An integrated model of customer loyalty in the Macedonian mobile service market
Abstract
The main objective of this paper is to develop and test an integrated model of customer loyalty in the Macedonian mobile service market, analysing its determinants and their interrelationships and effects on customer loyalty. The proposed research model integrates service quality, customer satisfaction, switching costs and switching barriers as the main determinants of customer loyalty, further analysing their different effects (direct/mediating/moderating) on customer loyalty. Confirmatory factor analysis (CFA) is used in order to assess the validity of the measurement model, and the developed structural model is tested with structural equation modelling (SEM). The reported results indicate that the proposed research model as a whole has a satisfactory level of adequacy and most of the hypothesized realtionships are confirmed. The results suggest that customer satisfaction partially mediates the relationship between service quality and customer loyalty and also service quality has a direct and positive effect on customer loyalty. The satisfaction-loyalty relationship is not mediated nor moderated by switching barriers, but switching barriers have a direct and positive effect on customer loyalty. The presented results would be of great importance for managers for the successful development of strategies for creating mobile service users’ loyalty. The empirical evidence obtained in this study suggest that Macedonian mobile operators should have in mind alternative strategies in creating customer loyalty, emphasising the fact that customer satisfaction has a stronger positive influence in creating customer loyalty than switching barriers. The present research provides an integrated approach in examining customer loyalty and explores the different roles of switching barriers in the relationship between customer satisfaction and loyalty. In order to further develop the integrated approach of customer loyalty modelling, some additional concepts should be incorporated and examined in the future.
Full text
199
2, XX, 2017
Ma ke ing and T ade
DOI: 10.15240/ ul/001/2017-2-015
In oduc ion
In oday’s compe i i e en i onmen , c ea ing
and main aining cus ome sa is ac ion and
consequen ly cus ome loyal y is o i al
impo ance o he companies. A he same
ime, i is a challenging ask because li le is
known abou how cus ome s di e in hei
ela ional p e e ences (Palma ie e al., 2006)
and because he e a e a lo o a iables ha
in l uence cus ome loyal y. Al hough cus ome s
end o be mo e demanding and less loyal
(Reichheld, 1996), se ice ma ke e s a e
con inced ha building endu ing ela ionships
wi h he cus ome s is inc edibly impo an
because loyal cus ome s a e less cos ly o
se e, buy mo e, pay p emium p ices and
sp ead posi i e wo d-o -mou h (Zei haml, 2000;
O’B ian & Jones, 1995).
Among he a ious ac o s ha di ec ly o
indi ec ly in l uence cus ome loyal y, pe cei ed
se ice quali y and cus ome sa is ac ion
a e equen ly iewed as key de e minan s
o cus ome loyal y (Vildo á e al., 2015;
Se ó-Pamies, 2012; C onin e al., 2000) wi h
a con i med media ing e ec o cus ome
sa is ac ion in he se ice quali y-cus ome
loyal y ela ionship (S i as a a & Rai, 2013;
Ca uana, 2002). Addi ionally, swi ching cos s
and swi ching ba ie s a e also iewed as key
de e minan s o cus ome loyal y (de Ma os
e al., 2009; Lee e al., 2001). Namely, when
swi ching cos s a e high, cus ome s may
s ay loyal e en i hey a e no highly sa is i ed
(San ou idis & T i ellas, 2010). Al hough
he esea ch in e es on pe cei ed se ice
quali y, cus ome sa is ac ion, swi ching cos s,
swi ching ba ie s and cus ome loyal y has
been inc easing egula ly, he numbe o
esea ch s udies in eg a ing all hese a iables
in explo ing cus ome loyal y is a he sca ce.
Recen ly, he e has been an exponen ial
inc ease in esea ch on se ice quali y,
cus ome sa is ac ion and cus ome loyal y
in he mobile se ice indus y. Speci i cally,
a numbe o esea che s ha e a emp ed
o e eal he ela ionship among a ious
independen a iables and cus ome loyal y
as a dependen a iable in he mobile se ice
indus y (San ou idis & T i ellas, 2010;
Ge po e al., 2001). E en hough all hese
esea ch e o s ha e ad anced ma ke e s’
unde s anding o mobile se ice consume
beha iou and pa icula ly o he ela ionships
among pe cei ed se ice quali y, cus ome
sa is ac ion, and cus ome loyal y (Lee, 2013),
ye an in eg a ed model o loyal y in he mobile
se ice con ex wi h swi ching ba ie s and
swi ching cos s as supplemen a y a iables is
s ill a e. This s udy s i es o i ll his esea ch gap
by p oposing and es ing an in eg a ed model o
cus ome loyal y. Mo eo e , he di e en e ec s
o swi ching cos s and swi ching ba ie s on he
sa is ac ion-loyal y ela ionship a e examined,
since he e is a lack o empi ical s udies es ing
hese app oaches simul aneously (de Ma os e
al., 2009).
The esea ch con ex o ou s udy is he
mobile se ice indus y in Macedonia. A p esen ,
h ee mobile se ice p o ide s and one mobile
i ual ne wo k ope a o ope a e in Macedonia,
gi ing use s a choice when conside ing hei
mobile communica ions needs. Wi h a numbe
o ac i e mobile se ice use s highe han he
popula ion, and a low mobile se ice use s g ow h
a e o only 1.05% in he hi d qua e o 2014
compa ed o he hi d qua e o 2013 (Agency
o Elec onic Communica ion o he Republic
o Macedonia, 2014), he Macedonian mobile
se ice ma ke is eaching i s ma u i y. In such
condi ions, he mobile ope a o s a e ocusing
hei ma ke ing e o s on e aining he cu en
cus ome s h ough c ea ing cus ome loyal y. I is
gene ally accep ed ha a high le el o cus ome
sa is ac ion will esul in high le el o cus ome
AN INTEGRATED MODEL OF CUSTOMER
LOYALTY IN THE MACEDONIAN MOBILE
SERVICE MARKET
Nikolina Palamido ska-S e jado ska, Ani a Ciuno a-Shuleska
EM_2_2017.indd 199EM_2_2017.indd 199 14.6.2017 9:29:5014.6.2017 9:29:50
200 2017, XX, 2
Ma ke ing a obchod
loyal y, bu some imes, cus ome s can mani es
high le el o epea pu chases as a esul o
some swi ching ba ie s hey a e acing.
