How brand post content contributes to user's Facebook brand page engagement The experiential route of active participation

Texto completo

(1)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s

www.elsevier.es/brq

BRQ

Business

Research

Quarterly

REGULAR

ARTICLES

How

brand

post

content

contributes

to

user’s

Facebook

brand-page

engagement.

The

experiential

route

of

active

participation

Jesús

Gutiérrez-Cillán,

Carmen

Camarero-Izquierdo

,

Rebeca

San

José-Cabezudo

FacultaddeCienciasEconómicasyEmpresariales,UniversityofValladolid,AvenidaValleEsgueva,6,47011Valladolid,Spain

Received6June2016;accepted1June2017

JEL

CLASSIFICATION

M15; M30

KEYWORDS

Facebookbrandpage; Brandpostcontent; Experientialvalue; Userbehavioural engagement

Abstract Creatinganddevelopingafirm-hostedvirtualbrandcommunityformspartofa rela-tionshipmarketingstrategy;therefore,itmakessensetoevaluateitseffectivenessinterms ofrelationaloutcomes.Inanattempttoknowhowmarketerscanfostertherelationshipwith thebrandthroughvirtualcommunities,we positandestimateamodelofrelationalefficacy forafirm-managedFacebookbrandpage(FBP)inwhichthebrandpostscreatedbythefirm influencethebehaviouralengagementofindividualusersthroughtheutilitarianandhedonic valuesderivedfromtheirinteractiveexperienceswithintheFBP.Thefindingshighlightthat informationpostsstimulateuserbehaviouralengagementthroughtheutilitarianexperiential route.Asidefromanyexperientialrouteandadoptingamoredirectpath,interactionpostsare themaindriversofengagementbehaviour.Imagepostscontributetowardstheperceptionof utility,butinnowayaffectengagement.Finally,inordertogainadeeperinsight,weexplore themoderatingeffectofuserbrandpurchaseintensityontherelationspositedinthemodel. ©2017ACEDE.PublishedbyElsevierEspa˜na,S.L.U.ThisisanopenaccessarticleundertheCC BY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

The marketing landscapeis changing. The new marketing environment, dominated to a large degree by the emer-gence of the Internet and social networks, is imposing a shift from conventional relationship marketing towards

Correspondingauthor.

E-mailaddress:camarero@eco.uva.es(C.Camarero-Izquierdo).

a ‘‘transcending view of relationships’’ (Vargo, 2009) or an ‘‘expanded view of relationship marketing’’ (Brodie etal.,2011; Vivek etal., 2012), in whichthe customer’s interactiveexperiencesand‘‘customerengagement’’play a central role and in which engaged customer involve-mentinthefirm’sactivities ismoreproactive,interactive and co-creative (e.g., Brodie et al., 2011; Prahalad and Ramaswamy,2000).

Virtual brand communities constitute an exceptional research context in which to explore firms’ capacity to

http://dx.doi.org/10.1016/j.brq.2017.06.001

(2)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s produce interactive experiences and promote relational

engagementamongcustomers(Rellingetal.,2016).Inthese onlinespaces,individualscometogetheraroundsome dis-tinctinterest (e.g.,abrand)tocontactandinteract with eachotherinordertoexchange,shareandpoolresources, such as information, knowledge, experiences, entertain-ment, socio-emotional support and friendship, through diverse computer-mediatedcommunication systems(e.g., Jinetal.,2010a;Preece,2001).Inthissense,avirtualbrand communityis,firstofall,an‘‘onlinecommunitybasedon socialcommunicationsandrelationships’’(deValcketal., 2009,p.185),a‘‘webofpersonalrelationships’’(Rheingold, 1993) or a ‘‘fabric of relationships’’(McAlexander etal., 2002,p.38).Attheveryleast,avirtualbrandcommunity providesa‘‘socialstructuretotheC2Crelationshipinthe consumer-brand-consumer triad’’(Wu and Fang, 2010, p. 573).

With the rapid diffusion and widespread use of social network sites (SNS), more and more firms are investing inSNS-basedbrandcommunitiestobuildrelationshipsand to encourage users to exchange knowledge about their experienceswith thebrand or the firm(Ruiz-Mafe etal., 2014).Millionsofconsumersareconnectedtotheirfavourite brandsthroughsocialnetworkssuchasFacebook,Twitter, YouTube,andothers(Statista,2016).Unfortunately,thisin nowaymeansthatallthesecommunitiesaresuccessfuland that all their individual members areactive participants. ThereportbyHamptonetal.(2012)onThePewResearch Center’sInternet&AmericanLifeProjectconcludesthat, onaverage, Facebookusers‘‘get more fromtheir friends onFacebookthantheygivetotheirfriends’’andthatthe typicalFacebookuseris‘‘moderatelyactive’’inperforming specific Facebook activities. Suchresults concur withthe findingsobtainedbyGummerusetal.(2012,p.870),Pöyry etal.(2013,p.232)andvanVarikandvanOostendorp(2013, p.456),amongstothers,inthecase ofvirtualbrand com-munities.Infact,byadoptinganaivetechnically-oriented approachandbypayinginsufficientattentiontothesocial interactionsrequiredtobuildatruecommunity,many ini-tially attractive online communities may have failed to retainandenduringlyengagetheirmembersand,not sur-prisingly, as a result have become ‘‘cyber ghost towns’’ (Preece,2001).Inanycase,regardlessofitslevelofsuccess, whatmakes a company’s social site (such as a company-hostedFacebookpage)recognizableasacommunityisthat itisastructuredsetofsocialrelationshipsamongmembers that share a common interest, i.e., they areadmirers of abrand(e.g.,CvijikjandMichahelles,2013; Pöyryetal., 2013;Zaglia, 2013; Munnukkaetal., 2015; Relling etal., 2016).

Given such a general framework, ourresearch focuses ontherelationalcontextdefinedbyaFacebook-pagebrand community(whichisfounded,managed,andcontrolledbya firm)inordertopinpointthekeydriversofuserbrand-page engagement (as a good indicator of community success). Specifically, our research question is how marketers can reach customers and stimulate their community engage-ment in the Facebook pagecontext. Thus, we centre on thosedeterminantsofrelationshipmarketingstrategy suc-cessinFacebookthatcanbedirectlycontrolledbythefirm, suchasso-called ‘‘brand posts’’. The present work seeks toestimatearelationalefficacymodelofaFacebookbrand

page(henceforth,FBP),whichwillallowusto(1)gaugethe extenttowhichbrandposts(henceforth,BP)orposts cre-ated bythefirm runningthe Facebookpagehelpimprove pageusers’overallrelationalexperienceand,throughthis indirect pathway, actually serves to foster user engage-ment behaviourand(2) ascertainwhetherthe orientation or contenttype of BP can determine the kind of experi-entialvalue (utilitarian or hedonic)obtained by theusers and theirlevel of pageengagement.We thus explore the degreetowhicheachtypeofBPencouragesuserbrand-page engagementthroughtheutilitarian andhedonic experien-tialroutes.Inaddition,weexplorethemoderatinginfluence whichtheuser’sbrandpurchaseintensity (i.e.,howmuch ofthebrandtheybuy)mighthaveonthemodel’sstructure andpaths,oneaspectnottotakenintoaccountthusfarin thecontextofFBPrelationalefficacy.

PreviousliteratureonthecontextofFBPshasempirically analyzedtheresponseofpageuserstobrandposts.For ins-tance,deVriesetal.(2012),CvijikjandMichahelles(2013), Sabateetal.(2014)andLuarnetal.(2015),amongstothers, examinehowthecontenttype(amongothercharacteristics) ofaBPdirectlyimpactsonthe‘‘popularity’’ofthatBP(as indicatedbythenumberoflikes,commentsandshareson the BP).Compared tothe worksmentioned, ourresearch evidencestwosignificantdifferencesintermsofapproach andmethodologywhichshouldbehighlighted.

First,whiletheunitofanalysisofprevious worksisthe BP,theunitofanalysisinthecurrentworkisthebrandpage andthepage-userbehaviour.Ourresearchfocusiscustomer behaviourintherelationalcontextofavirtualbrand com-munity,inwhicheachindividualmemberholdslinkswiththe communityasawhole,withtheothermembersofthegroup, withthebrandandwiththefirm(McAlexanderetal.,2002). Thedependentvariableinourmodelisnot,therefore,the community’s directandglobalresponsetoeach BPor the efficacy per brand post, but each user’s overall response to the brand page (i.e. user brand-page engagement) or therelationalefficacyperpageuser.Moreover,contraryto previous works,which evaluateobjectiveindicators of BP contenttypeandengagement,theindependentvariablesin thecurrentresearchcorrespondtousers’appraisalsvis-à-vis theinterestarousedinthembyvarioustypesofBP depend-ing on their content, and the dependent variable is the users’subjectiveevaluationoftheiractiveinvolvementin thepage’srelationalactivities.Inotherwords,wemeasure the numberof likes, commentsand sharesper page-user, notasanobjectivefigure,butratherastheuser’ssubjective evaluationoftheiractualbehaviour.

