ModuleOverview
November8,2007
NotesforInstructors
TheseslidesareavailablefordownloadasafilenamedPMN
Slides.pptathttp://www.people.hbs.edu/teisenmann/
Iveprovidedcommentsonmanyoftheslides,visibleasNotes
(fromtheViewmenu,selectNotesPage).
Feelfreetoadapttheslidesinanywayyouseefit.Inthespiritofopen
source,ifyoumakechangesoradditions,pleaseforwardacopytome
atteisenmann@hbs.edusoIcanpostthemforothers.Imalsohappy
toanswerquestionsabouttheslidesbyemail.
Theslidespresentconceptscoveredinanintroductorymoduleon
platformmediatednetworks,designedforuseinexistingcourseson
strategy,technologymanagement,marketing,entrepreneurship,or
ecommerce.ThemoduleisdescribedinPlatformMediated
Networks:ModuleNoteforInstructors,HBS#807067.
Forinstructorswhowishtodelvedeeper,themodulesmaterialsare
drawnfromacasebasedMBAelectivecourse,describedin
ManagingNetworkedBusinesses:CourseOverviewforEducators
HBS#807104.
Topics
1.
2.
3.
4.
Overview,Definitions,andCoreConcepts
NetworkMobilization
PlatformControl
PlatformEnvelopment
1995
NetscapeIPO
RobGlaserlaunchesRealNetworks
VocalTecintroducesVoIP
Firstcablemodems
FCCauctionsdigitalmobilespectrum
PalmlaunchesPilot
1999
IBMembracesLinux
XMLiscreated,pavingthewayforweb
services
802.11b(WiFi)takesoff
Napsterlaunched
TiVOlaunched
6
2004:ExplosiveGrowthin
Googlesvaluation
Skype
RSS,podcasting
Blogs
Opensourcesoftware
Wikipedia
BitTorrent
7
PlatformMediatedNetworks
Networkusersaccessacommonplatformthat
facilitatestheirinteractions
Platforms=subsetofcomponentsandrules
employedbyusersinmostoftheirtransactions
Components=hardware,software,services
Rules=technicalstandards,protocolsforinformation
exchange,policies,andcontractsthatgoverntransactions
Usersrelyonaplatformwhendoingsoismore
efficientthanunmediatedbilateraldealings
Network
User
User A
A
User B
Platform
Components
- Hardware
- Software
- Services
Rules
- Standards
- Protocols
- Policies
- Contracts
Architecture
eBaysPlatform
Components
BrowserandInternetaccess(from3rdparties)
Websitedesign
Bidtrackingsoftware
Shippingservices(from3rdparties)
LinkstoPayPal(aneBayownedplatform)
Etc.
Rules
Registrationrequirements
Disputeresolutionprocesses
Feedbacksystem
Policiesforbidding
Etc.
10
NetworkEffects
Definition:Networksvaluetoauserdependsonthe
numberofothernetworkusers
Value=willingnesstopayfornetworkparticipation=
WTPforplatformaffiliation=caponplatformfees
Properties
Canbenegative,e.g.,duetocongestion
WTPtendstoincreaseasSshaped(logistic)functionof
networkgrowth
Networkeffectsaredemandsideeconomiesofscale,i.e.,
theyimpactrevenues
Networkeffectsarenotsupplysidescaleeconomies,whichimproveunitmargins
throughfixedcostleverage
However,manynetworkedbusinessesdoenjoystrongsupplysideeconomies
11
Platformmediatednetworksdonot
necessarilyencompass:
Physicalnetworks(e.g.,gasutilities)
Socialnetworks(e.g.,yourgolfbuddies)
Supplynetworks,akaecosystems,
comprisedoflargefirmsandtheirmost
committedsuppliersandcustomers
(e.g.,WalMart,Toyota)
12
WhyStudy
PlatformMediatedNetworks?
