Servitization through Human-Data Interaction: A Behavioural Approach

Abstract

This paper proposes a new approach to servitization and business models by understanding behavioural aspects of human interactions with technology, specifically, with “smart” devices, connected devices, autonomous systems, and internet of things (IoT) through understanding and interacting with data which these devices and systems generate. Proposed approach, Behavioural Human Data Interaction Hypothesis (Behavioural HDI Hypothesis), which differs from existing literature, leverages on research in behavioural science, data-driven business models, multi-sided markets, and Human-Data Interaction (HDI). Behavioural HDI Hypothesis can offer a new approach to future markets for data because it helps to (a) predict consumer choice of product and services; (b) suggest new and improved interaction mechanisms between consumers and their self-generated data; and (c) propose a new approach for building and evaluating business models.
To date, very little has been known about whether and how consumers and households accumulate, review and use self-generated data about consumption decisions and how this affects market relationships between consumers and providers of goods and services. This paper shows how Behavioural HDI Hypothesis can make markets for data more efficient through better personalisation and servitization. It also has implications for data collection visibility, data ownership and structure, platform trade-off, security and other ICT-related challenges which negatively affect current business models in the digital economy.

Full text

Pogrebna

ProceedingsoftheSpringServitizationConference(SSC2015)

0

SERVITIZATIONTHROUGHHUMAN‐DATAINTERACTION:ABEHAVIOURALAPPROACH

GannaPogrebna

April2015

ABSTRACT

Purpose: This paper proposes a new approach to servitization and business models by
understandingbehaviouralaspectsofhumaninteractionswithte chnology,specifically,with“smart”
devices, connected devices, autonomous systems, and internet of things (IoT) through
understandingandinteractingwithdatawhichthesedevicesandsystemsgenerate.
Design/methodology/approach: Proposed approach, Behavioural Human Data Interactio n
Hypothesis (Behavioural HDI Hypothesis), which differs from existing literature, leverages on
researchinbehaviouralscience,data‐drivenbusinessmodels,multi‐sidedmarkets,andHuman‐Data
Interaction(HDI).
Findings:BehaviouralHDIHypothesiscanofferanewapproachtofuturemarketsfordatabecauseit
helps to (a) predict consumer choice of product and services; (b) suggest new and improved
interaction mechanisms between consumers and their self‐generated data; and (c) propose a new
approachforbuildingandevaluatingbusinessmodels.
Originality/value: To date, very little has been known about whether and how consumers and
householdsaccumulate,reviewanduseself‐generateddataaboutconsumptiondecisionsandhow
thisaffectsmarketrelationshipsbetweenconsumersandprovidersofgoodsandservices.Thispaper
shows how Behavioural HDI Hypothesis can make markets for
data more efficient through better
personalisationandservitization.Italsohasimplicationsfordatacollectionvisibility,dataownership
and structure, platform trade‐off, security and other ICT‐related challenges which negatively affect
currentbusinessmodelsinthedigitaleconomy.

Keywords:servitization,dataasaservice,Human‐DataInteraction(HDI),
newbusinessmodels


Pogrebna

ProceedingsoftheSpringServitizationConference(SSC2015)

1

1.INTRODUCTION
The development of informationandcommunication  technology (ICT)inthemodern economy has
created opportunities for businesses to provide customised products, services and experiences to
theircustomers.Thiscustomisationbecamepossibleduetolargevolumesof(personal)data
which
customersgenerateonaday‐to‐daybasisandwhichbusinessescollect,storeandanalyse.Formany
businesses, the future relies on their ability to process  the data in order to accurately predict
consumerpreferencesandcreatepersonalisedproducts, servicesandexperiencesinthemostcost‐
effectiveway.

Yet,
atthemoment,data‐drivenbusinessmodelsthroughpersonalisationarestillintheirinfancyas
evencompanieswithaccess tolargeamountsofdata struggletocreatereliableforecastsoffuture
customer wants and needs to quickly react to changes in market trends. One of the most notable
examples of forecasting
 inefficiency are so‐called recommendation systems (available via  major
retailers)whicharesupposedtomakesuggestionsaboutwhatacustomermightliketopurchasein
the future, but which are in fact rarely used. Furthermore, we also do not see a development of
effectivemarkets for data where consumers of goods and services (henceforth,users)would trade
theirself‐generateddatawithproducersofgoodsandservices(henceforth,providers)whichinhibits
aneffectiveuseofdataasaservice.

