Hao, Jingjing; Zou, Xin; Chen, Yu eng; Liu, Yuan
A icle
The in luence o e e ence knowledge on he digi al
se ice quali y and incen i e mechanism
Jou nal o Inno a ion & Knowledge (JIK)
P o ided in Coope a ion wi h:
Else ie
Sugges ed Ci a ion: Hao, Jingjing; Zou, Xin; Chen, Yu eng; Liu, Yuan (2025) : The in luence o
e e ence knowledge on he digi al se ice quali y and incen i e mechanism, Jou nal o Inno a ion
& Knowledge (JIK), ISSN 2444-569X, Else ie , Ams e dam, Vol. 10, Iss. 4, pp. 1-27,
h ps://doi.o g/10.1016/j.jik.2025.100745
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/327640
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-nc-nd/4.0/
The in luence o e e ence knowledge on he digi al se ice quali y and
incen i e mechanism
Jingjing Hao , Xin Zou , Yu eng Chen
*
, Yuan Liu
*
College o Economics and Managemen , Zhejiang No mal Uni e si y, 688 Yingbing Road, Wucheng Dis ic , Jinhua Ci y, Zhejiang P o ince, PR China
ARTICLE INFO
JEL codes:
M11
C73
M31
O33
M15
O31
Keywo ds:
Benchma k knowledge
Va ying weigh
Cos -sha ing incen i e
P ospec heo y
P incipal-agen model
ABSTRACT
In he digi al se ice supply chain, digi al se ice p o ide s unde ake he ma ke ing pla o m de elopmen
p og ams wi h sa is ying digi al se ice quali y (DSQ) o help adi ional e ail en e p ises (TREs) achie e digi al
echnologies and me chandizing inno a ion. To illus a e he in luence o ex e nal e e ence knowledge on
p incipals’ eeling on DSQ and imp o emen s, his s udy de elops a no el amewo k o assess TREs’ pe cei ed
DSQ and design incen i e s a egies conce ning di e en ial coope a ion scena ios. The s udy in eg a es he
p ospec heo y and he p incipal-agen model o e eal he DSQ incen i e s a egies in luenced by ex e nal
e e ences. I s con ibu ions include (1) p oposing DSQ app aisal indica o sys em and s anda dized compu a ion
p ocess, (2) e ealing he in luence o ex e nal e e encing knowledge on psychological u ili y in DSQ coope -
a ion, and (3) explo ing he incen i e amewo k and equilib ium o DSQ u ili y in he di e en ial deg ees o
in o ma ion asymme y. The manage ial implica ions can assis he TREs in selec ing digi al ma ke ing KPIs,
de e mining p ope benchma ks, con i ming he dynamical dominance o ex e nal e e ences, educing he de-
g ee o in o ma ion asymme y, and implemen ing e ec i e incen i es.
In oduc ion
In he digi al economy e a, he cus ome s’ shopping habi s and ac-
i i ies a e d ama ically changing o global e-comme ce sales ia mobile
de ices (Dolega e al., 2021). As an e ec i e business-d i en hand
(Azemi e al., 2022), digi al ma ke ing pla o m can p o ide he p o-
essional unc ions (e.g., p ecise p omo ion, eal- ime in e ac ion, and
da a secu i y) (Kashyap e al., 2025), enhancing cus ome s’ shopping
expe ience and e ailing b and epu a ion. In he end o digi al ans-
o ma ion, adi ional e ail en e p ises (TREs) ha e an u gen demand
o building he digi al ma ke ing pla o m o achie e ma ke ing inno-
a ion, such as da a mining (Royle & Laing, 2014), business ope a ion
(Reim e al., 2022), and sale p edic ion (De Caigny e al., 2020). Digi al
ma ke ing echnologies ha e become he d i ing o ce in ealizing
e ec i e p ecision ma ke ing (Chou e al., 2022), exac ma ke posi-
ioning (Palmi´
e e al., 2022), po en ial cus ome ecogni ion (Yang e al.,
2021), and clea cus ome segmen a ion (Kalia e al., 2022). Rega ding
pla o m compe i i eness, digi al se ice quali y (DSQ) p esen s he
cus ome s’ comp ehensi e sa is ac ion wi h he digi al ma ke ing ech-
nologies and is equi ed o de e mine he assessmen c i e ia and s an-
da diza ion measu emen .
In he de elopmen p ocess o digi al ma ke ing pla o m, TREs al-
ways coope a e wi h digi al se ice p o ide s (DSPs) o conduc join
de elopmen . As he agen , DSPs gua an ee he digi al pla o m wi h a
ce ain DSQ (Anand & Goyal, 2019). The DSQ can desc ibe he e ec o
join de elopmen . Mo e impo an ly, he TRE’s pe spec i e on DSQ is
signi ican ly in luenced by ex e nal e e ence poin s, such as he main
compe i o ’s DSQ and indus ial DSQ. I he TRE’s ecei ed DSQ is be e
han an ex e nal e e ence le el, i will gain addi ional posi i e psy-
chological bene i s caused by he leading posi ion. I he TRE’s ecei ed
DSQ is lowe han an ex e nal e e ence le el, i will expe ience addi-
ional nega i e psychological e ec s because o he lagging DSQ. The
in luence o ex e nal e e ence knowledge on he TRE’s pe cei ed DSQ
should be explo ed o show i s psychological u ili y. Pa icula ly in he
si ua ion o many e e ences, he a ying weigh s o di e en bench-
ma ks should be app op ia ely de e mined o desc ibe he agg ession
e ec .
In he digi al supply-chain coope a ion, he deg ee o in o ma ion
asymme y de e mines he pa icipan ’s dominan posi ion, which a -
ec s he DSQ incen i e s a egy selec ion. In digi al coope a ion, he
TRE is he p inciple and asks he DSP o de elop he digi al ma ke ing
pla o m wi h compe i i e ad an ages. As he agen , he DSP, i s ,
* Co esponding au ho s.
E-mail add esses: [email p o ec ed] (J. Hao), [email p o ec ed] (X. Zou), [email p o ec ed] (Y. Chen), [email p o ec ed] (Y. Liu).
Con en s lis s a ailable a ScienceDi ec
Jou nal o Inno a ion & Knowledge
jou nal homepage: www.else ie .com/loca e/jik
h ps://doi.o g/10.1016/j.jik.2025.100745
Recei ed 9 No embe 2024; Accep ed 27 May 2025
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
A ailable online 18 June 2025
2444-569X/© 2025 The Au ho (s). Published by Else ie España, S.L.U. on behal o Jou nal o Inno a ion & Knowledge. This is an open access a icle unde he
CC BY-NC-ND license (
h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/ ).
always de ises he ini ial decision-making on de elopmen in es men ,
which is o en in luenced by sel -in e es owa d i s own op imal s a -
egy (G eens one e al., 2022). La e , he TRE has o accep he DSP’s
in es men esul and inal digi al pla o m wi h DSQ, which leads o an
ad e se selec ion consequence (Pa lo e al., 2022). Because o in o -
ma ion asymme y, mo al haza d is gene a ed and may be de imen al
o he TRE’s in e es s (Zhang e al., 2023). Consequen ly, es ablishing an
e ec i e incen i e mechanism is a c ucial managemen solu ion o he
abo e e hical dilemma (Liu e al., 2022). Al hough p e ious s udies
con ibu ed o he di ec ansac ion e ec , he incen i e ocused on he
independen imp o emen . Acco ding o he p ospec heo y, he TRE’s
pe spec i e on DSQ is highly elian on ex e nal e e ences. The incen-
i e objec i e can be expanded o psychological u ili y.
To p e en he DSP’s mo al haza d in he de elopmen coope a ion,
his s udy me ges he p ospec heo y in o a p incipal-agen model
conside ing he TRE’s psychological u ili y ega ding compe i ion po-
si ion and in o ma ion asymme y. The p oposed cos -sha ing incen i e
me hod can assis he TRE in achie ing sa is ied pe cei ed DSQ in he
digi al se ice supply chain. As indica ed in Figu e 1, he s eps o he
s udy a e indica ed o highligh i s gene al pe spec i e.
The con ibu ions o he esea ch include he ollowing
(1) The de ini ion and a se o indica o s o inno a ion-d i en DSQ
ha e been es ablished, p o iding a ounda ional analy ical
amewo k o measu ing digi al ma ke ing pe o mance.
(2) How ex e nal e e ence poin s a ec he TRE’s psychological
u ili y is examined wi h a ocus on op imis ic p e e ences.
(3) The incen i e solu ion unde symme ic and asymme ic in o -
ma ion si ua ion is explo ed.
The con en s o he s udy a e o ganized as ollows. Sec ion 2 sum-
ma izes exis ing esea ch and esea ch gaps. Sec ion 3 p esen s he
heo e ical amewo k o he esea ch. In Sec ion 4, he de ini ion o
DSQ and he pe cei ed u ili y conside ing dual e e ence poin s a e
p esen ed, and he pe cei ed DSQ unde a single e e ence poin is
gi en. In Sec ion 5, he incen i e mechanism is conduc ed h ough he
“p incipal-agen ” heo y. In Sec ion 6, he incen i e solu ion is p o ided
unde symme ic and asymme ic in o ma ion si ua ions. In Sec ion 7, a
nume ical s udy is conduc ed o p o e he e ec i eness o he abo e
me hod. Sec ion 8 p o ides he esea ch conclusions and u u e esea ch
opics.
Li e a u e e iew
The s udy conside s he o e lapping digi al ma ke ing domains,
e e ence poin s, and incen i e me hods. Consequen ly, his sec ion
p o ides a b ie o e iew o he ollowing h ee aspec s. The subsec ion
desc ibes he esea ch gap o p ecisely assessing he TRE’s o e all
psychological u ili y and designing ela ed incen i e solu ions.
The impac o digi al echnology on digi al ma ke ing
In he new e ail e a, eme ging digi al echnologies a e apidly
changing he ma ke ing en i onmen in he a eas as consume beha io
(Ra ch o d e al., 2022), social media wi h use -gene a ed con en
(Babi´
c e al., 2020; ˇ
Sola e al., 2022), digi al ma ke ing pla o ms (Veile
e al., 2022) and online sea ching s a egy (Lin e al., 2020; Agniho i,
2020)). Ra ch o d e al. (2022) examined how digi al echnology
a ec ed consume s’ sea ch cos s and sea ch beha io h ough online
shopping, which e ealed ha he In e ne sho ened he cus ome s’
conside a ion ime. Babi´
c e al. (2020) explo ed he gene a ion p ocess
o consume s’ elec onic wo d-o -mou h in social media ela ed o
associa e mone a y alue. Veile e al. (2022) conduc ed an explo a o y
nume ical s udy o analyze how digi al pla o ms changed indus ial
i ms’ business models and ma ke ing s a egies. Lin e al. (2020)
e ealed ha paid sea ch engines could p omo e use s’ pu chasing e-
quency and inc ease cus ome s’ li e ime alue by e ec i ely iden i ying
high- alue cus ome s. The abo e ela ed esea ch con ibu es o un-
de s anding he a ious digi al ma ke ing p ocesses a ec ed by ech-
nology and p o ides he suppo o iden i ying he ele an DSQ
indica o s ha e lec he cu en digi al ma ke ing landscape.
The ole o psychological u ili y and weigh ing me hods in he supply chain
P io esea ch in his domain p o ides a solid ounda ion o be e
unde s anding how e e ence knowledge impac s supply chain ac i -
i ies, p o iding a basis o inco po a ing psychological ac o s in o he
TRE’s economic bene i .
Figu e 1. S udy s eps.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
2
Psychological u ili y in supply-chain coope a ion unde e e ence poin s
In supply-chain coope a ion, e e ence poin s signi ican ly in luence
he pa icipan s’ psychological compe i i e u ili y, which is widely used
in o de ing decisions (Wei e al., 2019; Uppa i & Hasija, 2019)), p icing
s a egies (Das e al., 2021), supply-chain coo dina e ac ics (Qiu e al.,
2022; Vipin & Ami , 2021), isk con ol (Liu & Chen, 2019), supplie
selec ion (Chai e al., 2023), and supply-chain pe o mance e alua ion
(Liang e al., 2022). Wei e al. (2019) s udied he impac o ex e nal
indus y le el and minimum p o i expec a ions on he newsboy decision
p oblem. Das e al. (2021) p oposed a manu ac u e ’s loss a e sion
model unde he e e ence le el o design a p icing s a egy in a g een
supply chain. Qiu e al. (2022) add essed he coo dina ion p oblem
conce ning dynamic e e ence poin s in online-o line channels wi h
di e en s uc u es. Liu and Chen (2019) in oduced a p o i e e ence
poin in o he decision-make s’ psychological pe cep ion o isk. Chai
e al. (2023) es ablished a sus ainable supplie selec ion decision
amewo k in which posi i e and nega i e ideal e e ence poin s we e
conside ed o desc ibe decision-make s’ isk p e e ences. Liang e al.
(2022) conside ed decision-make s’ bounded a ionali y unde andom
subjec i e e e ence poin s and cons uc ed a supplie pe o mance
e alua ion model in i e dimensions.
The weigh ing me hod o psychological compe i i e u ili ies unde mul iple
e e ence poin s
To conside he con ibu ions o mul iple e e ence poin s, he
weigh s o di e en e e ence poin s should be sui ably designed o
agg ega e he pa icipan s’ psychological compe i i e u ili ies. Exis ing
weigh ing me hods include subjec i e designa ion (Zhong e al., 2022;
Liang e al., 2020), addi i e me hod (Wei e al., 2019; Wang e al.,
2020), and a i ude e alua ion (Zhu e al., 2017). Zhong e al. (2022)
employed subjec i e equal weigh s o in eg a e psychological u ili y
wi h mone a y and ime e e ence poin s. Wei e al. (2019) simply added
he in luence o he bo om line and he s a us quo e e ence poin when
measu ing he psychological u ili y o newsboys. Zhu e al. (2017)
ea ed decision-make s’ a i udes owa d he e e ence poin s as he
coe icien o de e mining he weigh s.