Largeandgrowingshareofglobal
economy
Distinctivemanagementchallenges
13
LargeandGrowingShareof
GlobalEconomy
Notjustdigitalindustries,also:
Financialservices,e.g.,ATMs,creditcards,stockexchanges
Transportation,e.g.,packagedelivery,airlines,travelagents,
reservationsystems,fuelcellpoweredcars
Retail,e.g.,shoppingcenters,barcodes/RFID
Energy,e.g.,grid+appliances,energytrading
Realestate,e.g.,homebuying
Healthcare,e.g.,HMOs
Enterpriseadministration,e.g.,recruiting,B2Bprocurement
60oftheworlds100largestcompaniesearnmostof
theirrevenuefromplatformmediatednetworks
14
DistinctiveChallenges
Conceptsarenew:firstacademicpapersontwo
sidednetworkspublishedin2003
Platformmediatednetworksareverycomplex,but
oversimplificationiscommon
Overorunderestimatestrengthofnetworkeffects
Pricetonetworkssidesasiftheywereseparatemarkets
Errorscanbefatal
Yahoo,AmazonfailedinU.S.auctions
eBayexitedJapanandChina;failedtodislodgePayPal
NASDAQcollusion=$1billionfine;ECNentry
IBMspent$1+billiononOS/2
AppleforfeitedMicrosoftsposition
15
DistinctiveChallenges
BusinessModelDesign
WinnerTakeAllDynamics
16
BusinessModelDesign
ThisisliketheCambrianexplosion550millionyears
ago,whenmulticelledlifefirstappearedonthescene.It
wasthegreatestspeciationeverseen,butitwasalso
whichpeopleforgetthegreatestrateofextinctionever
seen.Weregoingtoseeallkindsofideastried,andthe
majorityofthemareprobablygoingtofail.
JeffBezos,CEOAmazon.com
(BusinessWeek,9/13/99)
17
18
DesigningBusinessModels
isDifficultDueTo:
Twosidednessofmostnetworks
Bifurcationofplatformroles
Rapidgrowth,precludingtrialanderror
19
OneandTwoSidedNetworks
Onesidednetworks:Transactionpartners
alternateroles,e.g.,emailerssend&receive,
tradersbuy&sell
Twosidednetworks:Usersarepermanent
membersofonedistinctgroupaside
whichtransactswithasecondgroup,e.g.,
Jobseekers+recruiters
Cardholders+merchants
20
Atwosidednetworkhas
fournetworkeffects
Side 1
Side 2
Platform
Asamesideeffectfor
eachside,i.e.,
preferenceregarding
numberofotherusers
onownside
Acrosssideeffectin
eachdirection,i.e.,
preferenceregarding
numberofuserson
otherside
Eacheffectcanbe
positiveornegative
21
PlatformRoles:SponsorsandProviders
Providersareusersprimarypointofcontact
withplatform
Sponsorsdonotdealdirectlywithusers;rather,
sponsorsholdpropertyrightsthatdetermine:
Whomaychangeplatformtechnology
Whomayparticipateinnetworkasaplatform
providerornetworkuser
22
VISA =
JOINTLY SPONSORED
SHARED PLATFORM
CARD
HOLDERS
MERCHANTS
ISSUING
BANKS
ACQUIRING
BANKS
VISA INTERNATIONAL
XBOX =
SOLE SPONSORED
PROPRIETARY PLATFORM
GAMER
DEVELOPER
CONSOLE
SDK
MICROSOFT
XBOX
23
DistinctiveChallenges
BusinessModelDesign
WinnerTakeAllDynamics
24
SchumpeterianCompetition:SerialWTABattles
VCRDVD+PVR+VOD
LPCDdownloadand/orsubscription
HumanmarketmakersECNs
1G2G3G
AnalogPOTSVoIP
BroadcastcableIPTV
AtariNintendoPlaystationXbox?
WiFiWiMax?
GasolinepoweredcarsFuelcellpoweredcars?
BarcodesRFID?
25
BluRayvs.HDDVD
26
WinnerTakeAll
Networkeffectsincreasingreturns
winnertakeall,i.e.,oneplatformprevails
Examples:Windows,eBay,PDF,DVD,fax,
realestateMLS
Winnertakeallimplieslosertakesnothing!
So,shouldyoushareplatformwithrivals?
Ifnot,shouldyouracetoacquirenetwork
users?
27
WTAImplications
Sharingandracingarebetthecompany
decisions,soorganizationaldesigniscrucial
Growthopportunitiescapitalmarket
bubbles
WTAmonopolypowergovernment
intervention
28
PlatformStructure
Winnertakeall:oneplatformservesthe
maturenetworkedmarket
Monohoming:mostusersonagivenside
affiliatewithasingleplatform
Multihoming:mostusersonagivenside
affiliatewithmultipleplatforms
29
PlatformStructureExamples
Network
Fax
DVDs
Mobile FeliCa
Real estate MLS
Windows desktop O/S
Online auctions
PDF
Instant messaging
Credit cards
Yellow Pages
Recruitment sites
Video games
Shopping mall
Subscription music
Multichannel TV (DBS
vs. cable)
Side 1
Users
Sender/receivers
Consumers
Cell phone users
Home buyers
PC users
Buyers
Readers
Sender/receivers
Consumers
Consumers
Job seekers
Consumers
Consumers
Consumers
Consumers
Side 1
Structure
Side 2
Structure
WTA
WTA
WTA
WTA
WTA
WTA
WTA
Mixed-mode
Mixed-mode
Mixed-mode
Mixed-mode
Mixed-mode
Mixed-mode
Mixed-mode
MonoMixed-mode
MonoMultiMonoMultiMonoMulti-
Side 2
Users
(Homogeneous network)
Movie studios
Retail stores, etc.
