Application Software for Insurance
in the 1960s and Early 1970s
JoAnne Yates •
MIT Sloan Schoolof Management
As Martin Campbell-Kelly[5] notesin his paperon this panel,dataon the
emergenceof the software industry is sparse,and "In the absenceof hard
quantitativedata, the besthistoricalapproachappearsto be one basedon case
studies." While his paperprovidescasestudiesof specificsoftwarefirms, this
paper providesa different type of case study:a study of the emergenceof
applicationsoftware,especiallypackagedsoftware,from the pointof view of one
userindustry,life insurance.
Fromthefirstadoptionsof UnivacsandIBM 650sbeginningin 1954/55into
the early 1960s, firms who bought or rented computerswere faced with the
formidabletaskof programming
them.Fromthe very beginningof insuranceuse
of computers,programmingtime and costs,consistentlyunderestimated,
were a
majorpartof the expenseof installingcomputersystems.Earlyestimatesranged
from onethird to two thirdsthe costof equipmentrental[ 11]. The representative
of oneinsurancecompany[30, p. 16] explainedthatmanyinsurance
firms turned
to computersin responseto a costsqueeze;yet, "This interestin usingcomputers
to performmanyof theroutinetasksof dataprocessing,
in turn,introduceda new
factor into the price cost squeeze":"the tremendouscostsbeing incurredby
companiesin the area of systemsdesignand programmingrequiredto install
computersto performthesevariousfunctions."He went on to claim that in many
cases,companies"were pointingout that programmingcostshad far exceeded
predictions,and the anticipatedcostsavingswere not forthcoming."
Initial attemptsto reduceprogrammingcosts(for otheruserindustriesas
well asinsurance)
camein theadoptionof programming
languages
easierto useand
morepowerfulthanmachinelanguage,ultimatelyincludingCOBOL for business
programming
[11, p. 235; 2, p. 364]. At thesametime,hardwarevendorswerealso
providingincreasedsupportto all usersin the form of programmedroutinesgiven
• I am gratefulfor theresearch
assistance
of Bob Hancke,HansGodfrey,CindyCollins,
HowardHecht,andPatMurrayin thisresearch.I alsoappreciatethe archivalassistance
of
Hal Keinerat thearchivesof CIGNA, DanielMay andthelibrarystaffof theMetropolitan
Life Archives,of Paul Lasewiczat the archivesof AetnaLife & CasualtyCompany,and
PhyllisSteeleat The New England.Hal KeinerandPhyllisSteelewerealsovery helpfulin
settingup interviewswith employees
for me. My work wasalsoaidedby accessto the
BostonInsuranceLibrary.
BUSINESS AND ECONOMIC HISTORY, Volume twenty-four,no. 1, Fall 1995.
Copyright¸1995 by the BusinessHistoryConference. ISSN 0849-6825.
123
124
to their customerswith the machines.Eventuallysuchcollectionsof software
routinestook on the identity of operatingsystems. Like improvementsin
programming
languages,
operating
systems
greatlyfacilitatedprogramming.
This paperfocuses,however,on softwareapplicationpackagesdesigned
specificallyto accomplishinsurancefunctions. For this story, IBM's 1969
unbundling
wasa crucialinflectionpoint. I will firstsurveytheemergence
anduse
of insuranceapplicationsoftwarebefore IBM's unbundlingin 1969, focusing
primarilyon IBM's bundledinsurancesoftwareandsecondarily
on the emerging
insurancecomputerserviceindustry.Then, I will look at the eventsfollowing
IBM's decisionto unbundle
hardwareandsoftwarein 1969,openingthewayfor
the rapidexpansion
of packagedsoftwareofferedby a varietyof newlyemerged
softwarespecialtyfirms.
Insurance Software before Unbundling
The earliestpre-developed
andtestedapplicationsoftwareavailableto the
insurance
industrywascreatedby thehardwarevendorsthemselves.Initially, the
vendorsprovidedtraining and aid to insurancefirm programmersto support
applicationprogramming.In addition,they suppliedlibrariesof routinesfor all
users[41]. As it becameclearthat costsof programmingwerea majorfactorin
decisions
to rentor buycomputers,
however,vendors
sawtheneedto do morethan
providegenericprogramssuchas reportgeneratingprograms.They beganto
supplyindustry-specific
applicationprograms.