This paper first considers reasons for the current data market inefficie ncies and then develops a
model of market for data where users and providers interact to develop new business models
utilisingdifferenttypesofdataaswellasdifferentwaysinwhich thisdataisperceivedbytheusers.
The proposed model – Behavioural Human‐Data Interaction Hypothesis – is based on Data‐Driven
Business Models approach which explains how business models can bedeveloped using data (e.g.,
Hartmannetal. 2014);an openmulti‐sidedmarkets approachwhichoffersanaccount ofhownew
marketswithmultipleplayerscan
becreatedinthedigitaleconomy(Ng2014);aswell asresearchin
Human‐DataInteractions(HDI)researchwhichexplainshowusersinteractwithdata(Mortieretal.
2014).ThisnewBehaviouralHDIHypothesisisalso rootedinbehavioural scienceliteratureandhas
significantimplicationsfornewbusinessmodelsin
thedigitaleconomyaswellasofferingimportant
solutions for the currently existing ICT‐related s ervitization problems such as data collection
visibility,dataownershipandstructure,platformtrade‐offs,andsecurity.

2.MARKETSFORDATA:PRESENTANDFUTURE

2.1CurrentMarketforData:ValueandWorth
Let us first consider the current market for data. In this market,  users supply data and providers
demanddataasdescribedonFigure1below.Forthepurposesofthispaperwewillconcentrateon
userself‐generateddatawhichmayincludepersonaldata(datareflectingbehaviourofanindividual
user) or social data (the data for the whole household, etc.). Providers demand the data and are
willing to pay  the demand price PD for the data (this is how much the data is worth to providers).
This price is relatively high as it allows providers to offer better (more personalised) goods and
services to users and increase providers’ profitability via better  understandi ng user demand for
goodsandservicesaswellasviaincreasinguservalue.Wedefineprovidersbroadly–thiscouldbe
companieswhichtradedata,dataanalysts,appdevelopersandprovidersofgoods/services.

Usersarewillingtoofferdataatasupply pricePSwhichisperceivedbythemasverylow.OnFigure
1wechooseapricelevel closeto0inordertodescribethelevelofPS(thisishow muchthedatais
worthtousers).Inpra ctice,thispriceisnotexpressedinmonetaryterms,i.e.,usersdonotdirectly

Pogrebna

ProceedingsoftheSpringServitizationConference(SSC2015)

2

receiveanymoneyfromtheproviders.Instead,itreflectsthe“cost” ofdatatousersintermsof,e.g.,
lossofprivacy,etc.

Abstractingfromdifferenttypesofdataaswellasfromdifferentwaysinwhichthedataisperceived
by users and providers, the level of PD and PS (shown using the vertical axis) remains stable
irrespective of the qualityof the data as a service(shown using the horizontal axis). The dataasa
service variable depicts how effectively available data can be converted into meaningful business
models (provision mechanisms). In other words, it reflects the value of the data for providers and
usersonthemarket.



Figure1CurrentModelofMarketforData

We assume that the value of data is  the same for providers and users for the following reason. If
providers receive valuable data about user behaviour, they will be able to provide better (more
personalised) goods and services to the users. Therefore, data of higher quality which produces
betterpredictions of behaviourand lead toan increasein userwellbeingand providerprofitability
should be valued higher by both sides of the marke t (users and providers). In practice, there is, of
course, a lotof uncertainty as to the value of the data (see, e.g. Ng etal. 2015). Yet, this question
requiresaseparateinvestigationandforthepurposesofthispaperwedono tconsideruncertainty
aboutthedata.

Figure 1 shows that the current market is inefficient: since the disparity between  the supply and
demandprice fordataisverylarge,thedataisnottraded. Inprinciple,providersarewillingtopay
PD to obtain the data, but users are offering the data at a  very low price PS which means that
providerscaneither(a)obtainthedatathemselves ataverylow(orevenzero)priceinwhichcase
they receive a profit margin of PD –PS > 0 (e.g., Google, Facebook, etc.); or (b) purchase  the data
from other providers (intermediaries) at PD in which case intermediaries receive a profit margin
PD–PS>0.Notethattheobtained/purchaseddatacanbeofloworhighqualityascapturedbythe
dataasaservicevariableandthedemand/supplypricedoesnotdepend on it.

2.2FutureMarketsforDataIgnoringBehaviouralHDI
In recent years, various issues were raised with regard to supply price for data. Specifically, the
developmentofnewtechnologies(e.g.,EcklandMacWilliams2009)resultinginconcernsaboutdata

Pogrebna

ProceedingsoftheSpringServitization Conference(SSC2015)

3

ownership(e.g.,Evans2011),datapriv acy(Itanietal.2009),aswellastheinequalitybetweenusers
and providers in terms of profit distribution from data usage. Under these circumstances, user
perceptions of data markets have changed giving rise to scepticism about the potential of trading
data with providers. This sceptical view which ignores the fact that people interact with different
typesofdatainadifferentwayisdepictedonFigure2.