The incen i e solu ions in supply chain coope a ion
In supply-chain coope a ion, he esea ch on incen i e mainly o-
cuses on he impac o incen i es on supply-chain pe o mance (Gao
e al., 2023), incen i e mechanism design (Chak abo y e al., 2019; Liu
e al., 2022), and in o ma ion managemen in incen i e sys em (Li &
Zhang, 2021; Fu & Xing, 2021). Gao e al. (2023) analyzed he esponse
decisions o he supply chain unde he incen i e s a egies, which
e ealed ha incen i es could enhance in en o y ca yo e capabili y in
decen alized supply chains. Chak abo y e al. (2019) add essed he
cos -sha ing mechanism be ween he e aile and he manu ac u e ,
which in ensi ied he alue o a cos -sha ing con ac on imp o ing
supply-chain pe o mance. Li and Zhang (2021) explo ed he in luence
o eal- ime in o ma ion and he pa icipan s’ o ecas ing abili y on he
design o incen i e mechanism in he supply chain, in which in o ma ion
acquisi ion could p omo e supply-chain membe s o dis o op imal
decisions. These s udies on incen i e solu ions in supply-chain coope -
a ion can p o ide a basic amewo k o us o design mo e e ec i e
incen i e mechanisms ha conside comp ehensi e psychological u ili y
and ex e nal DSQ e e ence poin s in ou esea ch.
In summa y, exis ing esea ch has e ealed a ious incen i e
me hods o imp o ing he pa icipan s’ economic bene i s and mi i-
ga ing he agen s’ mo al haza d. Howe e , in he digi al economy e a,
he p incipal’s psychological eeling on DSQ is signi ican ly in luenced
by he ex e nal e e ence knowledge and u u e po en ial. The incen i e
ocus should be ex ended om single economic bene i o comp ehen-
si e psychological u ili ies, conside ing he e ec o ex e nal DSQ
e e ence poin s and a ying weigh s. Fi s , he de ini ion, c i e ia, and
compu a ion p ocess o DSQ should be upda ed acco ding o he ea u e
o digi al economy. Second, he TRE’s psychological u ili y ega ding
he DSQ needs o be easonably measu ed o illus a e he in luence o
ex e nal benchma ks. Mo e impo an ly, he explo a ion o psycholog-
ical u ili y was no employed in he p incipal-agen amewo k. In dig-
i al se ice coope a ion, he DSP p io i izes i s own bene i s because o
in o ma ion asymme y, which may be de imen al o he TRE’s in-
e es s. The deg ee o in o ma ion asymme y can a ec he equilib ium
o he p incipal-agen analysis. Consequen ly, an incen i e me hod,
which easonably e alua es he TRE’s psychological u ili y and
Figu e 2. Theo e ical amewo k.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
3
conside s he deg ee o in o ma ion asymme y, should be explo ed o
di ec ly suppo he se ice supplie ’s managemen ac i i ies.
Theo e ical amewo k
This pape selec s he digi al se ice supply chain o explo e he
incen i e solu ion o imp o ing DSQ wi h ega d o dual ex e nal e -
e ences and a ying weigh s. Combining digi al ma ke ing (Ra ch o d
e al., 2022), he p ospec heo y (Wei e al., 2019), and weigh ing
(Zhong e al., 2022) and incen i e me hods (Liu e al., 2022), he
heo e ical amewo k is p oposed in Figu e 2. The logic o he s udy is
de eloped as “Ac ual DSQ and pe cei ed DSQ” → “Agg ession o
comp ehensi e DSQ wi h a ying weigh ” → “p incipal-agen ame-
wo k” → “Incen i e solu ions in speci ic scena ios” → “Manage ial im-
plica ions.” In pa icula , he no el equi emen s o digi al ma ke ing
p o ide he guidance on designing he de ini ion and po en ial in-
dica o s o DSQ, which is he answe o he i s esea ch ques ion. Nex ,
he p ospec heo y is used o ans e he DSQ o he TRE’s pe cei ed
DSQ, which conside s ex e nal e e ence poin s and a ying weigh s o
demons a e u u e po en ials, as he explo a ion o he second esea ch
ques ion. Fu he mo e, o ul ill he hi d esea ch gap, he
p incipal-agen model is es ablished o desc ibe he imp o emen
equi emen o he TRE’s pe cei ed DSQ, and he incen i e heo y ex-
plains he cos -sha ing incen i e solu ions o p e en he DSPs’ mo al
haza d caused by in o ma ion asymme y. The TRE can use he incen i e
solu ions o design sui able supply-chain con ac o sol e he ad e se
selec ion p oblem.
S age 1: Ac ual DSQ. Re e ing o he digi al ma ke ing e-
qui emen s, i s , DSQ is de ined, and mul iple indica o s in he di-
mensions o digi al angibles, digi al us , digi al in e ac ion, cus ome
cen ici y, and eliabili y a e p oposed as he KPIs (Büyük¨
ozkan e al.,
2020). Second, he aw pe o mance da a, qi, as well as he expec a ions
and ole an in e als o KPIs, can be ob ained. Nex , he KPIs can be
di ided in o he smalle - he-be e , he la ge - he-be e , and he
nominal- he-be e , which p o ides a amewo k o s anda dizing and
agg ega ing he DSQ (Taguchi, 1986). A he echnology le el, A i icial
In elligence (AI) can be u ilized o de e mine app op ia e DSQ indica o s
decons uc ed om he TRE’s ision and mission. AI can handle a la ge
amoun o da a and iden i y he key ac o s ela ed o DSQ, which is in
line wi h he insigh s in o ma ke and cus ome needs in he ma ke ing
heo y. To a oid human ope a ing e o , Machine Lea ning (ML) could
be employed in he da a s anda dized p ocess and ensu e au oma ed
compu a ion. A de ailed calcula ion p ocess is conduc ed in Sec ion 4.1.
S age 2: Comp ehensi e DSQ unde dual e e ence knowledge.
Guided by he p ospec heo y (Ba be is, 2013), he TRE’s alue unc-
ion unde a single e e ence knowledge is de eloped. Addi ionally,
op imis ic p e e ence, ki, is p oposed o e lec he TRE’s a i ude on he
u u e imp o emen po en ial o DSQ. Conside ing he ie ce ma ke
compe i ion (Llopis-Albe e al., 2021), he main compe i o ’s DSQ, c,
and indus ial DSQ, h,a e selec ed as dual e e ences poin s. Mo eo e ,
o calcula e he TRE’s comp ehensi e pe cei ed DSQ, he in luence o
op imis ic p e e ence and a ying DSQ dis ances a e in eg a ed, which
induces dynamic weigh in he agg ega ion p ocess. The p ospec heo y
ocuses on an indi idual’s eliance on e e ence poin s when making
decisions and hei di e en pe cep ions o gains and losses. In his
s age, da a mining can be applied o ob ain uns uc u ed knowledge (e.
g., indus y su ey and compe i o ’s ope a ion epo ) online, iden i-
ying “indus ial-le el” and “compe i o -le el” DSQ benchma ks.
Th ough da a mining, speci ic and quan i iable e e ence poin s a e
ound, making he applica ion o he p ospec heo y in DSQ e alua ion
mo e speci ic and ope able, and achie ing a deep in eg a ion o heo y
and cu ing-edge echnologies in a quan i a i e pe cep ion o DSQ.
Sec ion 4.2 p esen s he abo e speci ic ope a ions.
S age 3: P incipal-agen amewo k and incen i e mechanism.
Fi s , based on he comp ehensi e calcula ion o he TRE’s pe cei ed
DSQ, i s psychological u ili y can be ob ained, which conce ns he in-
luence o ex e nal e e ence poin s wi h economic bene i s. Second, a
p incipal-agen model con aining incen i e compa ibili y cons ain s
and indi idual a ionali y cons ain is p oposed. Nex , he incen i e
p ocess o cos -sha ing is designed o explo e he op imal solu ion o he
coope a ion. In his s age, ein o cemen lea ning can be applied o
dynamically op imize incen i e pa ame e s (e.g., cos -sha ing a ios) in
he “p incipal-agen ” model by analyzing eal- ime da a on ma ke
condi ions and psychological u ili y p e e ences. The ial- eedback-
adjus men cycle can e ine he ela ed incen i e s a egies h ough
i e a i e imp o emen and main ain obus ness ac oss di e en ial sce-
na ios con aining isk p e e ence shi o compe i i e in ensi ica ion.
This enables he “p incipal-agen ” heo y o adap o he dynamically
changing ma ke en i onmen in p ac ical applica ions and ans o m
he abs ac incen i e mechanism in o speci ic s a egies ha can be
adjus ed and op imized in eal ime. The de ailed ope a ions a e illus-
a ed in Sec ion 5.
S age 4: Incen i e solu ions in speci ic scena ios. Acco ding o
he deg ee o he DSP’s p i a e in o ma ion, he op imal solu ions in
di e en supply-chain dominan condi ions a e discussed. Fi s , unde
he condi ion o comple e symme ic in o ma ion, he TRE has a domi-
nan posi ion in he supply-chain coope a ion, which p io i izes
Table 1
Va iables and hei meanings.
Va iable Meaning Va iable Meaning
Y DSQ cha ac e is ic ylThe minimum bounda ies o
he ole ance in e al o
cha ac e is ic
μ
The op imal a ge alue
o cha ac e is ic
yuThe maximum bounda ies
he ole ance in e al o
cha ac e is ic
Ac ual DSQ le el qLThe s anda dized quali y
le el o he la ge- he-be e
qSThe s anda dized quali y
le el o he small- he-
be e
qNThe s anda dized quali y
le el o he nominal- he-
be e
ω
iThe weigh o indica o in
DSQ
k The op imis ic p e e ence
d The dis ance be ween
ac ual DSQ and e e ence
DSQ
P( )TRE’s pe cei ed DSQ
hThe a e age DSQ le el in
he indus y
min The smalle o he wo
e e ence poin s
cThe main compe i o ’s
DSQ le el
max The g ea e o he wo
e e ence poin s
k1The op imis ic p e e ence
o min
P1The pe cei ed DSQ unde
min
k2The op imis ic p e e ence
o max
P2The pe cei ed DSQ unde
max
λThe weigh o he
e e ence poin in TRE’s
Pe cei ed DSQ
α
The isk p e e ence
coe icien
βThe isk a e sion
coe icien
θThe loss a e sion coe icien
λ1The weigh o min d1The dis ance be ween ac ual
DSQ and min
λ2The weigh o max d2The dis ance be ween ac ual
DSQ and max
PI
iThe comp ehensi e
pe cei ed DSQ when ≥
max > min
λI
iThe weigh o he
i h e e ence poin when
≥ max > min
PII
iThe comp ehensi e
pe cei ed DSQ when <
min < max
λII
iThe weigh o he
i h e e ence poin when
< min < max
PIII
iThe comp ehensi e
pe cei ed DSQ when
min ≤ < max
λIII
iThe weigh o he
i h e e ence poin when
min ≤ < max
φ( )Value-added bene i s o
TRE
φ[P( )] Pe cei ed alue-added
bene i s o TRE
∅( )Incen i e ee B Fixed ee
CS( )De elopmen cos o DSP V[
π
T( )] TRE’s bene i unc ion o
pe cei ed DSQ
V[
π
S( )] DSP’s bene i unc ion o pe cei ed DSQ
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
4
achie ing i s expec ed psychological u ili y. Second, he cos -sha ing
incen i e solu ion o he “p incipal-agen ” model in he case o com-
ple e asymme ic in o ma ion is analyzed. Nex , unde pa ly asym-
me ic in o ma ion, he equilib ium o he ansac ion is explo ed whe e
he pa icipan s a e igh ing o he dominance. The incen i e heo y is
cen e ed on de ising e icien incen i e mechanisms o d i e agen s
owa d ul illing p incipals’ objec i es. In his s age, he blockchain
echnology is ha nessed o ensu e he anspa ency o in o ma ion low
by eco ding unal e able ansac ion da a, which se e as a dependable
basis o e alua ing he le el o in o ma ion symme y. Blockchain can
be le e aged o ensu e anspa en in o ma ion low by eco ding
immu able ansac ion da a. Assessing he deg ee o in o ma ion sym-
me y (e.g., symme ic, pa ially asymme ic) could enable ampe -
p oo alida ion on DSP pe o mance and educe mo al haza d in
incen i e s a egy design. The calcula ion p ocess can be ound in Sec-
ion 6.
S age 5: Manage ial implica ions. A e he empi ical s udy, some
policy implica ions, such as, he selec ion o DSQ KPIs, de e mina ion o
compe i i e benchma ks, adop ion o a ying weigh s and designa ion
o incen i e solu ions, a e p o ided, which can di ec ly help he TRE o
enhance he DSQ and imp o e he ope a ion pe o mance.
In he ollowing sec ions, he main e minologies and no a ions used
in his pape a e summa ized in Table 1.
Digi al se ice quali y and pe cei ed u ili y
As shown in Figu e 3, his sec ion illus a es he de ini ion o DSQ
and i s compu a ion me hod in Sec ion 4.1, which con ains he
indica o s, s anda diza ion and agg ega ion. The pe cei ed DSQ con-
ce ning a e e ence poin and he op imis ic p e e ence a e pu o wa d
in Sec ion 4.2. The TRE’s comp ehensi e pe cei ed DSQ unde dual
e e ence poin s is calcula ed wi h ega d o he dynamic dominance o
ex e nal e e ence poin s in Sec ion 4.3.
Digi al se ice quali y and i s compu a ion me hod
Digi al se ice quali y and i s indica o s
T adi ional Se ice Quali y (SQ) p ima ily desc ibes he cus ome s’
sa is ac ion wi h he se ice (Chen e al., 2022). In he SERVQUAL
amewo k, SQ e lec s he cus ome s’ expec a ions and needs ega ding
he se ice us in o line channels (Ba akabi ze e al., 2019), which can
be measu ed wi h he ollowing i e dimensions: angibles, eliabili y,
esponsi eness, assu ance, and empa hy (Pa asu aman e al., 2002).
Because o he apid de elopmen o mobile In e ne , elec onic SQ
(e-SQ) ocuses on he cus ome s’ in e ac i e assessmen s on online SQ,
p ima ily e alua ing independen online se ice p ocesses (e.g., web
b owsing and online ansac ions) h ough In e ne echnologies wi h
me ics such as esponse ime and usabili y (Chao e al., 2024). E-SQ can
be measu ed in ou dimensions as ollows: e iciency, sys em a ail-
abili y, ul illmen , and p i acy (Chao e al., 2024). In he new e ailing
e a, he in eg a ion o online se ice is eme ging, and digi al ma ke ing
pla o m can p o ide sys ema ic digi al solu ions co e ing he o e all
ma ke ing li e cycle, which no only ocuses on echnological imple-
men a ion (such as pla o m s abili y) bu also emphasizes he deep
in eg a ion o echnology and business. Fo example, AI-d i en p ecision
ma ke ing push di ec ly imp o es he ma ke ing e u n on in es men
Figu e 3. Pe cei ed DSQ calcula ion p ocess.
