Olo a, Oluwayomi Omo ayo; Balogun, Ebeneze Oluwadamila e; Babawale,
Opeyemi Emmanuel
A icle
Gaining cus ome insigh s in big da a o SMEs ma ke
segmen a ion decisions in eme ging ma ke s
Economic Fo um
P o ided in Coope a ion wi h:
Lu sk Na ional Technical Uni e si y
Sugges ed Ci a ion: Olo a, Oluwayomi Omo ayo; Balogun, Ebeneze Oluwadamila e; Babawale,
Opeyemi Emmanuel (2025) : Gaining cus ome insigh s in big da a o SMEs ma ke segmen a ion
decisions in eme ging ma ke s, Economic Fo um, ISSN 2415-8224, Lu sk Na ional Technical
Uni e si y, Lu sk, Uk aine, Vol. 15, Iss. 2, pp. 18-28,
h ps://doi.o g/10.62763/e /2.2025.18
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323731
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Gaining cus ome insigh s in big da a
o SMEs ma ke segmen a ion decisions in eme ging ma ke s
Oluwayomi Omo ayo Olo a*
PhD
Uni e si y o Ilo in
240003, P.M.B. 1515, Ilo in, Nige ia
h ps://o cid.o g/0009-0008-6633-9919
Ebeneze Oluwadamila e Balogun
Bachelo
Uni e si y o Ilo in
240003, P.M.B. 1515, Ilo in, Nige ia
h ps://o cid.o g/0000-0003-0419-188X
Opeyemi Emmanuel Babawale
Bachelo
Uni e si y o Ilo in
240003, P.M.B. 1515, Ilo in, Nige ia
h ps://o cid.o g/0009-0006-8782-3043
Sugges ed Ci a ion:
Olo a,O.O., Balogun, E.O., & Babawale, O.E. (2025). Gaining cus ome insigh s in big da a o SMEs ma ke segmen a ion decisions
in eme ging ma ke s. Economic Fo um, 15(2), 18-28. doi:10.62763/e /2.2025.18.
Copy igh © The Au ho (s). This is an open access a icle dis ibu ed unde he e ms o he
C ea i e Commons A ibu ion License 4.0 (h ps://c ea i ecommons.o g/licenses/by/4.0/)
*Co esponding au ho
Abs ac . Small and medium en e p ises a ound he wo ld, and especially in eme ging ma ke s, ace challenges
when i comes o ma ke segmen a ion. They ha e limi ed knowledge o he impo ance o big da a cus ome
insigh s o making conc e e ma ke segmen a ion decisions. The pu pose o his s udy was o assess gaining
cus ome insigh s in big da a o ma ke segmen a ion decisions o SMEs in eme ging ma ke s. The esul s o
he s udy indica ed ha cus ome beha iou analysis s ongly a ec s ma ke segmen a ion decisions o among
small and medium scale en e p ises. A be a- alue o 0.344 was shown o indica e ha a uni change in cus ome
beha iou analysis will lead o a uni change in ma ke segmen a ion decisions among small and medium scale
en e p ises wi h he -s a is ics o 4.608 and p- alue o 0.001. I was speci ied ha cus ome p e e ence analysis
s ongly a ec ed ma ke segmen a ion decisions o among small and medium scale en e p ises. The esul s
showed ha he be a- alue was 0.379, indica ing ha a uni change in cus ome p e e ence analysis will lead o a
uni change in ma ke segmen a ion decisions o among small and medium scale en e p ises wi h he -s a is ics o
6.654 and p- alue o 0.001. I was e ealed ha cus ome eedback analysis s ongly a ec ed ma ke segmen a ion
decisions o among small and medium scale en e p ises, he esul s showed ha he be a- alue was 0.215 also
indica ed ha a uni change in cus ome eedback analysis will lead o a uni change in ma ke segmen a ion
decisions o among small and medium scale en e p ises wi h he -s a is ics o 3.155 and p- alue o 0.002. I was
concluded ha gaining cus ome insigh in big da a was essen ial o small and medium en e p ises in eme ging
economy o make e ec i e ma ke segmen a ion decisions
Keywo ds: complex da a; cus ome beha iou analysis; cus ome eedback analysis; cus ome pe spec i e;
cus ome p e e ence analysis
ECONOMIC FORUM
Jou nal homepage: h ps://e- o um.com.ua/en
Vol. 15, No. 2, 2025, 18-28
A icle’s His o y: Recei ed: 13.01.2025 Re ised: 01.04.2025 Accep ed: 24.04.2025
UDC 65>005.5 DOI: 10.62763/e /2.2025.18
In oduc ion
Ma ke segmen a ion is impo an as i helps compa-
nies o alloca e esou ces be e , imp o e cus ome
engagemen , and he eby inc ease o e all p o i abil-
i y. Howe e , T. Ta o e al. (2023) c i icised ma ke
Olo a e al.
Economic Fo um, 2025, Vol. 15, No. 2
19
li e a u e in ol ed la ge en e p ises wi h abundan e-
sou ces, hence lea ing SMEs unde ep esen ed, espe-
cially hose in esou ce-cons ained se ings. Mo eo e ,
O. Abdul-Azeez e al. (2024) no ed abou he speci ic
scena ios o SMEs in eme ging ma ke s; his may includ
e in as uc u e limi a ions and di e ences in consum-
e beha iou ac oss egions. I was impo an o poin
ou hese gaps o c ea e solu ions ha can be scaled up
and implemen ed o p ac icali y by SMEs. This esea ch
will seek o p opose a new amewo k ha will in eg a e
cus ome beha iou analysis, cus ome p e e ence anal-
ysis, and cus ome eedback analysis in o ma ke seg-
men a ion decisions o SMEs in eme ging ma ke s. Em-
phasising hese h ee co e a eas, he s udy will p o ide
a comp ehensi e pe spec i e on how SMEs migh make
use o big da a in e ining segmen a ion s a egies.