(3)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s In summary, the current paper contributes to the

lit-erature on the relational efficacy of SNS-based brand communities.Byadoptingadifferentperspective,it anal-yses the individual users’ experiential and behavioural responsestoaFBPandthenexplainsitsefficacyper page-user.Inaddition,thestudyoffersinsightsaboutthewayin whichthe firm-generatedcontentcanfosterusersto par-ticipatein theFBP activities.The empiricalfindings show thattheutilitarianexperientialrouteclearlydominatesthe hedonicroute.Thus,brand-relatedcontentpostedbyfirm ultimatelyresultsinactiveengagementwiththeFBPifusers perceiveit astruly useful insolving problemsand making decisions.Contrarytoexpectations,entertainmentcontent, evenifappreciatedandenjoyedbyusers,doesnotpromote their active participation.Anyway, while the experiential route isperhaps themostconsistentwiththerelationship marketingorientation,itisnottheonly(northeshortest) routeofuserbrand-pageengagement.

Theoretical

background

and

research

hypotheses

Userbrand-pageengagementasanindicatorof relationalsuccess

Digital technology applications may supply the relational online space, but member-generated content and inter-personal communication provide the essence (Hagel and Armstrong,1997;Ridingsetal.,2002;WuandFang,2010) as well as the formative and shaping force (Bagozzi and Dholakia, 2002) of a virtual brand community. In other words, the value of a virtual community stems from the content produced and shared during member interaction and conversation (Jinet al., 2010a). In light of this, the activeandongoingparticipation(i.e.,engagement)ofeach member in the community’s public forum is regarded as a keyingredientin weavingthe relational fabric ofa vir-tual brand community (Rheingold, 1993; Rothaermel and Sugiyama, 2001) and crucial to ensuringthe community’s survivalandendurance(Algesheimeretal.,2005;Kohand Kim, 2004). In addition, creating and developing a firm-hosted community formspart of arelationship marketing strategy (e.g., Pitta and Fowler, 2005; Rothaermel and Sugiyama, 2001), so it makes sense to evaluate the effi-cacy of a virtual brand community in terms of relational outcomes.Sinceabrandcommunitybringstogether individ-ualsaroundabrandasafocalpointfortheirinterestsand discussions,individuals’communityengagementwillalsobe perceivedbythefirmrunningthecommunityasareliable indicatorofhowsuccessfulitsrelationshipmarketing strat-egyis(WiertzanddeRuyter,2007).

Althoughtherearemanyandwide-rangingdefinitionsof engagement(seeBrodieetal.,2011;Dessartetal.,2015; Hollebeeketal.,2014;vanDoornetal.,2010orViveketal., 2012forareviewofrelevantliterature),broadlyspeaking, engagementmaybeconceivedastheintensityof an indi-vidual’sconnectionandinteractionwithaparticularobject, activity, agent, idea, value or symbol. In the customer relationshipmarketingcontext,the‘‘engagementsubject’’ (i.e., the engaged individual) may be either a current customer or a potential customer, and the ‘‘engagement

object’’mayrefertoabrandorproduct,anorganization’s offering or activity, or the others in the social network (e.g.,abrandcommunity)createdaroundthebrand, prod-uct, offering or activity (Vivek et al., 2014). Customer engagementis basedonthe existenceoffocalinteractive customerexperienceswiththespecificengagementobject (Brodie et al., 2011), reaches ‘‘beyond purchase’’ (MSI, 2010, p. 4) and embraces the customer’s overall partici-pationwithinandoutsidethe specificexchangesituations (Vivek etal.,2012).Finally, customerengagementis cha-racterizedbyitsdynamicandinteractivelygeneratednature (Hollebeek etal.,2014)aswellasbyitscomplex multidi-mensionalstructure(Brodieetal.,2013),whichcomprises relevantcognitive,affective,behaviouralandsocial compo-nents(Viveketal.,2012).

In the current research, our interest focuses on the social-behavioural dimension of customer engagement, known as ‘‘customer behavioural engagement’’ (e.g., Gummerusetal.,2012;Zheng etal., 2015)or ‘‘customer engagement behaviour’’ (e.g., van Doorn et al., 2010). Specifically, in the case of a Facebook brand page(FBP), users’behaviouralengagementismanifestedthroughtheir activeparticipationin the functionalitiesFacebook offers (Gummerusetal.,2012;Luarnetal.,2015):clicking,liking, commentingandsharingbehaviours(Wallaceetal.,2014), whichweterm‘‘userbrand-pageengagement’’.

Effectofbrandpostsonuserbrand-page engagementbytheexperientialroute

Intheliteraturerelatingtothedeterminantsofcustomer’s integrationandactiveparticipationinsocialnetworksand onlinecommunities,therearetwomainstreamsofresearch although distinguishing between them does not always proves an easy task.The first group encompasses studies which focus on the role of the ‘‘benefits sought’’ (Cotte etal., 2006) as factors which determine community user behaviour (e.g., Hennig-Thurau et al., 2004; Liao et al., 2013; Pöyry et al., 2013), by considering that the pur-suit of these benefits is the main motivational driver of behaviour.Closelylinkedtothis,theothergroupofworks (e.g.,Dholakiaetal.,2004;Jinetal.,2010a,b;Mathwick et al., 2008; Wang et al., 2007; Zhou et al., 2013) cen-tresoncommunitymembers’relationalexperienceandthe ‘‘benefitsderived’’fromtheirintegrationinthesocial net-workasdeterminantsoftheirbehaviour.Thepresentwork isframedwithinthissecondcategory.

(4)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s Adapting the Brakus et al. (2009) definition of brand

experiencetoourcase study,theuser’sbrand-page expe-rienceis aterm employedtorepresentthe cognitiveand affectivereactionssparkedintheuserbypage-related stim-uli. Page experience occurs during the whole process of interactionwiththeFBPandstemsfromthesetof interac-tionsinvolvingtheuserovertime.Thisisnot,therefore,a one-offexperience(e.g.,ashort-livedemotion)broughton byaparticularstimulus(e.g.,aBP),butanoverallrelational experienceresulting from an accumulation of interactive experiences.Insofarastherelationalexperienceitselfcan bemoreorlessrichinvalue,usersevaluatetheoverall expe-rienceintermsofthebenefitsitprovides.Experientialvalue thusemergesasaconsequenceofcumulativeexperiences. By joining a virtual community, users may obtain dif-ferenttypesofrelationshipbenefits(i.e.benefitsresulting fromintegration),suchaspracticalbenefits,socialbenefits, social enhancement, entertainment and economic bene-fits(Gummerusetal.,2012),or differenttypesof values, suchaspurposive(includinginformationalandinstrumental) value,self-discovery, maintaininginterpersonal connectiv-ity,socialenhancementandentertainmentvalue(Dholakia et al., 2004). Reflecting the utilitarian versus hedonic dichotomy(HirschmanandHolbrook,1982),Mathwicketal. (2001)conceptualize the ‘‘experientialvalue’’ withinthe context of online purchasing and distinguish between an intrinsicorhedoniccomponent(e.g.,visualappeal, enter-tainment, escapism and enjoinment) and an extrinsic or utilitariancomponent (e.g.,service excellence,efficiency and economic value). In the case of virtual communities (Dholakiaetal.,2004;Jinetal.,2010a),theextrinsic expe-rienceislinkedtothepurposiveor utilitarianvalue,that is,tothe informationalandinstrumentalbenefits derived fromaccessinginformationontheproductanditsuses, solv-ingspecificproblems,andlearningfromothers.Incontrast, theintrinsic experienceis linkedtothe social hedonic or entertainmentvaluederivedfromsearchingforplayfulness, fun,amusement,fantasy,sensorystimulationorrelaxation byinteractingwithothers.Asanactivesourceofintrinsic value,playfulnessinvolvesenjoymentandescapism,andis reflectedinthepleasurethatcomesfromengagingin activ-itiesthatareabsorbingandprovidean opportunitytoget awayfromdailyroutine(Mathwicketal.,2001).

Inbothutilitarianandentertainmentexperiential dimen-sions,if communitymembers arenot satisfied,therewill benoincentivetoparticipate(ortocontinueparticipating) inthecommunity.Onthecontrary,ifindividualmembers’ experiencesinprevious interactionswithinthecommunity provepositive,theywillfeelsatisfiedandmotivatedto par-ticipate(ortoparticipatemoreactively)inthecommunity activities(Casalóetal.,2010).Inthisline ofthought,Jin etal.(2010a)evidencetheinfluenceofpositive disconfirma-tionof bothutilitarian andhedonicexperientialvalueson individualmembersatisfaction andaffective commitment and,inturn,on‘‘continuanceintentiontoparticipate’’in anonlinecommunity.Likewise,Loureiroetal.(2015)note howasatisfiedmemberismorewillingtoengagedeeply.