Conside ing ha , o he bes o ou
knowledge, no a icle examining he cus ome
loyal y o he mobile se ice indus y in
Macedonia has been published so a , his
a icle is he i s a emp o analyse he
e ec s o pe cei ed se ice quali y, cus ome
sa is ac ion, swi ching cos s and swi ching
ba ie s on cus ome loyal y. The o e all
esea ch objec i e is o gain insigh in o he
ac o s a ec ing cus ome loyal y, o de elop
a model acco dingly, and o es he model
empi ically. The e o e, he pu pose o his
s udy is o p esen a heo e ical amewo k o
he ac o s ha po en ially in l uence cus ome s
o emain loyal o hei exis ing mobile se ice
p o ide and o i nd he app op ia e model
o measu ing cus ome loyal y in he mobile
se ice ma ke in a de eloping coun y.
Thus, he main objec i es o his s udy a e:
a) o examine he ela ionship be ween se ice
quali y and cus ome sa is ac ion; b) o assess
he ela ionship be ween cus ome sa is ac ion
and cus ome loyal y; c) o explo e he media ing
ole o cus ome sa is ac ion in he ela ionship
be ween se ice quali y and cus ome loyal y; d)
o explo e he di ec , media ing and mode a ing
ole o swi ching cos s and swi ching ba ie s on
he ela ionship be ween cus ome sa is ac ion
and cus ome loyal y; and e) o alida e he
p oposed concep ual model.
The a icle sugges s ha cus ome loyal y
is posi i ely de e mined by se ice quali y,
cus ome sa is ac ion and swi ching ba ie s,
u he explo ing hei addi ional e ec s
on cus ome loyal y. This o ms he basis
o he de elopmen o a esea ch model
and hypo heses, using SEM. The a icle is
s uc u ed as ollows: i s ly, he concep s o
se ice quali y, cus ome sa is ac ion, cus ome
loyal y, swi ching cos s and swi ching ba ie s
a e elabo a ed, and hen hypo heses ega ding
hei ela ionships a e de eloped. Secondly, he
esea ch me hodology is p esen ed ollowed,
by a p esen a ion o i ndings and conclusions.
The las sec ion ou lines he heo e ical and
manage ial implica ions, limi a ions, and
di ec ions o u u e esea ch.
Ou i ndings ha e clea implica ions o
ma ke ing manage s in he mobile se ice
indus y in e es ed in main aining a loyal
cus ome base. The i ndings could se e well
as guidelines when modelling, measu ing
and acking cus ome loyal y in he mobile
se ice indus y, and should be conside ed
when designing ma ke ing s a egies aimed a
building long- e m ela ionships wi h cus ome s.
1. Concep ual Model and Hypo heses
De elopmen
The concep ualiza ion and ope a ionaliza ion
o he main concep s in his s udy is based on
ele an li e a u e, p esen ed below. Addi ionally,
hypo heses ega ding he ela ionships among
he concep s a e de eloped.
1.1 Se ice Quali y
The concep ualiza ion o se ice quali y is
de i ed om he p oduc quali y li e a u e
and cus ome sa is ac ion li e a u e (B ady
& C onin, 2001). Gene ally, he de i ning o
se ice quali y is based on he compa ison
be ween cus ome s’ expec a ions and hei
pe cep ions ega ding he deli e ed se ice
(Leh inen & Leh inen, 1991; Pa asu aman e
al., 1988; 1985; G ön oos, 1984). G ön oos
(1984, p. 37) has de i ned se ice quali y as “ he
ou come o an e alua ion p ocess, whe e he
consume compa es his expec a ions wi h he
se ice he pe cei es he has ecei ed, i.e., he
pu s he pe cei ed se ice agains he expec ed
se ice”. In ha di ec ion, Pa asu aman e al.
(1988, p.16) has de i ned se ice quali y as
“a global judgmen , o a i ude, ela ing o he
supe io i y o he se ice”.
Some au ho s ha e a gued he
concep ualiza ion o cus ome expec a ions
(Teas, 1993,) and he concep ualiza ion
and measu emen o se ice quali y as
a di e ence be ween cus ome pe cep ions and
expec a ions (C onin & Taylo , 1992). Based
on his, some s udies ha e ope a ionalized
se ice quali y as a pe o mance-based
concep (C onin, 2003; B ady e al., 2002;
Dabholka e al., 2000). Al hough hese s udies
a oid cus ome expec a ions in de i ning and
measu ing se ice quali y, hey do no a gue
he implici e ec o cus ome expec a ions
in c ea ing cus ome pe cep ions ega ding
pe o mance o he se ice.
E en hough mos o he s udies sugges
mul i-dimensionali y o he se ice quali y
concep (Kang & James, 2004; Pa asu aman
e al., 1985; 1988), he e a e s udies analysing
i as a uni-dimensional concep (C onin &
Taylo , 1992). Mo eo e , he e is a a ia ion
EM_2_2017.indd 200EM_2_2017.indd 200 14.6.2017 9:29:5014.6.2017 9:29:50
201
2, XX, 2017
Ma ke ing and T ade
o he numbe and na u e o he iden i i ed
dimensions (O el & Ka a, 2014; Ma ko ić &
Raspo Janko ić, 2013; Ciuno a-Shuleska e
al., 2013; Pe idou e al., 2007; Gla eli e al.,
2006; B ady & C onin, 2001), con ingen on he
cul u al and se ice con ex o he s udy. The
mos ecognized a e he concep ualiza ions
p oposed by Pa asu aman e al. (1988;
1985), iden i ying i e SERVQUAL dimensions
( angibles, eliabili y, esponsi eness,
assu ance and empa hy), and he wo-
dimensional concep ualiza ion o se ice quali y
( echnical and unc ional dimension), p oposed
by G ön oos (1984).
Fo he pu pose o his s udy, se ice quali y
is concep ualized and ope a ionalized as an
a i ude esul ing om cus ome pe cep ions
o se ice pe o mance. The dimensionali y o
pe cei ed se ice quali y was u he es ed in
o de o iden i y he unde lying dimensions o
se ice quali y in he mobile se ice sec o .