Figure2FutureModelofMarketforDatawithoutBehaviouralHDI

Accordingtothisview,providersinthefuturewillstillbewillingtopurchasedataatademandprice
P
D
.Atthesametime,thesupplypriceP
S
foruserswillrangefromverylowforlessvaluabledatato
high for more valuable data. Therefore, users will only trade the data with  providers at an
equilibriumpriceP
E
attheintersectionofsupplyanddemandpricefunctionsonFigure2.Effectively,
this means that in order to trade, users would need to provide data of high quality, exert a
significantamountofefforttoaccumulatethedata,andengagewithproviders.Thiscreatesserious
objections to direct user‐provider  markets for data since the potential logistical costs of users
engagingwithprovidersisveryhighandveryfew userswouldbeabletoengagewithtradingdata.
However, applying such a model of market relationships would not be correct because it  does not
capturethecomplexhuman‐datainteractionswithinthedigitaleconomy.

3.BEHAVIOURALHDIHYPOTHESISANDITSIMPACTONBUSINESSMODELS

3.1BehaviouralHDIHypothesis
Themarketstructurespresentedinsections2.1and2.2donottakeintoaccountthatdifferenttypes
ofdatawhichmaybeperceivedbyusersdifferently.Yet,byapplyingBehaviouralHDIHypothesiswe
can show how different types of data (with different value to users and providers) can be
successfully traded on the market for data. Behavioural HDI Hypothesis distinguishes  between
traditionaldata,invasivedata,andinventivedata(seeFigure3).







Loweffort
•Traditional
Data
•aggregatedas
ContentData
Mediumeffort
•Invasive
Data
•aggregatedas
ContentData
Higheffort
•Inventive
Data
•aggregatedas
Metadata

Pogrebna

ProceedingsoftheSpringServitizationConference(SSC2015)

4



Figure3DataTypesaccordingtoBehaviouralHDIHypothesis

Thedata typespresented onFigure3  differbytheamount ofeffortwhich a userneedstoexertin
ordertoengagewitheachtypefromloweffort(traditionaldata)tohigheffort(inve ntivedata).Due
to the fact thatusers need to exerta different am ount ofeffortto engage witheachtype of data,
theywillperceivethe3typesofdatadifferently.

Traditional data involves minimum/low user effort because it is accumulated by technology which
existsinthehouseholdsofthemajorityofusers.Thedatageneratedbythistechnologyisreviewed
byusersonaregularbasis andalluserscaneasily assessthisdata(e.g.,datafromelectricitymeters,
watermetersandother“traditional”devices).

Invasive data involves medium user effort because it is accumulated by technology which is
accessibleand yetnon‐“standard”.For example,datafrom mobileapplications(apps),smart home
sensors, etc. requires for the user to install the apps or devises and learn how to read and
understand self‐generated data obtained through this techno logy.  This type of data is called
“invasive” because this data often influences human behaviour (e.g., fitness apps may make an
individualexercisemore).

Inventive data involves maximum/high user effort because it requires for the user to add relevant
content to existing data accumulated through Internet‐of‐Things (IoT). Particularly, inventive data
mayrequirefortheusertoaddcontexttothedatacollectedthroughotherdevices.Inotherwords,
inventive data does not only tell an individual that electricity was used but also stores important
information about who used it, when and which device was turned on. This type of  data is called
“inventive”becauseitrequirestheusertoinnovateorco‐createtogetherwithprovidersinorderto
receivethebest‐qualityinformativedata.

While traditionaland invasive data is used, aggregated and analysed by providers as Content Data
(datawhichprovidesinformationaboutactioneventsbutgivesnocontextabouttheseeventssuch
as, e.g., Big Data or Connected Data), inventive data is accumulated as Metadata (data which
providesinformationabouteventsinconjunctionwiththeircontexts).

3.2PerceivedMarketforDatawithBehaviouralHDIHypothesis
Since different types of data under Behavioural HDI Hypothesis are not perceived by users in the
sameway,wecanmodifyFigure2tointroducedifferenttypesofdataandshowhowfuturemarkets
fordatamaybeaffectedbythesedifferentperceptions.