Home sellers
Application providers
Sellers
Document creators
(Homogeneous networks)
Merchants
Marketers
Recruiters
Game developers
Stores
Artists/labels
Programming networks
30
NetworkedMarketisMoreLikelytobe
ServedbyaSinglePlatformWhen:
Theplatformisanaturalmonopoly
OR
MultihomingcostsarehighAND
NetworkeffectsarepositiveandstrongAND
Demandfordifferentiatedfeaturesisweak
ORdominantplatformcanoffersuchfeatures
selectivelytouserswillingtopaypremium
31
Naturalmonopoly:minimumefficient
scaleofproduction>maturemarketsize
gure 2: Unit Cost Curves Showing Constant vs. Increasing Returns To Scale
Constant Returns
Long-Run ReturnRReturns
Unit Cost RReturns
Quantity
Increasing Returns
Long-Run
Unit Cost
Quantity
32
Naturalmonopoliesarerare,butevidentin
somenetworkedbusinesses
Past:localutilities(phone,power,
cableTV),railways,postaldelivery
Present
Internetcontentdeliverynetworks(e.g.,
Akamai,whichhas85%marketsharebasedon
hugeinvestmentinservers)
FailedLEOsatellitecommunications(e.g.,
Iridium,Teledesic)
33
StrengthofNetworkEffects
Networkeffectsarestrongerwhen:
Usersdemandnoveltyfromrepeatedtransactions(e.g.,DVDs)
Mobileusersrequiregeographiccoverage(e.g.,ATMs,WiFi,
refueling)
Participantsinamatchingnetworkhaveidiosyncraticneedsand
offers(e.g.,homebuying,executiverecruiting)
Accesstonetworkusersisnotvaluedequally
Someareextremelyvaluable(e.g.,friends,family)
Othersareworthlittle/nothing(e.g.,strangersinanothercountry)
Somehavenegativevalue(e.g.,telemarketers,stalkers)
Strongpreferenceforvarietyyieldslongtail
34
MeasuringNetworkEffects:ConjointAnalysis
Interviewselicitreactionstoproductattributes
includingnetworksizeusuallythroughpaired
comparisons
Hence,CONsideredJOINTly
Requiresconsumerstounderstandattributes,so
maynotbereliableforbreakthroughproducts
Studiesshouldbedesignedbymarketresearch
professionalsandcostbetween$50k$250k
35
Homingcosts:costs/investmentsincurredby
userduetoplatformaffiliation
Upfront
Searchandnegotiation
Accountsetup,e.g.,softwareconfiguration
Initialhardware&softwareinvestment;systemintegration
Training
Ongoing
Membershipandtransactionfees
Maintenancecosts;customerservicehassles
Tenureorvolumebasedbenefits
Exit
Accountterminationhasslesandcosts,e.g.,changing
emailaddress,movingfundsbetweenbrokerageaccounts
Contractseverancepenalties
Salvagevalueofhardware,software
36
SwitchingCosts
Outofpocketexpensesandinconveniencesincurredby
networkusers(orbyplatformprovidersontheirbehalf)
whenusersswitchfromoneplatformtoanother
Networkeffectsmaydeterorencourageswitching,
dependingontherelativesizesofrivalplatforms.
However,networkeffectsandswitchingcostsare
conceptuallydistinctandshouldnotbeconfused
37
Homingvs.SwitchingCosts
Mono-homing
Switching
1 SETUP + 1ONGOING
2 SETUPS + 2 ONGOING
Multi-homing
38
Homingvs.SwitchingCosts
Usuallytheymovetogether,e.g.,satelliteradio(bothhigh),
packagedelivery(bothlow)
Sometimes,switchingcostishighbutmultihomingcostis
low
eBookreaders:costlytoswitchbecauseyoumustreplaceentire
library,butnotverycostlytomultihome
Emailaccount:costlytoswitchbecauseyoumustnotifyallyour
contacts,butnotverycostlytomultihome
39
Homingvs.SwitchingCosts
Whenpredictingwhetheranewnetworkedmarket
willbeservedbyasingleplatform,focusonmulti
homingcosts
Whenpredictingwhethertoracetoacquirenetwork
users,focusonswitchingcosts
Whenpredictingwhetheranestablishedplatformis
vulnerabletodisplacementbyanewplatform,focus
onbothmultihomingcostsandswitchingcosts
40
UsersPreferencesforDifferentiated
PlatformFunctionality
Usersegmentshavedifferentneeds
DBSpictureissharperthancables
Appealingforsportsandmovielovers?