By using cus ome insigh s wi h big da a, SMEs will
signi ican ly enhance hei ma ke segmen a ion deci-
sions. Big da a echnologies p o ided en e p ises wi h
he capaci y o collec la ge olumes o in o ma ion on
cus ome beha iou s, p e e ences, and eedback in eal
ime. Ad anced analy ical ools hen p ocessed his da a
o pa e ns and ends ha allowed much ine -g ained
segmen a ion. This mean ha SMEs, e en wi h hei
limi ed esou ces in eme ging ma ke s, we e able o
de elop a deep unde s anding o hei cus ome base
and cus omise hei o e s o mee e y speci ic needs.
Acco ding o E.Gence al.(2019), his app oach will no
only enhance compe i i eness, bu also enable dynamic
segmen a ion adap ed o ma ke changes.
Cus ome beha iou analysis helps SMEs unde -
s and no only who hei cus ome s a e, bu how hey
in e ac wi h p oduc s and se ices. This can include
pu chasing pa e ns, b and in e ac ions, and e en social
media beha iou , which was o en o e looked in con-
en ional segmen a ion models. Cus ome p e e ence
analysis will allow businesses o ack changing as es
and p e e ences o e ime, p o iding a dynamic seg-
men a ion model ha e ol es wi h he ma ke . The da a
om big da a helps o analyse eedback h ough online
e iews, su eys, and commen s abou se ices on so-
cial media, se ing i s -hand insigh s in o cus ome sa -
is ac ion and a eas needing change. The pu pose o his
s udy was o ill he esea ch gap by p o iding a p ac ical
and scalable solu ion o SMEs in eme ging ma ke s o
imp o e hei ma ke segmen a ion decisions.
Li e a u e Re iew
Ma ke segmen a ion decisions a e hose in which a
b oad ma ke is di ided in o dis inc subse s o cus om-
e s wi h common needs, cha ac e is ics, o beha iou s,
and hen a ge ed wi h an app op ia e ma ke ing s a -
egy. Acco ding o Y.Cuie al.(2025), p e e ence-guided
segmen a ion models add essed he e ogeneous cus-
ome decision-making, allowing companies o make
close - o-accu a e p edic ions o a ious cus ome s’ be-
ha iou s and op imise ma ke ing s a egies acco dingly.
segmen a ion o being da ed, when i elied on gene al-
ised o pa i ioned models, since hey do no ca e o he
needs o di e en cus ome g oups in an age ha was
apidly changing and illing up wi h compe i ion. T ea ing
cus ome s as a single g oup gene ally esul s in e y wide
ma ke s ha o e look c ucial in elligence on consume s,
which, in u n, back i ed on ma ke ing s a egies, which
hen do no u n up high p o i s. This p oblem was agg a-
a ed o SMEs (Small and Medium En e p ise), in he case
o eme ging ma ke s as hey ha e low amoun s o da a
and canno pull oge he sophis ica ed analy ical ools.
When compa ed on a global scale, especially agains
eme ging ma ke s, SMEs we e ce ainly up agains a
challenge while ying o di ide hei ma ke du ing hei
segmen a ion p ocess. Limi ed esou ces o ce hem o
cu back on hei expendi u es o ma ke esea ch ools,
which esul s in elying on ins inc s o incomple e da a.
On op o ha , S.Odu o(2020) posi ed ha SMEs, espe-
cially in A ican and Asian coun ies, ha e low exposu e
o he digi al wo ld and uns eady ma ke condi ions,
which ende ed he basic segmen a ion me hods use-
less. The esea che s a ed ha his misma ch be ween
he ools ha SMEs ha e access o and i immensely ca-
e s o a place, whe e imp o emen was needed. Unlike
o he s udies ha ocused p ima ily on cus ome de-
mog aphics, his esea ch emphasised beha iou al and
p e e ence da a o p o ide mo e ac ionable insigh s.
The e we e e o s a he global le el o imp o e
ma ke segmen a ion o SMEs, including digi al ans-
o ma ion and capaci y building in da a analy ics. Va -
ious p og ammes by go e nmen s and in e na ion-
al o ganisa ions aim a inc easing access o big da a
echnologies and ma ke in elligence ools o SMEs.
Mo eo e , Z.Sami ae al.(2024) ag eed ha la ge co -
po a ions and ech i ms ha e s a ed o e ing scalable
solu ions ha allowed SMEs o pe o m mo e g anula
segmen a ion using cus ome insigh s om digi al chan-
nels. Howe e , hese e o s o en emained agmen ed
and ailed o each he mos esou ce-cons ained SMEs,
limi ing hei o e all e ec i eness.