This effectofpositive experiencesonactive participa-tionmight alsobeexplainedasthe resultof ‘‘thecallof duty’’(Wiertz and deRuyter, 2007) or a sense of indebt-edness(ChanandLi,2010),accordingtowhichcommunity memberstend toreciprocate thesocial supportreceived.

Jin etal. (2010b)show that, in return for the social and functional benefits gained from community participation, members reciprocate with their affective and calculative commitment tothe virtual community. In our case, users whogainutilitarianandhedonicexperientialvaluesby par-ticipating in the FBP are more likely to reciprocate. In this regard, active engagement is understood as a recip-rocating behaviour to reduce the sense of indebtedness. Furthermore,participationmayalsobeduetopatternsof imitation and contagion. Based on observational learning theory,Zhouetal.(2013)explorethetransformation mech-anismthatconvertsvisitors intoactivemembersin online brand communities. If, by viewing posts, visitors realize the benefits of activemember participationand perceive posts ashaving informational and social values,they will beinterestedinimitatingmemberbehaviourandinseeking greaterbelongingnesstothecommunitythroughastronger intentiontoparticipate.Finally,inline withtheflow the-ory (Novak et al., 2000), if community users experience ‘‘flow’’intheironlineinteractions(morelikelywhenthey areinvolvedinenjoyableactivitiesthatprovidethemwith utility and entertainment), they willbe more inclined to spendtimeandeffortincontributingtothecommunity.In fact,Hall-Phillipsetal.(2016)reportadirectandpositive linkbetweenescapismandengagementinsocialnetworks. All of this drives us to advocate the experiential ori-gin ofengagement andsuggest apositive directeffect of theexperientialvalueobtainedonthelevelofengagement expressedbypageusers.Inturn,users’experientialvalue comes from connecting to a brand page and stems from the global evaluation of their relationship history within the community(i.e.,their experientialhistoryof interac-tionwiththebrand,thefirm andother page-users).Even so,in thisprocessofproducing both utilitarianand hedo-nicexperiential values,pagepostsplaya particularlykey role.WhenusersareexposedtoFacebookpageposts,the elicitedaffectiveandcognitiveelaborationsdetermine atti-tudestowardstheposts(Chenetal.,2015),whicharelater transferredtoexperiencesandperceptionsofexperiential value.Asaresult,theinterestwhichusersperceiveinthe page posts will determine the quality of their relational experiencewiththebrandpage.

(5)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s routine (entertainment posts). Assuming, therefore, that

BPsaredesignedtoofferarangeofbenefits(e.g., entertain-mentorfirm-andbrand-relatedinformation)ortosparkan arrayofdifferentresponses,theinterestwhichsuchcontent arousesisexpectedtoshapeuserexperienceintheFBPand thetypeandlevelofexperientialvaluetoemerge.

Taken together, the preceding arguments suggest that experiential value mediates the effect of brand posts (BPs)onusers’pageengagementbehaviour.Thedistinction between ‘‘utility’’ and ‘‘entertainment’’ as experiential valuesleadsustoconsiderthepossibletwo-foldinfluence ofBPsonuserbehavioural engagement.Indeed,asshown in Fig. 1, we propose two experiential routes of active engagement:theutilitarianroute,whichderivesfrom infor-mationalvalue,andtheentertainmentroute,whichderives from enjoyment and escapism values. Such a considera-tion of experiential values as mediating variables entails the assumption that (1) BPs are factors inducing expe-riential value and (2) active engagement is a behaviour induced by experientialvalues.Accordingly, thefollowing hypothesis(brokendownintomorespecificandoperational sub-hypotheses)issubmittedforempiricaltesting:

H1. OnaFBP,experientialvaluemediatestheinfluenceof brandpostsona user’spageengagementbehaviour, inas-muchasbrandpostscontributetowardstheperceptionof bothutilitarian(H1a)andhedonic(H1b)experientialvalues (albeitdifferently,dependingontheorientationorcontent typeof brandposts) and,in turn, the perceptionof both utilitarian(H1c)andhedonic(H1d)experientialvalues con-tributespositivelytoauser’spageengagementbehaviour.

Themoderatingeffectofauser’sbrandpurchase intensity

Giventhatthepresentstudyfocusesonanalysingthe expe-riencesandengagementofindividualswhodisplayapositive attitudetowardsthebrand(inferredbecausetheyarefans ofthebrandpage)andwhoalsotendtopurchasethebrand, wewerekeentofindoutwhethertheamounttheypurchase inanywayinfluencestheeffectsconsidered.Empirical evi-dence has shown that more intense purchasers are more engagedandmaintainmorestableandlonger-lasting rela-tions with the firm (e.g., Reinartz and Kumar, 2003). In the present context, we might also speculate that brand purchase intensity could affect page-users’ perceptions, attitudesandbehavioursand,therefore,mightinfluencethe extenttowhichinteractingwiththeBPsgeneratesutilityor entertainmentand,inturn,fostersengagement.

AsregardstheeffectofthetypeofBPontheperceived experience (H1a and H1b), it could be argued that more intensepurchasers,thankstotheirgreaterbrandpurchase and consumption experience, are more familiar with the brandandwiththefirm’sproductsand,therefore,perceive lessutility inposts thatofferinformativecontent. Should thisbethecase,theeffectofinformationalpostson utili-tarianvaluewillbeless,thehigherthepurchaseintensity (negativemoderatingeffectonH1a).Yet,regularbuyersof the brand might also be thought tobe more sensitive to thelatest informationonthefirm’sproductsandits inno-vations,which would allow usto conjecture theopposite

informationalposteffecttotheonesuggestedearlier(i.e., apositivemoderatingeffectonH1a).Similarobservations mightalsobemadewithregardtootherkindsof BPsand theothertypeofexperientialvalue.

Likewise,itisalsopossibletoputforwardcontradictory hypothesesforthemoderatingeffectofpurchaseintensity onthe relationssuggestedinH1candH1d between expe-rientialvaluesandengagement.Ontheonehand,itseems reasonabletoconsiderthatthemostintensepurchaserswill be more engaged with the brand and will, therefore, be moreinclinedtowardspageengagementirrespectiveofhow theyevaluatetheirexperienceintheFBP.Insuchinstances, pageengagementwouldbeshapedlessbyexperiential val-ues when the user is a more intensepurchaser (negative moderatingeffectonH1candH1d).Nevertheless,itmight alsobearguedthat purchaserswhoaremoreintenseand involved with the brand behave motivated more by the callofreciprocityand,therefore,adjusttheirengagement behaviourbettertothelevelofvaluetheyreceive.By con-trast,purchaserswhoarelessintenseandlessinvolvedwith thebrandsmightactlessmotivatedbyreciprocityandmore by opportunism,leading them to displaya lower level of engagementwiththe page,somewhat irrespectiveof the valuereceived.Inthiscase,thepositiveeffectofthe expe-rientialvalueonpageengagementwillbeclearerformore intensepurchasers(positivemoderatingeffectonH1cand H1d).

Sincewe put forward this purchase-intensity effect as an exploratory question, we will test empirically for the moderatingeffect,but do not posit a specific directional hypothesis.

H2. The page-user’sbrandpurchase intensity moderates thedirectandindirecteffectsconsideredinhypothesisH1, i.e.,the directinfluence ofBP onexperientialvalue, the directinfluence ofexperientialvalueonengagement,and theindirectinfluenceofBPonengagement.

AlltheproposedrelationsareshowninFig.1.

Data

Sample

Theinformationrequiredtoestimatethemodelandtotest thehypotheseswastakenfromasampleoffansina Face-bookpage of a Spanish brand of women’s fashion run by thefirm itself.The fashionbrand,whichwishes toremain anonymous,currently boasts some 27,000 Facebookfans. AlthoughthebrandiscommercializedindifferentEuropean countries,the fanpageis addressedtoSpanishcustomers (itisinSpanish).

(6)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s

H1a UTILITY

(experiential value)

ENTERTAINMENT

(experiential value)

CONTENT

TYPE OF BRAND POST

USER BRAND-PAGE ENGAGEMENT H1d

H1c

H1b

Utilitarian experiential route of engagement

Hedonic experiential route of engagement

H2 (moderating effect)

Brand purchase intensity

Figure1 ResearchModel.

to find both hedonic and utilitarian experiences. In the case of the Facebookpage chosen for the present study, thecontentgeneratedbythefirm(BP)includes catalogue photos,descriptionsofclothes,informationonpromotions, ideas for matching clothes and accessories, ‘‘the making of’’videos,photosoffansandfamouspeoplewearingthe brand’sclothes,opinionpollsaboutnewdesigns,prizedraws amongst fans aswell asinvitations to take part in activ-ities organized by the page, information concerning the brand’spresenceat fairsandfashion shows togetherwith appearancesinthe socialmedia(television, magazines or blogs), etc. On said brand page, the firm does not post messagesthatdonotrefertothefirmorthebrand,in refe-rencetothetypeofcontentwhichCvijikjandMichahelles (2013),deVriesetal.(2012)orLuarnetal.(2015)termas ‘‘entertainmentposts’’.