1.2 Cus ome Sa is ac ion
Al hough cus ome sa is ac ion li e a u e has
e ol ed o e he yea s, he e is inconsis ency
in de nining his concep . Gene ally, cus ome
sa is ac ion is de i ned as a p ocess o
e alua ion (Va a, 1997; Fo nell, 1992) o as
an ou come ( esul ) o his p ocess (Oli e ,
2010; Yi, 1990). While Fo nell (1992, p. 11)
de i nes cus ome sa is ac ion as “an o e all
pos -pu chase e alua ion”, Oli e (2010,
p. 8) sugges s ha cus ome sa is ac ion is
“ he consume ’s ul i llmen esponse”. Oli e
(1997, p. 13) has s a ed ha “sa is ac ion is
a judgmen ha a p oduc o se ice ea u e,
o he p oduc o se ice i sel , p o ided (o is
p o iding) a pleasu able le el o consump ion-
ela ed ul i llmen , including le els o unde - o
o e - ul i llmen ”.
Ini ially, cus ome sa is ac ion esea ch
was ocused on de i ning and measu ing i on
a ansac ional le el as a cus ome ’s expe ience
wi h a p oduc episode o se ice encoun e
(Oli e , 1997; Yi, 1990;). In he ea ly 1990s,
as he ma ke ing pa adigm changed om
ansac ion-o ien ed o ela ional ma ke ing,
cus ome sa is ac ion s udies a e emphasised
he cumula i e pe spec i e o concep ualisa ion
and measu emen o cus ome sa is ac ion
(Ande son & Fo nell, 2000; Fo nell e al., 1996;
Ande son e al., 1994). The ecen na ional
cus ome sa is ac ion index models a e based
on he cumula i e pe spec i e concep ualizing
cus ome sa is ac ion as a cus ome ’s o e all
e alua ion o a p oduc o se ice p o ide o
da e (Fo nell e al., 1996).
In his pape , se ice quali y and cus ome
sa is ac ion a e analysed and measu ed as
di e en concep s a a cumula i e le el, as
his pe spec i e is c ucial in unde s anding
he impo ance o hese wo concep s and
he ela ionship be ween hem (Ande son &
Fo nell, 1994). Te ps a and Ve bee en (2014,
p. 499) assume ha “cus ome sa is ac ion is
e oked by he cus ome ’s en i e expe ience
wi h he company”. In addi ion, measu ing
cus ome sa is ac ion a a cumula i e le el
p o ides be e esul s in p edic ing he e ec on
cus ome loyal y (Yang & Pe e son, 2004) and
p esen s a be e indica o o he pas , cu en
and u u e company’s pe o mances (Ande son
e al., 1994). Ac ually, cus ome sa is ac ion
i sel is o en used as an essen ial ma ke ing
pe o mance indica o (Rego e al., 2013).
1.3 Cus ome Loyal y
Cus ome loyal y is a complex concep
and di e en esea che s p o ide di e en
de i ni ions and ope a ionaliza ions ega ding
his concep . In he pas , cus ome loyal y was
s udied using only he beha io al app oach
(Rus & Zaho ik, 1993), which iden i i es
cus ome loyal y wi h epea pu chases, i.e.
cus ome e en ion (Zei haml e al., 1996).
This concep ualiza ion does no p o ide an
explana ion o he psychological aspec s
and causes o cus ome loyal y. Many imes
cus ome s epu chase he company’s p oduc s
and se ices no because hey ha e s ong
p e e ences, bu because hey a e simply ine ,
hey pe cei e a high le el o swi ching cos s,
o because o some o he swi ching ba ie s. In
ha di ec ion, Dick and Basu (1994) assume
ha ue cus ome loyal y is no based only on
he high le el o epu chasing, bu also on he
posi i e ela i e cus ome a i udes ega ding
a ec i e b and e alua ion. In ac , a i udinal
loyal y is e l ec ed in he p e e ences o he
company, compa ed o hei compe i o s. Dick
and Basu (1994, p. 99) ha e de i ned cus ome
loyal y as “ he s eng h o he ela ionship
be ween an indi idual’s ela i e a i ude and
epea pa onage”.
The concep ualiza ion o cus ome loyal y
as a combina ion o beha io al and a i udinal
dimension is addi ionally suppo ed by many
au ho s (Picón e al., 2014; de Ma os e al.,
EM_2_2017.indd 201EM_2_2017.indd 201 14.6.2017 9:29:5014.6.2017 9:29:50
202 2017, XX, 2
Ma ke ing a obchod
2009; Wong & Sohal, 2003; Ganesh e al.,
2000). Some au ho s add a cogni i e dimension
o concep ualizing cus ome loyal y (Bloeme e
al., 1999; de Ruy e e al., 1998), e e ing o
he b and’s a ibu es e alua ion in he buying
decision-making p ocess and o bene i s om
he epea pa onage (Ca uana, 2002; Lee &
Cunningham, 2001). In his s udy, he concep
o cus ome loyal y is ope a ionalized as
a combina ion o a beha io al and a i udinal
app oach, measu ing epu chase in en ions,
in en ions o ecommenda ions o iends, e c.
1.4 Swi ching Ba ie s and Swi ching
Cos s
Swi ching ba ie s makes i mo e di i cul o
cos ly o cus ome s o change p o ide s,
and e e s o in e pe sonal ela ionships
(Colga e & Lang, 2001; Jones e al., 2000);
pe cei ed swi ching cos s ( ime, money and
e o ) (La i iè e, 2008; Jones e al., 2000);
a ac i eness o compe ing al e na i es (Wu,
2011; La i iè e, 2008; Jones e al., 2000), and
se ice eco e y (Colga e & Lang, 2001). Fo
a gi en le el o cus ome sa is ac ion, he le el
o cus ome loyal y can a y depending on he
le el o swi ching ba ie s (Kim e al., 2004;
Colga e & Lang, 2001; Jones e al., 2002; Lee
& Cunningham, 2001). The e o e, companies
should s i e o unde s and ac o s which ha e
a posi i e e ec on cus ome e en ion, and
hus apply a combined s a egy, complemen ing
he e ec s o cus ome sa is ac ion (Tes om &
Bi ch, 2011).
In his s udy, ou ocus is on swi ching cos s,
since hey a e among he mos analysed ac o s
in l uencing he ela ionship be ween cus ome
sa is ac ion and loyal y (de Ma os e al., 2009;
Bu nham e al., 2003; Jones e al., 2002; Lee
& Cunningham, 2001; de Ruy e e al., 1998).