These p og ammes ha e had a spo adic impac on
SMEs. Whe eas he digi al ans o ma ion p og ammes
ha e gi en some he capaci y o enhance hei ma ke
segmen a ion using be e da a cap u e and analy ics,
many s ill all in o bad o w ong segmen a ion decisions
due o inconsis ency o inaccu acies due o he absence
o a coo dina ed s a egy o cus ome insigh . All his
exace ba es he limi ed applica ion o da a-d i en ap-
p oaches, gi en ha consume beha iou in di e se
ma ke s– whe he hey we e u al o unde se ed– was
gene ally no well comp ehended. Acco ding o O.Ab-
dul-Azeeze al.(2024), SMEs mos ly ailed o op imise
ma ke oppo uni ies. Wi h big da a p omising many
changes in almos e e y o he ope a ional aspec , he
esea ch gap on how SMEs in eme ging ma ke s can
ake ull ad an age o cus ome insigh s owa ds e ec-
i e ma ke segmen a ion s ill emains signi ican . Mos
Gaining cus ome insigh s in big da a...
Economic Fo um, 2025, Vol. 15, No. 2
20
A.Aouade al.(2023) indica ed he de elopmen o SMEs,
which we e conside ed as a new app oach o in eg a -
ing segmen a ion wi h esponse modelling by o e ing
compu a ionally e icien and in e p e able amewo ks
o segmen ing la ge da a.
Cus ome insigh s can be de ined as unde s anding
cus ome beha iou s, p e e ences, and eedback, which
can hen be used om da a o d i e business s a egy.
Acco ding o B.M.Omowolee al.(2024), big da a o e ed
scalable solu ions o he compe i i e ad an age o SMEs
by means o s a egies pe aining o he analysis o cus-
ome beha iou and p e e ences. Y.Zhonge al.(2024)
ha e shown how big da a can be used o imp o e cus-
ome sa is ac ion by le e aging online e iews in e in-
ing se ice o e s, especially o ou ism indus ies.
The in eg a ion o cus ome insigh s om big da a
in o ma ke segmen a ion decision-making enables
businesses o a ge hei cus ome s wi h unp eceden -
ed accu acy. Acco ding o X. Li & Y.S. Lee (2024), big
da a allowed companies o mo e beyond s a ic demo-
g aphic o geog aphic segmen a ion by in eg a ing e-
al- ime beha iou al and p e e ence da a in o hei seg-
men a ion models. This helped in he de elopmen o
cus omised ma ke ing s a egies ha will lead o mo e
sa is ac ion and loyal y among cus ome s. Acco ding o
Z.Zhang(2024), neu al ne wo k-based models played a
c ucial ole in e alua ing segmen a ion and ma ke ing
s a egies, showing how hese echnologies will help o
op imise segmen a ion o be e esul s.
Cus ome beha iou analysis helped o iden i y he
pa e ns and e e y s ep in ol ed in eaching pu chasing
decisions. Acco ding o S.Ga g & A.Khandha (2024),
consume beha iou analysis indeed p o ided an op-
po uni y o imp o e businesses h ough adjus men s o
ma ke s a egies wi h da a-d i en insigh s on pu chas-
ing ends and spending pa e ns. The ela ionship be-
ween beha iou analysis and ma ke segmen a ion was
implemen ed h ough he iden i ica ion o di e en be-
ha iou al ends ac oss a ious segmen s o cus ome s.
R.Y.Daulaye al.(2024) indica ed ha p e e ence analysis
played a signi ican ole in he iden i ica ion o cus ome
pe sonas and he elabo a ion o co esponding ma ke -
ing s a egies; his may e e o he use o K-Means clus-
e ing o segmen consume s o he co ee shop ma -
ke by li es yle, p e e ence, and pu chasing beha iou .
Cus ome eedback analysis is he p ocess o sys-
ema ic collec ion and analysis o cus ome opinions,
e iews, and expe iences abou a p oduc o se ice.
Th ough he iden i ica ion o common hemes in cus om-
e eedback, businesses can add ess pain poin s and e-
ine p oduc o e ings o imp o e cus ome sa is ac ion
and loyal y. A.Gopakuma e al.(2024) no ed ha he
in eg a ion o consume beha iou analysis wi h clus e -
ing algo i hms was e y impo an o enhance segmen-
a ion s a egy in bo h e-comme ce and con en ional
e ail se ings. Acco ding o Z.Sami ae al.(2024), CRM
(Cus ome Rela ionship Managemen ) in eg a ed wi h
AI-d i en ools helped SMEs o op imise hei ma ke -
ing s a egies o be e cus ome in e ac ion. The basic
p inciple o CRM was ha companies c ea e mo e alue
o hemsel es, when hey ocused on cus ome e en-
ion a he han sho - e m sales, wi h cus ome insigh
d i ing pe sonalisa ion and e ec i e segmen a ion.
In he con ex o SMEs, CRM heo y explained how
da a exploi a ion o inc eased cus ome insigh would
lead o mo e accu a e ma ke segmen a ion. D.Gam-
ba(2022) explained ha se ice-o ien ed segmen a ion
models allowed he SME o il e ou only he p o i able
clus e s o cus ome s, achie e esou ce op imisa ion,
and adap o ope a ional ba ie s. A.Singhe al.(2024),
in es iga ed machine lea ning me hods, which used o
de elop a ma ke segmen a ion model using K-Means
clus e ing and a consume beha iou p edic ion mod-
el using a Random Fo es om la ge-scale e-comme ce
da ase s. The Random Fo es model was much be e in
p edic ing cus ome habi s in con as o he K-Means
clus e ing model. The s udy included p ecision, ecall,
and F1 sco es as assessmen c i e ia. Acco ding o he
indings, machine lea ning echniques will be use ul in
ma ke analysis, and businesses may u ilise his mod-
el o c ea e e icien ma ke ing plans and comp ehend
cus ome beha iou .