DBSrequiresasettopboxforeveryTV
Aliabilityforlargefamilies?
Canfunctionalitybeofferedselectivelyto
userswillingtopayapremium?
TomatchDBSpicturequality,cablewould
havetoconverttoalldigitalarchitecture,so
selectivityisnotanoption
41
HomingLogic:AffiliateWithPlatform
WhenVP+IH
VP = expected value from ability to interact
with a potential transaction partner
I = value from platform, independent of
platform size
H = homing cost
42
MultiHomingLogic
ForplatformsAandB,calculateVP + I H
Considerincrementalvaluefromaccesstopotentialtransaction
partnersfrommultihomingonAandB,comparedtomono
homingwiththebestsingleplatform,A
BmayofferaccesstosomepartnersnotaffiliatedwithA
Duetodifferentiation,Bmayoffermorevalueforcertaintransactions
thanA
Compareincrementalvaluetoincrementalhomingcost
43
WTAPotential?
DVD:
WTA
CreditCards:
MultiHoming
Strengthof
NetworkEffect
Highformostuserson
Highformostuserson
bothsides,i.e.,consumers bothsides,i.e.,card
andstudios
holdersandmerchants
MultiHoming
Cost
Highforbothsides
Lowforbothsides
Demandfor
Inimitable
Features
Lowduetotechnical
standardizationofTV
High:revolvevs.
charge(i.e.,payinfull
withnopresetlimit)
44
MonoHoming+Multihoming
Example:OnlineSubscriptionMusic
Side1MonoHoming:
Consumers
Strengthof
NetworkEffect
High:wantaccessto
allmusic
MultiHoming
Cost
High:monthlyfees,
playlistmanagement
Demandfor
Inimitable
Features
Side2MultiHoming:
MusicCompanies
High:revenueincreases
indirectproportionwith
userbase
Low:duplicatedlegal
work,butzeroinventory
andmodestincremental
productioncostsdueto
digitaldistribution
Moderate:differences Low:musiccompanies
includequalityof
havesimilarneedsfor
editorialcontent,IM
DRM,promotional
integration
support,etc.
45
Topics
1.
2.
3.
4.
Overview,Definitions,andCoreConcepts
NetworkMobilization
PlatformControl
PlatformEnvelopment
46
BarrierstoNetworkUserAdoption
Businessstealingrisk:toomanyrivalsonyourside
Unlikeotherrisks,doesnotinvolveprovider/usertrust
Holduprisk:dominantplatformextractstoomuchvalue(e.g.,
MicrosoftOSforcablesettopboxormobilephones)
Strandingrisk:1)userbackswrongsideinWTAbattle;2)
platformfailstoinvest;3)nobackwardcompatibility
Integrationrisk:platformproviderintegratesintonetwork
userrole(e.g.,MicrosoftwithOffice&Halo)
Favoritismrisk:toattractusers,extractrent,orserveself
interest,platformproviderskewstermsinonesidesfavor(e.g.,
Covisint)
Relationshiprisk:In2sidednet,whocontrolsaccesstousers
onotherside?(e.g.,Xboxvs.EA;Autobytelloans)
47
WillingnessToPayandNetworkSize
WTP
D2
Biggernetworkismoreuseful
Existingcustomerswillconsumemore
Newprospectswilljoinnetwork
D3
P2
P3
P1
N2
Outwardshiftsbecomesmallerasexpected
networksizenearssaturationpoint,Ns,dueto
Lateradoptersnotasvaluableasearlyadopters
Congestion
D1
N1
D1istraditionaldemandcurveif
customers/prospectsexpectanetworksize
equaltoN1
Ifcustomers/prospectsexpectedabigger
network,N2,thendemandcurvewouldshift
outwardtoD2
N3
Ns
Network Size
48
BewareMetcalfesLaw
AccordingtoMetcalfesLaw,thevalueofanetworkgrows
withthesquareofgrowthinitsuserbase
InameshnetworkwitheachofNnodesconnectedtoeveryothernode,
thereareNx(N1)links,i.e.,~N2
But,Value=N2onlyif:
PercapitatransactionvolumesalsoincreasebyN2
EachusersWTPpertransactionremainsconstantashertransaction
volumegrows
Intherealworld:diminishingreturnssetinandconstraintson
attention/budgetprevail
49
UpwardSlopingDemand
inNetworkIndustries
WTP
D2
D3
Fulfilled Expectations
Demand Curve
P2
P3
P1
Fulfilledexpectationscurve
tracespricethatwouldexactly
fulfilldemand,foragiven
expectednetworksize
Curveisupwardslopingin
domainwherenetworkeffects
dominatepriceelasticity
effects
D1
N1
N2
N3
Network Size
50
MultipleEquilibriainNetworkIndustries
Assumecompetitivemarketwith
Price=MarginalCost
Markethasthreeequilibriumpoints:
N0,N1,andN2
N1isnotastableequilibrium
WTP
Fulfilled Expectations
Demand Curve
Marginal
Cost
N0
N1
N2
Network Size
Ifexpectednetworksize<N1,
firmscannotmakemoney,soN0
prevails
Ifexpectednetworksizeis>N1,
supplywillsurgetoN2
Implication:networkmarketsare
tippytheycanstall,andtheycan
growexplosivelyand
expectationsmatter!