'62 CFO
IBM's '62 CFO (Consolidated
FunctionsOrdinary,1962),designedfor the
1401tapecomputer,
seemsto havebeenthefirst--andwascertainlythebestknown-of suchapplication
programs
designed
for theinsurance
industry.The software(it
wasactuallya setof programs)wasbasedon the Societyof Actuaries'notionof a
Consolidated
Functions
planfbr handlingordinaryinsurance
billing,accounting,
andrelatedfunctions,initiallyconceivedin the 1950sin preparation
for the first
computers(thoughbased on an earlier, tabulatorera notion of integrated
applications)
butonlyrarelyachieved
in thesoftwaredeveloped
by insurance
firms
in thefirstdecadeof computerization
[52, 53]. A fewyearslatertherepresentative
of oneinsurancefirm describedCFO's inceptionasfollows:
IBM, one of the computermanufacturers,
beingcloseto the
problem[of high applicationprogramming
costs]andawareof its
impactontheirmarketsetoutto do somethingaboutit. In addition
to engineering
changes
in computers
to simplifyprogramming,
they
assembleda systemsand programmingteam to develop a
ConsolidatedFunctionsSystemfor OrdinaryLife Insurancewith
threeprimaryobjectives:
1. To developa systemsapproachfor computers
that will allow
automaticprocessingof all scheduledtransactions
and requested
processing
for all non-scheduled
transactions
involvingOrdinary
Life contracts.
125
2. To writea seriesof computerprogramsfor the 1401 whichwill
performall record-keeping
functionson a daily, monthly,or annual
cycle for Ordinary Life Insurancecontractsand miscellaneous
accountingapplicationsthat are directlyrelatedto OrdinaryLife
Insurance contracts.
3. To test and documentthesecomputerprogramsto such an
extentthat theycan be usedbroadlythroughoutthe Life Insurance
Industry as operationalcomputerprogramsor as a guide in the
development
of personalized
totalsystems
on an individualcompany
basis[30, pp. 16-17].
In springof 1962,IBM helda one-weekcourseat theirEndicottfacility
to presenttheunfinished
softwarethencalledCFO '62 to interested
representatives
of thelife insurance
industry.Accordingto an attendee,in developingtheprograms,
"The PAL [ProgramApplicationLibrary] unit of IBM has been guidedby a
consultingactuaryand severallife companies"[34]. '62 CFO, as it cameto be
called,wasnotyetreadyfor customers,
but wasintendedto be field testedstarting
laterthatyear. Indeed,thefirsttworeferences
to CFO '62 in theproceedings
of the
Insurance
Accounting
andStatisticalAssociation(IASA--oneof severalinsurance
trade associationsto devote considerableattention to the use of computers)
appearedin 1963,andin neithercasehadit yet beenfully implemented[49, 50].
By the 1964IASA meetings,
however,'62 CFO wasclearlybeingusedin
variouswaysat severalmedium-sized
companies.Three presenters
discussed
it
extensively.Onepresented
a paperon "TheCFO Packageasa Guide"in whichhe
explainedthathiscompany's
specificequipmentconfiguration
precludedits using
thepackage
directlyaswritten;nevertheless,
thefirm wasfindingit veryusefulas
a guide in its own development
of a similarsystem[14]. The packageprovided
guidancein areasfrom systemdefinitionto documentation
techniques,
andfrom
recordformatdesignto training.In spiteof thislimiteduseof it, he announced
that
"Frankly, my impressionof the packageis rather favorable"[14, p. 13]. The
representativeof anotherfirm describedhis firm's adoptionof the packagewith
only minormodifications[30]. He was alsoenthusiastic
aboutit, but notedthat it
did noteliminateall workinvolvedin convertingoperations
ontoa 1401computer:
In closingI wouldlike to illustratemy appraisalof '62
CFO in this manner....We now have an objective we wish to
reach-- automaticprocessing
of ordinarylife insurance
-- '62
CFO is theSantaFe Trail fbr thisobjective.It is clearlymarked
and documented.There will be somehardshipsalongthe way.