Figu e 4. The e olu iona y oadmap.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
5
(ROI) in e ail en e p ises, a he han me ely op imizing online in e -
ac i e expe iences. DSQ ep esen s a holis ic e alua ion o end- o-end
digi al solu ions in he new e ail e a, ex ending o he en i e ma ke -
ing li e cycle om eal- ime da a collec ion and AI-d i en pe sonaliza-
ion o cus ome engagemen in inno a ion. The e olu ion o SQ, e-SQ,
and DSQ is illus a ed in Figu e 4.
To ealize digi al ans o ma ion om sepa a ed o line business o
sma e-comme ce, TREs always ou sou ce he de eloping asks o dig-
i al ma ke ing pla o ms o digi al echnology companies (Anand &
Goyal, 2019). In his digi al coope a ion, he adi ional SQ should be
expanded o sma sys em le el, hus, he no el comp ehensi e concep
o DSQ was bo n (Büyük¨
ozkan e al., 2020), emphasizing eme ging
digi al ma ke ing echnologies, such as eal- ime da a in eg a ion,
AI-d i en pe sonaliza ion, and alue co-c ea ion (Buyukozkan e al.,
2020). Di e en om he SERVQUAL ocusing on o line eliabili y and
esponsi eness (Collie & Biens ock, 2006), DSQ exhibi s a se ies o
digi al capabili ies, such as dynamic cus ome insigh s, ailo ed expe i-
ences, and cus ome engagemen in inno a ion.
De ini ion 1. DSQ e e s o use s’ subjec i e sa is ac ion in how digi al
se ices mee hei digi al needs. I is a dynamic cons uc ha no only
e lec s he in eg a ion e ec i eness in digi al ma ke ing li e cycle bu
also conce ns he ou comes o he digi al se ice.
The DSQ o digi al ma ke ing pla o m ep esen s he TRE’s sa is-
ac ion wi h he digi al se ice unc ions o a sys ema ic solu ion, such as
eal- ime da a collec ion, sma p e e ence ecogni ion, accu a e
cus ome po ai , p ecise ma ke ing push, and AI. Eme ging in o ma-
ion echnologies ha e made he mos impo an de elopmen s in he
ield o in e ac i i y h ough he shi ing om gene al mass ma ke ing o
indi idual p ecise ma ke ing (Demi el, 2022). Acco ding o use ex-
pec a ions on digi al pla o m, DSQ highly elies on ex ac ing in o -
ma ion om big da a o enhance ma ke ing pe o mance.
Fo example, du ing Alibaba’s Double Ele en campaign, DSQ man-
i es ed as he seamless alignmen o echnical pe o mance and cus ome
expe ience on i s digi al ma ke ing pla o m. Real- ime da a collec ion
ensu es sys em eliabili y and sub-3-second la ency. Sma p e e ence
ecogni ion uses ML models o e ec i ely p edic pu chase in en ion and
signi ican ly educe cus ome s’ acquisi ion cos s. P ecise ma ke ing
push se e s’ dynamic ad e isemen s o high- alue segmen s iden i ied
ia he Recency-F equency-Mone a y Value analysis. These DSQ-d i en
capabili ies no only align wi h cus ome expec a ions bu also demon-
s a e how big da a ex ac ion and AI in eg a ion con ibu e o
measu able business ou comes (Kamble & Gunaseka an, 2020; Elia e al,
2022).
Acco ding o Büyük¨
ozkan (2020), he indica o s o DSQ comp ise
i e dimensions as ollows: digi al angibles, digi al us , digi al in e -
ac ion, cus ome cen ici y, and eliabili y. Some indica o s e lec he
applica ion o digi al echnologies in he supply chain, such as a ic,
ime spen on page isi , cus ome in o ma ion asse s, and packe loss
a e. O he indica o s p esen he imp o emen o ma ke ing pe o -
mance because o he applica ion o digi al echnology, such as a e age
ansac ion alue, ROI, and cus ome acquisi ion cos .
The indica o s in he digi al angibles dimension ep esen digi ized
equipmen , acili ies, and hei digi al p ope ies (Büyük¨
ozkan e al.,
2019). In pa icula , unc ionali y e lec s he a ailabili y o digi al
channel and se ice cha ac e is ics (Chan e al., 2020; Melo i´
c e al.,
2021). E iciency e lec s he capabili y o p o iding sui able p oduc s
and in o ma ion wi h minimum e o (Liu e al., 2022; Va ada ajan,
2020).
The indica o s in he digi al us dimension ep esen he pe o -
mance and he s abili y o he digi al pla o m, which can be measu ed
by ne wo k pe o mance indica o s such as packe loss a e, ans-
mission delay, and h oughpu (Skaka-ˇ
Ceki´
c e al., 2023; Huang e al.,
2018; Alnawas & Al Kha eeb, 2022).
The indica o s in he digi al in e ac ion dimension conside he
digi al communica ion ne wo ks be ween companies and he supply
chain membe s h ough digi al pla o ms (Büyük¨
ozkan e al., 2019).
Collabo a ion and mobile communica ion ha e been selec ed in his
s udy o e lec he capabili y o digi al in e ac ion, which will be
enhanced using digi al echnology in he ma ke ing ac i i ies.
The indica o s in he cus ome cen ici y dimension ep esen an
“ou side-in” app oach h ough inno a i e se ice deli e y expe ience o
ul ill he cus ome ’s emo ional needs by pu ing hem a he hea o an
o ganiza ion (Büyük¨
ozkan e al., 2019). In pa icula , cus ome seg-
men a ion measu es he abili y o unde s and p ecisely he cus ome s’
p e e ences and shopping beha io , such as equency, ecency, and
mone a y (Si e al., 2015). Cus ome insigh s a e used o measu e he
ans o ming abili y om cus ome analysis in o ma ke ing
pe o mance.
The indica o s in he eliabili y dimension conside he ole o digi al
echnology in achie ing ma ke ing objec i es and educing inpu cos s
(Büyük¨
ozkan e al., 2019). Some inancial indica o s, such as ma ke ing
ROI, ma ke ing cos , and cus ome acquisi ion cos , a e selec ed as he
second-g ade indica o s. Some ma ke pe o mance indica o s, such as
in e na ional ma ke sha e, ade compe i i eness index, and e ealed
compa a i e ad an age index, a e included o e lec he imp o ed
pe o mance a e he ial ope a ion on he digi al pla o m.
This indica o sys em is designed o p o ide an e alua ion ame-
wo k o he co e ope a ional dimensions o digi al pla o ms. Howe e ,
as di e en en e p ises may ha e a ying s a egic p io i ies and ope -
a ional en i onmen s, o ganiza ions can adap o supplemen ele an
indica o s acco ding o hei de elopmen goals and business needs in
p ac ical applica ions, ensu ing ha he indica o sys em aligns closely
wi h s a egic objec i es.
No e: Ns: Numbe o cus ome s a he s a o he pe iod; Ne: Numbe
o cus ome s a he end o he pe iod; Na: Numbe o cus ome s acqui ed
du ing he pe iod; Nc: Numbe o chained cus ome s in he gi en pe iod;
N : To al numbe o cus ome s a he s a o he pe iod; Sc: o e all
ma ke ing campaign cos s spen on acquisi ion; S : ma ke ing eam
sala y; Ss: he cos o ma ke ing so wa e; So: o e head ela ed o ma -
ke ing (e.g. designe s, consul an s); CE: Company’s Expo ; TCE: To al
Company’s Expo ; GE: Global Expo ; TGE: To al Global Expo
S anda diza ion and agg ega ion o DSQ
To ensu e addi i i y and consis ency ac oss di e se DSQ indica o s
(wi h a ying uni s, anges, and de ini ions), da a s anda diza ion is
equi ed. Addi ionally, he e a e h ee kinds o quali y cha ac e is ics:
he la ge- he-be e (L ype), he small- he-be e (S ype), and he
nominal- he-be e (N ype). Suppose ha he ole ance in e al o
cha ac e is ic Y is [yl,yu], in which yl and yu e e o he minimum and
maximum bounda ies,
μ
is he op imal a ge alue. Le
ε
be an in ini-
esimal posi i e numbe ex emely close o 0. The s anda dized quali y
le el o he la ge- he-be e , small- he-be e and nominal- he-be e
quali y cha ac e is ics, qL,qS and qN, can be designed as
qL=
⎧
⎪
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎩
y−yl
μ
−yl
,y∈ (yl,yu)
ε
,y=yl
0,y<yl
,qS=
⎧
⎪
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎩
yu−y
yu−
μ
,y∈ (yl,yu)
ε
,y=yu
0,y>yu
,
qN=
⎧
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎩
y−yl
μ
−yl
y∈ [yl,
μ
]
yu−y
yu−
μ
y∈ (
μ
,yu)
ε
y=ylo yu
0 y ∕∈ [yl,yu]
(1)
Fo agg ega ion, le n DSQ indica o s ha e weigh s
ω
i(
ω
i≥0,
∑n
i=1
ω
i=1). The DSQ can be comp ehensi ely desc ibed as he
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
6
weigh ed sum o qL, qS and qN.
=⎧
⎨
⎩∑
n
i=1
ω
iqi,∀qi∕= 0
0,∀qi=0
(2)
He e, ∈ [0,1]. I any quali y pe o mance o an indica o is no
loca ed in he quali y in e al, =0. Only i he pe o mances o all he
quali y cha ac e is ics mee he quali y equi emen >0. Especially, i
any me ic is loca ed a he bounda y, =
ε
→0+.
TRE’s pe cei ed DSQ and op imis ic p e e ence conce ning dual DSQ
e e ence knowledge
TRE’s pe cei ed DSQ based on p ospec heo y
Acco ding o he p ospec heo y, he pa icipan ’s gain-loss s a e
signi ican ly impac s psychological u ili y, especially when i aces
e e ence le el (Ba be is, 2013). The essence o a use ’s pe cei ed u ili y
is he ou come o he psychological accoun ing p ocess de e mining he
gain/loss in e al h ough e e ence poin s and an asymme ic alue
unc ion (Tian e al., 2022). Di e en om he “absolu e u ili y maxi-
miza ion” assump ion, pe cei ed u ili y in he p ospec heo y e lec s
he di e en ial psychological eelings ega ding bene i s and losses (Jin
e al., 2024). In his s udy, TREs’ pe cei ed DSQ ep esen p incipals’
subjec i e u ili ies—gain and loss—because o he dynamic compa ison
be ween DSQ and he ex e nal e e ence poin s.
In he digi al se ice supply chain, he p ospec heo y can e eal
how ex e nal e e ence knowledge in luence TREs’ pe cei ed DSQ. TREs
may ha e e e ence knowledge based on pas expe iences, indus y
no ms, o an icipa ed u u e scena ios. Consequen ly, he p ospec he-
o y can help TREs gain deepe insigh s in o he psychological dynamics
d i ing he p e e ences and choices. The p ospec heo y is essen ially
applied o model TREs’ p e e ences o adop ing digi al echnologies in
p ecision ma ke ing in he digi al e ailing e a. Consequen ly, pe cei ed
DSQ is concep ualized as he TREs’ subjec i e e alua ion o he digi al
pe o mance de i ed om digi al ma ke ing pla o ms, caused by he
cogni i e compa isons be ween ac ual DSQ ou comes and ex e nal
e e ence poin s. Guided by he p ospec heo y (T e sky & Kahneman,
1981), he o ma ion o pe cei ed DSQ in ol es wo in e dependen
psychological elemen s as ollows: e e ence dependence and loss
a e sion. Re e ence dependence means ha he TRE e alua es he DSQ
no only in absolu e digi al pe o mance bu also ela i e o ex e nal
knowledge as benchma ks. Loss a e sion shows ha nega i e de ia ions
om e e ence knowledge can gene a e s onge psychological impac s
han equi alen gains.
In pa icula , i he TRE’s DSQ is la ge han an ex e nal e e ence, i
will ob ain he addi ional posi i e psychological bene i caused by
leading posi ion, and ice e sa. Addi ionally, he dis ances be ween
ac ual DSQ and ex e nal e e ence poin s can show he u u e de elop-
men space and in luence he TRE’s pe cei ed DSQ as well. I he ac ual
DSQ is lowe han a e e encing le el, he TRE’s pessimis ic pe spec i e
damages i s pe cei ed DSQ. As he gap is con inuously changing, he
TRE’s op imis ic a i udes ega ding ex e nal e e ence poin s a e
dynamically swi ching, which c ea es he a ying dominance o ex e nal
e e ence poin s.
I he TRE’s ecei ed DSQ le el is , le us assume ha he DSQ
e e ence poin , such as indus ial DSQ o compe i o ’s DSQ, is 0. The
TRE’s pe cei ed DSQ unde he in luence o DSQ e e ence poin 0 is as
ollows:
P( ) = {( − 0)
α
, 0≤ ≤1
−θ( − 0)β,0≤ < 0(3)
Acco ding o empi ical da a,
α
=β=0.88, and θ=2.25 (Ba be is,
2013). The ela ionship be ween TRE’s pe cei ed DSQ, P( ), and DSQ
le el, , can be ep esen ed as shown in Figu e 5, le P( = 0) = 0.
I he e is no DSQ e e ence poin , assume ha TRE’s psychological
u ili y is a linea unc ion L( ),
∂
L
∂
>0, L( 0) = 0. As illus a ed by
Figu e 5, he DSQ e e ence poin enhances he TRE’s pe cei ed DSQ
b ough by he gap be ween ac ual quali y and e e enced le el 0. In
pa icula , when = 1< 0, he isk a e sion coe icien , β, will le TRE
eel mo e dep essed due o he backwa d si ua ion, P( 1)<L( 1). Once
= 2> 0, he isk p e e ence coe icien ,
α
, will le TRE be mo e p oud
because o quali y leading, P( 2)>L( 2).
TRE’s op imis ic p e e ence-d i en dynamic agg ega ion o pe cei ed DSQ
In addi ion o i s e e ence poin s, he TRE exhibi s an op imis ic
p e e ence—a beha io al p opensi y o p io i ize u u e de elopmen
po en ial o e cu en pe o mance gaps, which can be o malized
h ough he second de i a i e o he alue unc ion
De ini ion 2: Op imis ic p e e ence is he decision-make s’ posi i e
pe cep ion o u u e de elopmen , which is induced by he gaps be-
ween ac ual pe o mance and e e ence le els. The TRE’s op imis ic
p e e ence can be designed as ollows:
k=
∂
2P( )
∂
2(4)
Figu e 5. TRE’s psychological u ili y cu e.