51
ThePenguinProblem
52
ThePenguinProblem
(Farrell&Saloner,1986)
Innewmarkets,uncertaintyishigh;individualshave
differentexpectationsaboutlongtermdemand
Ifuserbaseisfragmented,userscannotcommunicate
expectationsorcoordinatebehavior.Networkmay
stall,evenwithstrongnetworkeffects!
Noonemovesunlesseveryonemoves,sonoone
moves,likepenguinsafraidtobefirsttodiveforfood
53
Twowaystocategorizeplatforms:
Structure:Proprietary,Shared,Joint
Venture,Licensor
Function:Connectivity,Variety,Matching,
PriceSetting
54
CategorizationBasedonPlatformFunction
Connectivity:platformfacilitatespointtopointexchangeofinformation,
physicalgoods,orpassengers(e.g.,fax;Fedex)
Morepointsconnectedmorevaluablenetworkmorepointsetc.
Variety:platformelicitssupplyofavarietyofcomplementsfromone
networksideandpromotescomplementconsumptionbytheotherside
(e.g.,console+games;creditcard+merchantlocations)
Morecomplementsmoredemandsideusersmorecomplementsetc.
Matchingnetwork:platformmatchestransactionpartnerswith
heterogeneousneedsandoffers(e.g.,realestate;onlinedating)
Morebuyersmoresellersimprovedoddsofsuitabletransactionmore
buyersetc.
Pricesettingnetwork:platformfacilitatesexchangeofwelldefined
itemsvalueddifferentlybytradingpartners(e.g.,stockexchange,
gamblingbookmaker)
Morebuyersmoresellersreducedpricevolatilitymorebuyers
etc.
55
CategorizationBasedonPlatformFunction
One-Sided
Multi-Sided
Partners Identity
Generally Known in Advance
Connectivity
Price-Setting
Variety
Matching
56
KeyChallengesinNetworkMobilization
Connectivity
Ensure reliability
Optimize reach (infrastructure investment
to extend network vs. incremental revenue)
Price Setting
Balance users different needs for:
Anonymity
Transparency with respect to most recent
price and best current offer
Transaction speed
Flexibility to trade with preferred partners
Variety
Matching
57
InternalizingExternalities
Externality=benefitorharmexperiencedbyBduetoAs
actions,withnocompensatingpayment
Positiveexternality:Iadmireyourflowers
Negativeexternality:yourpigpenisupwind
Networkeffectsareexternalities
Threeapproachesforinternalizingexternalities
#1:Sidepaymentsbetweenusers,e.g.,Ibuygrandmasvideophone
#2:Intwosidednetwork,permanentsubsidizationofoneside
#3:Intertemporalinternalization:sponsorprimesthepumpwith
discountstoearlyadopters;thisleadsotherstoreviseexpectations/WTP
For#2and#3,proprietaryplatformistypicallyrequired;
otherwise,freeridersexploitsubsidy
Exception:sponsorscandictatepricingtermsthatunderwritesubsidies,as
withcreditcardsandrealestate
58
StrategiesforNetworkMobilization
Pricing
Permanent:shouldonesidebesubsidized?
Penetration:shouldplatformprovidersraceto
acquirenetworkusers?
Participation
Exclusivity:shouldplatformprovidersseek
exclusiverelationshipswithsomeusers?
Integration:shouldplatformprovidersbecome
networkuserstoavoidchicken&eggproblem?