However, if you follow the trail, I'm sure you will reach the
objective[30, p. 18].
In general,thepackageseemedto be well on its way towardswidespread
acceptanceby 1964. In that year the first of threeregionalCFO usergroupswas
established
by insurance
companies
in the westernpart of the country[46]. Papers
givenat IASA in the nextfew yearsillustratedthat acceptanceandgavethe clear
impression
thatbothinsurance
usersandIBM benefitedfrom it. One speaker,for
example,notedthat"If a companywereto adoptCFO in its entirety,about35 manyears of research,planning,and programmingwould be saved";even with his
126
company'smodifications
to CFO, "We realizeda savingsof approximately
22 manyearsin our installation"
[43, p. 13]. The representative
of anotherinsurancefirm
notedthathisfirm "decidedto orderanIBM 1401,specificallyto utilizetheIBM
'package'of programsfor Life Insuranceknown as '62 CFO" [48, p. 225].
Indeed, insuranceindustrysourcesfrom that era and from the present
consider'62 CFO the mostsuccessfulinsurancesoftwarepackageever, with the
estimatednumberof usersrangingfrom 200 to 300 insurancecompanies,manyof
whichhad multiple1400 seriescomputersin their operations[46, 19]. In 1972,
seventy-fivelife insurancecompaniesattendedthe Midwest-AtlanticCFO user
groupmeeting,one of threeregionalusergroups.
The popularityof '62 CFO wasclearlyinfluenced
by thepopularityof the
1401 computer,which,accordingto IBM accounts,"quicklybecameone of the
mostimportantandsuccessful
productsIBM hadeverannounced,"
the "Model-T
Ford of the computerindustry"with over 12,000 (far over the initial estimates)
produced[2, pp.473,676n]. But thehardware'spopularitywasin turn influenced
by industry-specific
softwarepackagessuchas '62 CFO, which allowedmedium
andsmallinsurance
firmsto computerizefor a reasonable
cost. Firmssuchasthe
onecitedaboveadoptedthe 1401becausetheCFO softwaremadedoingsofeasible
evenfor a smallfirm; theconversionprocessfor the firm's life andhealthpolicies
(thelatterrequiringsomemodifications
to theprograms)tookonly threeemployees
andwascompletedin one andone half months[48, p. 225]. While the insurance
industrywasonly oneindustryusingcomputers,
onesoftwarevendorservingthat
industryin 1970assertedthatit was"probablythe largestuserof computersoutside
thebankingindustry"[3, sec.2, p. 55]. Clearly,providingsuchsoftwarefree with
the hardwaremadethe 1401 moreattractiveto smallerinsurancefirms that might
otherwisehavedelayedcomputerizingor goneto other,lessexpensivevendors.
The largestinsurancefirms did not generallyuseCFO, for at leasttwo
reasons.First,firms suchas the Prudential,MetropolitanLife, andJohnHancock
dependedon much larger computersthan the 1401 to handle their enormous
numbersof policies[53]. In addition,thesefirmstendedto preferto programtheir
computersthemselves,
with only limitedhelpfrom vendorsor othersources.As
someoneworkingin the systemsareaat thePrudentialsaid,"We liked to do things
the Prudential way" [28]. Thus the vendorsmight provide supportand utility
programsto suchlargeinsurancefirms, but the targetsof '62 CFO were medium
and small firms.
TIP and ALIS
By the late 1960sIBM's successwith '62 CFO had beenobservedand
imitated by at least one other vendor and by itself with its next generationof
hardware. The three hardwarevendorsthat, accordingto insuranceindustry
sources,put the mosteffort intoattractingcustomers
from the insuranceindustry,
mountingexhibitsat IASA meetings,
LOMA (Life OfficeManagementAssociation)
triennialSystems
Forums,andotherindustryevents,wereIBM, Honeywell,andthe
UNIVAC Division of Sperry Rand Corporation[15, pp. 140, 155]. While
UNIVAC appearsonly to haveprovidedgeneralpurposeprogramsfor functions
suchas input-outputand reportgeneration(its lack of applicationprogramswas
undoubtedly
oneof itscompetitiveproblems),Honeywellwasmorecompetitiveon
the applicationprogrammingdimension.