Figu e 6. The TRE’s alue unc ion unde double e e ence poin s.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
7
Suppose ha indus ial DSQ e e ence is h and he compe i i e DSQ
e e ence is c, max =max{ h, c}, min =min{ h, c}. As shown in
Figu e 6, TRE’s pe cei ed DSQ caused by min and max a e P1( )and
P2( ), which beha es he weigh as λ1 and λ2. The second de i a i e o
he alue unc ions k1, k2 a e
k1=
∂
2P1( )
∂
2,k2=
∂
2P2( )
∂
2(5)
The weigh s o DSQ e e ence poin a e ela ed o he op imis ic
p e e ence coe icien s, which e lec he TRE’s a i ude on he u u e
de elopmen po en ial o a ious DSQ e e ence poin s. I he TRE is in
he lagging posi ion, i.e. he pe cei ed DSQ is posi i e bu has a mo e
p omising u u e po en ial, i is mo e inclined o endu e cu en un-
pleasan e en s and alle ia e i s sense o app ehension, which leads o a
lowe weigh o his e e ence knowledge. I he TRE is in he dominance
compe i ion posi ion, i expe iences a mo e declining end, he
es ic ed oppo uni y o quali y de elopmen will esul in educed
p ospec s and less posi i e a i ude on pe cei ed DSQ. The changing
dominan posi ion o e e ence poin s c ea es a ying weigh s o in-
dus ial and compe i i e psychological u ili ies, which can be designed
as
λ1=|k2|
|k1| + |k2|,λ2=|k1|
|k1| + |k2|(6)
The comp ehensi e u ili y unc ion is
P( ) = ∑
i=1,2
λiPi=λ1P1+λ2P2
=
⎧
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎩
−θk2
k1+k2
( − min)β−θk1
k1+k2
( − max)β, < min < max
k2
−k1+k2
( − min)
α
+θk1
−k1+k2
( − max)β, min ≤ < max
k2
k1+k2
( − min)
α
+k1
k1+k2
( − max)
α
, ≥ max > min
(7)
When < min < max, he TRE’s DSQ is weake han all e e ence
le els, which gene a es dual nega i e u ili ies, P1= − θ( − min)β <0,
P2= − θ( − max)β<0. Acco ding o he dual e e ences min and max,
he co esponding e e ence weighs a e λ1=k2
k1+k2 and λ2 =k1
k1+k2.
Because he compa a i e smalle imp o ing space o P2 c ea es mo e
panic eeling, P2 occupies mo e dominan posi ion han P1, λ2 >λ1.
When min ≤ < max, he TRE’s DSQ is loca ed be ween he ex e nal
e e ences. The leading on min gene a es a posi i e u ili y, P1 =
( − min)
α
≥0 and he lag on max causes a nega i e u ili y, P2 = −
θ( − max)β<0. Addi ionally, he ela ed e e ence weighs a e λ1 =
k2
−k1+k2 and λ2=−k1
−k1+k2. When inc eases om min, because P2 c ea es
mo e panic eeling, P2 occupies mo e dominan posi ion han P1, λ2
>λ1. I is a om min, he de elopmen space o P2 is highe han he
de e io a ion space o P1, which induces TRE’s comp ehensi e ocus on
he posi i e u ili y P1, λ2<λ1.
When ≥ max > min, he TRE’s DSQ is highe han all e e ence
le els, which gene a es dual posi i e u ili ies, P1= ( − min)
α
>0,P2=
( − max)
α
≥0. Acco ding o he dual e e ences min and max, he
co esponding e e ence weighs a e λ1=k2
k1+k2 and λ2=k1
k1+k2. Because
he compa a i e la ge P2 c ea es mo e sa is ying eeling, P2 occupies
mo e dominan posi ion han P1, λ2>λ1.
Fu he discussion on geog aphic loca ion and i m size
Le us suppose ha a a ge ed TRE, i, in a ce ain ci y, has de e mined
i s indus ial DSQ, h, and compe i o ’s DSQ, c; i s annual sale e enue
is s i, and he ci y’s annual GDP is gdpi. To expand he e e ence
knowledge o ano he i m in ano he ci y, he p ospe i y e icien and
he scale e icien can be used o desc ibe he in luence o he TRE’s
geog aphic loca ion and i m size on ex e nal e e ence le els.
P ospe i y e icien , θj, shows he in luence o geog aphic loca ions,
such as ci y j, on indus ial DSQ, h, compa ing o ha o he a ge ed
ci y i. Le us assume ha he annual GDP o ci y j is gdpj. The p ospe i y
e icien equals he a io o annual GDPs: θj=gdpj
gdpi. I ano he TRE in ci y
j wan s o de e mine i s indus ial DSQ, j
h, i will be di ec ly based on
ci y’s indus ial DSQ, h, j
h=θj h=gdpj
gdpi h.
Scale e icien , ϑk, deno es he in luence o i m size, such as sale
e enue s k, on he compe i o ’s DSQ h, compa ed o he a ge ed TRE i.
Le us assume ha TRE k’s annual sale e enue is s k. The size e icien
can be se as he a io o annual sale e enue, ϑk=s k
s i. I ano he TRE k
de e mines i s compe i o ’s DSQ, k
c, i will be di ec ly based on ci y’s
indus ial DSQ, c, k
c=ϑj c=s k
s i c.
Incen i e mechanism o imp o ing TRE’s DSQ
A e achie ing he TRE’s pe cei ed DSQ, a “p incipal-agen ” model
desc ibing he ou sou cing coope a ion be ween i and he DSP is
es ablished o explo e he incen i e mechanism, and he analysis p ocess
is demons a ed in Figu e 7. Sec ion 5.1 illus a es he bene i s and cos s
o TREs and DSPs. In Sec ion 5.2, he “p incipal-agen ” model con aining
he cos -sha ing incen i e solu ion, indi idual a ionali y cons ain
(IR), and incen i e compa ibili y cons ain (IC) is designed o explo e
he coope a i e ela ionship. In Sec ion 5.3, he incen i e solu ion
design p ocess is explo ed.
The incen i e mechanism plays a c ucial ole in mo i a ing he DSPs
o imp o e he DSQ o he digi al ma ke ing pla o m and p o iding
s onge ma ke ing suppo o he TREs. Conce ning he in luence o he
incen i e mechanism, he DSPs ac i ely s i e o imp o e he DSQ om
mul iple dimensions. Specially, in he digi al angible dimension, DSPs
can inno a i ely op imize he in as uc u e and unc ional modules o
inc ease he online e iew ime and page isi ing leng h. Acco ding o
he digi al us dimension, hey could op imize algo i hms o enhance
da a secu i y and se ice s abili y. In he digi al in e ac ion dimension,
Figu e 7. S eps o incen i e mechanism.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
8
in which he TRE’s and he DSP’s psychological u ili y a e V[
π
T( )] =
0.178, V[
π
S( )] = 0.982, espec i ely. The compa ison is p esen ed in
Table 4.
As shown in Table 4, h ough cos sha ing incen i e, he second-bes
will be ob ained a ∗
2=0.5. The DSP will ac i ely imp o e he DSQ
le el om ∗
2(0.26) o ∗
2(0.5), achie ing an inc ease o 92.31%. The
TRE’s psychologically u ili y is inc eased om 0.178 o 1.81, eaching a
p omo ion o 916.85%. And he DSP’s psychologically u ili y is
imp o ed om 0.982 o 1.25, ob aining an ele a ion o 27.29%.
Op imal solu ion o he ou sou cing coope a ion unde pa ly asymme ic
in o ma ion condi ion
In his si ua ion, le V[
π
S( )] =
π
S( ) = 1− 3+g∗ , V1( )
= (1−
ω
)V[
π
S( )] +
ω
V[
π
T( )] and
ω
=0.4. The TRE’s pe cei ed DSQ
inc eases as he enhancemen o he ecei ed DSQ le el. In his si ua ion,
V1( ) = 0.6∗(1− 3+g∗ )+0.4{ln[P( ) + 1] + 2−g∗ }
The in o ma ion ad an age o he DSP is weake han ha o
comple ely incomple e in o ma ion bu s onge han ha o comple e
in o ma ion, esul ing in ∗
3 mee s
∂
V[
π
S( ∗
3)]
∂
∗
3=0. When ∈[0, ∗
3),
∂
V[
π
S( ∗
3)]
∂
∗
3
>0; When ∈[ ∗
3,1),
∂
V[
π
S( ∗
3)]
∂
∗
3<0. A he same ime, he TRE’s psy-
chological u ili y should no be lowe han ha in he comple ely
asymme ic in o ma ion condi ion, i.e. V[
π
T( )] >1.81. Re e encing o
Equa ion (8) he “p incipal-agen ” model unde pa ly asymme ic in-
o ma ion can be designed as
maximize 0.6∗(1− 3+g∗ )+0.4{ln[P( ) + 1] + 2−g∗ },(obj −I)
s. .⎧
⎨
⎩
∈ [0,1],(C1)
ln[P( ) + 1] + 2−g∗ >1.81,(C2)
1− 3+g∗ >0.657,(C3)
The equilib ium solu ion o his model is ob ained h ough Lingo.
The cos -sha ing a io, g3, is 0.554, ∗
3=0.643, and V[
π
T( ∗
3)]=2.015.
In summa y, he compa ison o incen i e e ec s in di e en si ua-
ions is shown in Table 5.
Acco ding o Table 5, high DSQ le el, ∗
1=0.7, is achie ed wi hou
addi ional incen i e ees in he comple ely symme ic in o ma ion si -
ua ion, i.e. g1=0. In he asymme ic in o ma ion, i no incen i e ees,
g−
2=0, a e gi en o he DSP, he deli e ed DSQ was a a low le el as ∗
2
=0.26. Howe e , he DSQ le el was signi ican ly imp o ed as
∗
3(0.643)> ∗
2(0.5)> ∗
2(0.26)a e he implemen o incen i es.
Especially, he DSQ le els in he asymme ic in o ma ion a e lowe han
ha in he comple ely symme ic in o ma ion, ∗
1, which is mainly
caused by he TRE’s dominance posi ion.
Compa ison analysis
To agg ega e he speci ic psychological u ili ies in di e en e e ence
poin s, he subjec i e weigh ing me hod is commonly used by a aching
ixed weigh s o di e en e e ence poin s (Wei e al., 2019; Uppa i &
Hasija, 2019; Zhong e al., 2022; Wang e al., 2020; Weinga en e al.,
2019; Tu e al., 2022). Howe e , he ixed weigh s ailed o desc ibe he
p incipal’s psychological luc ua ion in he dominance o e e ence
poin s. Especially when = i, i =c,h, Vi=0, he p incipal will pay
comple e no ice on ano he e e ence poin by gi ing absolu e dominan
weigh .
Compa ison o he pe cei ed DSQ
Fo example, Zhong in oduced equal weigh s o dual e e ence
poin s (Zhong e al., 2022), Pʹ( ) = 0.5P1+0.5P2, which is desc ibed as
a ed line in he absolu e middle o P1 and P2. The esul compa ison
be ween Zhong’s equal and dynamic weigh s p oposed in his s udy
p o ides he cu e displays in Figu e 12.
In he pa o 0 ≤ <0.24, he comp ehensi e psychological u ili y
PII( )is la e han Pʹ( ). Due o PII
2<PII
1<0, P2 c ea es mo e panic
eeling, λII
1<0.5<λII
2. Because he DSQ gap d1 is educing, nega i e PII
1
is inc easing up o 0, and he dominance posi ion o PII
2 is con inuously
enhanced.
In he pa o 0.24 ≤ <0.62, he posi i e PIII
1 is inc easing om 0,
which s a s o play a ole in comp ehensi e u ili y. D i en by he in-
c ease in posi i e PIII
1 and i s weigh λIII
1, PIII( )is apidly ising. Addi-
ionally, when is close o h, DSQ gap d2 is elimina ing and nega i e
PIII
2 is inc easing up o 0, which causes he absolu e dominance o PIII
1.
In he pa o 0.62 ≤ ≤1, bo h PI
1 and PI
2 a e posi i e, PI
1>PI
2>0,
which con ibu es o he inc ease o comp ehensi e u ili y. The p incipal
de e mines he di e en weigh s on dual e e ence poin s acco ding o
u u e de elopmen p ospec s. Due o k2<k1<0, he de e io a ing
space eeling on PI
2 is la ge han ha on PI
1, which leads o
λI
2<0.5<λI
1, PI( )>Pʹ( ). Wi h he gap be ween PI
2 and PI
1 is educing,
λI
1 and λI
2 a e ending o equal condi ion.
Compa ison o he op imal solu ion
In he symme ic in o ma ion si ua ion. Unde dual e e ence poin s,
conside ing a ying weigh s, ∗
1=0.7, g1=0, V[
π
T( 1)] = 2.37. Simi-
la ly, o he ixed weigh s, λ1=λ2=0.5, he DSP’s IR is a ha d
cons ain , and i will coope a e only i i s ac ual bene i is no less han
i s oppo uni y bene i s,
π
S. The e o e, he op imal solu ion is loca ed a
∗
1
ʹ=0.7, gʹ
1=0, Vʹ[
π
T( 1)] = 2.27. Because o he conside a ion o
p ospec , he TRE’s comp ehensi e u ili y inc eases.
In he comple ely asymme ic in o ma ion si ua ion. In a comple ely
asymme ic in o ma ion si ua ion, conside ing he a ying weigh , he
op imal solu ion can be ob ained a ∗
2=0.47, g2=0.66, and
V[
π
T( ∗
2)]=1.61. Simila ly, when he ixed weigh s, λ1=λ2=0.5, he
“p incipal-agen ” model unde incomple e in o ma ion can be designed
as ollows:
maximize ln[Pʹ( ) + 1] + 2−g∗ ,(obj −I)
s. .⎧
⎪
⎪
⎨
⎪
⎪
⎩
3∗ 2=g,(C1)
∈ [0,1],(C2)
ln[P’( ) + 1] + 2−g∗ >0.178,(C3)
1+ 3−g∗ >0.657,(C4)
The equilib ium solu ion o his model is ob ained h ough Lingo.
Using he ixed weigh s, he op imal solu ion is ob ained a ∗
2
ʹ=0.47,
Figu e 12. Compa ison o di e en psychological u ili y measu emen me hod.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
15
gʹ
2=0.66, and Vʹ[
π
T( ∗
2)]=1.61.