59
60
PricingPuzzles
WhatexplainsLadiesNight?
Whydosimilarindustrieschargedifferent
sides?
Videogames:endusersubsidized;developer
pays
PCs:enduserpays;developersubsidized
61
AssumemonopolistservesmarketsAandC
Marginalcost=0
WithnonetworkeffectbetweenAandC,monopolistpriceseachproductto
maximizePxQ
PA
PC
P*A
Market C
P*C
Q*A
QA
Q*C
QC
62
P^A
Side B
P^B
P*B
P*A
Q*A Q^A
QMAXA
Q*B
Q^B
QMAXB
63
StrategiesforNetworkMobilization
PermanentSubsidies
AcceleratedGrowthStrategies
Exclusivity
VerticalIntegration
64
65
66
RacingtoAcquireUsers
Firmsracetoacquirecustomerswhentheyface:
Increasingreturnstoscale
Highswitchingcosts
Latemoveradvantagese.g.,reverseengineering,
freeridingonpioneersmissionarymarketingmay
discouragepreemptiveinvestment
Willinvestorsfundracingstrategies?
67
RacingtoExploitIncreasingReturns
Withoutincreasingreturns:
Maximize:Profit=RevenueCost
Implies:MarginalRevenue(MR)MarginalCost(MC)=0
MR=MC
Withincreasingreturnsduetonetworkeffects:
Willingnesstopay(WTP)=f(Numberofotherusers)
So,addingauserboostseveryotherusersWTP
MR+NPVofextrarevenuefromotherusers=MC
Tomaximizeprofit,youshouldpricelower(reducingMR)
orinvestmoreinmarketing(increasingMC)
Inotherwords,youshouldracetoacquirecustomers!
68
Highswitchingcostsalsoencouragefirms
toracetoacquirecustomers
Inanotherwisecompetitivemarket,afirmshouldearnprofit
equaltothesumofswitchingcostsconfrontingitscustomers
$100pricejustcoverscostsofproducing/deliveringoneunit
Assumeconsumersbuyoneunitofproductinperiods1and2
Ifswitchingcostsare$50,thenfirmAcanpriceat$149.99toits
existingcustomersinperiod2.
TostealAscustomers,FirmBmustfullycompensatecustomersfor
switchingcosts
However,$150isthebestpriceBcanofferandstillrecoveritscost
Theseeconomicswillfuelaracetoacquireunaffiliated
customersinperiod1:eachfirmwillpriceat$50.01
69
ThePayofffrom
RacingtoAcquireCustomers
I*
Asinvestmentrises,NPVofa
newcustomereventually
declinesdueto:
Broadeningbeyondnatural
segments
Prematurelysolicitingprospects
Retaliatoryresponses
Scalabilityconstraints
I*correspondstoLifetime
ValueofCustomer=0forthe
lastcustomeryouacquired
AverageLVCshouldbepositive
70
LifetimeValueofaCustomer
Rule:acquirecustomersuntilLifetimeValueofaCustomer(LVC)=0for
lastcustomeracquired,i.e.,PVofvariablecontributionmargin=customer
acquisitioncost(includingincrementalcapitalexpenditures)
LVCdoesnotallocatefixedcosts.IfLCVispositive,butyoudonthaveenough
customerstocoverthepresentvalueoffixedcosts,youllneverearnaprofit!
LVCprovidesaguidelineformarketingspending;itisnotasubstitutefor
detailedcashflowanalysis
IssueswithLVCinnetworkedindustries
Issue#1:LVCshouldincludenewcustomersimpactonothercustomers
willingnesstopay
Issue#2:HardertocalculateLVCintwosidednetworks(e.g.,Monster.com)
Issue#3:Withsomeconnectivitynetworks(e.g.,FederalExpress),variable
costsdependonuserdensity,soyoumustprojectcustomerbasetocalculate
LVC
71
Whatmightexplaindeparturesfromlong
termvaluemaximizinginvestmentlevels?
Investment<I*
I*
Resourceallocationprocesses
Fundingconstraints
Antitrustconstraints
Investment>I*
Inadvertent:Overconfidence
Deliberate:Exploitovervaluation
72
Managingthrougha
speculativevaluationcycle
Can/shouldmanagersexploitovervaluation?
Currencyforacquisitions(Google)
Buffer(Akamai)
Howwilldifferenttypesofinvestorsreacttobust?
Privatevs.publicequity(e.g.,Fedexvs.UPS)
VCvs.strategicinvestor
Family(e.g.,Cox)
73
StrategiesforNetworkMobilization
PermanentSubsidies
AcceleratedGrowthStrategies
Exclusivity
VerticalIntegration
74
Topics
1.