127
At the end of 1963 when Honeywellannouncedits forthcomingH-200,
whichwasplug-compatible
with theIBM 1401(but much faster),it alsopromised
a programcalledLiberatorthatwouldtranslate1401programsfor the H-200 [38,
p. 162]. Liberatorwasintendedto makeit easyfor IBM 1401 usersto upgradeto
Honeywell'sfasterH-200. Thus this programwould translate'62 CFO for useon
the H-200. In addition,however,Honeywellsoonprovidedits own life insurance
applicationpackagecalledtheTIP (Total InformationProcessing)
System[15, p.
161]. Thispackagewas,according
to onesoftwarevendora few yearslater,"The
CFO system... putintoHoneywelllanguage
withfrontandrearextensions"
[46, p.
141]. He wenton to notethatit wasan exampleof "IBM competitionrecognizing
the importanceof industry-oriented
software."Judgingby thelack of discussion
of thispackagein paperspresented
at theIASA andLOMA conferences,
however,
its popularitydid not approachthat of '62 CFO.
A moreimportantdevelopmenteffort wasIBM's ALIS (AdvancedLife
InformationSystem),designed
to takeadvantage
of theIBM 360's newcapabilities.
While IBM alsoprovidedwaysof adapting'62 CFO to the 360 architectureand
languagethroughemulationandadaptation[26, p. 327], thesewaysdid not take
advantageof the 360 series' speednor its direct accessstorageand inquiry
capabilities.Thusin 1964a groupof insuranceindustryrepresentatives
andIBM
industrymarketers
cametogetherto discuss
a projectfor developing
a consolidated
functionsinsurance
packagefor the 360/30, andin 1965IBM beganworkingon the
project,with the participationof monitorgroupsof insurancerepresentatives
[46,
p. 140]. Scheduledfor releaseat the end of 1966, ALIS wasdelayedrepeatedly,
eventuallyappearingin Januaryof 1969 [36, 4].
ALIS wasdesigned
to process
a singledailyrunto accomplish
all routine
functionsconcerninga policy, as well as to processspecialperiodic runs for
reportingpurposes[26, 16]. Most importantly,as one IBM representative
explainedto membersof IASA, it madeuseof the360's directaccessstorageand
telecommunications
capabilities"to providestatusandseveraltypesof quotations
via a teleprocessing
terminal"[16, p. 123;seealso29, p. 119]. Thiscapabilitywas
perhapsto be the mostinnovativeandattractiveaspectof ALIS to potentialusers,
thoughit tooka longtimeto realize. For example,onecompanyinterested
in ALIS
noted that "the availability of mass storagefacilities with immediateaccess
capabilities"was particularlyattractiveto it [36, p. 319].
Someof the firmspreparingfor ALIS werequiteenthusiastic
aboutits
possibilities
in their presentations
to IASA andLOMA in 1967and 1968 [e.g.39,
40]. The views expressedby the representatives
from two of the five insurance
companyBetatestsites,however,weretempered[35, 36, 31]. Oneof thempointed
out severalareasthat might poseproblemsfor his company. Perhapsthe most
revealingproblemand its originwasdescribedasfollows:
The ALIS rate file will not be carried on disk. This will make it
necessaryto create an anniversaryextract and provide for a
specialrun againstthe rate tape to obtainanniversaryvalues.
The monitorcompaniesdid not agreewith thisphilosophy,but
IBM felt thedisk approachwouldraisethe minimummachine
configurationneededto install and operateALIS. Herein lies
one of the basic problems that must have contributed
substantially
to the unexpected
delaysin the completionof the
128
ALIS project. The companywascommittedto providesoftware
for a minimum configurationand this restriction worked
hardshipson the projectteam [36, p. 320].
While the360 serieswouldincludea widerangeof machines,
IBM wasattempting
to reachthe samesmallto mediumsizedsegmentthatit hadreachedwith '62 CFO.