Gene ally, swi ching cos s can be de i ned
“as he pe cei ed economic and psychological
cos s associa ed wi h changing om one
al e na i e o ano he ” (Jones e al., 2002,
p. 441). Recen ly, swi ching cos s a e analysed
as a mul idimensional concep , consis ing o
he ollowing dimensions: economic (Aydin e
al., 2005; Pa e son, 2004; Lee & Cunningham,
2001), i.e. i nancial cos s (de Ma os e al., 2009;
Bu nham e al., 2003); p ocedu al swi ching
cos s (de Ma os e al., 2009; Aydin e al.,
2005; Bu nham e al., 2003), and ela ionship
cos s (Bu nham e al., 2003; de Ma os e al.,
2009), i.e. psychological cos s (Aydin e al.,
2005; Pa e son, 2004). Blu e al. (2014) ha e
sugges ed a wo-dimensional app oach o
swi ching cos s, comp ising he in e nal and
ex e nal pe spec i e o swi ching cos s.
In his s udy, swi ching cos s a e analysed
as di ec de e minan s o swi ching ba ie s.
Fu he mo e, he dimensionali y o swi ching
cos s is analysed in o de o iden i y he
unde lying dimensions o swi ching cos s in he
mobile se ice con ex in Macedonia.
1.5 Hypo heses De elopmen
Se ice quali y is analysed as one o he se ice
ac o s con ibu ing o cus ome sa is ac ion
judgemen s (Gouna is & Boukis, 2013;
Ca uana, 2002). In ac , mos o he s udies
explo ing he ela ionship be ween se ice
quali y and cus ome sa is ac ion ha e p o ed
he exis ence o a s ong posi i e ela ionship
whe e cus ome sa is ac ion is a consequence
o se ice quali y (O el & Ka a, 2014; Rajic
e al., 2013; Pisnik Ko da e al, 2012; Kang
& James, 2004; B ady & Robe son, 2001;
Dabholka e al., 2000; C onin e al., 2000).
Addi ionally, he causal o de o se ice quali y
and cus ome sa is ac ion is consis en ac oss
cul u es (Ma inko ić e al., 2011; Pisnik Ko da
& Snoj, 2010; Kang & James, 2004; B ady &
Robe son, 2001; Dabholka e al., 2000) and
he posi i e e ec o se ice quali y on cus ome
sa is ac ion has been also con i med by se e al
s udies in he mobile se ice sec o (San ou idis
& T i ellas, 2010; Kim e al., 2004; Lee e al.,
2001; Bayol e al., 2000). Based on his, he
ollowing hypo hesis is p oposed:
H1: Pe cei ed se ice quali y has a di ec
and posi i e e ec on cus ome sa is ac ion.
Imp o emen o se ice quali y suppo s
de ensi e ma ke ing s a egies ocused on
e aining cu en cus ome s (Bell e al., 2005).
Al hough cus ome loyal y is implici ly se as an
ul ima e dependen a iable in se ice quali y
models (Fo nell e al., 1996), heo e ical and
empi ical e idence is lacking ega ding his
ela ionship (de Ruy e e al., 1998). Nume ous
empi ical s udies sugges a posi i e ela ionship
be ween cus ome sa is ac ion and cus ome
loyal y (Picón e al., 2014; O el & Ka a, 2014;
Kuo e al., 2013; C onin e al., 2000; Dabholka
e al., 2000) as well as, be ween pe cei ed
se ice quali y and cus ome loyal y (Chen e
al., 2013; Mi al & Lassa , 1998; Zei haml e al.,
1996). Mos o he s udies analyse he in l uence
o se ice quali y on cus ome loyal y as a di ec
EM_2_2017.indd 202EM_2_2017.indd 202 14.6.2017 9:29:5014.6.2017 9:29:50
203
2, XX, 2017
Ma ke ing and T ade
e ec (Wong & Sohal, 2003; Bloeme e al.,
1999; de Ruy e , 1998; Zei haml e al., 1996)
o as an indi ec e ec , media ed by cus ome
sa is ac ion (P en ice & King, 2014; O el &
Ka a, 2014; Kuo e al., 2013; Bee li e al., 2004;
Dabholka e al., 2000). Some au ho s indica e
ha se ice quali y has a simul aneous di ec
and indi ec in l uence on cus ome loyal y, i.e.
on epu chase in en ions (Kim e al. 2014; Bei
& Chiao, 2006; Va ki & Colga e, 2001; C onin e
al., 2000), sugges ing an in eg a ed app oach in
concep ualizing cus ome loyal y. Simila esul s
ha e been ound in he mobile se ice s udies,
sugges ing he posi i e e ec o se ice quali y
on cus ome loyal y, media ed by cus ome
sa is ac ion (San ou idis & T i ellas, 2010; Kim
e al., 2004; Lee e al., 2001). Based on he
p esen ed li e a u e, he ollowing hypo heses
a e de i ed:
H2: Se ice quali y has a di ec and posi i e
e ec on cus ome loyal y.
H3: Cus ome sa is ac ion has a di ec and
posi i e e ec on cus ome loyal y.
H4: Se ice quali y has an indi ec and
posi i e e ec on cus ome loyal y, media ed by
cus ome sa is ac ion.
Since he cus ome sa is ac ion-loyal y
ela ionship is complex, he esea ch goes
u he in explo ing i s na u e and in analysing
o he ac o s ha in l uence cus ome loyal y.
The e ec o cus ome sa is ac ion on cus ome
loyal y and epu chase in en ions depends on
he le el o swi ching ba ie s in he analysed
se ice sec o (Jones e al., 2000). In his s udy,
swi ching ba ie s a e analysed as con ac ual
obliga ions which a ec cus ome loyal y, while
swi ching cos s a e analysed as he mos
impo an ac o and a di ec de e minan o
swi ching ba ie s (Kim e al., 2004). This s udy,
in ac , examines he ole o swi ching ba ie s in
he ela ionship be ween cus ome sa is ac ion
and cus ome loyal y, assuming he di ec
e ec o swi ching cos s on swi ching ba ie s.
The e o e he ollowing hypo hesis is p oposed:
H5: Swi ching cos s has a di ec and
posi i e e ec on swi ching ba ie s.