2.
3.
4.
DefinitionsandCoreConcepts
NetworkMobilization
PlatformControl
PlatformEnvelopment
75
PlatformRoles:SponsorsandProviders
Providersareusersprimarypointofcontactwith
platform
Sponsorsdonotdealdirectlywithusers;rather,
sponsorsholdpropertyrightsthatdetermine:
Whomaychangeplatformtechnology
Whomayparticipateinnetworkasaplatform
providerornetworkuser
Eachrolemaybefilledbyonefirmormany;
sometimes,asinglecompanyfillsbothroles
76
PlatformTypes
One
Sponsor
Many
Sponsors
OneProvider
ManyProviders
Proprietary
Licensor
Macintosh
Monster.com
Playstation
JointVenture
CareerBuilder
Orbitz
Covisint
Windows
AmericanExpress
ScientificAtlanta
Shared
Linux
Visa
RealEstateMLS
77
KeyDesignDecision
Preserveproprietarycontrolofnewplatform,e.g.,eBay,
FederalExpress,Google?
Or
Shareplatformwithrivalswhooffercompatiblebut
differentiatedplatformproductsandservices,e.g.,
barcodes,WiFi,DVD?
78
OutcomesVary
Proprietaryplatformsmayprevail
Akamaivs.ContentBridge
eBayvs.FairMarket
Sharedplatformsmayprevail
Citibankvs.CirrusATMnetworks
Proprietaryandsharedplatformsmaycoexist
MacintoshandLinux
AmericanExpress(until2004)andVisa
79
DecisionRules
Aspiringplatformproviderswithabigedge(e.g.,via
strongpatents)shouldalmostalwayspursueproprietary
approach
Whenmanyevenlymatchedfirmscansimultaneously
launchanewplatform:
Winnertakeallpropensityencouragessharedapproach
Freeriderproblemsencourageproprietaryapproach
80
IfaProprietaryPlatform
ServestheEntireMarket
Userswillworryaboutholdupbymonopolist
Prospectiveproviderswillworryaboutlosing
100%oftheirinvestmentiftheyarenotthe
winner
81
FreeRiderProblemswith
BigUpfrontInvestments
UserSubsidization
CentralizedInfrastructure
82
WTAPotential?
Yes
Free
Rider
Issue?
Yes
No
ProprietaryFavored
ProprietaryFavored
Proprietaryexamples:
PayPal,YellowPagesin
smallercities
Sharedexamples:WWW,
realestateMLS
SharedFavored
No
Sharedexamples:DVD,
fax,barcodes,WiFi,SMS
Proprietaryexamples:video
games,paidsearch
Coexistenceexamples:
NYSE/ECNs;creditcards
CoexistenceCommon
Coexistenceexamples:
Symbian+Blackberry;Linux
+Mac
83
CompetingonaSharedPlatform
84
CompetingonaSharedPlatform
Pulledtogetherbypositiveimpactonplatformadoptionand
eachfirmssurvivalodds
Fearofpenguineffect
FearofWTAoutcome(e.g.,mobileFeliCa)
FearofBalkanization(e.g.,prestandard11n)
Fearofoutsidethreats(e.g.,mobilephoneOEMsvs.Microsoft)
Pulledapartbyreductioninappropriabilityofplatformrent
Concernabouttoomanyrivals
Desiretoincorporateowntechnologyinstandard(802.11)
Desiretodifferentiateastandardizedproduct(e.g.,J2EE;802.11)
FrictionbetweenIPownersandfreeriders(e.g.,RANDZ)
Propensitytoformcoalitionsthatexcludecloserivals(e.g.,KDDI+
JCBvs.DoCoMo+Sumitomo)
85
ValueCaptureStrategiesfor
SharedPlatformProviders
RestrictMembership
Canbackfireifexcludedpartiessponsorcompetingplatform,e.g.,Sunvs.
IBM/Microsoftinwebservices
ProfitfromIP
Iftoolow,IPrichfirmsmayavoidplatform
Iftoohigh,IPpoormayavoidplatformandstalematesmayoccuras
firmsvietocontributetechnology
ProfitfromImplementation
FreelycontributeIPtosecuretimetomarketadvantageordevelop
proprietaryextensionstoindustrystandards
But,proprietaryextensionsmaysplinteraplatform,e.g.,Unix
86
KeySkills
Timing:Toosoon=ignored;toolate=blocked
IPManagement:Avoidsubmarinepatentsandpatentthicketsvia:
Standardsettingorganization(SSO)disclosurerules
Patentpools
Patenthoardingtoensuremutualassureddestruction
Diplomacy:Tacticsincludeappeasement,intelligencegathering,
logrolling,bluffing,burningbridges,irrationalcommitmenttodogma,etc.