K.Kuma e al.(2025) in es iga ed he cha ac e is ics
ha de e mine cus ome p e e ence o OTT (O e he
Top) ideo s eaming. Logis ic eg ession was applied
o he mul i a ia e analysis o su ey da a a o dabili y,
quali y, and accessibili y ha e eme ged as c ucial ac o s
a ec ing p e e ence. Demog aphic da a also played an
impo an ole in subsc ip ion decisions. M.M.Ib ahim
& H.A. Mamdouh (2025) in es iga ed he in luence o
online cus ome e iews on consume buying decision.
The me hodological app oach included a quan i a i e
online su ey conduc ed in Egyp , ocusing on OCR (On-
line Cus ome Re iews) dimensions such as alence and
olume. The indings indica ed ha online e iews ha e
an immense impac on pu chasing decisions, hough
mode a ed by demog aphics. Howe e , no e idence has
been obse ed linking cus ome eedback wi h ma ke
segmen a ion. I was concluded ha hough OCRs im-
pac pu chasing decisions, i s use ulness in ma ke seg-
men a ion equi es mo e esea ch.
Ma e ials and Me hods
The esea ch ques ions de i ed om he speci ic s udy
objec i es led o he use o a quan i a i e esea ch de-
sign, which used a su ey echnique. The main ques ions
we e: 1)how does cus ome beha iou analysis a ec s
ma ke segmen a ion decisions; 2) in wha ways does
cus ome p e e ence analysis a ec ma ke segmen a-
ion decisions; 3)wha is he e ec o cus ome eedback
analysis on ma ke segmen a ion decisions. This mean
ha ia su eys, nume ical da a was collec ed om small
business owne s o unde s and how hey use cus ome
da a o ma ke segmen a ion. This design was chosen
Olo a e al.
Economic Fo um, 2025, Vol. 15, No. 2
21
because i allowed ga he ing speci ic, measu able in o -
ma ion om many SME owne s and analysing i s a is-
ically o d aw eliable conclusions abou how big da a
insigh s in luenced hei ma ke segmen a ion decisions
(Ragab & A isha,2018).
Selec ion consis ed o 1,628 egis e ed SME own-
e s in Kwa a S a e, Nige ia, who main ained cus om-
e da abases and ha e de ined ma ke segmen s. The
2023 da ase o egis e ed SMEs main ained by he
Minis y o Comme ce & Co-Ope a i es (2025) p o id-
ed his igu e. These SMEs we e selec ed because hey
e lec ed companies ha ac i ely pa icipa e in o i-
cial ma ke segmen a ion p ocedu es and keep ack
o consume da a. To de e mine he sample size, he
s udy used T.Yamane’s(1969) equa ion o ini e popu-
la ion: n=N/(1+N(e)²), whe e: n– sample size; N– pop-
ula ion size (1,628); e– ma gin o e o (0.05); n–1,628/
(1+1,628(0.05)²); n=1,628/5.07; n=321 esponden s. To
accoun o po en ial non- esponses, he sample size
was inc eased by 10%, esul ing in 353 esponden s. The
s udy employed a sys ema ic andom sampling ech-
nique, whe e e e y i h SME owne om he alphabe -
ically a anged lis o 1,628 eligible SMEs was selec ed.
This me hod ensu ed unbiased selec ion, while main-
aining ep esen a i eness ac oss di e en business
sec o s and sizes. The sys ema ic app oach p o ided a
s uc u ed way o selec ing pa icipan s, while main ain-
ing andomness.
The uni o inqui y was indi idual SME owne s o
manage s who: 1)ha e egis e ed businesses in Kwa a
S a e; 2)main ained cus ome da abases wi h a leas
one yea o da a; 3)ha e implemen ed some o m o
ma ke segmen a ion in hei business ope a ions. This
speci ic ocus ensu ed ha esponden s ha e ele an
expe ience wi h bo h da a managemen and ma ke
segmen a ion p ac ices. The uni o analysis was he
indi idual SME owne ’s esponses ega ding hei use
o cus ome da a insigh s o ma ke segmen a ion de-
cisions. This included hei p ac ices in da a collec ion,
analysis me hods, and how hey applied hese insigh s
o segmen hei ma ke s and make business decisions.
A s uc u ed ques ionnai e was used, di ided in o sec-
ions co e ing demog aphics, cus ome insigh , and
ma ke segmen a ion decisions. The ques ionnai e em-
ployed a 5-poin Like scale anging om “S ongly disa-
g ee” (1) o “S ongly ag ee” (5).
Pa ial Leas Squa es S uc u al Equa ion Modelling
(PLS-SEM) using Sma PLS 3.2.9 was employed o da a
analysis. I was used o examine he measu emen mod-
el and s uc u al model o he s udy da a. Con en alid-
i y was es ablished h ough expe e iew by h ee busi-
ness managemen p o esso s and wo SME consul an s.
Cons uc alidi y was assessed h ough con e gen a-
lidi y (AVE >0.5) and disc iminan alidi y (Fo nell La ck-
e ). Reliabili y was measu ed using composi e eliabili y
and C onbach’s Alpha ( h eshold >0.7). A pilo es wi h
35 SME owne s (10% o sample size) was conduc ed o
e ine he ins umen .