In he li e a u e, he e is inconsis ency
in analysing swi ching cos s as a di ec
de e minan o cus ome loyal y (Blu e al.,
2014; Pick & Eisend, 2014; de Ma os e al.,
2009; Bee li e al., 2004; Bu nham e al.,
2003; Jones e al., 2002); as a mode a o o
he ela ionship be ween cus ome sa is ac ion
and loyal y (Picón e al., 2014; de Ma os e al.,
2009; Yang & Pe e son, 2004; Pa e son, 2004;
Jones e al., 2000), o as a media o (de Ma os
e al., 2009; Aydin & Öze , 2006).
The gi en a iable (e.g. swi ching cos s)
is a mode a o i , only on ce ain condi ions o
his a iable, he p edic o (e.g. sa is ac ion) is
ela ed o he ou come a iable (e.g. loyal y)
and he mode a o a ec s he o m o he
s eng h o he ela ionship (Walsh e al., 2008).
The i ndings o he mode a ing ole o swi ching
cos s show ha “as swi ching cos s inc ease,
he ela ionship be ween sa is ac ion and
epu chase in en ions (loyal y) dec eases” (de
Ma os e al., 2009, p. 509).
A media o ansmi s he e ec o an
independen a iable o a dependen a iable,
explaining he ela ionship be ween hem
(MacKinnon & Fai child, 2009). A media ing
e ec o swi ching cos s in he sa is ac ion-
loyal y ela ionship is p esen i “ he e is no
di ec e ec be ween sa is ac ion and loyal y,
bu he e is a di ec e ec o sa is ac ion
on swi ching cos s and his one on loyal y
(sa is ac ion-swi ching cos s-loyal y)” (de Ma os
e al., 2009, p. 509). Al hough he ela ionship
be ween cus ome sa is ac ion and loyal y
p o es o be s ong (Picón e al., 2014; O el
& Ka a, 2014; Kuo e al., 2013), i is u he
expec ed swi ching cos s o be de e mined by
sa is ac ion and o di ec ly a ec loyal y (Picón
e al., 2014).
The mobile se ice s udies ha e analysed
he di ec e ec o swi ching cos s on cus ome
loyal y (Aydin e al., 2005; Kim e al., 2004), he
mode a ing ole o swi ching cos s (Aydin e
al., 2005; Kim e al., 2004; Lee e al., 2001), as
well as he media ing ole o his a iable in he
ela ionship be ween cus ome sa is ac ion and
loyal y (Aydin & Öze , 2006). In his s udy, he
in l uence o swi ching cos s on cus ome loyal y
is analysed indi ec ly, h ough swi ching ba ie s
and he di e en e ec s o swi ching ba ie s
(i.e. di ec , media ing and mode a ing) on he
ela ionship be ween cus ome sa is ac ion
and loyal y a e es ed. The e o e, he ollowing
hypo heses a e p oposed:
H6: Swi ching ba ie s has a di ec and
posi i e e ec on cus ome loyal y.
H7: Swi ching ba ie s has a media ing
ole in he ela ionship be ween cus ome
sa is ac ion and cus ome loyal y.
H8: Swi ching ba ie s has a mode a ing
ole on he ela ionship be ween cus ome
sa is ac ion and cus ome loyal y.
EM_2_2017.indd 203EM_2_2017.indd 203 14.6.2017 9:29:5014.6.2017 9:29:50
204 2017, XX, 2
Ma ke ing a obchod
Since swi ching ba ie s a e analysed as
a media o in he ela ionship be ween cus ome
sa is ac ion and cus ome loyal y, implici ly,
i is assumed ha cus ome sa is ac ion has
a di ec e ec on swi ching ba ie s. The e o e,
he ollowing hypo hesis is sugges ed:
H9: Cus ome sa is ac ion has a di ec and
posi i e e ec on swi ching ba ie s.
Based on he p e iously e iewed li e a u e,
he au ho s ha e de eloped he concep ual
model (Fig. 1), simul aneously linking se ice
quali y, cus ome sa is ac ion, cus ome loyal y,
swi ching cos s and swi ching ba ie s.
2. Resea ch Me hodology
2.1 Sampling and Measu ing
The ques ionnai e used in his s udy was
designed o collec empi ical da a e e ing o
mobile se ice use s’ pe cep ions o se ice
quali y, cus ome sa is ac ion, cus ome loyal y,
swi ching cos s and swi ching ba ie s. Each o
hese concep s could no be di ec ly measu ed
and he e o e mul i-i em scales we e de eloped
o each la en cons uc . In o de o ensu e
con en alidi y, he p e iously analysed ele an
li e a u e was used o he ope a ionaliza ion
o each cons uc (Kim e al., 2004; B ady &
Robe son, 2001; Bayol e al., 2000; Ganesh e
al., 2000; Bloeme e al., 1999). All o he i ems
we e measu ed on a i e-poin Like scale (1 =
“s ongly disag ee” o 5 = “s ongly ag ee”). The
ques ionnai e also comp ised demog aphic
ques ions. Since he ques ionnai e was
ansla ed in o and adap ed o Macedonian,
a backwa d ansla ion in English was done,
in o de o a oid any misunde s andings. The
ques ionnai e was p e- es ed by 8 mobile
se ice use s and mino e i nemen s o he
wo ding we e made. The i nal esea ch was
conduc ed on unde g adua e s uden s a he
Facul y o Economics – Skopje, collec ing
224 esponses ou o which 209 we e alid.
Al hough modes , he sample size is abo e
he ecommended minimum le el o a leas
en imes as many obse a ions as he numbe
o a iables o be analysed (Hai e al., 2006).
The a e age age o he esponden s is 21.85
and 66.51% o hem a e emales, e l ec ing
he a e age age and gende s uc u e o he
s uden s a he Facul y o Economics – Skopje.
Conside ing ha young people a e he hea ies
use s o mobile se ice (no only o alking, bu
also o ex ing and In e ne usage), hey a e in
ac he mos po en ial ma ke segmen . Wi h
he apid p oli e a ion o sma phones, young
people a e using hei mobile de ices o online
ac i i ies h ee imes longe han he elde s
(McKinsey & Company, 2013). Addi ionally, in
iew o c ea ing long- e m cus ome ela ionship
and building loyal y, young people a e an e en
mo e p ominen segmen which mobile se ice
p o ide s should be ocused on, since hey
and hei upcoming amilies will cons i u e he
cus ome base o mobile se ice p o ide s in
he u u e.