In he pa ly asymme ic in o ma ion si ua ion. In a pa ly asymme ic
in o ma ion si ua ion, conside ing he a ying weigh , he op imal so-
lu ion can be ob ained a ∗
3=0.643, g3=0.554, and V[
π
T( ∗
3)]=
2.015. Simila ly, when he ixed weigh s, λ1=λ2=0.5, he “p incipal-
agen ” model unde incomple e in o ma ion can be designed as
maximize 0.6(1− 3+g∗ )+0.4{ln[Pʹ( ) + 1] + 2−g∗ },(obj −I)
s. .⎧
⎨
⎩
∈ [0,1],(C1)
ln[P’( ) + 1] + 2−g∗ >1.61,(C2)
1+ 3−g∗ >0.657,(C3)
The equilib ium solu ion o his model is ob ained h ough Lingo.
Unde ixed weigh s, he op imal solu ion is loca ed a ∗
3
ʹ =0.62, gʹ
3 =
0.94, and Vʹ[
π
T( ∗
3)]=1.62. Acco ding o he esul in Table 6, in he
symme ic in o ma ion si ua ion, he op imal solu ion is ob ained a he
same DSQ le el unde he dual weigh ing me hod, which is mainly
caused by he same ha d cons ain . Howe e , al hough he DSQ le el
and cos -sha ing a io is equal, he TRE’s psychological u ili y is
di e en , V[
π
T( 1)] >Vʹ[
π
T( 1)], due o he conside a ion on he de el-
opmen po en ial. In he comple ely asymme ic in o ma ion si ua ion,
he incen i e a io is lowe han he si ua ion unde a ying weigh s.
Howe e , he TRE’s psychological u ili y, V[
π
T( ∗
2)], is highe han ha
in ixed weigh me hod, Vʹ[
π
T( ∗
2)], which is consis en wi h he
conclusion o he pe cei ed DSQ compa ison abo e.
In he pa ly asymme ic in o ma ion si ua ion, he DSQ le el in he
ixed weigh ing me hod is lowe han ha unde a ying weigh s, i.e., ∗
3
> ∗
3
ʹ, which is is mainly caused by he impac o u u e de elopmen
po en ial on he TRE’s psychological u ili y. In gene al, conside ing he
a ying weigh s, highe DSQ le els and psychological u ili y can be
ob ained wi h a lowe incen i e cos .
Sensi i i y analysis on incen i e e ec
Sensi i i y analysis o e e ence knowledge
Le us suppose ha he indus y-a e age e e ence le el is ixed, h =
0.24 and c> h. The sensi i i y analysis o he cos -sha ing incen i e
a io on he changing e e ence le el can be conduc ed. In pa icula ,
when he e e ence le el g adually changes 10% om 0.62, he se o c
alues is ob ained by inc easing and dec easing i by 10%,20%, 30%,
and 40%, espec i ely ( c=0.372, 0.434, 0.496, 0.558, 0.62, 0.744,
0.806, and 0.868). Repea ing he incen i e design p ocess abo e, he
op imal solu ion unde di e en in o ma ion si ua ion can be ob ained.
Symme ic in o ma ion si ua ion. In a comple ely symme ic in o ma ion
si ua ion, e e ing o Theo em 1, he op imal solu ion is loca ed whe e
∗
1=0.7, g1=0. I so, he TRE’s compe i i e u iliza ion is V[
π
T( 1)] =
2.37.
Simila ly, epea ing he s eps, one can ob ain he op imal solu ion
unde di e en in o ma ion si ua ions, shown in Table 7.
In comple ely symme ic in o ma ion, he op imal DSQ is no ela ed
o he compe i o ’s e e ence le el, which depends only on he DSP’s
oppo uni y bene i
π
S. As he TRE handles he in o ma ion ad an age, i
can di ec ly obse e he DSP’s DSQ in es men and o ce he la e o
choose i s mos expec ed le el. In his si ua ion, he TRE does no need o
pay addi ional incen i e ee.
Comple ely asymme ic in o ma ion si ua ion. When c=0.558, h=
0.24, e e ing o Equa ion (8), he “p incipal-agen ” model can be
designed as
maximize ln[P( ) + 1] + 2−g∗ ,(obj −I)
s. .⎧
⎪
⎪
⎨
⎪
⎪
⎩
3∗ 2=g,(C1)
∈ [0,1],(C2)
ln[P( ) + 1] + 2−g∗ >0.178,(C3)
1− 3+g∗ >0.657,(C4)
The equilib ium solu ion o his model is ob ained h ough Lingo. In
he op imal condi ion, he cos -sha ing a io, g2=0.288, ∗
2=0.31, and
V[
π
T( ∗
2)]=1.874.
Simila ly, epea ing he abo e s eps, one can ob ain he op imal so-
lu ion unde di e en in o ma ion si ua ions when c=0.372, 0.434,
0.496, 0.558, 0.62, 0.744, 0.806, and 0.868, espec i ely, which a e
shown in Table 8.
In comple ely asymme ic in o ma ion si ua ion, he DSQ equilib-
ium is posi i ely co ela ed o he main compe i o ’s e e ence le el.
Because he DSP absolu ely handle he he in o ma ion ad an age, i can
hide i s DSQ in es men . In his si ua ion, he TRE needs o pay addi-
ional incen i e ee o imp o e he DSQ le el. The TRE’s comp ehensi e
u ili y is nega i ely co ela ed o he main compe i o ’s e e ence le el,
which is mainly due o he inc easing incen i e ee as he imp o emen
o DSQ le el.
As shown in Figu e 13, in luenced by he comple ely asymme ic
in o ma ion, he g aphical ends illus a e he dynamic in e play
among he compe i i e e e ence knowledge, c, coope a i e equilib-
ium, ∗
2, cos incen i e coe icien , g, and he TRE’s comp ehensi e
u ili y, V. As he e e ence knowledge, c, inc eases, he blue cu e ex-
hibi s a g adual upwa d ajec o y wi h decele a ing g ow h, indica ing
diminishing ma ginal imp o emen o DSQ ega ding incen i es on
coope a ion. Concu en ly, he yellow cu e ises sha ply ( om 0.415 o
0.961) e lec ing he addi ional cos s equi ed o sus ain coope a ion.
The in e ac ion be ween hese ends d i es he g een cu e in o a
pe sis en decline ( om 1.999 o 1.493), sugges ing accele a ed e osion
o he TRE’s comp ehensi e u ili y by incen i e cos s. The e o e, he
TRE needs o iden i y a easonable compe i i e e e ence le el o p e-
en he blind imp o emen in DSQ and pay mo e incen i e cos s, which
may lead o a educ ion in o e all u ili y.
In pa ly asymme ic in o ma ion si ua ion. In pa ly asymme ic in o -
ma ion si ua ion, when c=0.558, h=0.24, e e ing o Equa ion (8),
he “p incipal-agen ” model can be designed as ollows:
maximize 0.6∗(1− 3+g∗ )+0.4{ln[P( ) + 1] + 2−g∗ },(obj −I)
s. .⎧
⎨
⎩
∈ [0,1],(C1)
ln[P( ) + 1] + 2−g∗ >1.874,(C2)
1− 3+g∗ >0.657,(C3)
The equilib ium solu ion o his model is ob ained h ough Lingo.
The cos -sha ing a io, g3, is 0.278, ∗
3=0.513, and V[
π
T( ∗
3)]=2.114.
Simila ly, epea ing he abo e ope a ions, we can ob ain he op imal
solu ions unde di e en in o ma ion si ua ions when c=0.372, 0.434,
Figu e 13. Sensi i i y analysis o c in comple ely asymme ic in o ma-
ion si ua ion.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
16
0.496, 0.558, 0.62, 0.744, 0.806, and 0.868, espec i ely, shown in
Table 9.
In pa ly asymme ic in o ma ion si ua ion, he DSQ equilib ium and
he TRE’s comp ehensi e u ili y a e lowe han ha a he comple ely
symme ic in o ma ion si ua ion bu highe han ha a he comple ely
asymme ic in o ma ion. The incen i e equilib ium solu ion, ∗
3, and he
TRE’s comp ehensi e u ili y, V[
π
T( ∗
3)], is p ima ily ela ed o he ba -
gaining powe be ween he DSP and he TRE, whe e nei he can p io -
i ize hei own u ili y. In his case, he DSP’s ela i e in o ma ion
ad an age p io i izes i s comp ehensi e u ili y in he incen i e goal,
esul ing in an o e all decline in he TRE’s comp ehensi e u ili y.
Pa icula ly, a a ce ain e e ence le el ( c=0.62), he TRE’s incen i e
ee dec eased because o pa icipa ion cons ain s, esul ing in a em-
po a y ebound in i s comp ehensi e u ili y.
As shown in Figu e 14, he inc ease in compe i i e e e ence
knowledge, c, d i es complex in e ac ions among he coope a i e
equilib ium, ∗
3, cos incen i e coe icien , g, and he TRE’s comp e-
hensi e u ili y, V. The adjus men o incen i e cos , g, keeps nonlinea
co ela ion o he compe i i e e e ence knowledge, c, which is p i-
ma ily caused by he TRE’s pa ly in o ma ion ad an age.
The sensi i i y o ∗
3 on he e e ence knowledge, c, is cha ac e ized
by non–mono onic luc ua ions wi h an o e all upwa d end. In he
ini ial s age ( c=0.372→0.496), he coope a i e equilib ium, ∗
3 d ops
om 0.495 o 0.434 and hen ebounds o 0.496, e lec ing sho - e m
supp ession o coope a ion by compe i i e p essu e. In he mid-phase
( c=0.558→0.744), he coope a i e equilib ium, ∗
3, ises s eadily o
0.836, indica ing signi ican e iciency gains om incen i es unde
mode a e compe i ion. In he la e phase ( c>0.744), he g ow h o c
slows (0.836→0.925) as he su ge in incen i e ee o se ing he posi i e
e ec s o compe i ion. O e all, c exhibi s phase-speci ic op imiza ion,
bu incen i e e iciency diminishes unde high compe i ion in ensi y.
The sensi i i y o V o he e e ence knowledge, c, ollows a
“decline-b ie ebound-accele a ed decline” pa e n. In he ini ial s age
( c=0.372→0.558), he TRE’s comp ehensi e u ili y, V, d ops om
2.100 o 1.895 because o he ising incen i e ee, which e odes i s
u ili y. In he mid-s age ( c=0.620), V b ie ly ebounds o 2.015,
consis en wi h he dip in he incen i e ee, g, sugges ing ansien e -
iciency gains om pa ly asymme ic in o ma ion. In he la e s age ( c
>0.620), he incen i e ee, g, ebounds sha ply (up o 0.974), d i ing
he TRE’s comp ehensi e u ili y down om 1.892 o 1.675. O e all, he
TRE’s comp ehensi e u ili y, V, is domina ed by he nonlinea luc ua-
ions o incen i e ee, whe e high compe i ion in ensi y leading o he
u ili y dec eases because o p ohibi i e incen i e cos s.
Sensi i i y analysis o he deg ee o in o ma ion symme y
To explo e he in luence o he deg ee o in o ma ion symme y on
incen i e e ec , le
ω
inc ease om 0.1 o 0.9. Repea ing he calcula ion
p ocess as
ω
=0.4, he co esponding exci a ion equilib ium solu ion
can be ob ained.
Fo example, when
ω
=0.1, he TRE’s pe cei ed DSQ inc eases as
he enhancemen o he ecei ed DSQ le el. V1( ) = 0.9∗(1− 3+
g∗ )+0.1{ln[P( ) + 1] + 2−g∗ }.
Re e encing o Equa ion (8) he “p incipal-agen ” model unde pa ly
asymme ic in o ma ion can be designed as ollows:
maximize 0.9∗(1− 3+g∗ )+0.1{ln[P( ) + 1] + 2−g∗ },(obj −I)
s. .⎧
⎨
⎩
∈ [0,1],(C1)
ln[P( ) + 1] + 2−g∗ >1.81,(C2)
1− 3+g∗ >0.657,(C3)
The equilib ium solu ion o his model is ob ained h ough Lingo.
The cos -sha ing a io, g3, is 0.721, ∗
3=0.708, and V[
π
T( ∗
3)]=1.871.
In summa y, he compa ison o incen i e e ec s in di e en si ua-
ions is shown in Table 10.
Acco ding o Table 10, when
ω
is in he ange o 0.1–0.3, ∗
3 emains
s able a 0.708, and he incen i e ee keeps a 0.721 because bo h pa ies
ha e limi ed in o ma ion. The DSP occupies he dominance because o
i s la ge weigh . As
ω
inc eases, ∗
3 dec eases and, hen, emains s able
wi hin a ce ain ange and inally shows a small-scale changing end.
The eason is he TRE cons an ly adjus s i s incen i e s a egy because o
he change in i s dominan powe . The incen i e ee dec eases signi i-
can ly when
ω
inc eases. When he deg ee o in o ma ion symme y is
low, high incen i es a e needed o mobilize en husiasm. Howe e , when
he deg ee o in o ma ion symme y inc eases, he TRE can guide he
coope a ion h ough a easonable mechanism by i ue o i s dominan
powe , and no high-le el incen i es a e equi ed.
Rega ding on he TRE’s u ili y, when
ω
is lowe han 0.3, due o he
s abili y o he coope a ion model and he incen i e cos , he u ili y
emains unchanged a 1.871. As
ω
inc eases, he TRE’s u ili y is posi-
i ely ela ed o he deg ee o in o ma ion symme y. The adjus men o
he equilib ium solu ion and he dec ease in he incen i e cos enable
he en e p ise o be e in eg a e esou ces, educe cos s and isks, and
c ea e mo e alue. Tha is he eason o con inuous p omo ion and
inc easing coope a ion u ili y.
In he comple ely symme ic in o ma ion si ua ion, he TRE’s psy-
chological u ili y V[
π
T( ∗
1)]=2.37, is highe han ha in o he si ua-
ions, which is caused by i s dominance posi ion in he supply-chain
coope a ion. In comple ely asymme ic in o ma ion si ua ion, wi h he
implemen a ion o incen i es, he TRE’s psychological u ili y,
V[
π
T( ∗
2)], is g adually imp o ed om 0.178 o 1.81.In pa icula , he
TRE’s psychological u ili y, V[
π
T( ∗
3)], is highe han V[
π
T( ∗
2)], bu
lowe han V[
π
T( ∗
1)], which is consis en wi h he TRE’s in o ma ion
ad an age.
Sensi i i y Analysis o geog aphic loca ion
When he GDP o he a ge ed ci y is luc ua ing compa ed o
Hangzhou GDP, he p ospe i y coe icien , θj=gdpj
gdpHangzhou, g adually
changes om 1. Le he se o θj alues inc ease and dec ease by 10%,
20%, 30%,and 40%, espec i ely ( j
h=0.144, 0.168, 0.192, 0.216,
0.24, 0.264, 0.288, 0.312, and 0.336). Repea ing he abo e incen i e
p ocess, he op imal solu ion unde di e en in o ma ion si ua ion can
be ob ained in he ollowing si ua ions.