Resul s and Discussion
The analysis began wi h an o e iew o he esponse a e
o he adminis e ed ques ionnai e. A majo i y o pa ici-
pan s p o ided comple e and alid esponses, ensu ing
a solid ounda ion o he s udy’s indings. This s ong
le el o engagemen e lec ed he ele ance and cla i-
y o he esea ch ins umen . The subsequen analysis
included desc ip i e s a is ics and es s o no mali y, o -
e ing deepe insigh in o he da a dis ibu ion (Table1).
Table 2. Desc ip i e analysis and no mali y es
Table 1. Ques ionnai e adminis e ed esponse a e
Sou ce: de eloped by he au ho s
Validi y F equency Pe cen age Valid pe cen age Cumula i e pe cen age
Fully submi ed esponses 254 71.9 71.9 71.9
Remaining sample size 99 28.1 28.1 100.0
To al 353 100 100
Table 1 showed ha 71.9% o esponden s an-
swe ed he ques ionnai e comple ely and accu a ely o
which hei esponses we e alid o his s udy. The high
esponse a e helped o achie e eliable indings om
he s udy. The desc ip i e esul , which showed he
mean o he measu es o he s udy in depic ed in Ta-
ble2, also ga e he s anda d de ia ion o he measu es,
he no mali y es , including he ku osis and skewness.
Mean S anda d
de ia ion
Excess
ku osis Skewness Numbe o
obse a ions used
Cus ome beha iou analysis 1 3.236 1.383 -1.164 -0.305 254.000
Cus ome beha iou analysis 2 3.457 1.356 -0.928 -0.566 254.000
Cus ome eedback analysis 1 2.866 1.193 -0.873 -0.019 254.000
Cus ome eedback analysis 2 3.339 1.305 -1.022 -0.284 254.000
Cus ome p e e ence analysis 1 3.433 1.290 -0.911 -0.450 254.000
Cus ome p e e ence analysis 2 3.213 1.234 -0.730 -0.310 254.000
Gaining cus ome insigh s in big da a...
Economic Fo um, 2025, Vol. 15, No. 2
22
The s udy conside ed ma ke segmen a ion and
cus ome insigh s. A numbe o impo an indica o s
we e e alua ed, each o which p o ided insigh in o a
dis inc ace o he ma ke segmen a ion and consum-
e insigh s. The mean sco es, s anda d de ia ions, and
he numbe o obse a ions used o each indica o p o-
ided aluable insigh s and implica ions o esea che s
and p ac i ione s. The ela i ely high mean sco e, which
we e abo e 3 o he ques ions sugges ed ha espond-
en s pe cei e cus ome insigh s o be highly ele an o
ma ke segmen a ion decisions. Wi h low s anda d de i-
a ion in each cases, indica ing ha he e was low de ia-
ion o he esponses om he mean. These desc ip i e
esul s unde sco ed he mul i ace ed na u e o cus om-
e insigh s on ma ke segmen a ion decisions. They em-
phasised he impo ance o ma ke segmen a ion deci-
sions h ough success ul cus ome insigh s.
The no mali y esul s o he dis ibu ion e ealed
ha he sample size was abo e 100, which implied ha
an absolu e alue o skewness o +1.0 o below was
expec ed o he da a o be no mal. In addi ion, o
ku osis, an absolu e alue o ±3.0 was expec ed o a
no mal peak, as any alue ou side he h eshold could
be a se ious signal o conce n. The no mali y esul s
showed ha all he a iables we e wi hin he h eshold
o he absolu e alue o ±1.0 and he ku osis esul s
we e also wi hin he absolu e alue o ±3.0. The implica-
ion om he no mali y es esul s showed ha all he
da a inpu ed o he analysis we e no mally dis ibu ed
and can be used o u he analysis and in e ences. This
implied ha all he a iables used o measu e esou ce
op imisa ion ha e mode a e mean wi h low de ia ion
om he mean and he a iables we e all no mally dis-
ibu ed indica ing he use ulness o he a iables in de-
e mining he causali y be ween cus ome insigh s and
ma ke segmen a ion decisions. Fo his, he a iables
used o measu e cus ome insigh s we e cus ome be-
ha iou analysis, cus ome p e e ence analysis, and cus-
ome eedback analysis agains ma ke segmen a ion
decisions. Figu e 1 showed he s uc u al pa h model
ha assesses he e ec o cus ome insigh s on ma ke
segmen a ion decisions.
Mean S anda d
de ia ion
Excess
ku osis Skewness Numbe o
obse a ions used
Ma ke segmen a ion decision 1 3.780 1.380 -0.463 -0.900 254.000
Ma ke segmen a ion decision 2 3.929 1.393 -0.118 -1.122 254.000
Ma ke segmen a ion decision 3 3.874 1.403 -0.240 -1.065 254.000
Figu e 1. Model o he pa h o cus ome insigh s and ma ke segmen a ion decision
Sou ce: de eloped by he au ho s
Table 2, Con inued
Sou ce: de eloped by he au ho s
Cus ome beha iou
analysis 1
Cus ome beha iou
analysis 2
Cus ome p e e ence
analysis 1
Cus ome p e e ence
analysis 2
Cus ome eedback
analysis 1
Cus ome eedback
analysis 2
Cus ome
beha iou analysis
Cus ome
p e e ence analysis
Cus ome
eedback analysis
Ma ke segmen a ion
decision 1
Ma ke segmen a ion
decision 2
Ma ke segmen a ion
decision 3
Ma ke
segmen a ion
decision
Th ee independen a iables– analysis o consum-
e beha iou , p e e ences, and eedback– and one de-
penden a iable– choice o segmen he ma ke we e
included in he model. Acco ding o he model’s ind-
ings, ma ke segmen a ion decisions we e signi ican ly
in luenced a ou ably by all h ee independen ac o s.