2.2 Measu emen Assessmen
The explo a o y ac o analysis was pe o med
by using SPSS 20, while o pe o ming
con i ma o y ac o analysis s uc u al
equa ion modelling AMOS 20 was applied.
Fi s , he mul iple i ems measu ing he
espec i e cons uc s we e ac o -analysed
in o de o de e mine he dimensionali y
o each cons uc (Chu chill, 1979). The
esul s indica ed uni-dimensionali y o all he
p oposed cons uc s. Fu he mo e, in e nal
consis ency o he i ems ep esen ing each
cons uc was assessed by using C onbach’s
alpha. The eliabili y o each cons uc was as
ollows: pe cei ed quali y = 0.880; cus ome
sa is ac ion = 0.894; cus ome loyal y
= 0.907; swi ching cos s = 0.877 and swi ching
ba ie s = 0.672. All he alues we e abo e 0.6,
exceeding he h eshold alues ecommended
by Hai e al. (2006) and indica ing high in e nal
eliabili y o he da a collec ed. Hence, he i ems
we e sui able o use as e l ec i e indica o s
in u he con i ma o y ac o analysis (CFA)
and s uc u al equa ion modelling (SEM). The
alidi y o he p oposed measu emen model
was assessed by conduc ing CFA (Tab. 1).
I was obse ed ha all he i ems
appea ed o ha e a signi i can ela ionship
wi h he hypo hesized cons uc s. Namely,
he mani es indica o s we e good e l ec ions
o hei espec i e la en cons uc s wi h
mode a e o s ong s anda dized loadings
(0.58-0.899). The - alues we e all signi i can
(p < 0.01). On inspec ion o he ou pu i les,
he pa ame e modi i ca ion indices showed no
need o loosening cons ain s on ce ain model
pa ame e s in o de o imp o e he o e all model
i , p o iding suppo ha he model speci i ca ion
was app op ia e. Mo eo e , because o he la ge
s anda dized esidual co a iances be ween
some o he a iables, disc epancies be ween
he p oposed model and es ima ed model we e
EM_2_2017.indd 204EM_2_2017.indd 204 14.6.2017 9:29:5014.6.2017 9:29:50
205
2, XX, 2017
Ma ke ing and T ade
no ed. Consequen ly, i e i ems we e emo ed
om he model: h ee i ems on se ice quali y
( echnical quali y o he ne wo k; ange o
se ices and p oduc s o e ed; cla i y and
anspa ency o in o ma ion p o ided); one i em
on cus ome loyal y (in en ion o ecommend
he ope a o ); and one i em on swi ching cos s
(cos o eplacing he mobile de ice). The
i nal model consis s o 16 i ems, measu ing
i e cons uc s: pe cei ed quali y, cus ome
sa is ac ion, cus ome loyal y, swi ching cos
and swi ching ba ie s. The con i ma o y ac o
analysis indica ed ha he alidi y e alua ion
s anda ds we e sa is i ed (Tab. 1). In addi ion,
he chi-squa e/deg ees o eedom a io was
well below Ma ch & Hoce a ’s (1985) c i e ion.
The e o e, he p oposed esea ch model can
be ecommended as an accep able model.
Cons uc s/I ems Mean SD
S anda dized
eg ession
coe i cien s
CR
Pe cei ed se ice quali y 0.782
Cus ome se ice and pe sonal ad ice o e ed 3.823 0.833 0.805***
Quali y o he se ices one uses 3.813 0.843 0.765***
Reliabili y and accu acy o he p oduc s and se ices p o ided 3.656 0.864 0.638***
Cus ome sa is ac ion 0.900
Sa is ac ion wi h he decision o be a use o his mobile ope a o 4.038 0.796 0.805***
The choice o be a use o his mobile ope a o was a wise one 4.053 0.873 0.890***
Conside ing he igh hing was done when s a ing o use his
mobile ope a o 4.072 0.877 0.899***
Cus ome loyal y 0.907
Conside ing he mobile ope a o as he i s choice among
mobile ope a o s 3.785 1.196 0.881***
In en ions o use he se ices o his mobile ope a o e en
i i s p ices somewha inc eased 3.297 1.184 0.866***
In en ions o complain o o he cus ome s i a p oblem
is expe ienced 3.646 1.147 0.876***
Swi ching cos s 0.856
Di i cul y o changing numbe 3.474 1.144 0.751***
Loss o bene i s, such as mileage p og am 3.120 1.256 0.672***
Incon enience o ha ing o lea n a new se ice 3.258 1.193 0.810***
Need o sea ch o in o ma ion on a new mobile ope a o
when swi ching 3.459 1.172 0.791***
Subsc ip ion ee o new membe ship 3.297 1.212 0.652***
Swi ching ba ie s 0.701
Economic loss associa ed wi h swi ching mobile ope a o s 2.737 1.153 0.873***
Psychological bu den associa ed wi h swi ching mobile
ope a o s 2.067 1.191 0.580***
Sou ce: au ho s’ calcula ions
Fi indices: x2/d = 1.128; GFI = 0.942; AGFI = 0.916; CFI = 0.993; RMR = 0.042; RMSEA = 0.025, NFI = 0.94
Recommenda ion c i e ia: x2/d < 2; GFI (goodness-o - i -index) > 0.90; AGFI (adjus ed goodness-o - i index) > 0.90; CFI
(compa a i e i index) > 0.90; RMR ( oo mean squa ed esidual) < 0.05; RMSEA ( oo mean squa e e o o app oxima-
ion) < 0.05; NFI (no ma i e i index) > 0.90
*** signi i can a p < 0.01
Tab. 1: I ems, s anda dized loadings and CR
EM_2_2017.indd 205EM_2_2017.indd 205 14.6.2017 9:29:5114.6.2017 9:29:51
206 2017, XX, 2
Ma ke ing a obchod
The composi e eliabili y alues (CR) o
all he cons uc s a e abo e 0.70 ( anging om
0.701 o 0.907), con i ming in e nal consis ency
o he measu es (Tab. 1). Cons uc alidi y
was assessed by examining he con e gen
and disc iminan alidi y o each cons uc .
In he assessed measu emen model, he
a e age a iance ex ac ed alue (AVE) o each
cons uc is abo e he h eshold alue o 0.50
(Fo nell & La cke , 1981), sugges ing adequa e
con e gen alidi y o he cons uc s (see
Tab. 2). Disc iminan alidi y o each cons uc in
he p oposed model is p o ed, by using squa e-
oo s o he AVEs which should be highe han
he squa ed co ela ion coe i cien s o each pai
o cons uc s (Hai e al., 2006) (Tab. 2).
