowledge management and talent management, therefore employees are required to express their thoughts, ideas and abilities to achieve all of this. This research adds to knowledge about how logistics service companies can manage talented employees and knowledge management that can lead to increased job satisfaction which has an impact on employee performance.
... The identification of responsible entrepreneurship derives in part from the friendly relationship established between companies, especially new and small ones, in the environment in which they operate [46,47]. The underlying philosophy is none other than the consideration of the business system as part of a broader system than the social system in which it is particularly immersed [48]. ...
Full-text available
One of the particularities of companies with a social purpose is that, through their business model of B companies, they have incorporated into their processes the necessary mechanisms to obtain, simultaneously, the profits to ensure the existence of the organization in the market. At the same time, social value is generated, which is necessary to address the problems of the social crisis caused by COVID-19 and the environmental problems affecting the community. The current global health and economic crisis has opened up the possibility of adopting business model B and focusing more on the individual. Based on the grounded theory method, we have examined 3500 B Corporations in Latin America, of which 57 were examined in 10 countries listed in the Directory of B Corporations for Latin America. The main conclusions are that B Corporations dedicated to tourism through responsible entrepreneurship develop a more inclusive, sustainable and environmentally friendly economy for the benefit of society, go beyond the notion of CSR and move away from traditional business, as B Corporations combine social development and economic growth.
This research seeks to verify whether perceived crowding, perceived service climate, positive customer-to-customer (C2C) interactions, and dysfunctional customer behavior influence the perception of experiential consumption value in self-service restaurants. We surveyed by applying a questionnaire to 337 consumers. The results showed that the perceived service climate and positive C2C interactions could influence the experiential value of consumption in a self-service restaurant. We found that perceived crowding can influence positive C2C interactions and dysfunctional customer behavior. The results also demonstrate that positive C2C interactions can mediate the relationship between the perceived service climate and the experiential value of consumption.
Det är sedan länge känt att bakgrundsmusik kan påverka hur konsumenterna upplever och agerar på en marknadsplats. Vi vet däremot väldigt lite om personalens upplevelser av bakgrundsmusiken, samt vad effekterna blir om personalen i högre utsträckning kan påverka den musik som spelas i butiken. Det övergripande syftet med detta projekt har varit att studera personalens upplevelser av bakgrundsmusik i detaljhandeln, samt om försäljningen påverkas av att personalen får möjligheter att påverka bakgrundsmusiken i butiken.
Full-text available
Purpose: This study examined the links between office types (cellular, shared-room, small and medium-sized open-plan) and employees’ subjective well-being regarding cognitive and affective evaluations, and the role perceived noise levels at work has on the aforementioned associations. Design/methodology/approach: A survey with measures of office types, perceived noise levels at work, and the investigated facets of subjective well-being (cognitive vs. affective) was distributed to employees working as real estate agents in Sweden. In total, 271 usable surveys were returned and were analyzed using analyses of variance (ANOVAs) and a regression-based model mirroring a test of moderated mediation. Findings: A significant difference was found between office types on the well-being dimension related to cognitive, but not affective, evaluations. Employees working in cellular and shared-room offices reported significantly higher ratings on this dimension than employees working in open-plan offices, and employees in medium-sized open-plan offices reported significantly lower cognitive evaluation scores than employees working in all other office types. This pattern of results was mediated by perceived noise levels at work, with employees in open-plan (vs. cellular and shared-room) offices reporting less satisfactory noise perceptions and, in turn, lower well-being scores, especially regarding the cognitive (vs. affective) dimension. Originality/value: This is one of the first studies to compare the relative impact of office types on both cognitive and affective well-being dimensions, while simultaneously testing and providing empirical support for the presumed process explaining the link between such aspects. Keywords: office type; cellular office; shared-room office; open-plan office; noise; subjective well-being; cognitive evaluation; affective evaluation; positive activation; negative deactivation.
Full-text available
Organizations’ pursuit of increased workplace collaboration has led managers to transform traditional office spaces into ‘open’, transparency-enhancing architectures with fewer walls, doors and other spatial boundaries, yet there is scant direct empirical research on how human interaction patterns change as a result of these architectural changes. In two intervention-based field studies of corporate headquarters transitioning to more open office spaces, we empirically examined—using digital data from advanced wearable devices and from electronic communication servers—the effect of open office architectures on employees' face-to-face, email and instant messaging (IM) interaction patterns. Contrary to common belief, the volume of face-to-face interaction decreased significantly (approx. 70%) in both cases, with an associated increase in electronic interaction. In short, rather than prompting increasingly vibrant face-to-face collaboration, open architecture appeared to trigger a natural human response to socially withdraw from officemates and interact instead over email and IM. This is the first study to empirically measure both face-to-face and electronic interaction before and after the adoption of open office architecture. The results inform our understanding of the impact on human behaviour of workspaces that trend towards fewer spatial boundaries. This article is part of the theme issue ‘Interdisciplinary approaches for uncovering the impacts of architecture on collective behaviour’.
Full-text available
Retail atmospherics is becoming an increasingly important strategic tool for stores and restaurants. Ambient music and background noise are especially important atmospheric elements given their ubiquity in retail settings. However, there is high variation in the volume of ambient music and background noise, with some stores/restaurants having very loud ambience and others having very quiet ambience. Given the variation in loudness levels at stores/restaurants, and the managerial ease of adjusting volume level, we investigate the consequences of ambient music (and background noise) volume on food choices and sales. A pilot study, two field experiments, and five lab studies show that low (vs. high or no) volume music/noise leads to increased sales of healthy foods due to induced relaxation. In contrast, high volume music/noise tends to enhance excitement levels, which in turn leads to unhealthy food choices.