This indica ed ha o ganisa ions should alue consum-
e insigh s since hey can aid in making be e ma ke
segmen a ion decisions. The pa icula impac s demon-
s a ed ha e e y independen a iable signi ican ly
Olo a e al.
Economic Fo um, 2025, Vol. 15, No. 2
23
in luences he choice o segmen he ma ke . This im-
plied ha in o de o imp o e ma ke segmen a ion de-
cisions, o ganisa ions should concen a e on c ea ing
consume insigh s. Impo an s a is ical indica o s pe -
aining o he alidi y and cons uc eliabili y o he ou
la en a iables in his s udy we e shown in Table3.
C onbach’s Alpha Composi e eliabili y A e age Va iance Ex ac ed (AVE)
Cus ome beha iou analysis 0.828 0.920 0.852
Cus ome eedback analysis 0.759 0.819 0.693
Cus ome p e e ence analysis 0.754 0.891 0.803
Ma ke segmen a ion decision 0.948 0.966 0.905
Table 3. Cons uc eliabili y and alidi y
Sou ce: de eloped by he au ho s
These me ics aid in e alua ing how well hese a i-
ables quan i y he undamen al ideas hey a e mean o
e lec . C onbach’s Alpha and composi e dependabili y
we e he wo main measu es used o assess cons uc
dependabili y. C onbach’s Alpha assesses a la en a i-
able’s in e nal consis ency by de e mining he ex en o
which each i em was ela ed o e e y o he i em. Good
quali y was shown by he in e nal consis ency sco es o
he ou la en a iables, which we e abo e 0.7. Since
hese alues we e a highe han he widely accep ed
cu o limi o 0.7, hey sugges ed ha he i ems wi hin
each a iable we e eliable ma ke s o he ela ed s uc-
u es. Composi e eliabili y was ano he cons uc elia-
bili y s a is ic ha conside ed bo h in e nal consis ency
and he ela ionships be ween he i ems and he la en
a iable. All o he a iables in his s udy showed s ong
composi e dependabili y, p o iding a mo e us wo hy
measu e o eliabili y, wi h all alues o e 0.7. The la en
a iables’ high alues sugges ed ha hey we e us -
wo hy p edic o s o he cons uc s hey s and o .
Table3 also displayed he A e age Va iance Ex ac -
ed (AVE), which e alua ed each la en a iable’s con e -
gen alidi y. The deg ee o which i ems in a a iable
measu e he same unde lying no ion and we e connec -
ed o one ano he was known as con e gen alidi y.
All o he AVE alues in he able we e highe han he
sugges ed cu o o 0.5. This sugges ed ha each la en
a iable’s i ems we e con e ging nicely and measu ing
hei espec i e cons uc s as a whole. The choice o
hese a iables as alid and dependable measu es in he
s udy was suppo ed by hei s ong composi e eliabil-
i y, high in e nal consis ency, and good con e gen a-
lidi y (Oseie al.,2024). S ong e idence o disc iminan
alidi y among he la en a iables– cus ome eedback
analysis, cus ome beha iou analysis, ma ke segmen-
a ion decision, and cus ome p e e ence analysis– was
shown by he indings o he disc iminan alidi y s udy
in Table 4. Whe he hese cons uc s we e sepa a e and
no s ongly associa ed wi h one ano he was de e -
mined by disc iminan alidi y.
Cus ome
beha iou analysis
Cus ome eedback
analysis
Cus ome
p e e ence analysis
Ma ke segmen a ion
decision
Cus ome beha iou analysis 0.923
Cus ome eedback analysis 0.723 0.833
Cus ome p e e ence analysis 0.642 0.601 0.896
Ma ke segmen a ion decision 0.743 0.691 0.729 0.952
Table 4. Disc iminan alidi y
Sou ce: de eloped by he au ho s
I was clea om examining he co ela ions be-
ween hese a iables ha he o -diagonal alues– he
co ela ions be ween o he a iables– we e signi ican -
ly lowe han he diagonal alues, which ep esen he
co ela ions o each a iable wi h i sel . This suppo -
ed he no ion ha each la en a iable was unique and
measu es a sepa a e ea u e o he o e all cons uc by
indica ing ha each la en a iable has a s onge ela-
ionship wi h i sel han wi h he o he cons uc s. Com-
pa ed o i s co ela ions wi h cus ome beha iou , cus-
ome eedback, and cus ome p e e ence analysis, he
ma ke segmen a ion decision has a s onge connec-
ion wi h i sel . In a simila ein, he connec ion be ween
cus ome p e e ence analysis and i sel was s onge
han ha be ween he o he ac o s. Howe e , his was
also ue o o he a iables in hei own con ex s. These
indings demons a ed ha a he han being me ely
a ious exp essions o he same unde lying cons uc ,
he la en a iables in his s udy we e measu ing unique
ideas. Gi en ha i success ully dis inguished be ween
hese c ucial elemen s – cus ome eedback analysis,
cus ome beha iou analysis, ma ke segmen a ion de-
cision, and cus ome p e e ence analysis– i appea ed
ha he measu emen model was app op ia e o he
goals o his in es iga ion.