Symme ic in o ma ion si ua ion. In he comple ely symme ic in o ma-
ion si ua ion, e e ing o Theo em 1, he op imal solu ion is de e -
mined by he DSP’s oppo uni y bene i
π
S, whe e ∗
1=0.7, g1=0.
TRE’s compe i i e u iliza ion is V[
π
T( 1)] = 2.37.
Simila ly, one can ob ain he op imal solu ion when j
h=0.144,
0.168,0.192,0.216,0.24,0.264,0.288,0.312,and 0.336, espec-
i ely (Table 11). In comple ely symme ic in o ma ion, he op imal
DSQ is no ela ed o he p ospe i y coe icien , θj, which shows he
alida ion o Theo em 1.
Figu e 14. Sensi i i y analysis o c in pa ly asymme ic in o ma ion si ua ion.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
17
Comple ely asymme ic in o ma ion si ua ion. In his si ua ion, when h =
0.24, j
h=90% h=0.216, e e ing o Equa ion (8), he “p incipal-
agen ” model can be designed as ollows:
maximize ln[P( ) + 1] + 2−g∗ ,(obj −I)
s. .⎧
⎪
⎪
⎨
⎪
⎪
⎩
3∗ 2=g,(C1)
∈ [0,1],(C2)
ln[P( ) + 1] + 2−g∗ >0.178,(C3)
1− 3+g∗ >0.657,(C4)
The equilib ium solu ion o his model is ob ained h ough Lingo. In
he op imal condi ion, he cos -sha ing a io, g2=0.288, ∗
2 =0.484,
and V[
π
T( ∗
2)]=1.836.
Simila ly, one can ob ain he op imal solu ion unde di e en in-
o ma ion si ua ions when j
h=0.372, 0.434, 0.496, 0.558, 0.62, 0.744,
0.806, and 0.868, espec i ely, which a e shown in Table 12. The ends
o equilib ium, incen i e cos and he TRE’s u ili y a e shown in
Figu e 15.
As shown in Figu e 15, in luenced by he comple ely asymme ic
in o ma ion, he g aphical ends illus a e he dynamic in e play o he
p ospe i y coe icien , θj, coope a i e equilib ium, ∗
2, incen i e coe i-
cien , g, and he TRE’s comp ehensi e u ili y, V. As he GDP o
geog aphic loca ion, θj, inc eases, he blue cu e (equilib ium end)
exhibi s a g adual upwa d ajec o y wi h decele a ing g ow h, indi-
ca ing diminishing ma ginal imp o emen on he DSQ o incen i es on
coope a ion. In addi ion o ha , he yellow cu e (incen i e in ensi y)
ises sha ply ( om 0.654 o 0.872), e lec ing he addi ional cos s
equi ed o main ain he coope a ion. The in e ac ion be ween hese
ends d i es he g een cu e, he TRE’s u ili y end, in o a pe sis en
decline ( om 1.91 o 1.694), p esen ing he accele a ed e osion o he
TRE’s comp ehensi e u ili y in luenced by he addi ional incen i e cos .
To sum up, in a comple ely asymme ic in o ma ion si ua ion, he
DSQ equilib ium is posi i ely co ela ed o he p ospe i y coe icien , θj,
because o he inc easing indus ial le el in he p ospe i y a ea. In his
si ua ion, he TRE needs o pay an addi ional incen i e ee o imp o e
he DSQ le el. The TRE’s comp ehensi e u ili y is nega i ely co ela ed
o he p ospe i y coe icien , θj, because o he inc easing incen i e ee
o imp o ing he DSQ le el.
In pa ly asymme ic in o ma ion si ua ion. In he pa ly asymme ic in-
o ma ion si ua ion, when c=0.62, j
h=90% h=0.216, e e ing o
Equa ion (8), he “p incipal-agen ” model can be designed as
maximize 0.6∗(1− 3+g∗ )+0.4{ln[P( ) + 1] + 2−g∗ },(obj −I)
s. .⎧
⎨
⎩
∈ [0,1],(C1)
ln[P( ) + 1] + 2−g∗ >1.836,(C2)
1− 3+g∗ >0.657,(C3)
The equilib ium solu ion o his model is ob ained h ough Lingo.
The cos -sha ing a io, g3, is 0.848, ∗
3=0.62, and V[
π
T( ∗
3)]=1.845.
Simila ly, epea ing he abo e ope a ions, he op imal solu ions
unde di e en si ua ions can be gained when j
h=0.372,0.434,0.496,
0.558,0.62,0.744,0.806,and 0.868, espec i ely, shown in Table 13.
The ends o changing equilib ium, incen i e cos and TRE’s u ili y a e
shown in Figu e 16.
As shown in Figu e 16, he coope a i e equilib ium sol-
u ion ∗
3 emains a 0.62 when θj anges om 0.6 o 0.9, consis en wi h
he compe i i e e e ence knowledge ( c=0.62). Tha indica es ha ∗
3
is absolu ely in luenced by he compe i i e e e ence poin du ing his
ange, and he e is no ob ious e ec caused by a ying θj∈ [0.6,0.9].
When θj≥1, he inc easing θj leads o he aise o indus ial knowledge,
and ∗
3 jumps o 0.643and emains s able. In addi ion o ha , he
incen i e ee, g, inc eases when θj∈ [0.6,0.9]bu d ops sha ply o
0.554when θj≥1 and emains cons an . Addi ionally, he TRE’s
comp ehensi e u ili y, V, dec eases om 1.921 o 1.845 acco ding o he
inc ease in incen i e a io when θj∈ [0.6,0.9]. When θj=1, because o
he jump o ∗
3 and he sha p d op in incen i e ee, g, he TRE’s
comp ehensi e u ili y, V, ises o 2.015. When θj>1, he TRE’s
comp ehensi e u ili y dec eases om 2.015 o 1.934, esponding o he
inc ease o θj, which is p ima ily caused by he DSQ lag.
To sum up, in pa ly asymme ic in o ma ion, he p ospe i y coe i-
cien , θj, does no di ec ly a ec he ou comes, including he incen i e
equilib ium solu ion, ∗
3, incen i e a io, g, and he TRE’s comp ehensi e
u ili y, V. Howe e , he e is a u ning poin on ends. When he p os-
pe i y coe icien is lowe han 1 (i.e., θj<1), he incen i e equilib ium
solu ion equals o he main compe i o ’s DSQ le el. Mo eo e , when he
p ospe i y coe icien is highe han 1 (i.e., θj>1), he TRE’s psycho-
logical u ili y is imp o ed because o he imp o ing DSQ and lowe
incen i e ee.
Sensi i i y analysis o i m size
When he scale o he a ge ed en e p ise is luc ua ing compa ed o
ha o he Jiebai g oup, he scale coe icien , ϑk=s k
s Jiebai, g adually
changes om 1 . Le he se o alues inc ease and dec ease by 10%,
20%, 30% and 40%, espec i ely. The compe i o ’s e e ence le el, k
c=
0.372,0.434,0.496,0.558,0.62,0.682,0.744,0.806 and 0.868, as
shown in Table 14. The de ailed analysis o equilib ium, incen i e
pa ame e , and he TRE’s u ili y a e he same as shown in Sec ion 7.5.1.
Sensi i i y analysis o incen i e budge
To conside he incen i e budge ’s impac on he TRE’s incen i e
solu ion, he incen i e ee unde he op imal solu ion, φ( ∗
i), can be used
o se he cons ain o incen i e budge ,
ρ
=φ( ∗
i), which can be ea ed
as he incen i e h eshold.
When he incen i e budge ,
ρ
, g adually changes om he o iginal
op imal le el, φ( ∗
i). Le he se o
ρ
alues inc ease and dec ease by
Figu e 15. The sensi i i y in luence o geog aphic loca ion (comple ely
asymme ic in o ma ion).
Figu e 16. The sensi i i y in luence o geog aphic loca ion (pa ly asymme ic
in o ma ion).
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
18
10%, 20%, 30% and 40%, espec i ely. Repea ing he abo e incen i e
p ocess, he op imal solu ion unde di e en in o ma ion si ua ion can
be ob ained in he ollowing si ua ions.
Symme ic in o ma ion si ua ion. In he comple ely symme ic in o ma-
ion si ua ion, e e ing o Theo em 1, he op imal solu ion is loca ed a
whe e ∗
1=0.7, g1=0. The TRE’s comp ehensi e u iliza ion is
V[
π
T( 1)] = 2.37. Simila ly, epea he s eps, we can ob ain he op imal
solu ion unde di e en in o ma ion si ua ions can be shown in
Table 15.
In he comple ely symme ic in o ma ion si ua ion, he op imal DSQ
is no ela ed o he incen i e budge , in which he TRE need no o pay
addi ional incen i e ee because o he in o ma ion ad an age.
Comple ely asymme ic in o ma ion si ua ion. In his si ua ion, when
ρ
=
0.9∗φ( ∗
2)=0.9∗0.375 =0.338, e e ing o Equa ion (8), he “p in-
cipal-agen ” model can be designed as ollows:
maximize ln[P( ) + 1] + 2−g∗ ,(obj −I)
s. .
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
3∗ 2=g,(C1)
∈ [0,1],(C2)
ln[P( ) + 1] + 2−g∗ >0.178,(C3)
1− 3+g∗ >0.657,(C4)
g∗ <0.338,(C5)
The equilib ium solu ion o his model is ob ained h ough Lingo. In
he op imal condi ion, he cos -sha ing a io, g2=0.696, ∗
2 =0.482,
and V[
π
T( ∗
2)]=1.808.
Simila ly, epea he s eps, we can ob ain he op imal solu ion unde
di e en incen i e budge s can be shown in Table 16. The ends o
equilib ium, incen i e cos and he TRE’s u ili y a e shown in Figu e 17.
As shown in Figu e 17, in luenced by he comple ely asymme ic
in o ma ion, he g aphical ends illus a e he dynamic in e play o he
incen i e budge ,
ρ
, coope a i e equilib ium, ∗
2, incen i e coe icien , g,
and he TRE’s comp ehensi e u ili y, V. When he incen i e budge ,
ρ
, is
lowe han he o iginal le el (i.e.,
ρ
≤0.375), bo h he coope a i e
equilib ium solu ion, ∗
2, and he incen i e a io, g, will inc ease. The
inc easing incen i e budge boos s he inc ease o he TRE’s comp e-
hensi e u ili y, V, om 1.761 o 1.81. Once he incen i e budge ,
ρ
, is
highe han he o iginal le el (i.e.,
ρ
>0.375), he coope a i e equilib-
ium solu ion ∗
2=0.5, incen i e a io g =0.75, and he TRE’s
comp ehensi e u ili y V =1.81, emain s able, indica ing ha he
incen i e budge ’s impac on incen i es becomes ine ec i e unde such
condi ions.
To sum up, in a comple ely asymme ic in o ma ion si ua ion, when
he incen i e budge is changing lowe han he op imal condi ion, he
DSQ equilib ium will be posi i ely co ela ed o he incen i e budge .
Once he incen i e budge is enhanced mo e han he op imal condi ion,
he DSQ equilib ium, incen i e a io, and he TRE’s comp ehensi e
u ili y will keep s able, which equal o ha a he o iginal op imal le el.
The e o e, he e is an op imal budge h eshold o incen i e when
ρ
=
0.375, whe e he DSQ equilib ium, ∗
2, and he TRE’s op imal psycho-
logical u ili y, V, can be achie ed a he lowes budge .
In pa ly asymme ic in o ma ion si ua ion. In pa ly asymme ic in o -
ma ion, when
ρ
=0.9∗φ( ∗
3)=0.9∗0.356 =0.322, e e ing o
Equa ion (8), he “p incipal-agen ” model can be designed as ollows:
maximize 0.6∗(1− 3+g∗ )+0.4{ln[P( ) + 1] + 2−g∗ },(obj −I)
s. .⎧
⎪
⎪
⎨
⎪
⎪
⎩
∈ [0,1],(C1)
ln[P( ) + 1] + 2−g∗ >1.808,(C2)
1− 3+g∗ >0.657,(C3)
g∗ <0.322
The equilib ium solu ion can be achie ed h ough sma algo i hms.
In he op imiza ion condi ion, he cos -sha ing a io g3=0.518, he
equilib ium DSQ ∗
3=0.62, and he TRE’s u ili y is V[
π
T( ∗
3)]=2.034.
Simila ly, epea ing he abo e ope a ions, we can ob ain he op imal
solu ions, shown in Table 17. The ends o changing equilib ium,
incen i e cos and u ili y a e shown in Figu e 18.
As shown in Figu e 18, when he incen i e budge ,
ρ
, inc eases om
0.214 o 0.498, he incen i e a io, g, is enhanced om 0.345 o 0.804.
Howe e , he incen i e equilib ium solu ion, ∗
3, emains s able a 0.62,
only when
ρ
=0.356, ∗
3 has a small inc ease o he equilib ium poin
0.643. In his si ua ion, he TRE’s comp ehensi e u ili y, V, keeps
dec easing om 2.142 o 1.855, inducing by he inc easing incen i e
a io, g, and s able DSQ le el, ∗
3.
In summa y, in a pa ly asymme ic in o ma ion si ua ion, he
incen i e a io, g, is posi i ely ela ed o he incen i e budge ,
ρ
. How-
e e , he coope a i e equilib ium solu ion, ∗
3, is almos una ec ed by
incen i e budge s, p ima ily because he TRE and he DSP implemen a
ba gaining game in pa ial asymme ic in o ma ion. The TRE’s
comp ehensi e u ili y, V, is nega i ely ela ed o he incen i e budge ,
ρ
,
which is p ima ily caused by he inc easing incen i e a io as he aise o
o al cos .
Fu he discussions on he applica ions ac oss egions
To ex end he applica ion ac oss egions, 20 Chinese TREs a e
selec ed and hei de ailed in o ma ion is shown in Table 18. The TREs’
sale e enue da a a e om he en e p ises’ annual inancial epo s in
2023, and hei loca ion ci y GDP da a a e collec ed om Na ional Bu-
eau o S a is ics in 2023. Compa ing o benchma k TRE da a and
Hangzhou GDP, he p ospe i y coe icien s and he scale coe icien s can
be calcula ed o he c oss- egional applica ions o he incen i e me hod.
The in luence o geog aphic loca ion on incen i e e ec
To ex end he applica ion om he cu en ci y Hangzhou o mo e
Figu e 17. The sensi i i y in luence o incen i e budge s (comple ely asym-
me ic in o ma ion).