Inordertogatherinformation,alinktoanonline ques-tionnairewasinsertedinapostoftheFacebookfan-page.In anefforttoencouragepeopletoreply,fanswereoffereda 10Eurovoucherfortheirnextonlinepurchase.Weobtained 266responses.Sincethefirmdisposedoftheemailsaddress providedbyrespondents,itwasnotedthatsomeindividuals submittedthequestionnairetwice.Afterdeletionof dupli-catedanswers,afinalsampleof252responseswasaccepted for analysis. Given that thosesurveyed chose themselves whenresponding,thesample cannotbeconsidered repre-sentative.Nevertheless, we feel that it is a valid sample forstudyingthebehaviourofthemostactivefollowersand thosemostcloselylinkedtothebrand.Eventhoughthe sam-pleis not random, we calculated the sampling error and obtainedavalueof6.13%foraconfidenceintervalof95%, afinitepopulationsize (19,895fans)anda sample of252 interviewees.

As regards the description of the sample in terms of sex, age, qualifications and occupation, all of those who answeredwerewomen,52%wereagedbetween36and45, 69%heldauniversitydegreeand71%weresalariedworkers, reflectingfairlywelltheprofileofthebrand’spurchasers. Asfortheirpurchasebehaviour,overthepastyear,33%had boughtoverfourofthebrand’sgarments,37%betweentwo andfour,17%only oneand13%none. Finally,withregard to where they preferred to buy the brand, 52% stated a

preference for the physical store and 48% for the online store.

Measures

Measurement of the variables (Table 1) is based on a reviewoftheliteratureandonthescalesvalidatedinprior research, although some variables had to be adapted to the contextof thestudy.Exceptfor thenumber ofbrand articles purchased, all items were measured on 1 to 5 pointLikerttypescales.Userbrand-pageengagementwas measured using three indicators reflecting the frequency (from1:virtuallynever,to5:quiteoften)withwhichthose surveyed engaged withthe brand pagethroughthe three optionsofferedbyFacebook:like,commentandshare(e.g., CvijikjandMichahelles,2013;Luarnetal.,2015).Although theseindicatorsreflectdifferentbehaviours,theyare usu-allycloselyrelated,andsoweretreatedasreflectiveitems. Asregardsuserexperiencesinthesocialnetwork,the hedo-nicexperientialvalue(i.e.,entertainment)wasmeasured withtwoindicatorsofenjoymentandoneindicator reflect-ingescapism usedbyMathwicketal.(2001),whereasthe utilitarianexperientialvalue(i.e.,utility)scalewasbased onprevious scalesof perceived utility (e.g., Davis etal., 1989; Kulviwat etal.,2007) butadapted tothecase of a fashionsocialnetwork.

(7)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s Table1 Variablesmeasured.

Variablesandindicators Mean S.D. Loadings

Engagement(=0.88;CR=0.92;AVE=0.80)

Itakepartinthepostswith‘‘likes’’ 2.88 1.261 0.847***

Icommentontheposts 2.13 1.055 0.908***

Isharethepostswithotherpeople 2.31 1.114 0.942***

Utility(˛=0.92;CR=0.94;AVE=0.80) Thefan-page...

Helpsmetodecidewhattobuy 3.25 1.100 0.884***

Helpsmetodecidewhattowearorhowtomatchupclothes 3.12 1.137 0.877***

Offersmeusefulinformation 3.51 1.024 0.902***

Givesmeideas 3.63 0.991 0.917***

Entertainment(˛=0.92;CR=0.95;AVE=0.86) Thefan-page...

Providesawelcomedistractionfrommydailyroutine 3.27 1.145 0.899***

Isfun 3.44 1.060 0.952***

Reallyentertainsme 3.15 1.101 0.925***

Informationposts(˛=0.83;CR=0.88;AVE=0.66) Ipayattentiontoandaminterestedin...

Postsinformingmeaboutspecialoffers 4.25 0.960 0.793***

Postswithcataloguephotos 4.08 1.009 0.801***

Postswithdescriptionsofgarments 3.87 1.078 0.818***

Postswithideasabouthowtomatchupclothes 3.77 1.115 0.831***

Imageposts(˛=0.92;CR=0.94;AVE=0.76) Ipayattentiontoandaminterestedin...

Postsabout[theBrand]appearingontelevision 3.37 1.147 0.943***

Postsabout[theBrand]appearinginblogs 3.30 1.142 0.920***

Postabout[theBrand]appearinginmagazines 3.37 1.130 0.919*** Postswithphotosoffamouspeoplewearing[theBrand] 3.33 1.149 0.867***

Postswithvideosonthemakingof 3.18 1.207 0.701***

Interactionposts(˛=0.88;CR=0.92;AVE=0.69) Ipayattentiontoandaminterestedin...

Postsencouragingmetouploadfanphotos 3.06 1.193 0.839***

Postsencouragingmetocomment 3.09 1.147 0.920***

Postsaskingmetovote 2.94 1.157 0.852***

Postswithspecificmentionsoffansortheircomments 3.08 1.123 0.895***

Postsannouncingprizedrawsforfans 4.15 1.086 0.622***

*p<0.05(bilateral). **p<0.01(bilateral).

*** p<0.001(bilateral).

orideasformatchingupclothes).Forthesubsequent anal-yses,the itemrelating to‘‘prize draws’’,which factorial analysislocatedinthefactorwereferredtoas‘‘information content/posts’’, wasrelocatedtothe factor ‘‘interaction content/posts’’.Thisis becauseinthepageanalyzed,the prize draws appear as an incentive to take part or as a reward for userswho engagein the pageby liking, shar-ing or commenting. Finally, brand purchase intensitywas measured asthe number of garmentspurchased over the lastyearonafour-pointscale:(1)nopurchase,(2)one gar-ment,(3)betweentwoandfourgarmentsand(4)morethan fourgarments.Thisvariablewasdichotomizedsoasto cre-atetwousergroups:lower-intensity(LI)group,comprising the76fanswhopurchasedoneornogarments,and higher-intensity(HI)group,madeupofthe196fanswhopurchased twoormoregarments.

(8)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s Table2 Correlationmatrix.

Informationposts Imageposts Interactionposts Utility Entertainment

Informationposts 0.811

Imageposts 0.660 0.874

Interactionposts 0.698 0.755 0.832

Utility 0.684 0.603 0.630 0.895

Entertainment 0.684 0.584 0.678 0.749 0.925

Engagement 0.396 0.456 0.571 0.465 0.440

ThediagonalshowsthesquarerootoftheAVEofthereflectiveconstructs(allexceptengagement).

performed TypeI second-orderconfirmatory factor analy-sis(Jarvisetal.,2003)thatconfirmedthereflectivenature ofutility andentertainmentasdimensions ofexperiential value [2(13)=32.82, p=0.000, GFI=0.964, AGFI=0.923, CFI=0.987,NFI=0.979,RMSEA=0.078].

Inordertoavoid,oratleastminimize,commonmethod variance bias, we followed some recommendations made by Podsakoff et al. (2003) when designing the question-naire:respondentswereexplicitlyassuredthattherewere noright or wrong answers and that the information they providedwouldbetreatedconfidentially;itemwordingwas revisedsoastoavoid ambiguousorunfamiliarterms; dif-ferentresponseformatswereused;andquestionorderdid notmatchthecausalsequence inthe model.In addition, Harman’s single-factor test was performed. Evidence for commonmethodbiasexists whena singlefactor emerges fromthefactoranalysisorwhenonegeneralfactoraccounts for the majority of covariance among the measures. In ourcase,exploratoryfactoranalysiswithalltheindicators producedfivefactorswithan eigenvaluegreater than1.0 (accountingfor 76%ofexplainedvariance)andafirst fac-torexplainingonly22%ofvariance.Insum,theprocedural remediesappliedandthefindingsoftheabove-mentioned testsuggestthatcommonmethodbiasisnotamajor con-cerninthisstudy.

The unmeasured latent methods factor test (Podsakoff etal.,2003)wasalsoperformed.WeintroducedaCMV fac-torthatincludesalltheprincipalconstructs’indicatorsand calculated the degree to which each indicator’s variance wasexplainedby itsprincipal construct(i.e., substantive variance)andbytheCMVfactor.Whilesubstantivevariance averaged0.777,theaveragemethod-basedvarianceis0.023 (Table3). As the ratio of substantivevariance tomethod varianceisabout34:1,andmostofthemethodfactor load-ingsareinsignificant,thisanalysisalsoindicatesthatCMVis unlikelytobeacriticalfactorinthestudy.