This allowed e alua ing he independen a iable’s
co ela ion. The pu pose was o de e mine, whe he wo
independen a iables we e no associa ed and yielding
same esul s. In his s udy, he expec ed associa ion be-
ween he independen a iables was e alua ed using
Gaining cus ome insigh s in big da a...
Economic Fo um, 2025, Vol. 15, No. 2
24
he a iance in la ion ac o (VIF). The VIF alues o he
la en a iables pe aining o he choice o ma ke seg-
men a ion we e shown in Table5. Cus ome eedback,
cus ome beha iou , and consume p e e ence analysis
all ha e VIF alues ha we e much below he 10-poin
cu o , which was encou aging. I implied ha hese la-
en a iables do no exhibi signi ican mul icollinea i y.
Since he e was li le co ela ion be ween hese a ia-
bles, mul icollinea i y was no a majo p oblem, when
hey we e included in his s udy. The coe icien o de e -
mina ion, o R-squa ed, which is a measu e o a model’s
quali y o i , was displayed in Table6.
Cus ome
beha iou analysis
Cus ome
eedback analysis
Cus ome
p e e ence analysis
Ma ke segmen a ion
decision
Cus ome beha iou analysis 2.444
Cus ome eedback analysis 2.246
Cus ome p e e ence analysis 1.822
Ma ke segmen a ion decision
Sou ce: de eloped by he au ho s
Table 5. Inne VIF alues
Table 6. Coe icien o de e mina ion sco e
R-squa e R-squa e adjus ed
Ma ke segmen a ion decision 0.680 0.676
Sou ce: de eloped by he au ho s
App oxima ely 68.0% o he a iabili y seen in he
dependen a iable (ma ke segmen a ion decision) can
be explained by he independen o la en a iables in-
cluded in he model, acco ding o he ma ke segmen-
a ion decision model’s R-squa ed sco e o 0.680. This
sugges s ha he model cap u es and explained he
obse ed a ia ions in he buying expe ience. The co -
ec ed R-squa ed alue was 0.676. This esul s in a mo e
ca e ul e alua ion o he model’s deg ee o i . The mod-
i ied R-squa ed alue was almos he same as he con-
en ional R-squa ed alue, indica ing ha he inclusion
o he independen a iables in he model was unlikely
o cause o e i ing o excessi e complexi y. This implied
ha e en when conside ing any p oblems ela ing o
model complexi y, he explana o y powe o he model
was s ill s ong. Acco ding o he R-squa ed and mod-
i ied R-squa ed alues, he ma ke segmen a ion deci-
sion model explained ma ke segmen a ion decision
a iabili y a he well, and adding mo e la en a iables
does no seem o deg ade he model’s pe o mance. In
s a is ical analysis, he e ec size, which was commonly
ep esen ed as -squa e and shown in Table7, quan i-
ied he s eng h o he co ela ion o in luence o inde-
penden a iables on a dependen a iable.
Cus ome
beha iou analysis
Cus ome
eedback analysis
Cus ome
p e e ence analysis
Ma ke segmen a ion
decision
Cus ome beha iou analysis 0.152
Cus ome eedback analysis 0.064
Cus ome p e e ence analysis 0.246
Ma ke segmen a ion decision
Table 7. Assessmen o he e ec size ( 2)
Sou ce: de eloped by he au ho s
I e alua ed he impac sizes o se e al la en ac-
o s on ma ke segmen a ion decision. E e y inde-
penden a iable had a alue g ea e han 0.02, which
was ega ded as a mino e ec size. This implied ha
e e y a iable had a mode a e e ec size, meaning ha
each one had a disce nible in luence on he choice o
segmen he ma ke . Va iabili y in ma ke segmen a ion
decisions can be mode a ely explained by changes o
a ia ions in any o he ac o s. The null hypo hesis ha
cus ome insigh s ha e no disce nible impac on ma ke
segmen a ion decisions was es ed using he boo s ap
ou e coe icien analysis shown in Table8.
O iginal
sample (O)
Sample
mean (M)
S anda d
de ia ion (STDEV)
T-s a is ics
(|O/STDEV|) P- alues
Cus ome beha iou analysis ->
Ma ke segmen a ion decision 0.344 0.346 0.075 4.608 0.000
Cus ome eedback analysis ->
Ma ke segmen a ion decision 0.215 0.213 0.068 3.155 0.002
Table 8. Boo s apping esul s showing pa h coe icien o s uc u al model
Olo a e al.
Economic Fo um, 2025, Vol. 15, No. 2
25
Acco ding o he indings, ma ke segmen a ion de-
cisions we e signi ican ly impac ed by cus ome eed-
back, cus ome beha iou , and consume p e e ence
analyses as componen s o cus ome insigh s. An ex-
amina ion o he sequence om Cus ome eedback,
cus ome beha iou , and cus ome p e e ence analy-
sis o he decision abou ma ke segmen a ion e eals
a s a is ically signi ican ela ionship be ween hese
h ee ypes o analysis and he decision o segmen he
ma ke . S ong e idence o ejec he null hypo hesis
was sugges ed by he -s a is ics being mo e han 1.96
and he p- alues being less han he adi ional signi -
icance le el o 0.05. As a esul , he choice o segmen
he ma ke was g ea ly in luenced by he cha ac e is ics
o consume insigh s; cus ome eedback, cus ome be-
ha iou , and cus ome p e e ence.