3. Resea ch Findings
In he i nal s age o he analysis, hypo hesised
ela ionships be ween cons uc s we e analysed
h ough s uc u al equa ion modelling (SEM).
The esul s (Tab. 3) con i m mos o he
hypo hesised posi i e and di ec ela ionships,
excep he ela ionship be ween cus ome
sa is ac ion and swi ching ba ie s (H9).
The s onges posi i e ela ionships a e
hose be ween se ice quali y and cus ome
sa is ac ion (0.574***), suppo ing H1 and
be ween swi ching cos s and swi ching ba ie s
(0.535***), suppo ing H5. Addi ionally, se ice
quali y and cus ome sa is ac ion, sepa a ely,
ha e a signi i can in l uence on cus ome loyal y
(0.34*** and 0.333***, espec i ely). These
Squa e oo o AVEs in diagonal cells
AVE Swi ching
cos s
Se ice
quali y
Cus ome
sa is ac ion
Cus ome
loyal y
Swi ching
ba ie s
Swi ching cos s 0.544 0.738
Se ice quali y 0.547 0.123 0.739
Cus ome sa is ac ion 0.749 0.162 0.571 0.866
Cus ome loyal y 0.764 0.199 0.544 0.553 0.874
Swi ching ba ie s 0.549 0.543 0.067 0.132 0.260 0.741
Sou ce: au ho s’ calcula ions
Tab. 2: Con e gen and disc iminan alidi y
Fig. 1: S uc u al model
Sou ce: own
Fi indices: x2/d = 1.113; GFI = 0.941; AGFI = 0.917; CFI = 0.993; RMR = 0.046; RMSEA = 0.023, NFI = 0.939
EM_2_2017.indd 206EM_2_2017.indd 206 14.6.2017 9:29:5114.6.2017 9:29:51
207
2, XX, 2017
Ma ke ing and T ade
esul s suppo H2 and H3. The hypo hesis
e e ing o he posi i e ela ionship be ween
swi ching ba ie s and cus ome loyal y, i.e. H6
was also con i med (0.19**).
As an ul ima e a iable in he model,
cus ome loyal y is s ongly de e mined by
se ice quali y, cus ome sa is ac ion and
swi ching ba ie s. On he one hand, cus ome s
will emain loyal when hey pe cei e a high le el
o se ice quali y and expe ience a high le el
o sa is ac ion, exp essing a i udinal loyal y as
well (Picón e al., 2014). On he o he hand,
he exis ence o swi ching ba ie s could o ce
cus ome s o e ain he cu en se ice p o ide
e en when hey a e dissa is i ed, in o de o
economize on swi ching cos s (Lam e al., 2004).
Cus ome s pe cei e swi ching o an al e na i e
p o ide as a po en ially ime-consuming and
isky ask and he e o e hey a e mo i a ed o
s ay loyal (Woise schläge e al., 2011).
The absence o in l uence o cus ome
sa is ac ion on swi ching ba ie s ( ejec ion o
H9) is con a y o he sugges ions o Picón e
al. (2014) ha a a ce ain le el o cus ome
sa is ac ion, cus ome s pe cei e g ea e
oppo uni y cos s in swi ching he p o ide . This
could be explained by he ela i ely low le el
o swi ching cos s ha cus ome s pe cei e on
he Macedonian mobile se ice ma ke , hus
minimizing he pe cei ed loss o sa is ac ion
when swi ching. Namely, when swi ching
cos s a e low, a ela i ely s ong sa is ac ion-
loyal y ela ionship is expec ed, i.e. cus ome ’s
decision o s ay o lea e is based on he
sa is ac ion le el (Woise schläge e al., 2011).
The hypo heses which e e o he indi ec
e ec o se ice quali y and cus ome sa is ac ion
on cus ome loyal y (H4) and he media ing/
mode a ing e ec o swi ching ba ie s on he
cus ome -loyal y ela ionship (H7 and H8) we e
u he es ed by conduc ing media ion and
mode a ion analysis, app op ia ely.
The media ion analysis p o ed he media ing
ole o cus ome sa is ac ion in he ela ionship
be ween se ice quali y and cus ome loyal y,
suppo ing H4. Since he eg ession coe i cien
declined conside ably ( om 0.529 o 0.34),
bu is s ill signi i can , and he indi ec e ec
is signi i can (0.197**), a pa ial media ion o
cus ome sa is ac ion was p o ed. This esul
sugges s ha besides he di ec in l uence,
se ice quali y has a simul aneous indi ec
in l uence on cus ome loyal y. Ac ually,
pe cei ed se ice quali y ansmi s some o
he e ec s o cus ome sa is ac ion on loyal y,
i.e. in c ea ing cus ome loyal y, mobile se ice
p o ide s should s i e no only o c ea e
cus ome sa is ac ion, bu also o deli e an
adequa e le el o se ice quali y.
Al hough i was p oposed ha swi ching
ba ie s ha e a media ion ole in he ela ionship
be ween cus ome sa is ac ion and cus ome
loyal y (H7), his hypo hesis was no con i med.
S ep 2 ( he co ela ion be ween he causal
a iable and he media o ) p oposed by Ba on
& Kenny (1986) was no comple ed, and also
he indi ec e ec o swi ching ba ie s was
insigni i can , sugges ing ha swi ching ba ie s
do no media e he ela ionship be ween
cus ome sa is ac ion and cus ome loyal y.
Mo eo e , he mode a ing ole o swi ching
ba ie s on he ela ionship be ween cus ome
sa is ac ion and cus ome loyal y was
assessed, by impu ing an in e ac ion cons uc
S anda dized
eg ession
coe i cien
S.E. C.R.