Analysis

and

results

In orderto test the hypotheses posited, the partial least squares (PLS)method wasapplied usingSmartPLS (Ringle etal.,2005)software.Thelevelsofstatisticalsignificance ofthecoefficientsofthemeasurementandstructuralmodel werecalculatedusingbootstrappingwith500sub-samples. The results of estimating the structural model (Model 1 or full-mediation model) are shown in the first column of Table 4. In order totest the robustness of the results obtained, each latent constructs was reduced toa single

index(astheaverageofitsindicators) andthemodelwas estimatedbymeansofpathanalysis(usingAMOS23.0 soft-ware).Saidestimationallowsustoconsiderthereflective natureofthetwodimensionsofexperientialvalue (allow-ing the correlation of the measurement errors of utility and entertainment) and toobtain goodness-of-fit indices. AsshowninthesecondcolumnofTable4,thefitobtained did not prove satisfactory. The full-mediation model was then compared to several nested models considering the directeffectsofbrandpostcontentonengagement,thatis, thepartialmediationoftheutilitarianandhedonic experi-ential values in the relationship between type of BP and user engagement. This revealed that introducing a direct effect of the interaction posts on engagement significan-tly improved the model’s goodness-of-fit (third column in Table4).Lastly,thisquasi-mediationmodel(Model2)was estimated by means of PLS (fourth column in Table 4). ThesefinalresultswereusedtotesthypothesesH1a,H1b, H1c, and H1d. Despite the high correlation between util-ityandentertainment,thevaluesofthevarianceinflation factor (VIF) rule out the existence of problems of multi-collinearitybetweenthedeterminantsofengagement,both inthefullmediationmodel[VIF(Utility)=2.274;VIF (Enter-tainment)=2.274] and in the quasi-mediation model [VIF (Interaction posts)=1.976;VIF(Utility)=2.728;VIF (Enter-tainment)=2.711].

In light of the results, it can be confirmed that the content of BPsimpacts on the perception of experiential values. In fact, the three kinds of BP account for over fifty per cent of the variability of utility (R2=0.524) and entertainment(R2=0.547).Specifically,inagreementwith H1a, theinterest aroused byinformationposts (b=0.434, p<0.001),imageposts(b=0.161,p<0.05)andinteraction posts(b=0.206,p<0.05)positivelyandsignificantlyaffects the utilitarianexperientialvalue. Inagreement withH1b, informationposts(b=0.399,p<0.001)andinteractionposts (b=0.367,p<0.001)positivelyandsignificantlyaffect hedo-nicexperientialvalue.However,thesamecannotbesaidof imageposts,whoseeffectonentertainmentdoesnotprove significant (b=0.044, p>0.1). Aside from this exception, hypothesesH1aandH1bcanbeseentobesupported.

(9)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s Table3 Commonmethodvarianceanalysis.

Variablesandindicators Construct

loading(CL)

CL2 Method-Factor

Loading(MFL)

MFL2

Engagement

Itakepartinthepostswith‘‘likes’’ 0.887** 0.787 0.002 0.000

Icommentontheposts 0.921** 0.848 0.016 0.000

Isharethepostswithotherpeople 0.901** 0.812 0.018 0.000

Utility

Thefan-page...

Helpsmetodecidewhattobuy 0.870** 0.757 0.017 0.000

Helpsmetodecidewhattowearorhowtomatchupclothes 0.940** 0.884 0.073 0.005

Offersmeusefulinformation 0.936** 0.876 0.040 0.002

Givesmeideas 0.839** 0.704 0.091* 0.008

Entertainment Thefan-page...

Providesawelcomedistractionfrommydailyroutine 0.968** 0.937 0.080 0.006

Isfun 0.994** 0.988 0.049 0.002

Reallyentertainsme 0.817** 0.667 0.126** 0.016

Informationposts

Ipayattentiontoandaminterestedin...

Postsinformingmeaboutspecialoffers4 0.924** 0.854 0.135 0.018

Postswithcataloguephotos1 0.913** 0.834 0.132* 0.017

Postswithdescriptionsofgarments13 0.809** 0.654 0.015 0.000 Postswithideasabouthowtomatchupclothes15 0.612** 0.375 0.238** 0.057

Imageposts

Ipayattentiontoandaminterestedin...

Postsabout[theBrand]appearingontelevision 1.031** 1.063 0.095* 0.009 Postsabout[theBrand]appearinginblogs 0.997** 0.994 0.087 0.008 Postabout[theBrand]appearinginmagazines 0.892** 0.796 0.036 0.001 Postswithphotosoffamouspeoplewearing[theBrand] 0.982** 0.964 0.124* 0.015

Postswithvideosonthemakingof 0.379** 0.144 0.346** 0.120

Interactionposts

Ipayattentiontoandaminterestedin...

Postsencouragingmetouploadfanphotos 1.043** 1.088 0.222** 0.049

Postsencouragingmetocomment 0.982** 0.964 0.069 0.005

Postsaskingmetovote 0.814** 0.663 0.042 0.002

Postswithspecificmentionsoffansortheircomments 0.975** 0.951 0.086 0.007

Postsannouncingprizedrawsforfans 0.200 0.040 0.459** 0.211

Average 0.859 0.777 0.007 0.023

* Significancelevelp<0.05. ** Significancelevelp<0.01.

model(Model1)evidencesapositiveandsignificanteffect (b=0.209,p<0.01)ofentertainmentonengagement. Nev-ertheless,this effectloses allits powerand ceases tobe significant (b=−0.005, p>0.1) in the case of the quasi-mediation model (Model 2). We thus find support for hypothesis H1c, but are forced toreject hypothesis H1d. Finally,although itis notan effectwhichis consideredin hypothesisH1andintheoriginalmodel,weseehow inter-action posts have a strong direct impact on engagement (b=0.472,p<0.001).Indeed,thistypeofBPbecomesthe main driver of active participation, even above the two dimensionsofexperientialvalue.

By way of a complement to the results commented on, Table 5 shows the indirect effects of BP type on

(10)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s Table4 Results:hypothesizedandrespecifiedmodels.

Model1

Full-mediationmodel

Model2

Quasi-mediationmodel

PLSestimation Pathanalysis Pathanalysis PLSestimation

Antecedentsofutility

Informationposts→utility 0.434*** 0.418*** 0.418*** 0.434***

Imageposts→utility 0.162* 0.168* 0.168* 0.161*

Interactionpostsutility 0.206* 0.206*** 0.206*** 0.206*

Antecedentsofentertainment

Informationposts→Entertainment 0.399*** 0.391*** 0.391*** 0.399***

Imageposts→entertainment 0.044 0.044 0.044 0.044

Interactionposts→entertainment 0.367*** 0.370*** 0.370*** 0.367***

Antecedentsofengagement

Utility→engagement 0.309*** 0.303*** 0.173* 0.169*

Entertainmentengagement 0.209** 0.215** -0.009 -0.005

Interactionposts→engagement - - 0.472*** 0.472***

R2adjusted(Utility) 0.524 0.512 0.512 0.524

R2adjusted(Entertainment) 0.547 0.543 0.543 0.547

R2adjusted(Engagement) 0.236 0.234 0.347 0.347

Goodnessoffit

2(3)=41.27 2(2)=1.430 (p=0.000) (p=0.715) GFI=0.952 GFI=0.998 AGFI=0.663 AGFI=0.980 CFI=0.959 CFI=1.000 NFI=0.957 NFI=0.998 RMSEA=0.225 RMSEA=0.000

* p<0.05(bilateral). **p<0.01(bilateral). *** p<0.001(bilateral).

Table5 Indirecteffectsofbrandpostsonengagement.

Indirecteffects

Estimation P-values Informationposts→engagement 0.072 0.023 Imageposts→engagement 0.027 0.159 Interactionpostsengagement 0.033 0.310

aninsignificanteffect(bothdirectandindirect)inthecase ofimageposts,and(3)apurelydirecteffect(notmediated byexperientialvalues)inthecaseofinteractionposts.

With regardto exploratory hypothesis H2, namely,the moderatingeffectofbrandpurchaseintensityonthe rela-tionspositedinH1,weexaminethedifferencesbetweenthe twogroupscreatedwithregardtohowmanyofthebrand’s itemswerebought(i.e.,lowerandhigherpurchase inten-sity).Table6showstheresultsofthePLSestimationofthe multi-groupstructuralmodel. Inordertotest H2,we use theparametric significance testprovided bySmartPLS for thedifferenceofgroup-specificpathcoefficients.The para-metrictestalsoallowsustogaugemeasurementinvariance. Acrossthegroups,neitherthefactorloadingsnorthefactor weightsdiffersignificantly.Configuralandmetricinvariance canthusbeconfirmed(SteenkampandBaumgartner,1998).

Finally,Table7showstheindirecteffectsofthethreetypes ofBPonengagementforeachusergroup.

As observed when analysing the sample as a whole (Table4,model2),entertainmenthasnosignificantimpact onengagementin eitheroftheintensity groups(Table6). However, the effect of utility onengagement, which did provesignificantforthesampleasawhole(p<0.05),ceases tobesofortheLIgroup(b=0.160,p>0.1)andonlyproves marginallysignificantfortheHIgroup(b=0.175,p<0.1).In anycase,astheparametrictestreveals,inter-group differ-encesarenotintheleastsignificant.Asregardstheeffects of experiential values on engagement, the two intensity groupsthusfollowvirtuallythesamepatternofbehaviour, leadingustorejectthemoderatinginfluenceofbrand pur-chaseintensityontherelationspositedinH1candH1d.