Previous research (e.g., Parry et al. 2015; Ng et al. 2015) shows that context‐dependent data
provides important benefits for customisation, personalisation, and creating new business models.
Therefore, it is likely that the quality of data as a service will increase from
 traditional to invasive
dataandthenfrominvasivetoinventivedata.Userswoulddemandahigherandhigherprice P
Sas
they go from traditional to invasive and from invasive to inventive data because, according to
Behavioural HDI, they have to exert more and more effort to obtain the data. At the same time,
since under Behavioural HDI, users will not perceive traditional, invasive and inventive data in the
same
way,rationalproviderswillanticipatethischangeinuserpreferencesfordatawhichwillresult
in changes to demand function for data. Specifically, the demand function for data will follow a
pattern,atfirstincreasingandthenstationary.Traditionaldatawillbecomelessvaluedbyproviders

Pogrebna

ProceedingsoftheSpringServitizationConference(SSC2015)

5

andthedemandpricewillbeflatontheregioncoveringtraditionaldata.However,forinvasiveand,
especially, inventive data the demand price will be increasing intersecting with P
S on an interval
covering a large portion of inventive data and forming an interval of equili brium  prices P
E. Such
shape of P
D function even allows for a small portion of invasive data to be traded if this data is of
relativelyhighquality(seeFigure4).






Figure4FutureModelofMarketforDatawithBehaviouralHDI

Overall,underBehaviouralHDI,differentuserperceptionsoftraditional,invasive,andinventivedata
willresultinlargeportionsofdatabeingtradedonthedatamarketwhichwillbebeneficialforboth
usersandproviders.Afterreaching itsmaximum,
PDwillbeflatduetothefactthatprovidershave
budgetconstraintsandbeyond a certain point even extremely valuable inventive datawill become
toocostlyforproviders.

BehaviouralHDIprovidesasystemofmarketrelationshipsthroughwhichproviderscanbetterfulfil
users’ wants and needs  by better understanding their
preferences and offering better (more
personalised) services. It also suggests new and improved interaction mechanisms between users
and providers as they have an opportunity to dire ctly trade data on the market. It also may offer
new approaches for building and evaluating business models. Specifically, business  models can be
evaluated based on the user effort level necessary to engage with providers, the actual price at
whichthedataistraded(topofbottomofthePEinterval),etc.

4.IMPLICATIONSOFBEHAVIOURALHDI
The proposed approach has several important implications not only for new business models but
alsoforresearchandpracticeofdatacollectionvisibility,dataownershipstructure,platf o rmtrade‐
offsandsecurityofdata.

CurrentICTsystemsoften collectdatainwayswhicharesubtletousers:many peopledonotrealise
thattheirsupermarketorcoffeeshopclubcards,smartphonesorsocialmediawebpagesconstantly

Pogrebna

ProceedingsoftheSpringServitizationConference(SSC2015)

6

collectandaccumulatetheirpersonaldata.Eventhoughprovidersseemtobelievethatusersprefer
subtle data collection to visible(judging, for example, from the caution around the deployment of
GoogleGlass),itis notclearwhetherusersactuallypreferdeviceswhichcollect theirpersonaldata
in a subtle way to those which do it in a visible way. It is also not clear whether users are more
concerned about the visibility of data collection or about the possibility that a device maybe
collectinginformationwhichisunknown totheuser.BehaviouralHDIallowsustostudytheseissues
systematicallybyelicitinguserpreferen ces overdifferenttypesofdata.

Since the supply ofdata is dependenton the technology,  theownership of the data oftenremains
with the technology owner. For example,  Internet search data trends are owned by large
corporations(e.g.,Google)orsupermarketdataownedbylargesupermarkets(e.g.,Tesco) anditis
often difficult or even impossible for individual users to obtain their self‐generated data.
Furthermore,thedata collectionmechanism,structure,representation, storageand,  therefore,  the
potentialapplicabilityofthedataisdependentonthetechnology,i.e.,thenatureofhowthedatais
collectedaffectshow it couldbe used. Sincesuchdata often has a vertical structure, it  isprimarily
beneficialtocompanies andnottoindividualusers.However,itisnotclearwhetheruserswouldbe
interested in having access to theirowndata (shouldthey be able toview their data in adifferent
waythroughnovelvisualisationmechanisms)orprefertooutsourcedatamanagementandanalysis
toathirdpartywhichwouldthenpresentitinameaningfulwayandcommunicateittoeachuseras
a set summary statistics or recommendations. Understanding these individual preferences is very
importantand Behavioural HDI can provide novel data ownership solutions through increas ed user
participationindatamarkets.