H1: Cus ome s sa is ac ion <--- Se ice quali y 0.574*** 0.108 6.181
H2: Cus ome loyal y <--- Se ice quali y 0.340*** 0.178 3.647
H3: Cus ome loyal y <--- Cus ome sa is ac ion 0.333*** 0.139 3.929
H5: Swi ching ba ie s <--- Swi ching cos s 0.535*** 0.113 6.017
H6: Cus ome loyal y <--- Swi ching ba ie s 0.190** 0.076 2.630
H9: Swi ching ba ie s <--- Cus ome sa is ac ion 0.050ns 0.116 0.687
Sou ce: au ho s’ calcula ions
Whe e ns = non-signi i can ; ***p < 0.00; **p <0.01; *p < 0.05
Tab. 3: S anda dized eg ession coe i cien s wi hin he s uc u al model
EM_2_2017.indd 207EM_2_2017.indd 207 14.6.2017 9:29:5114.6.2017 9:29:51
214 2017, XX, 2
Ma ke ing a obchod
Vildo á, E., Ma inčík, D., Tlučhoř, J., &
Jakubíko á, D. (2015). Measu ing Cus ome
Sa is ac ion and Loyal y in Spa Companies.
E&M Ekonomie a Managemen , 18(1), 151-
168. doi:10.15240/ ul/001/2015-1-012.
Walsh, G., E anschi zky, H., & Wunde lich,
M. (2008). Iden i i ca ion and Analysis o
Mode a o Va iables: In es iga ing he
Cus ome Sa is ac ion-Loyal y Link. Eu opean
Jou nal o Ma ke ing, 42(9/10), 977-1004.
doi:10.1108/03090560810891109.
Woise schläge , D. M., Len z, P., &
E anschi zky, H. (2011). How Habi s, Social
Ties, and Economic Swi ching Ba ie s A ec
Cus ome Loyal y in Con ac ual Se ice
Se ings. Jou nal o Business Resea ch, 64(8),
800-808. doi:10.1016/j.jbus es.2010.10.007.
Wong, A., & Sohal, A. (2003). Se ice
Quali y and Cus ome Loyal y Pe spec i es
on Two Le els o Re ail Rela ionships. Jou nal
o Se ices Ma ke ing, 17(5), 495-513.
doi:10.1108/08876040310486285.
Wu, L. W. (2011). Sa is ac ion, Ine ia, and
Cus ome Loyal y in he Va ying Le els o he
Zone o Tole ance and Al e na i e A ac i eness.
Jou nal o Se ices Ma ke ing, 25(5), 310-322.
doi:10.1108/08876041111149676.
Yang, Z., & Pe e son, R. T. (2004). Cus ome
Pe cei ed Value, Sa is ac ion, and Loyal y: The
ole o Swi ching cos s. Psychology & Ma ke ing,
21(10), 799-822. doi:10.1002/ma .20030.
Yi, Y. (1990). A C i ical Re iew o Consume
Sa is ac ion. Re iew o Ma ke ing, 4(1), 68-123.
Zei haml, V. A., Be y, L. L., & Pa asu aman,
A. (1996). The Beha io al Consequences o
Se ice Quali y. Jou nal o Ma ke ing, 60(2),
31-46. doi:10.2307/1251929.
Zei haml, V. А. (2000). Se ice Quali y,
P o i abili y, and he Economic Wo h o
Cus ome s: Wha We Know and Wha We
Need o Lea n. Jou nal o he Academy
o Ma ke ing Science, 28(1), 67-85.
doi:10.1177/0092070300281007.
Nikolina Palamido ska-S e jado ska, Ph.D.
Ss. Cy il and Me hodius Uni e si y in Skopje
Facul y o Economics-Skopje
Depa men o Ma ke ing
[email p o ec ed]
Ani a Ciuno a-Shuleska, Ph.D.
Ss. Cy il and Me hodius Uni e si y in Skopje
Facul y o Economics-Skopje
Depa men o Ma ke ing
[email p o ec ed]
EM_2_2017.indd 214EM_2_2017.indd 214 14.6.2017 9:29:5214.6.2017 9:29:52
215
2, XX, 2017
Ma ke ing and T ade
Abs ac
AN INTEGRATED MODEL OF CUSTOMER LOYALTY IN THE MACEDONIAN
MOBILE SERVICE MARKET
Nikolina Palamido ska-S e jado ska, Ani a Ciuno a-Shuleska
The main objec i e o his pape is o de elop and es an in eg a ed model o cus ome loyal y in
he Macedonian mobile se ice ma ke , analysing i s de e minan s and hei in e ela ionships and
e ec s on cus ome loyal y. The p oposed esea ch model in eg a es se ice quali y, cus ome
sa is ac ion, swi ching cos s and swi ching ba ie s as he main de e minan s o cus ome
loyal y, u he analysing hei di e en e ec s (di ec /media ing/mode a ing) on cus ome loyal y.
Con i ma o y ac o analysis (CFA) is used in o de o assess he alidi y o he measu emen
model, and he de eloped s uc u al model is es ed wi h s uc u al equa ion modelling (SEM).
The epo ed esul s indica e ha he p oposed esea ch model as a whole has a sa is ac o y le el
o adequacy and mos o he hypo hesized eal ionships a e con i med. The esul s sugges ha
cus ome sa is ac ion pa ially media es he ela ionship be ween se ice quali y and cus ome
loyal y and also se ice quali y has a di ec and posi i e e ec on cus ome loyal y. The sa is ac ion-
loyal y ela ionship is no media ed no mode a ed by swi ching ba ie s, bu swi ching ba ie s ha e
a di ec and posi i e e ec on cus ome loyal y. The p esen ed esul s would be o g ea impo ance
o manage s o he success ul de elopmen o s a egies o c ea ing mobile se ice use s’ loyal y.
The empi ical e idence ob ained in his s udy sugges ha Macedonian mobile ope a o s should
ha e in mind al e na i e s a egies in c ea ing cus ome loyal y, emphasising he ac ha cus ome
sa is ac ion has a s onge posi i e in l uence in c ea ing cus ome loyal y han swi ching ba ie s.
The p esen esea ch p o ides an in eg a ed app oach in examining cus ome loyal y and explo es
he di e en oles o swi ching ba ie s in he ela ionship be ween cus ome sa is ac ion and loyal y.
In o de o u he de elop he in eg a ed app oach o cus ome loyal y modelling, some addi ional
concep s should be inco po a ed and examined in he u u e.
Key Wo ds: Cus ome loyal y, cus ome sa is ac ion, se ice quali y, swi ching cos s, swi ching
ba ie s, mobile se ice ma ke .
JEL Classi i ca ion: M31.
DOI: 10.15240/ ul/001/2017-2-015
EM_2_2017.indd 215EM_2_2017.indd 215 14.6.2017 9:29:5214.6.2017 9:29:52