2.2 Priing Internet aess in the presene of user loyalty
2.2.7 Disussion on servie bundling
0 500 1000 1500 2000 2500 3000 3500 4000 4500
0 20 40 60 80 100 120
Total profits
Round
ISP1 ISP2
(a)
0 500 1000 1500 2000 2500 3000 3500
0 20 40 60 80 100 120
Total profits
Round
ISP1 ISP2
(b)
0 20 40 60 80 100 120 140
0 20 40 60 80 100 120
Total profits
Round
ISP1 ISP2
()
Figure 2.8: Total individual prots for 2 ISPs (
L = 0.05, 0.30, 1.0
), 1 inumbent, 1 entrant,deterministi loyalty
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
0 20 40 60 80 100 120
Total profits
Round Threshold=0.1
Threshold=0.2 Threshold=0.3 Threshold=0.4
Figure2.9: Overall umulative prots of 2ISPs at dierent loyaltylevels
beingloyal to them,sine stronger loyalty results inhigher overallprots. Thisresultis inline
withreent empirialsurveys onuser loyalty [13℄.
harder toswithspeiserviesovertime. Thisusuallyinludesa mandatoryloyaltyperiodin
ontrat, inreased swithing ostsand vanishing disounts (if a single servieis swithed to a
dierent provider).
Being suh an important and omplex topi,there is abroad literatureon bundling. Bakos
and Brynjolfsson [47℄ introdue theonept of the preditive valueof bundling,that itis easier
forasellertopredithowaonsumerwillvalueaolletionofgoodsthanitistovalueanygood
individually. Inthisaseaseller anextratmore revenue fromabundle, andsheismorelikely
to innovate. Moreover, at theoptimal prie, moreonsumers will nd thebundle worthbuying
thanwouldhaveboughtthesamegoodssoldseparately. Authorsonludethatformostphysial
goods, the potential impat of large-sale aggregation is limited, however, these eets an be
deisive for the suess or failure of information goods. In [48℄ it is shown through a Hotelling
model thata pair ofrms an prot fromoering abundled disount to thedetriment ofrms
who do not bundle and onsumers whose preferenes are farther removed from the bundled
brands. Authors of [49℄study the impatof multiprodutnonlinear priingon prot, onsumer
surplus and welfare ina duopoly. Their main result is thatbundling an raiseprot but harm
onsumer surplus ompared to linear priing if there is substantial heterogeneity in onsumer
demand. Gilbrideetal. disusses[50℄theframingeets(howtheprieofthebundleispresented
to ustomers) inmixed prie bundling. Theynd that the integratedframe results inthemost
ustomershoosing the bundle. Considering tripleplay, authors of[51℄ derive thattheoutome
oftripleplayompetitionislikelytodependonthespeedofthedevelopmentofnewtehnologies
and theadaptation of the regulatory environment. In the short run, telephone ompanies will
enjoyanadvantageattributabletoswithingosts. However,thisadvantagewillerodeasyounger
subsribers swith to telephony on the Internet. In [52℄, authors propose a model of dening
themost protable bundles andtheir assoiated priesinthe teleommuniation market. They
onstrut andanalyze atwo-level game usingtheframework of Bayesian gametheory.
As evidened by the broadness of related work and a ertain lak of onsensus among
re-searhers, bundlingisalivelyeldforresearh. That said,a omprehensive investigationonthe
interplay of user loyalty and bundling is a timely, relevant and most ambitious (even foused
only on the ommuniation servies market) topi. While we onsider this investigation as a
very important future work for us, it annot t into the boundaries of this work. Nonetheless,
here we give a glimpse on thepossible extensions for both our game-theoretial and simulation
frameworkthatmaybringus losertothethoroughunderstandingofbundling anduserloyalty.
Regarding our analysis, a simple form of bundling an be inorporated to our framework.
Bundled servies an be pereived as a single super-servie with a prie
p B
. In this ase theprie ompetitiongamebetween ISPsmodelstheuserbuying substitutableserviebundles. Let
us onsidera triple play senario, whenonsumers an buythree dierent servies: broadband
eah userisharaterized bya vetor
v
,whih onsistsoftherespetiveustomer's valuationofindividual servies,hene
v = (v I , v T , v P )
, where∀ j, 0 ≤ v(j) ≤ 1
. For analytial tratability, we make the assumption thatevery userhasthe same valuationvetor.Furthermore, we believe thata dual-reservation prie senarioarises naturally here, aswell.
We dene the reservation priesfor singleservies (
α 1 = (α I 1 , α T 1 , α P 1 )
for swithers andα 2 = (α I 2 , α T 2 , α P 2 )
for loyals). Now, we an determine thereservation priesfor bundles asfollows:α B 1 = v × α 1 α B 2 = v × α 2
where
α B 1
(α B 2
)isthereservationprieforswithers (loyals). Furtheron,theanalysisinSetion 2.2.4 also holdsfor this extendedmodel.Theabove model doesnot apture some important aspets ofthe interplay ofbundling and
loyalty. First,byassumingonstant servievaluations arossusers,we losetheheterogeneityof
userdemandfordierentservies. Seond,temporalevolutionofthemarketannotbeanalyzed
asa one-shotgame. Third,network providers oerboth dierent bundles andsingle serviesat
thesametime. Addingalltheseextrafeaturesmakesthemodelveryomplex,henewepropose
simulation-based evaluation.
Whiledeveloping suha toolremains animportant future work forus, we arguethat itan
be done by extending our existing simulator presented in Setion 2.2.6. Heterogeneous servie
valuations an be introdued among users, and integrated into the user and deision model.
Simulationsarerepetitive,sotemporalharateristisof marketevolution anbeobserved.
Ad-ministrative ostsof swithing serviesisinherently present intheprie-dierenebasedmodel;
swithingostsproportionaltothenumberofswithedserviesanbeeasilyadded. Last,
opera-torsoeringsingleservies,anddierentbundles anbemodeled;inthisaseusers wouldmake
an integrated deisionarossallownedanddesiredservies. To sumitup,thesimulator should
alsomodelthevaluationsandserviesubsriptionsofindividualusers. Withthehelpofsuhan
improved tool, wean analyzehowustomersofheterogeneous loyaltyintentions would onvert
topurhasingserviepakagesindierentsenarios,andadditionally,howtheironversionaet
their loyaltylevel.
Another possible diretion for future work regarding bundling is onstruting multi-level
games (extended games), where users solve an optimization problem at a lower level (deiding
upon their beststrategy for purhasing the servies/bundles they need), then servieproviders
determine theiroered bundles,settheir priesandompeteagainsteahotheratthetoplevel.
A gameof this type an beanalyzed throughthe onept ofsub-game perfet equilibrium[9℄.
Note, that suh extended analyti models are still not suient to apture every aspet of
priing,demandandROIC(ReturnonInvestedCapital)forbundledteleomserviesofamajor
a full,operatingmodel issometimes more artthan siene, greatlyrelying on empirial market
studies and experiene. Nevertheless, our work an provide valuable input to suh a full model
withrespetto thesensitivityof onsumerdemand to priingstrategies.