The s udy de e mined he e ec o cus ome in-
sigh s on ma ke segmen a ion decisions, wi h he
hypo hesis being ha cus ome insigh s do no signi -
ican ly a ec ma ke segmen a ion decisions. The e-
sul s e ealed ha all h ee ac o s; cus ome eedback
analysis, cus ome beha iou analysis, and cus ome
p e e ence analysis, ha e s a is ically signi ican e ec s
on ma ke segmen a ion decisions. This inding aligned
wi h O.R.Amosue al.(2024), who demons a ed ha
eal- ime da a analy ics p o ides s a egic cus ome in-
sigh s c ucial o e ec i e e-comme ce segmen a ion.
Simila ly, M.E.Jalal & A.Elmagh aby(2024) ound ha
coun e ac ual analysis o e ed a new pe spec i e on
pe sonalised ma ke ing, enabling mo e p ecise cus-
ome segmen a ion. S.Pa ke al.(2024) suppo ed his
inding h ough hei impo ance-induced cus ome seg-
men a ion app oach using explainable machine lea n-
ing, which enhanced he accu acy o ma ke segmen-
a ion decisions. These s udies collec i ely emphasised,
how mode n analy ical echniques ans o med cus om-
e insigh s in o ac ionable segmen a ion s a egies.
The ejec ion o he null hypo hesis was u he sup-
po ed by bo h his o ical and con empo a y esea ch.
F.Qian(2008) es ablished a ounda ional unde s anding
o CRM and cus ome segmen a ion ou sou cing o small
and medium businesses, highligh ing he long-s and-
ing impo ance o cus ome insigh s in segmen a ion.
Mo e ecen ly, B.S.V.Reddye al.(2023) demons a ed
he e ec i eness o clus e ing algo i hms in cus ome
segmen a ion analysis, p o iding echnical alida ion
o he ela ionship be ween cus ome insigh s and seg-
men a ion decisions. L.Sanu(2024) p esen ed a p ac-
ical applica ion h ough Reliance Jio’s cus ome analy -
ics pla o m, showing how big da a can be collec ed o
meaning ul cus ome insigh s ha d i e segmen a ion
s a egies. D.K.Sha ma & M.Kuma (2023) con ibu ed
me hodological igou h ough hei ma ke segmen
e alua ion using g ey ela ional analysis, demons a ing
quan i a i e app oaches o ansla e cus ome insigh s
in o segmen a ion decisions.
The comp ehensi e e ec o cus ome insigh s
on ma ke segmen a ion was con ex ualised by P.Sin-
ghe al.(2023), who p o ided an in eg a i e e iew o
consume beha iou in he se ice indus y, es ablish-
ing he heo e ical ounda ion o why cus ome in-
sigh s ma e in segmen a ion decisions. The indings
o his s udy we e also consis en wi h he esul s o
M.K.Chaudha ye al.(2024), who ound cus ome be-
ha iou o be i al ma ke ing concep , M.Deng(2024)
ocused on cus ome p o iling o ma ke segmen a-
ion, and M.M. Ib ahim & H.A. Mamdouh (2025), who
demons a ed ha u ilising cus ome insigh s h ough
ad anced analy ics enabled businesses o c ea e mo e
p ecise cus ome p o iles, leading o mo e e ec i e seg-
men a ion s a egies. Th ough le e aging hese mul i-di-
mensional cus ome insigh s, o ganisa ions can de elop
highly a ge ed ma ke ing app oaches ha add essed
speci ic cus ome needs, signi ican ly imp o ing engage-
men and con e sion a es. This da a-d i en app oach
o segmen a ion allowed o con inuous e inemen and
adap a ion o changing ma ke condi ions, ensu ing sus-
ainable compe i i e ad an age in inc easingly dynamic
business en i onmen s.
Conclusions
In he dynamic landscape o eme ging ma ke s, small
and medium-sized en e p ises in Kwa a S a e, Nige ia,
ha e disco e ed he ans o ma i e powe o da a-d i -
en ma ke segmen a ion. Resea ch has highligh ed he
c i ical ole o cus ome insigh s de i ed om big da a
analy ics in shaping s a egic business decisions.
The comp ehensi e analysis in he s udy e ealed
h ee pi o al ac o s o cus ome insigh s ha sig-
ni ican ly impac ma ke segmen a ion: cus ome
eedback analysis, cus ome beha iou analysis, and
cus ome p e e ence analysis. These dimensions p o-
ided SMEs wi h a nuanced unde s anding o hei
a ge ma ke s, enabling mo e p ecise and e ec i e
s a egic posi ioning. Businesses may gain nume ous
signi ican bene i s by collec ing and analysing clien
da a in a sys ema ic manne . Fi s , i was de e mined
how o design highly ocused ma ke ing ac ics ha
esona e wi h ce ain clien ca ego ies. This app oach
allowed o mo e pe sonalised p oduc o e ings and
O iginal
sample (O)
Sample
mean (M)
S anda d
de ia ion (STDEV)
T-s a is ics
(|O/STDEV|) P- alues
Cus ome p e e ence analysis ->
Ma ke segmen a ion decision 0.379 0.379 0.057 6.654 0.000
Table 8, Con inued
Sou ce: de eloped by he au ho s