Figu e 18. The sensi i i y in luence o incen i e budge (pa ly asymme ic
in o ma ion).
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
19
a ge ed ci ies, p ospe i y coe icien , θj=gdpj
gdpHangzhou, can be used o
no malize he indus ial e e ence knowledge in he ci y j. The ex e nal
indus ial e e ence in ci y j can be j
h=gdpj
gdpHangzhou h.
Symme ic in o ma ion si ua ion. Based on he GDP o each egion and
Hangzhou, he p ospe i y coe icien s, θj, o each egion and he e ised
indus ial e e ence knowledge a e p esen ed in Table 11. In a
comple ely symme ic in o ma ion si ua ion, he op imal solu ion is
de e mined by he DSP’s oppo uni y bene i
π
S, whe e ∗
1 =0.7, g1 =0,
and V[
π
T( 1)] = 2.37. Simila ly, e e ing o Theo em 1, one can ob ain
he op imal solu ion when j
h a di e en le els, which a e shown in
Table 19.
Acco ding o he equilib ium DSQ, ∗
1, and incen i e pa ame e , g,
because he TRE absolu ely exploi s he in o ma ion ad an age, i can
o ce he DSPs o ake he basic oppo uni y bene i . The o cing con ac
can ensu e ha he op imal DSQ is no ela ed o egional GDP and no
addi ional incen i e is needed.
Comple ely asymme ic in o ma ion si ua ion. Le us conside he Ningbo
Zhongbai company in Ningbo ci y as an example. The GDP o Ningbo
ci y is 15704.30 and he p ospe i y coe icien s is 0.837. In his si ua-
ion, when he adjus ed indus ial e e ence j
h=0.837 ∗0.24 =0.201
and compe i o e e ence c=0.62, e e ing o Equa ion (8), he
“p incipal-agen ” model can be designed as ollows:
maximize ln[P( ) + 1] + 2−g∗ ,(obj −I)
s. .⎧
⎪
⎪
⎨
⎪
⎪
⎩
3∗ 2=g,(C1)
∈ [0,1],(C2)
ln[P( ) + 1] + 2−g∗ >0.178,(C3)
1− 3+g∗ >0.657,(C4)
The equilib ium solu ion o his model is ob ained h ough Lingo. In
he op imal condi ion, he cos -sha ing a io, g2=0.668, ∗
2=0.472,
and V[
π
T( ∗
2)]=1.856.
Simila ly, one can ob ain he op imal solu ions in o he ci ies. The
equilib ium condi ions a e p o ided in Table 20. Because o he
inc easing egional GDPs, he indus ial e e ence le els, j
h, will be
aised as well, c ea ing mo e se ious ma ke ing compe i ion and
educing he TREs’ eelings on hei DSQ le els. The ends o he
equilib ium DSQ, ∗
2, incen i e pa ame e , g, he TRE’s comp ehensi e
u ili y, V, a e shown in Figu e 19, which ha e a clea ela ionship wi h
he p ospe i y coe icien , θj.
1) The equilib ium DSQ, ∗
2, and he incen i e pa ame e , g, a e posi-
i ely ela ed o egional GDP le el. Because o he TRE’s pa icipan
cons ain , V[
π
T( )] ≥ V[
π
T], he inc easing indus ial e e ence im-
p o es i s oppo uni y p o i
π
T, which leads o he inc easing equi-
lib ium DSQ, ∗
2. To induce he DSP o p o ide highe DSQ, he e
should be mo e incen i e in ensi y, and he incen i e pa ame e , g,
should be inc eased o main ain coope a ion in mo e p ospe ous
egions.
2) The TRE’s comp ehensi e u ili y, V, is nega i ely ela ed o egional
GDP le el. The inc easing egional GDP imp o es he indus ial
e e ence le el and educes he TRE’s eeling on pe cei ed DSQ.
Highe egional GDP signi ies a mo e de eloped indus ial base and
s onge ma ke po en ial. Howe e , he posi i e associa ion is
accompanied by he diminishing ma ginal e ec . The inc emen al
gains in coope a ion s abili y slow down when θj eaches an
ad anced le el. The inc easing equilib ium DSQ, ∗
2, and he incen-
i e pa ame e , g, g ea ly imp o e he TRE’s ou sou cing paymen ,
which imposes hea ie cos bu den and ha ms i s coope a i e u ili y.
The TRE’s comp ehensi e u ili y, V, declines pe sis en ly om 2.011
o 1.368, which is caused by he g ow h a e o addi ional incen i e
cos s exceeding he g ow h a e o coope a i e equilib ium, ∗
2.
The in luence o i m size on incen i e e o
To ex end he applica ion om he cu en en e p ise o mo e a -
ge ed en e p ises, he scale coe icien , ϑk, is used o adjus he com-
pe i o ’s e e ence knowledge be ween he a ge en e p ise, k, and he
Jiebai g oup, whe e ϑk=s k
s Jiebai. The scale le el conside ing he in luence
o company size as k
c=ϑk c=s k
s Jiebai c.
Symme ic in o ma ion si ua ion. Th ough he ans o ma ion o TREs’
sale e enue, he scale coe icien s, ϑk, and he e ised compe i o
e e ence le els a e p esen ed in Table 21. In a comple ely symme ic
in o ma ion si ua ion, he op imal solu ion is de e mined by he DSP’s
oppo uni y bene i
π
S, whe e ∗
1=0.7, g1=0, V[
π
T( 1)] = 2.37.
Simila ly, e e ing o Theo em 1, one can ob ain he op imal solu ion
when k
c is in di e en le els. In comple ely symme ic in o ma ion, he
op imal DSQ is no ela ed o he p ospe i y coe icien , ϑk.
Comple ely asymme ic in o ma ion si ua ion
Le us conside Ningbo Zhongbai as an example. In his si ua ion,
when he sale e enue o 119030.84 is and i s scale coe icien is 0.587.
I s adjus ed compe i o e e ence k
c=0.587 ∗0.62 =0.364 and he in-
dus ial e e ence h=0.24. Re e ing o Equa ion (8), he “p incipal-
agen ” model can be designed as ollows:
maximize ln[P( ) + 1] + 2−g∗ ,(obj −I)
s. .⎧
⎪
⎪
⎨
⎪
⎪
⎩
3∗ 2=g,(C1)
∈ [0,1],(C2)
ln[P( ) + 1] + 2−g∗ >0.178,(C3)
1− 3+g∗ >0.657,(C4)
The equilib ium solu ion o his model is ob ained h ough Lingo. In
he op imal condi ion, he cos -sha ing a io, g2=0.389, ∗
2=0.360,
and V[
π
T( ∗
2)]=2.000.
Figu e 19. The in luence o he p ospe i y coe icien in comple ely asym-
me ic in o ma ion.
Figu e 20. The in luence o he scale coe icien in comple ely asymme ic
in o ma ion.
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
20
Repea ing he abo e ope a ion o he o he TREs, one can ob ain he
op imal solu ions wi h a ying compe i o e e ences, k
c, shown in
Table 22. Because o inc easing i m sizes, he TREs’ compe i o e e -
ences, k
c will be imp o ed and educe TREs’ subjec i e eeling on hei
DSQ alues. As shown in Figu e 20, he ends o he equilib ium DSQ,
∗
2, incen i e pa ame e , g, TRE’s comp ehensi e u ili y, V, ha e clea
ela ionships wi h he scale coe icien , ϑk.
Table 2
Some Indica o s o DSQ.
Dimension Fi s -g ade
indica o s
Second-g ade indica o s Equa ion Quali y
Cha ac e is ics
Sou ces
Digi al
Tangibles
Func ionali y T a ic Numbe o egis e ed membe s L ype (Chan e al., 2020; Melo i´
c e al., 2021)
Time Spen on Page
Visi
The ime ha use s spend b owsing
he web
N ype (Chan e al., 2020; Si e al., 2015)
Cus ome In o ma ion
Asse s
The amoun o cus ome in o ma ion L ype (Chan e al., 2020; Va ada ajan, 2020)
E iciency Membe Inc ease Ne−NsL ype (Liu e al., 2022; Va ada ajan, 2020
Con e sion Ra e Con e sion equency
To al isi o s
L ype (Melo i´
c e al., 2021; Mo gan e al., 2022;
Min z e al., 2021)
Digi al T us Da a In eg i y Packe Loss Ra e Loss pocke s
To al pocke s
S ype (Skaka-ˇ
Ceki´
c & Ba ako i´
c Husi´
c, 2023; Huang
e al., 2018; Alnawas and Al Kha eeb, 2022)
Se ice A ailabili y T ansmission Delay Channel leng h
T ansmission a e
S ype (Skaka-ˇ
Ceki´
c & Ba ako i´
c Husi´
c, 2023; Huang
e al., 2018; Alnawas & Al Kha eeb, 2022))
Th oughpu Inpu /Ou pu
To al seconds
L ype (Chan e al., 2020; Liu e al., 2022; Min z e al.,
2021)
Digi al
In e ac ion
Collabo a ion Social Media
In e ac ion
Numbe o coope a ion social media L ype (Chan e al., 2020; Mo gan e al., 2022; Sau a,
2021)
B and Men ions Numbe o b and men ions L ype (Chan e al., 2020; Melo i´
c e al., 2021)
Mobile
Communica ion
Use -gene a ed Con en Numbe o use -gene a ed con en L ype (Babi´
c e al., 2020)
Cus ome
Cen ici y
Cus ome Insigh s A e age T ansac ion
Value
To al sales
To al o de olume
L ype (Huang e al., 2018)
Cus ome Re en ion
Ra e
(Ne −Na)/Ns L ype (Melo i´
c e al., 2021)
Cus ome Chu n Ra e Nc/N S ype (J¨
a inen & Ka jaluo o, 2015)
Cus ome
Segmen a ion
F equency Numbe o pu chases a cus ome
makes in a pe iod
L ype (Si e al., 2015)
Recency Las pu chase in e al S ype (Si e al., 2015)
Mone a y To al amoun spen by a cus ome in
a pe iod
L ype (Melo i´
c e al., 2021; Min z e al., 2021)
Reliabili y Ope a ional
E iciency
Re u n on Ma ke ing
In es men
Ma ke ing e enue
Ma ke ing In es men
S ype (J¨
a inen & Ka jaluo o, 2015)
Ma ke ing Cos To al cos o he ma ke ing ac i i ies N ype (Melo i´
c e al., 2021; Min z e al., 2021)
Cus ome Acquisi ion
Cos
SC+S +Ss+SoS ype (Si e al., 2015)
Ma ke
Pe o mance
In e na ional Ma ke
Sha e
Company’s sales/Global ma ke
sales
L ype (F ench, 2017)
T ade Compe i i eness Expo alue-Impo alue/Expo
alue+Impo alue
L ype (F ench, 2017)
Re ealed Compa a i e
Ad an age
(CE/TCE) ∗ (TGE /GE)L ype (F ench, 2017)
Table 3
Se ice quali y equi emen and pe o mance in o ma ion.
KPI ①Type ②Weigh ③Tole ance in e al ④Op imal a ge alue ⑤Compe i o ’s DSQ Indus y a e age DSQ
Ac ual DSQ ⑥S anda d DSQ ⑦Ac ual alue ⑧S anda d DSQ ⑨
CRR L 0.4 [10,30] 30 25 0.75 15 0.25
CAC S 0.4 [0.01,0.05] 0.01 0.03 0.5 0.04 0.25
MC N 0.2 [25,35] 30 32 0.6 34 0.2
Table 4
Incen i e e ec analysis in he comple ely asymme ic in o ma ion si ua ion.
Comple ely asymme ic
in o ma ion
Fixed ee Cos -sha ing
incen i e
Inc ease
a e
∗
2=0.26 ∗
2=0.592.31%
V[
π
T( )] V[
π
T( ∗
2)]=
0.178
V[
π
T( ∗
2)]=1.81 916.85%
V[
π
S( )] V[
π
S( ∗
2)]=
0.982
V[
π
S( ∗
2)]=1.25 27.29%
Table 5
Incen i e e ec and cos sha ing a io unde di e en scena ios.
Scena io Comple ely
symme ic
in o ma ion
Comple ely
asymme ic
in o ma ion
wi hou
incen i e
Comple ely
asymme ic
in o ma ion
unde
incen i e
Pa ly
asymme ic
in o ma ion
unde
incen i e
∗
1=0.7 ∗
2=0.26 ∗
2=0.5 ∗
3=0.643
V[
π
T( )] V[
π
T( ∗
1)]=
2.37
V[
π
T( ∗
2)]=
0.178
V[
π
T( ∗
2)]=
1.81
V[
π
T( ∗
3)]=
2.015
g g1=0g−
2=0 g2=0.75 g3=0.554
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
21
The equilib ium DSQ, ∗
2, and he incen i e pa ame e , g, exhibi
posi i e co ela ion wi h he scale coe icien , ϑk. As he enla ged i m
size leads o he inc eased compe i o e e ence le el, he TRE should
change i s ocus on i s compe i o o each be e DSQ. La ge scale
signi ies a s onge ma ke in luence, mo e complex supply chain, and
highe esou ce in eg a ion capabili y. To compe e wi h a s onge i al,
he TRE needs o imp o e he incen i e in ensi y and induce he DSP o
gua an ee la ge DSQ o he digi al ma ke ing pla o m.
The TRE’s comp ehensi e u ili y, V, has a nega i e ela ionship wi h
scale coe icien , ϑk. The inc easing i m size makes he TRE ocus on a
mo e powe ul compe i o , and he la ge compe i o e e ence di ec ly
educes he TRE’s eeling on pe cei ed DSQ. To keep compa a i e
ad an age, he TRE has o imp o e i s equi emen s on DSQ le el and
incen i e in es men , which educe i s economic u ili y. The TRE’s
Table 6
Incen i e e ec compa ison o di e en weigh ing me hods.