Full-text available
Objectives: This cross-sectional study investigated the associations between office type (cellular office, shared-room office, small open-plan office, and medium-sized open-plan office) and employees' ease of interaction with co-workers, subjective wellbeing, and job satisfaction. Methods: A brief survey including measures of office type, ease of interaction with co-workers, subjective wellbeing, and job satisfaction was sent electronically to 1500 Swedish real-estate agents, 271 of whom returned usable surveys. The data were analyzed using a regression-based serial multiple mediation model (PROCESS Model 6), which tested whether the relationship between office type and job satisfaction would be mediated by ease of interaction and, in turn, subjective wellbeing. Results: A negative relationship was found between the number of co-workers sharing an office and employees' job satisfaction. This association was serially mediated by ease of interaction with co-workers and subjective wellbeing, with employees working in small and medium-sized open-plan offices reporting lower levels of both these aspects than employees who work in either cellular or shared-room offices. Conclusions: Open-plan offices may have short-term financial benefits, but these benefits may be lower than the costs associated with decreased job satisfaction and wellbeing. Therefore, decision-makers should consider the impact of office type on employees rather than focusing solely on cost-effective office layout, flexibility, and productivity.
Full-text available
Research suggests that in-store crowding can lower customers’ spending, thus limiting overall benefits of high store frequentation. Here, we propose that this negative effect can be mitigated by adjusting store ambiance, specifically by using certain types of in-store music. To test this idea, we conducted a longitudinal field experiment in which we manipulated in-store music tempo and measured social density in six European retail stores. Analyzing over 40,000 individual shopping baskets, we found that social density had an inverted u-shape effect on customer spending. This effect was moderated by in-store music tempo, such that fast music strongly increased spending under high-density conditions. The increase in shopping basket value was driven by customers buying more items rather than buying items that were more expensive. Fast music thus alleviated negative effects of social density. We discuss the theoretical implications of these findings and describe how practitioners can use in-store music to counter negative effects of high customer density.
A typology of service organizations is presented and a conceptual framework is advanced for exploring the impact of physical surroundings on the behaviors of both customers and employees. The ability of the physical surroundings to facilitate achievement of organizational as well as marketing goals is explored. Literature from diverse disciplines provides theoretical grounding for the framework, which serves as a base for focused propositions. By examining the multiple strategic roles that physical surroundings can exert in service organizations, the author highlights key managerial and research implications.
This paper critically reviews the literature available and presents an empirical study that examines the effects of background music on in-store shopping behavior. It finds that music tempo variations can significantly affect the pace of in-store traffic flow and dollar sales volume.
While research has shown the positive impact of sensory cues and cue- congruency on emotion and behavior in retail store atmospheres, these cues have primarily been investigated in isolation or in pairs. Consequently, little is known on how multi-sensory cues in interplay impact on consumer emotions and purchase behaviors. In addition, research has not yet provided any clear conceptualization of congruency in marketing when designing retail store atmospheres, other than stating that some cues are expected to match, therefore become pleasantly perceived. Thus, the main purpose of this research is to examine and show how multi-sensory cues in interplay and congruency can be utilized in creating a retail store atmosphere to enhance consumer emotions and purchase behaviors. To address the purpose, a sequential method was adopted with four essays. The first essay explores multi-sensory interplay in marketing contexts with a literature review that forms the basis for a research agenda. The second essay employs focus groups to highlight the congruency between cues, products and the retail setting, and identifies which category of cues is in need of investigation. The third essay uses field experiments to investigate two congruent visual, auditory and olfactory cues (six cues in total) in a retail setting, and their impact on consumer emotion and purchase behavior. The final essay, also use field experiments to examine and duplicate one cue from each sense, and employs these together in interplay, to show how multi-sensory cues in interplay impacts emotions and purchase behaviors. This research concludes that multi-sensory cues in interplay in a retail store atmosphere have a greater impact on consumer emotions and purchase behaviors than single visual, auditory and olfactory setting-congruent sensory cues. Among single sensory cues, those perceived as complementary in the atmosphere, specifically auditory and olfactory in an already visual dominated atmosphere, have the largest impact on consumer emotions and purchase behaviors. Overall, this research signifies that congruent multi-sensory cues in interplay emerge as reliable predictors for the influence on consumer arousal, valence, time spent, touching, browsing and purchasing Theoretical and managerial implications are discussed.
The current study looked at the distracting effects of ‘pop music’ on introverts' and extraverts' performance on various cognitive tasks. It was predicted that there would be a main effect for music and an interaction effect with introverts performing less well in the presence of music than extraverts. Ten introverts and ten extraverts were given two tests (a memory test with immediate and delayed recall and a reading comprehension test), which were completed, either while being exposed to pop music, or in silence. The results showed that there was a detrimental effect on immediate recall on the memory test for both groups when music was played, and two of the three interactions were significant. After a 6-minute interval the introverts who had memorized the objects in the presence of the pop music had a significantly lower recall than the extraverts in the same condition and the introverts who had observed them in silence. The introverts who completed a reading comprehension task when music was being played also performed significantly less well than these two groups. These findings have implications for the study habits of introverts when needing to retain or process complex information. © 1997 John Wiley & Sons, Ltd.