(11)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s Table6 PLSmulti-groupanalysis.

BrandpurchaseintensityComparisonsofpathcoefficients

Lowerintensity Higherintensity Parametrictesta

Antecedentsofutility

Informationposts→utility 0.427*** 0.419*** 0.958

Imageposts→utility 0.356*** 0.032 0.052

Interactionpostsutility 0.104 0.317*** 0.239

Antecedentsofentertainment

Informationposts→entertainment 0.482*** 0.359*** 0.331

Imagepostsentertainment 0.007 0.032 0.878

Interactionposts→entertainment 0.314* 0.422*** 0.518

Antecedentsofengagement

Utility→engagement 0.160 0.175 0.933

Entertainmentengagement 0.070 0.045 0.533

Interactionposts→engagement 0.600*** 0.382*** 0.148

R2adjusted(Utility) 0.531 0.562 0.740

R2adjusted(Entertainment) 0.631 0.501 0.236

R2adjusted(Engagement) 0.444 0.300 0.169

a Thismethodisaparametricsignificancetestforthedifferenceofgroup-specificPLS-SEMresultsthatassumesequalvariancesacross groups.Inthiscase,Levene’stestindicatesthatequalvarianceacrossgroupscanbeassumedforallthevariables.

* p<0.05(bilateral). **p<0.01(bilateral).

*** p<0.001(bilateral). p<0.10(bilateral).

Table7 Indirecteffectsofbrandpostsonengagementforeachgroup.

Brandpurchaseintensity

Lowerintensity Higherintensity Parametrictest

Informationposts→engagement 0.034 0.089* 0.408

Imagepostsengagement 0.056 0.007 0.288

Interactionposts→engagement −0.006 0.074 0.286

* p<0.05(bilateral). **p<0.01(bilateral). ***p<0.001(bilateral).

p<0.10(bilateral).

H2,brandpurchaseintensitydoesnotseemtomoderatethe directandindirecteffectsofinformationalcontent.

Inthecaseofimageposts,thesituationdiffersslightly. Asinthegeneralanalysis,thiskindofbrandedcontentdoes notcontributedirectlytoentertainmentvalueorindirectly tobehaviouralengagementineither ofthetwo purchase-intensitygroups.Nevertheless,thegeneraleffectofimage posts on the perception of utilitarian value (b=0.161, p<0.05)isnownolongerevidentinthecaseoftheHIgroup (b=0.032,p>0.1), but ismaintained andevenfurthered, bothintermsofpowerandlevelofsignificance,inthecase oftheLIgroup(b=0.356,p<0.001).Moreover,asevidenced by the parametric test, this differencebetween intensity groups is significant witha confidencelevel closeto 95%, whichprovidesonlyaverypartialsupportforthe hypoth-esisH2.Itwould appearthat BPswhich conveythebrand image prove useful for decision making, although only in thecaseoffanswhodisplaylowerpurchaseintensity (i.e.

negative moderating effect of purchase intensity on the relationpositedinH1a).

(12)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s thedirecteffectofinteractionpostsonengagement.Once

again, even though the difference between the intensity groups is not statistically significant, it does appear that said effect is greater amongst members of the LI group (b=0.600,p<0.001)thanamongstmembersoftheHIgroup (b=0.382,p<0.001),allowingustoventurethepossibility ofadampeningeffectofpurchaseintensityonthecapacity ofinteractionposts todirectly encourage participationin brandpageactivities.

In light of the resultsto emerge fromthis exploratory analysisandexercisingduecaution,weconjecturea possi-bletwo-foldmoderatingeffectofbrandpurchaseintensity ontheroleplayedbyinteractionposts:(1)apositive moder-atingeffectconcerningtheinfluencethesepostshaveonthe perceptionofutilityand(2)anegativemoderatingeffecton theirdirectinfluenceonengagement.Botheffectsmaybe seenassmallsignsinfavourofH2.

Discussion

In the new context of relationship marketing, social net-worksareanextremelyvaluabletoolforfirmswhenitcomes to handling their relations with clients and when creat-ingandmaintainingrealcommunitiesaroundtheirbrands. Specifically,beyondmerelyservingasawindowtopromote productsandbrands,FBPsmaycontributepositivelyto busi-nessperformanceprovidedthatrelationshipwithclientscan beenhancedandthatcustomerengagementin communica-tionand value creation through social interactioncan be furthered.Inthiscontext,firmsclearlyneedtodecidewhat theircommunicationstrategywillbeiftheyaretosecure therelationalengagementofpageusers.Basedonthis,in thepresent study we posit that moreactive commitment maybeobtainedfromusersbyofferingthemaninteractive experiencein the brand page,particularly an experience whichprovessufficientlyappealingandvaluable.

As contended by authors such as Brodieet al. (2011), Gummerusetal.(2012)orMalthouseandCalder(2011), cus-tomerengagementisattainedthroughrelationalcustomer experiences. The findings to emerge from the empirical analysisevidencethatexperienceswiththebrandpageand theexperientialvaluesobtained determineusers’levelof engagement,allowingustoconsideranexperientialroute ofengagement.Atthesametime,itseemsclearthatactive engagementcanbeunderstoodasareciprocatingbehaviour. Inorder toexplainthis general remarkmore clearly, two importantclarificationsshouldbemade.

Firstly, it should also be recognized that, contrary to whatwasconjectured,theexperientialrouteisnotthemain wayofpromotingengagement.Whilstnotdenyingthe rela-tiveimportanceoftheexperientialroute,engagementdoes appeartobean immediateresponse tothedirect callto participatemadebythefirmtopageusersthroughso-called interactionposts.

Secondly,weseehowtheexplanatorycapacityof expe-riential value is due exclusively to utility value, which means that the utilitarian experiential route of engage-ment fully prevails over the hedonic experiential route. Theinformationalsupportreceivedwhenmakingpurchase decisionsgeneratesafeelingofreciprocity whichcanonly bematchedbyreciprocation,abehaviourwhichisslightly

moreevidentinthecaseofpageuserswhodisplaygreater brand purchase intensity. Contrary to expectations, the particular case analyzed does not seem to indicate that entertainmentvaluemotivatesengagement.Itmaythusbe concludedthatperceived utilityencouragesactive partic-ipation, whereas entertainment produces receptive users but not active participants. One feasible explanation for thisresultcouldberelatedtothecontentoftheposts ana-lyzed.CivijikjandMichahelles’(2013)findingsrevealthat brandpostswhichcontainentertainingcontenttriggerthe highestlevelofactiveengagement(i.e.,liking,commenting andsharingbehaviours),followedbypostsproviding brand-related information and posts which offer remuneration. However,the FBP analyzed in our study does not contain specific‘‘entertainmentposts’’,evenifuserscanfind enter-tainment in other posts. Thus,the firm is responsible for perceived utilitarian value (informational BP), but is not responsible for perceived hedonic value. Therefore,users wouldreciprocateafirm’seffortswithactiveparticipation butwouldonlyreciprocateinformativeefforts.

Another possible explanation for this result concerns individuals’ motivations to use the FBP, which might help to ascertain whether engagement is determined by the perception of utility or the perception of entertain-ment.Unfortunately,thestudyoftherelationshipbetween users’ motivations and participation has yielded contra-dictory evidence.Pöyry,Parvinenand Malmivaara’s(2013) analysis reveals that hedonic motivations for using the company-hosted Facebook brand page relate to a higher propensity toparticipateinthe community,whereas utili-tarianmotivationsrelatemorestronglytomerelybrowsing thecommunitypage.Otherstudies,however,pointto par-ticipation linked to utilitarian-motivated individuals. For instance, Zaglia (2013, p. 219) affirms that, compared to Facebook groups, ‘‘in [company-hosted Facebook] fan pages, activities related to the community’s purpose are [more] central, and consumers participate mainly due to utilitarian(e.g.,gettinginformation)motives’’.Inthisline, ShaoandRoss(2015)exploredifferentstagesofconsumer interactionwithaFBPcommunityandconcludethat com-munity users require entertainment if their involvement with the FBP is to be sustained. However, as consumers become more sophisticated, ‘‘information seeking is the onlysignificantmotivefor anindividualtopost ona Face-bookbrandpage’’(p.253).Inourstudy,wedonotmeasure usermotivationalthoughwedofindthatFBPusersperceive bothutilityandentertainment(themeanvaluesofthe indi-catorsareover3).Therefore,inaFacebookfashion-brand page,both utilitarianandhedonicexperientialvaluesmay provetobe keyelementsvis-à-vis retainingindividuals in theFacebookcommunity.However,intermsofengagement, usersseemtobemotivatedbyutilitarianreasons,sinceonly theutilitarianexperientialvaluedeterminesactive partici-pationinFBPactivities.