Scena io Comple ely symme ic in o ma ion Comple ely asymme ic in o ma ion unde incen i e Pa ly asymme ic in o ma ion unde incen i e
Fixed weigh Va ying weigh Fixed weigh Va ying weigh Fixed weigh Va ying weigh
∗
1
ʹ=0.7 ∗
1=0.7 ∗
2
ʹ=0.47 ∗
2=0.5 ∗
3
ʹ=0.62 ∗
3=0.643
V[
π
T( )] Vʹ[
π
T( 1)] = 2.37 V[
π
T( 1)] = 2.37 Vʹ[
π
T( ∗
2)]=1.61 V[
π
T( ∗
2)]=1.81 Vʹ[
π
T( ∗
3)]=1.62 V[
π
T( ∗
3)]=2.015
ggʹ
1=0g1=0gʹ
2=0.66 g2=0.75 gʹ
3=0.94 g3=0.554
Table 7
Op imal solu ions caused by changing c in symme ic in o ma ion si ua ion.
c40% dec ease 30% dec ease 20% dec ease 10% dec ease 0.62 10% inc ease 20% inc ease 30% inc ease 40% inc ease
∗
10.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7
g 0 0 0 0 0 0 0 0 0
V 2.37 2.37 2.37 2.37 2.37 2.37 2.37 2.37 2.37
Table 8
Op imal solu ion by changing c in comple ely asymme ic in o ma ion si ua ion.
c0.372 0.434 0.496 0.558 0.62 0.682 0.744 0.806 0.868
∗
20.372 0.413 0.441 0.467 0.5 0.524 0.537 0.554 0.566
g 0.415 0.565 0.583 0.654 0.75 0.824 0.865 0.921 0.961
V 1.999 1.973 1.929 1.874 1.810 1.738 1.661 1.581 1.493
Table 9
Op imal solu ions caused by changing c in pa ly asymme ic in o ma ion.
c0.372 0.434 0.496 0.558 0.620 0.682 0.744 0.806 0.868
∗
30.495 0.434 0.496 0.558 0.643 0.708 0.836 0.843 0.925
g 0.464 0.551 0.674 0.783 0.554 0.721 0.733 0.974 0.910
V 2.100 1.984 1.932 1.895 2.015 1.892 1.844 1.688 1.675
Table 10
Incen i e e ec and cos sha ing a io unde di e en
ω
.
ω
0.1 0.2 0.3 0.4 0.5 >0.6 0.7 0.8 0.9
∗
30.708 0.708 0.708 0.643 0.62 0.62 0.62 0.621 0.679
V[
π
T( )] 1.871 1.871 1.871 2.015 2.170 2.355 2.355 2.359 2.368
g 0.721 0.721 0.721 0.554 0.3 0 0 0 0
Table 11
Op imal solu ions caused by changing θj unde symme ic in o ma ion si ua ion.
θj0.6 0.7 0.8 0.9 11.1 1.2 1.3 1.4
j
h0.144 0.168 0.192 0.216 0.24 0.264 0.288 0.312 0.336
∗
10.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7
g 0 0 0 0 0 0 0 0 0
V 2.37 2.37 2.37 2.37 2.37 2.37 2.37 2.37 2.37
Table 12
Op imal solu ions caused by changing θj in comple ely asymme ic in o ma ion.
θj0.6 0.7 0.8 0.9 11.1 1.2 1.3 1.4
j
h0.144 0.168 0.192 0.216 0.24 0.264 0.288 0.312 0.336
∗
20.467 0.475 0.483 0.484 0.5 0.506 0.525 0.532 0.539
g 0.654 0.677 0.699 0.703 0.75 0.768 0.827 0.849 0.872
V 1.91 1.886 1.861 1.836 1.810 1.783 1.751 1.723 1.694
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
22
comp ehensi e u ili y, V, declines pe sis en ly om 2.002 o 1.339,
which is caused by he g ow h a e o addi ional incen i e cos .
Conclusions, implica ions, and u u e wo k
Resea ch conclusions
In he digi al se ice supply chain, he DSP designs and de elops he
digi al ma ke ing pla o m o he TRE, ying o sa is y he la e ’s
digi al equi emen s, such as da a mining, business ope a ions, and sales
p edic ion. Rega ding he end o digi al ma ke ing ans o ma ion, he
TRE’s pe cei ed DSQ is g ea ly in luenced by ex e nal e e ence
knowledge, such as he main compe i o ’s DSQ and indus ial DSQ.
Rega ding he se ice supply chain con aining asymme ic in o ma ion,
he TRE lacks he in o ma ion ad an age and has o ace he DSP’s
ad e se selec ion. Consequen ly, DSQ incen i e solu ions conce ning
ex e nal e e ence knowledge should be sui ably exploded o egula e
he DSP’s DSQ-assu ance in es men and elimina e he nega i e in lu-
ence o in o ma ion asymme y.
This s udy con ibu es o he incen i e mechanism in digi al se ice
supply chain in luenced by ex e nal e e ence knowledge. In heo e ical
le el, wo Nobel P ize heo ies— he p ospec heo y and he p incipal-
agen heo y—a e e ec i ely in eg a ed o explo e he ole o ex e nal
e e ence knowledge on he u ili y equilib ium and incen i e s a egy.
The in oduc ion o he TRE’s psychological u ili y caused by ex e nal
e e ence knowledge is he co e d i ing o ce o e alua e he DSQ e ec
and u u e po en ial. The objec i e o he p incipal-agen model is
upda ed o psychological u ili y, and he ou come o incen i e s a egy
conce ns ex e nal e e ence cons ain s. Rega ding me hodology, he
s udy explo es a dynamic a ying weigh ing me hod o desc ibe he
op imis ic p e e ence and dynamic dominance o ex e nal e e ences.
The no el a ying weighing me hod can o e come he igidi y o
adi ional ixed weigh ing app oach. On a p ac ical le el, his s udy
p o ides he incen i e ools o TREs o app aise he DSQ le el and
design di e en ial DSQ incen i e s a egies. The incen i e e ec can
imp o e TREs’ psychological DSQ u ili y by mi iga ing he DSPs’ mo al
haza d unde asymme ic in o ma ion. The abo e heo e ical, me hod-
ological, and p ac ical explo a ions conce ning he ex e nal e e ence
knowledge no only ill gaps in e ealing he psychological DSQ u ili y
bu also o e no el incen i e me hod and p ac ical ool o enhance he
manage ial abili y in digi al se ice supply chain.
Manage ial implica ions
In he digi al se ice supply chain, TREs ou sou ce he digi al ma -
ke ing pla o m de elopmen missions o DSPs. A compe i i e digi al
Table 13
Op imal solu ions caused by changing θj in pa ly asymme ic in o ma ion si ua ion.
θj0.6 0.7 0.8 0.9 11.1 1.2 1.3 1.4
j
h0.144 0.168 0.192 0.216 0.24 0.264 0.288 0.312 0.336
∗
30.62 0.62 0.62 0.62 0.643 0.643 0.643 0.643 0.643
g 0.804 0.819 0.834 0.848 0.554 0.554 0.554 0.554 0.554
V 1.921 1.897 1.872 1.845 2.015 1.987 1.971 1.953 1.934
Table 14
Op imal solu ions caused by changing ϑk in symme ic in o ma ion si ua ion.
ϑj0.6 0.7 0.8 0.9 11.1 1.2 1.3 1.4
k
c0.372 0.434 0.496 0.558 0.62 0.682 0.744 0.806 0.868
Table 15
Op imal solu ions caused by
ρ
in symme ic in o ma ion si ua ion.
ρ
40% dec ease 30% dec ease 20% dec ease 10% dec ease
ρ
=φ( ∗
i)10% inc ease 20% inc ease 30% inc ease 40% inc ease
∗
10.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7
g 0 0 0 0 0 0 0 0 0
V 2.37 2.37 2.37 2.37 2.37 2.37 2.37 2.37 2.37
φʹ( )0 0 0 0 0 0 0 0 0
Table 16
Op imal solu ions caused by
ρ
in symme ic in o ma ion si ua ion.
ρ
40% dec ease 30% dec ease 20% dec ease 10% dec ease
ρ
=0.375 10% inc ease 20% inc ease 30% inc ease 40% inc ease
0.225 0.263 0.3 0.338 0.413 0.45 0.488 0.525
g 0.534 0.591 0.646 0.696 0.75 0.75 0.75 0.75 0.75
∗
20.422 0.444 0.464 0.482 0.5 0.5 0.5 0.5 0.5
V 1.761 1.783 1.804 1.808 1.81 1.81 1.81 1.81 1.81
Table 17
Op imal solu ions caused by caused by
ρ
in pa ly asymme ic in o ma ion.
ρ
40% dec ease 30% dec ease 20% dec ease 10% dec ease
ρ
=0.375 10% inc ease 20% inc ease 30% inc ease 40% inc ease
0.214 0.245 0.285 0.322 0.392 0.427 0.463 0.498
g 0.345 0.395 0.46 0.518 0.554 0.632 0.689 0.747 0.804
∗
30.62 0.62 0.62 0.62 0.643 0.62 0.62 0.62 0.62
V 2.142 2.110 2.070 2.034 2.015 1.964 1.928 1.893 1.855
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
23
ma ke ing pla o m can p ecisely ecognize po en ial cus ome s, e ec-
i ely de elop online cus ome adhe ence, highly enhance social media
in luence, and consis en ly imp o e cus ome s’ loyal y, which di ec ly
ep esen s he en e p ise’s co e compe i i eness. In a compe i i e
Table 18
The TREs’ sale e enue da a and u ban GDP in 2023.
NO En e p ise Ci y Ci y GDP
(Billion
RMB)
Sale e enue
(Ten housand
RMB)
1 Ningbo
Zhongbai
Ningbo 15704.30 119030.84
2 Li u G oup Xianggang 3808.10 134897.50
3 Zhongshang
G oup
Nanjing 17421.00 249697.27
4 Hanshang G oup Wuhan 20012.00 138960.13
5 Dongbai G oup Fuzhou 12928.00 188607.16
6 Ma ke ing
Dajiang G oup
Hainan 7551.18 140000.95
7 Huijia Times U umqi 4168.46 249420.53
8 Maoye
Comme cial
Chengdu 22075.00 241194.33
9 Huaguang
Comme cial
Jian 2735.07 606542.7
10 Go un G oup Lanzhou 3487.00 96962.74
11 Tongcheng
Holdings
Changsha 12332.00 213286.99
12 Shenzhen Seg Shenzhen 34606.00 194906.55
13 Ren enle Shenzhen 34606.00 285267.95
14 Hualian
Holdings
Beijing 43760.70 109946.23
15 You-A
Depa men
S o e
Changsha 14332.00 134245.19
16 Be e Li e
Comme ce
G oup
Xiang an 2741.84 310114.33
17 Wen eng G ea
Wo ld Chain
Nan ong 11813.30 216565.63
18 New Huadu
Re ail G oup
Quanzhou 12172.33 282392.16
19 Zhongxing
Comme cial
G oup
Shenyang 8122.00 80994.67
20 Jiebai G oup Hangzhou 18753.07 202731.78
A e age
alue
15156.47 195288.17
S anda d
de ia ion
11470.14 115188.81
Table 19
Op imal solu ions unde symme ic in o ma ion si ua ion.
θj j
h ∗
1g V
1 0.240 0.7 0 2.37
0.837 0.201 0.7 0 2.37
0.203 0.049 0.7 0 2.37
0.929 0.223 0.7 0 2.37
1.067 0.256 0.7 0 2.37
0.689 0.165 0.7 0 2.37
0.403 0.097 0.7 0 2.37
0.222 0.053 0.7 0 2.37
1.177 0.283 0.7 0 2.37
0.146 0.035 0.7 0 2.37
0.186 0.045 0.7 0 2.37
0.658 0.158 0.7 0 2.37
1.845 0.443 0.7 0 2.37
1.845 0.443 0.7 0 2.37
2.334 0.560 0.7 0 2.37
0.764 0.183 0.7 0 2.37
0.146 0.035 0.7 0 2.37
0.630 0.151 0.7 0 2.37
0.649 0.156 0.7 0 2.37
0.433 0.104 0.7 0 2.37
Table 20
Op imal solu ions unde comple ely asymme ic in o ma ion si ua ion.
θj j
h ∗
2g V
1 0.240 0.500 0.750 1.810
0.837 0.201 0.472 0.668 1.852
0.203 0.049 0.457 0.627 1.998
0.929 0.223 0.492 0.726 1.832
1.067 0.256 0.503 0.759 1.792
0.689 0.165 0.474 0.674 1.889
0.403 0.097 0.471 0.666 1.954
0.222 0.053 0.458 0.629 1.995
1.177 0.283 0.488 0.714 1.759
0.146 0.035 0.453 0.616 2.011
0.186 0.045 0.456 0.624 2.002
0.658 0.158 0.472 0.668 1.896
1.845 0.443 0.557 0.931 1.562
1.845 0.443 0.557 0.931 1.562
2.334 0.560 0.592 1.051 1.368
0.764 0.183 0.480 0.691 1.871
0.146 0.035 0.453 0.616 2.011
0.630 0.151 0.470 0.663 1.903
0.649 0.156 0.471 0.666 1.898
0.433 0.104 0.454 0.618 1.949
Table 21
Op imal solu ions unde symme ic in o ma ion si ua ion.
ϑk k
c ∗
1g V
1 0.24 0.7 0 2.37
0.587 0.364 0.7 0 2.37
0.665 0.413 0.7 0 2.37
1.232 0.764 0.7 0 2.37
0.685 0.425 0.7 0 2.37
0.930 0.577 0.7 0 2.37
0.691 0.428 0.7 0 2.37
1.230 0.763 0.7 0 2.37
1.561 0.968 0.7 0 2.37
1.190 0.738 0.7 0 2.37
0.478 0.297 0.7 0 2.37
1.052 0.652 0.7 0 2.37
0.961 0.596 0.7 0 2.37
1.407 0.872 0.7 0 2.37
0.542 0.336 0.7 0 2.37
0.662 0.411 0.7 0 2.37
1.530 0.948 0.7 0 2.37
1.068 0.662 0.7 0 2.37
1.393 0.864 0.7 0 2.37
0.400 0.248 0.7 0 2.37
0.587 0.364 0.7 0 2.37
Table 22
Op imal solu ions unde comple ely asymme ic in o ma ion si ua ion.
ϑk k
c ∗
2g V
1 0.620 0.5 0.750 1.810
0.587 0.364 0.360 0.389 2.000
0.665 0.413 0.398 0.475 1.984
1.232 0.764 0.539 0.872 1.638
0.685 0.425 0.409 0.502 1.978
0.930 0.577 0.481 0.694 1.855
0.691 0.428 0.418 0.524 1.975
1.230 0.763 0.538 0.868 1.639
1.561 0.968 0.604 1.094 1.339
1.190 0.738 0.525 0.827 1.671
0.478 0.297 0.302 0.274 1.996
1.052 0.652 0.505 0.765 1.775
0.961 0.596 0.494 0.732 1.834
1.407 0.872 0.581 1.013 1.486
0.542 0.336 0.326 0.319 2.001
0.662 0.411 0.396 0.470 1.985
1.530 0.948 0.594 1.059 1.371
1.068 0.662 0.511 0.783 1.763
1.393 0.864 0.576 0.995 1.498
J. Hao e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100745
24