SSOdelegatesrequireraremixoftechnicalprowessanddiplomacyskills;their
loyaltiesareoftendividedbetweenprofessionandfirm
OrganizationalDesign:Adhocvs.establishedSSOs;forumshopping;
SIGs;etc.
87
Topics
1.
2.
3.
4.
DefinitionsandCoreConcepts
NetworkMobilization
PlatformControl
PlatformEnvelopment
88
PlatformEnvelopment
PlatformEnvelopment:Entrybyoneplatformprovider
intoanothersmarket,combiningitsownfunctionalitywith
thetargetsinamultiplatformbundlethatleverages
commoncomponentsand/orshareduserrelationships
Envelopmentisapowerfulforceshapingtheevolutionof
platformmediatednetworks
Pathtoplatformleadershipchangethatdoesnotrequire
Schumpeteriancreativedestruction
Driverofindustryconvergence
Strategyforcrosslayercompetition
89
PlatformEntrenchment
Singleplatformoftendominatesduetodemandandsupplyside
scaleeconomies,e.g.,Windows,eBay,PDF
Dominantplatformsoftendifficulttodisplaceduetonetwork
effectsandswitchingcostsTosucceed,entrantsgenerallymust
offerrevolutionaryproducts,e.g.,ECNs,fuelcellcars,IPTV,
VoIP,WiMax
Hence,platformleadershipchangeoftenentailsserialwinner
takeallbattles,withsuperiornewplatformsdisplacingoldones
Platformenvelopmentisasecondpathtoplatformleadership
changethatdoesnotrequireSchumpeteriancreative
destruction
90
Platformprovidersservingseparatemarketsmay
shareusersandcomponents
Users
Platform
A
Platform
B
Users
Platform
A
Platform
B
91
PlatformEnvelopment
Platform
A
Platform
B
BentersAsmarket
withamultiplatform
bundle
Aisvulnerableifit
cannotmatchan
appealingbundle
92
EnvelopmentinSubscriptionMusic
AT&T
Apple
Yahoo
Rhapsody
Labels
Consumers
Priorto1999,RealPlayerdominates
streamingmedia
MicrosoftsWMPenvelopsRealPlayer,
bundlingstreamingserverintoNTatno
extracost
Reallaunchessubscriptionmusicservice
Rhapsody
Yahoodiscountssubscriptionmusic,
crosssellingviaportal
Applecouldbundlesubscriptionservice
intoiTunes,leveragingiPod
AT&TWirelessbundlesYahoo
subscriptionmusicwithcellphoneservice
Real
Microsoft
93
ConvergenceEnvelopment
94
Xbox360vs.PS3
95
Cablevs.Telco
96
CrossLayerCompetitionEnvelopment
Applications
Applications
Middleware
Operating Systems
Operating Systems
PCs
PCs
Components
Components
97
TelephoneandTelegraph
Onenewspaperaccount
describedAT&T,inits
slowbutsteady
envelopmentofthe
telegraphindustry,asa
quietoctopus.
Hochfelder,
BusinessHistoryReview,
2002
98
Mechanism:Bundling
EfficiencyGains:Transactionandproductioncostsavings;improved
userexperience
PriceDiscrimination:Reducesheterogeneityinconsumersaggregate
valuations
StrategicAdvantages:Extendmarketpowerintocomplementmarket;
protectcoremarketbyforeclosingrivalsaccesstokeycomplement
99
Foreclosure
Conglomeration
(Cross Layer)
(Convergence)
Intermodal
Relationship
Complements
Examples
Efficiency
Gains
Economies in
marketing
Avoid double
marginalization
Production cost
synergies
Some economies in
marketing
Price
Discrimination
Benefits
Limited
Moderate to strong
Limited
Strategic
Advantages
Extend market
Limited
power, defend core
Neutralize emerging
threat
100
DefendingAgainstEnvelopment
Matchbundle,butbundlevs.bundlecompetitioncanbefierce
Findabigbrotherviamerger(e.g.,ScientificAtlanta+Cisco;
Lotus+IBM)oralliance(e.g.,RealNetworks+Comcast,Cingular)
Cedeandredeploy,e.g.,RealNetworksshiftsfromstreamingto
subscriptionmedia
Sue,e.g.,Novell,RealNetworks,Netscape,Sunvs.Microsoft
101