In fact, at leastone of the largerBeta sites[35; 31] usingthe near-top-of-the-line
360/65, AetnaLife and Casualty,evidentlyultimatelyrejectedALIS in favor of
developingits ownsystem,LIAS (Life InsuranceAdministrationSystem),put into
place in 1972 [17].
This targetingdecisionclearlyhurtthepackage'schancesof succeeding
with firms buyingcomputers
in the middleand upperend of the 360 range.In
addition,thelatereleaseof ALIS, solongafterthefirst360 machines
wereshipped
in 1965 [38, p. 171], createdproblemsfor somepotentialusersof it [e.g., 36],
perhapsaccountingfor the large numberof userswho insteadopted for an
adaptationof •62 CFO to the 360 system,or who developedtheir own systems.
Prudential's
similarAdvancedOrdinarySystem(AOS), for example,wasconceived
before System360 was announced,built on that platformduringthe periodof
ALIS's development,and implementedin the sameyear that ALIS was finally
released[27, 35]. Moreover,ALIS's eventualreleaseoccurredonly monthsbefore
IBM's unbundlingof softwareand hardware,which, as will be shown below,
changed the nature of the software market permanently. One immediate
consequence
wasthataftertheunbundling,
ALIS wasno longerfree. Thesefactors
undoubtedlycontributedto its relatively modestlevel of successcomparedto that
of its predecessor,
'62 CFO. By 1972,whentheregionalmeetingof oneof three
usergroupsfor '62 CFO included75 companies,
the onlyALIS usergroupin the
U.S. includeda modest29 user firms. As a recentretrospectiveaccountof
computers and insurance described it, "While the 360 quickly became
overwhelminglysuccessful,
the insurancesoftwaresystemswrittenfor it by IBM
in 1965--ALIS (Advanced Life Insurance System) and PALIS (Property and
Liability InsuranceSystem)--didnot" [19, p. 28].
Service and Software Companies
In theera beforeIBM's 1969 unbundling,firms had a third alternativeto
developingtheirown softwareor adoptingthatofferedby theirhardwarevendor:
turning to outsidecompaniesas a sourceof data processingservicesand/or
software.Oneof theearliestandmostunusual
servicearrangements
beganin 1955,
whenthe relativelysmallAetna InsuranceCompany(a fire and casualtyinsurance
companynot at that time affiliated with the much larger Aetna Life Insurance
Company)persuadedthreeotherfire and casualtyinsurancecompaniesto buy a
computerto serve all of them [45]. SPAN (named for the four companies:
Springfield,Phoenix,Aetna, and National) startedwith the IBM 705 and later
progressed
to theIBM 7070 and7074, in eachcasedevelopingsoftwarepackages
that would serveall of the companiesowningandbuyingservicesfrom it. Thus
thesepackages
weremorewidelyusedthanthoseof a singlefirm, butnotaswidely
as'62. In 1963,SPAN was ultimatelyboughtby AetnaLife InsuranceCompany,
thoughit continuedto serveits former ownersas well as firms owned by Aetna
Life.
129
Oneof theearliestindependent
computerservicecompanies
targetingthe
life insurance
firmswasEDS (ElectronicData SystemsCorporation).Startedby
dissatisfiedformer IBM salesmanRoss Perot in 1962, EDS sold what came to be
calledfacilitiesmanagement
services:EDS wouldmakea contractto run a firm's
entiredataprocessing
operation[21, pp. 27-30]. Life insurancefirmswereamong
his early targets.and Mercantile SecurityLife in Dallas was one of his earliest
contractsandthe first amongmanyinsurancecompanyclients. A decadelater, a
member of anothersoftwareservicesfirm servingthe life insuranceindustry
referred to this contract as "One of the first (perhapsthe first) servicing
arrangements
in which a computerservicescompanyuseda packagesystemto
processthe businessof a life company"[46, p. 141]. This marketwas lucrative
enoughthat by the mid-1960s,EDS had hired two life insurancemen, who had
installed'62 CFO in theirformerlife insurance
firmsto developsoftwarefor and
sell servicesto life insurancefirms [42; 7, p. D4-1]. The EDS facilitiesmanagementapproachrequiredthat insurancefirmsgive up a greatdeal of control.