AsregardshowefficientBPsareas(eitherdirector indi-rect) determinantsof engagement,ourfindings pointtoa differenteffectofeachtypeofpostdependingonthekindof contentinquestion:anindirecteffectofinformationposts (throughutility),adirecteffectofinteractionposts,andno effectofimageposts.

(13)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s utilityandentertainment,althoughtheirindirecteffecton

engagement is only evident throughthe utilitarian route. Onemay,therefore,senseacertainwishtorespondtothe callofdutysparkedbytheperceivedutilityofbrand-related information which aids problem solving and/or decision making.Nevertheless,thefactthatthisindirecteffecton engagement is not asstrong as anticipatedmight bedue totheexistence ofusers whoaremere information seek-ersandactaslurkers,inotherwords,whobrowsewithout reciprocating(ChanandLi,2010;deValcketal.,2009).

Unliketheabove,imagepostsinnowayaffect engage-ment.Throughtheseposts,thebrand‘‘showsoff’’itssocial relevancetoitsfansbyappearingin variousmedia. How-ever, this sparks no feeling of reciprocity amongst fans, perhaps becausesuchcontentisseen asaformof adver-tising. Even so, image posts do make quite a relevant contributiontoutility,albeitonlyinthecase ofuserswho displaylessbrandpurchaseintensity.Whentakingor boost-ingtheirpurchasedecisions,theselessintensepurchasers need toknowthat thebrandhas asocial presenceandis recognized in the market. Forsuch buyers, the fact that thebrandappearsonTVorradioprogramsandthatfamous people wear its clothes proves useful when it comes to evaluation and choice in the purchase decision process. Forregularpurchasersofthebrand,namelythosewhoare probablymorefamiliar itandhaveamoreclearlydefined image thereof, such posts are not so useful. What does prove surprising is that image posts fail to contribute to theperceptionofhedonicexperientialvalue.The possibil-ityshouldthereforebeconsideredthatimagepostsarenot comparableto‘‘entertainmentposts’’,messagesdesigned for users to enjoy, entertain and amuse themselves with contentthatisnotspecificallyrelatedtothebrand(Cvijikj andMichahelles,2013;Luarnetal.,2015).

Finally,interactionpostscontributetowardsthe percep-tionofutility(albeittoagreaterextentinthecaseofthe higher purchase-intensity group) and entertainment (very clearly in the two groups), without this effect becoming noticeable,througheitherofthetwoexperientialroutes,in behaviouralengagement.Inparticular,theeffectof inter-actionpostsonutilitarianexperientialvaluemeritsabrief explanation,sinceit might be surprisingthat postswhich providenokindofinformationmakeasignificant contribu-tiontoperceivedutility.Inourview,foraspecificuser,BPs whichencourageparticipationarenotusefulinthemselves, but do prove valuable in that they fosterthe production of content generated by other users (such as comments, opinionsoranswerstoquestions),whichmightbehelpfulin decisionmaking.Thus,thegreatertheamountof informa-tionavailableonabrandpagethankstotheseuserposts,the greatertheperceivedutilityofthepage.Throughthis indi-rectroute,interactionpostscontributetousers’utilitarian experientialvalue.

Whatwouldappeartobemoreevidentistheparticularly strong direct effect of interaction posts on engagement, aphenomenon found inboth groups of users.As aresult, what is interesting in this case is not so muchits expla-nation, which is obvious given the natureand purposeof thistypeof BP,but howit isinterpreted. Inthisvein,we findtwopossibleinterpretationswhichcorrespond, respec-tively,toapessimisticandanoptimisticviewofrelational realityinbrandpages.Ontheonehand,correspondingto

themorepessimisticview,usersingeneral(butmoreclearly lessintense purchasers)might be felttorespond directly tointeractionposts since theypromise more or less def-inite rewards (e.g., the right to take part in a draw) in exchangeforactiveparticipation.AscanbeseeninTable1, BPswhich generatemore attentionandprovemore inter-estingarepreciselythosewhichprovideinformationabout specialoffers(akind of informationpostwithan average interestscoreof4.25outof5)andthosewhich announce prize draws amongst fans(a kind of interaction post dis-playinganaveragescoreof 4.15).Moreover,theveryfact that subjects in the sample answered the questionnaire serves to support such an interpretation, since it should berememberedthatthesurveywasputupin an interac-tivepostthatrewardedusercooperationbyofferingthema voucher.

Incontrasttothispessimisticview,onemightbeinclined tothinkthatthelessrelevantroleoftheexperientialroute (favouringthedirecteffectofinteractionposts)isnot nec-essarily indicative of a low level of commitment to the brandpage. Perhaps, users who receive experiential val-uesreciprocate with‘‘brand engagement’’ (e.g.,positive word-of-mouthin theirmost personalenvironment, brand advocacy, brand acceptance, brand loyalty, inhibition to switchbrand,propensity tostay inthe brandrelationship andother consumerbrand-supporting behaviours) outside the brand page, but not in the public forum of the vir-tualcommunity. The fact that they donot correspond to the value received with their active participation in the pageactivitiesmightbeduetotheinhibitingeffectof cer-tainpersonality traits, such as having a more withdrawn orintroverted character(e.g.,Kabadayi andPrice,2014), aversion to publicly expressing their opinions or drawing attention, a fear of being labelled as show-offs,concern that their opinions and commentsmight go unnoticed or unappreciated,orlackoffaithintheirabilitytomake con-tributionsthat proveoriginaland interestingtoothers.In such instances, interaction posts provide the opportunity andthe means of expression these userswere seeking to participateunselfishly in the FBP and torepay the bene-fitsreceived.Unfortunately,theinformationavailabledoes not allow us to gauge to what extent which of the two views(orindeedacombinationofthetwo)isthedominant one.

(14)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s Table8 Synthesisofcontributionsforacademicsandpractitioners.

Academiccontributionsandresearchfindings Managerialimplications

Researchpositsandestimatesamodelofrelational efficacyforaFacebookbrand-page(FBP)community inwhichthefirm-generatedcontentintheformof brandposts(BP)influencesthebehavioural engagement(i.e.,thelevelofparticipationinpage activities)ofindividualmembersthroughboth utilitarianandhedonicexperientialvaluesderived fromtheirinteractionhistorywithintheFBP.

Evaluatingtherelationalbehaviourandsuccessofa firm-hostedFBPcommunityisrecognizedasan importanttooltoimprovetheplanningofonline marketingstrategy.Onthisbasis,byclarifyingthe relationbetweencommunitymembershipandactive engagement,themodelilluminateshowafirmcan assesstherelationalefficacyofitsFBPandprovides guidanceonmanagingcommunityrelationships.

UnitofanalysisistheFBPandthebehaviourof individualuser.Thus,contrarytopreviousworks,the finaldependentvariableisnotthecommunity’sdirect andglobalresponsetoeachBP(i.e.,efficacyper brandpost),buteachuser’soverallresponsetothe FBP(i.e.,theso-called‘‘userbrand-page

engagement’’asreliableindicatorofrelational efficacyperpageuser).

FBPefficacycannotbeexaminedthroughtheimpactor popularityofeachparticularBPinisolation,butin conjunction.RelationalefficacyofaFBPcommunity shouldbemeasuredthroughtheirmembers’cumulative experientialandbehaviouralresponsestothewholeset ofcontentpostedovertimebyfirm.

Moreactivecommunityengagementmaybeobtained fromFBPusersbyofferingthemasufficiently appealingandvaluablerelationalexperience.

Ingeneral,firmholdingaFBPshouldpostcontentthat attractsparticularinterestfromusers,contributesto improvetheirinteractiveexperience,providesthem withexperientialvaluesandtrulyservestopromote theiractivecommunityengagement.

Theutilitarianexperientialrouteofengagementfully prevailsoverthehedonicexperientialroute.Since entertainmenthasnoimpactonengagement,the hedonicexperientialroutedoesnotappeartobe applicable.

Inthecontextoffashionbrandpages,evenifallBP typescontributetoproducingexperientialvalues,very specialattentionshouldbedevotedtoBPsconveying brand-relatedinformation,whichmaygeneratemore utilityandmakeuserstofeeltheneedtoreciprocate withtheiractiveengagement.

Theexperientialrouteofuserbrand-pageengagement isperhapsthemostconsistentwiththerelational spirit,butitisnottheonlyone.Asidefromany experientialroute,activeengagementisalsoseenas animmediateresponsetothedirectcallto

participatemadebythefirmtocommunitymembers throughso-called‘‘interactionBP’’.

Ifhandledjudiciously,interactionBPs(intheformof directinvitationstoparticipate,usuallyinreturnfor somepotentialreward)arethemaininstrumentto promoteactiveengagementandinvigoratecommunity life.However,thereisariskthatsuchBPtypemotivates aselfishandmerelyopportunisticengagement.