AllprovidershaveplatformsfortheirIoTdevicessuchas“smart”sensorswithinthehome,apps,and
wearabledevices.Increasingly,platformsemergewhichofferreportingservicesacrossmanyofthe
same provider’s products. This causes “vendor lock‐ins”. Consider an individual who owns  a
technology
producedbyacertainprovider(providerA).Whenauserisnextpresentedwithachoice
betweentwonewtechnologies,ofwhichoneismadebyproviderAandanotherbyanewprovider
(providerB),the“convenient”decisionfortheuseristooptfortechnologyfromproviderAbecause

it allows this user to stay with the current platform instead of using two different platforms or
switching to a new platform. As a result, users may not always choose the best or cheapest
technology or device weighing their decision more on their existing products and on how an
additional
technology benefits  the overall platform than how it performs  on its own. Behavioural
HDIallowsuserstodifferentiatebetweendatatypesandproviderpropositionsonthemarketwhich
cangiveusersmoreinformationabouthowtomakemosteffectivedecisions.

Privacy, confidentiality,and trust issues of data, especially invasive and inventive data, can impact
individual be haviour. While Behavioural HDI does notaim to influence the area ofprivacy directly, 
data protection mechanisms are significantly more manageable if the data is partitioned into
different types. Inventive data is collected and shared by the users under their own control and,
therefore,privateinformation  isunlikelytobesharedagainuser’ s will(e.g.,Ng2014).Atthesame
time,traditionalandinvasivedata,especiallywhencombinedthroughlinkingandre‐matchingdata
from different sources, may pose challenges for privacy. Behavioural HDI may offer a systematic
approach to policy regulation oftraditional and invasive data by identifying data types and market
relationshipswithhighriskofprivacyinfringement.

Behavioural HDI is useful for business practice. The understanding of the types of data as well as
differentwaysinwhichthesedataareperceivedbyconsumerscanallowbusinessesto(a)decrease
uncertainty about the value of the consumer‐generated data; (b) simplify consumer‐business

Pogrebna

ProceedingsoftheSpringServitizationConference(SSC2015)

7

interactions; and (c) motivate consumers to collect and supply high‐quality data to businesses. By
incorporating Behavioural HDI into their business models, companies can create systems which
wouldallowthemtoquicklyaggregateandusedatatoaccuratelyanticipateconsumerdemandand
produce customised products and services. Behavioural HDI can change recommendation systems
(availableviamajorretailers) to co ‐creationsystems where insteadofmakingrecommendations to
consumers,companiescancollectdataonfeaturesofproductswhichconsumersmayneedorwant
inthefutureandcatertoconsumerneedsmakingfulluse
ofdataasaservice.

REFERENCES
Eckl, Roland, and Asa MacWilliams. "Smart home challenges and approaches to solve  them: A
practicalindustrialperspective."InIntelligentInteractiveAssistanceandMobileMultimedia
Computing, pp. 119‐130. Springer Berlin Heidelberg, 2009. Evans, B. J. (2011). Much ado
aboutdataownership.Harv.JL&Tech.,25,69.
Hartmann,PhilippMax,MohamedZaki,NielsFeldmann,andAndyNeely."BigDataforBigBusiness?
ATaxonomyof Data‐driven Business Models used byStart‐upFirms."ATaxonomyof Data‐
DrivenBusinessModelsUsedbyStart‐UpFirms(March27,2014)(2014).
Itani,Wassim,AymanKayssi,andAli Chehab. "Privacy as a service: Privacy‐aware data storage and
processing in cloud computing architectures." InDependable, Autonomic and Secure
Computing, 2009. DASC'09. Eighth IEEE International Conference on, pp. 711‐716. IEEE
(2009).
Mortier,Richard,HamedHaddadi,TristanHenderson,DerekMcAuley,andJonCrowcroft."Human‐
Data Interaction: The Human Face of the Data‐Driv en Society."Available at SSRN 2508051
(2014).
Ng,IreneCL.CreatingNewMarketsintheDigitalEconomy.CambridgeUniversityPress(2014).
Ng,IreneCL,KimberleyA.Scharf,GannaProgrebna,andRogerMaull."Contextualvariety,internet‐
of‐things and the choice of tailoring over platform: mass customisation strategy in supply
chainmanagement."(2014).
Parry,Glenn,SaaraA.Brax,RogerMaull,andIreneCLNg."Visibilityofconsumercontextimproving
reversesupplywithInternet‐ofThingsdata."(2015).

ACKNOWLEDGMENTS
Ganna Pogrebna acknowledges  financial support via the EPSRC grant “Smart Me versus Smart
Things:TheDevelopmentofaPersonalResourcePlanning(PRP)SystemthroughHumanInteractions
withDataEnabledbytheIoT”.

AUTHORCONTACTDETAILS
Dr.GannaPogrebna
WarwickManufacturingGroup,
UniversityofWarwick,g.pogrebna@warwick.ac.uk