Managerialimplications

The findingsof ourresearchprovide firmswhich investin Facebook-pagebrandcommunitieswithsomeguidelinesto fosteruserbrand-pageengagementandsosucceedintheir SNS-basedrelationshipmarketingstrategy.

Thefirstimplicationisthatinformationandimageposts arenotasefficientasinteractionpostsintermsoftheir abil-itytoencourage activeparticipationin aFacebook brand page.Yetthisshould notleada communitymanagertoa feelingof frustration, asit is possible tomake a positive commentontherelationalefficacyofinformationandimage posts.If page-user interactionwithsuch posts helps gen-erate experiential value, the firm can feel well pleased. In our view, generating utilitarian and/or hedonic values might prove to be a relational result in itself, regardless of whether it is subsequently manifested in engagement behaviouronthepage.Althoughwehavenotexploredthe matterinthepresent analysis,wefeelthatthe experien-tial values to emerge from users’ elaboration process of

BPsandfromtheir interactionwiththe brandpagemight translate directly to more positive attitudes towards the brand(Chenetal., 2015)andindirectlytovarious brand-supportivebehaviourssuchasadvocacy andloyalty.Inany case,thefirmmightnotneedtoseekimmediaterewardfor havingsetupaFBPorvirtualbrandcommunity.Itshouldat least,however,considerthat(1)usersaccessthepageand actuallyreadtheposts,indicatingthattheyareinfactfans ofthebrand,and(2)usershaverelationalexperiencesand obtainexperientialvalues,whichwillstrengthentheir posi-tionasfansandhelptoenhancethebrand’ssocialimage. Thisdoesnotmeanthatthefirmshouldceaseinitsattempt tocreateaveritablecommunityfeelingsothatactive par-ticipationismorespontaneousandinformationflowsmore fluently.

(15)

Please cite this article in press as: Gutiérrez-Cillán, J., et al., How brand post content contributes to user’s engagementinthebrandpagemight,firstandforemost,be

opportunisticbehaviour spurredby thedesiretobe given a more or less certain and immediate reward. Neverthe-less, even if engagement is self-interested, what is true is thatactiveuser participationbreatheslife intothe vir-tualbrand community.The complementaryinterpretation isthatinteractionpostsactasamechanismtopreventthe reluctancetoparticipateandself-disclosethatsomepeople experience.Inthisway,thehurdlespreventingnaturaland spontaneousparticipationcanbeovercomebyusersbeing specificallyinvited(throughinteractionposts)toengagein someconcreteactivitybythefirm(e.g.,voting,commenting oruploadingphotos).

A third implication is related tocustomer relationship management through FBP. A firm should beaware of the segmentsofcustomerswhofollowitonitsFacebookbrand pageandthetypeofpostspreferredbyeachsegmentsoas tothenbeabletooffercontentadaptedtothedifferent cat-egoriesoffollowers.Evenifourfindingsarecontextualized inthecaseofaspecificfashionbrand,wehavepointedto someattitudinalandbehaviouraldifferencesbetweenpage usersdependingontheirbrandpurchaseintensity.Although all types of page users builttheir utilitarian andhedonic experienceswiththe FBP throughinformationposts,they differintermsoftheexperientialvalueextractedfromboth imageandinteractionposts.Thus,imagepostsmightprove moreappropriateforlower-intensitybuyerswhoare look-ingforquick-fireinformationaboutthebrand.Contrastingly, interactionpostsprovetobemoreexperientiallyeffective when targetingusers displaying acloser relationship with thebrand(higher-intensitybuyers).

Table8offersasummaryofthemostrelevantresearch contributionsforacademicsandpractitioners.

Limitationsandfutureresearch

To conclude thework, itsmost salient limitations aswell asthe possibilitiesit opens upfor futureresearch should bementioned.Firstly,thestudyhasbeencarriedoutusing data gatheredfrom asingle FBP representing thefashion industry, whichreduces the need tocontrol theeffectof certainbrand-orpage-specificvariables,butlimits gener-alizability.Validationofourresultsandconclusionsrequires themodelbetestedonotherFBPswithinandoutsidethe sector in question (Pöyry et al., 2013). Secondly, experi-ential values originate from the overall set of relational interactions involving users in the brand page, not only with BPs. Future inquiry should, therefore, posit a more inclusivemodelinwhichbrandpostsanduserpostscan com-pete freely and contemporaneously against each other to testtheirrespectiveabilitiestogenerateexperientialvalue andmotivateengagement.Thirdly,thereisaneedto rec-ognize that much of the variability in engagement is not accountedforbybrandpostsandexperientialvalues, lead-ingus toconsiderthatvariables whicharenot relatedto users’ interactionwith the FBP might lie at the heart of their decision tomake an active engagement.The model might befurtheredby includingsubjects’general person-ality traits which we have used to explain and interpret theresults. Certainpersonaltraits suchassociability,the senseofsocialinvolvement,attitudetowardscollaboration

withothers, aversiontoopportunism,extraversion, show-ingoff,self-esteemorperceivedcreativitymightwellhelp tobetterexplainengagement.Likewise,itmightbeworth including user variables that are more directly linked to the brand page (e.g., motivations to use the FBP and a senseof belongingtothecommunity),tothebrand(e.g., brandlove), the product (e.g.,knowledge and expertise, orattitudetofashion)orthesetting(e.g.,motivationsand attitudestowardstheInternetorsocialnetworks).Following onfromdeValcketal.(2009),Kabadayi andPrice(2014), vanDoornetal.(2010),WaskoandFaraj(2005)or Wiertz anddeRuyter(2007),individualuser traitscandirectly or indirectlyaffectthelikelihoodandlevelofuserengagement or moderatethe relationsposited in themodel. Fourthly, amodelwhichismoreunderstandingoftherealityshould considerother expressionsofuserengagementaside from brandpage(e.g.,brandadvocacyorbrandloyalty)as indi-catorsofthesuccess ofonline relationalstrategy. Finally, furthertoourinvestigation,evenifthismeanschangingthe approachadoptedhere,itwouldbedesirabletoexaminein detail,bypositingspecifichypothesesandapplyingtheright methodology(e.g.,experiments), howdiversefeatures of abrandpost(brand-orproduct-relatedversusnon-related content, informativeversus emotional orientation, formal versus informal language, serious versus amusing format, etc.)impactonuserresponsetotheBP(quantifiedasthe numberoflikes,sharesandcommentsonthisBP).

Acknowledgements

Theauthorsaregratefulfor thefinancialsupportreceived fromtheSpanishMinistryofEconomyandCompetitiveness (researchprojectECO2012-36275).

References

Algesheimer,R.,Dholakia,U.M., Herrmann, A.,2005.The social influence of brand community: evidence from European car clubs.J.Market.69(3),19---34.

Bagozzi,R.P.,Dholakia,U.M.,2002.Intentionalsocialactionin vir-tualcommunities.J.Interact.Market.16(2),2---21.

Brakus,J.J.,Schmitt,B.H.,Zarantonello,L.,2009.Brand experi-ence:whatisit?Howisitmeasured?Doesitaffectloyalty?J. Market.73(3),52---68.

Brodie,R.J.,Hollebeek,L.D.,Juric,B.,Ilic,A.,2011.Customer engagement:conceptualdomain,fundamentalpropositions,and implicationsforresearch.J.Serv.Res.14(3),252---271. Brodie,R.J.,Ilic,A.,Juric,B.,Hollebeek,L.D.,2013. Consumer

engagementinavirtualbrandcommunity:anexploratory anal-ysis.J.Bus.Res.66(1),105---114.

Casaló,L.V.,Flavián,C.,Guinaliú,M.,2010.Relationshipquality, communitypromotionandbrandloyaltyinvirtualcommunities: evidencefromfreesoftwarecommunities.Int.J.Inf.Manag.30 (4),357---367.

Chan,K.W.,Li,S.Y.,2010. Understanding consumer-to-consumer interactionsinvirtualcommunities:thesalienceofreciprocity. J.Bus.Res.63(9---10),1033---1040.

Chen,K.J.,Kim,J.,Lin,J.S.,2015.Theeffectsofaffectiveand cog-nitive elaborationsfrom Facebookpostsonconsumerattitude formation.J.Consum.Behav.14,208---218.

Figure

Figure 1 Research Model.

Figure 1

Research Model. p.6
Table 1 shows the indicators used and the descriptive statistics (mean and standard deviation) of each indicator, as well as the standardized factorial loadings in the  measure-ment model to emerge from partial least squares analysis (PLS)

Table 1

shows the indicators used and the descriptive statistics (mean and standard deviation) of each indicator, as well as the standardized factorial loadings in the measure-ment model to emerge from partial least squares analysis (PLS) p.7
Table 5 Indirect effects of brand posts on engagement. Indirect effects Estimation P-values

Table 5

Indirect effects of brand posts on engagement. Indirect effects Estimation P-values p.10
Table 7 Indirect effects of brand posts on engagement for each group.

Table 7

Indirect effects of brand posts on engagement for each group. p.11
Related subjects : BRAND ENGAGEMENT