Other service companiesalso targetedthis life insuranceniche, though not
necessarily
withsocomprehensive
a bundleof services[46, p. 141]. Thusfacilities
managementand other servicecontractswere anotherroute by which smaller
insurance
firmsor firmsthatdidn't wantto developtheir own expertisein this area
couldgetapplication
softwarewithouthavingto developor eveninstallit. In this
independent
servicecompanyoption,softwarefor life insuranceapplicationswas
bundled with services, not with hardware.
Such firms continued to sell their
servicesafterthe unbundlingdiscussed
in the nextsection[7].
Finally,softwarefirmsin thestill-youngbutgrowingsoftwareindustryof
themid-1960sprovidedpackaged
general-purpose
softwareaswell ascustomized
softwareto clients[8, pp. 322-325],includinga few life insurance
firms. As early
as 1956,ComputerUsageCorporation,whichhasclaimedto be "theworld'sfirst
computersoftwarecompany,"hadprogrammed
its "first nontechnical
application"
for the Prudential,an actuarialprogramthatranon an IBM 650 [20, pp. 65, 67].
Around 1965 it developedPru-COBOL for that firm [27, p. 8; 28]. By 1969,
Prudentialhadalsoturnedto anotherindependent
softwarefirm, Informatics,Inc.,
for thatfirm's packagedfile management
system,Mark IV, to usein its pension
work [47]. These contacts,though limited and predominantlycenteringon
customizedor general-purpose
software,indicatethatthe insuranceindustryhad
somedealingswith the youngsoftwareindustrybeforeunbundling.
Unbundlingand the Emergenceof PackagedInsuranceApplication Software
In addition to the delaysin and lack of enthusiasmfor ALIS discussed
earlier, the major eventtriggeringan explosionof softwarefirms and software
packages
for the life insuranceindustrybeginningin 1969wasIBM's unbundling
of softwarefrom hardware[19, p. 32; 8, p. 323]. Whetherin reactionto anti-trust
pressures,
or asa way of raisingprices,or both,IBM ceasedto includesoftwareand
servicesin the priceof hardware[44, pp. 250-252;8, p. 323]. This moveprovided
an immediateimpetusto the insurancesoftwareindustry,still in its infancy.
In 1969,majorplayersin thissectorof the softwaremarketbeganto be
founded,by a varietyof differentprocesses
andconstituencies.
CybertekComputer
Productswasfoundedby a teamthatincludeda memberof the IBM programming
team that developed'62 CFO [19. p. 28]. Accordingto a later description,
130
"Cybertekwascreatedin 1969 to providea broadspectrumof computer-related
servicesto the life insuranceindustry.[...] Its main productsare an automatic
policyissuesystemandsupportsystems
for CFO andALIS" [6, p. 73] Its founders
clearlysawan opportunityfor profit openingup, and took advantageof it. Tracor
ComputingCorporation(TCC), a softwareand servicecompanyalso startedin
1969, developedits ownconsolidated
functionssystem,Life 70, to competewith
'62 CFO [6, p. 73; 22]. By 1971,EDS haddecidedto offer its insurancecustomers
a standard,
ratherthancustomized
facilitiesmanagement
package,includingitsown
standardsoftware[13, p. 58].
The insurance
industrywasnot,however,sittingbackawaitingactionfrom
softwarefirms;assomeinsurance
firmsin earliererashadenteredintorelationships
withtabulatorandearlycomputervendorsin orderto shapeavailabletechnology,
some of insurancefirms now took action to assuresoftwareavailability in the
unbundledera. In 1969, 15 largelife insurancefirms combinedforcesto found
InsuranceSystemsof America,Inc. "to developapplications
systemsandrelated
installationservicesfor the life insuranceindustry. ISA offers applications
packages,
consultingservices,andcustomcontractservices"[6, p. 73]. Similarly,
EquitableLite Assurance
societyestablished
a joint venturewith Informatics,Inc.,
a generalsoftwarevendor;the joint venture,called Equimatics,was intendedto
design software applicationpackagestargeted at the life insuranceindustry.
Whetherfoundedby usersor by softwareand serviceproviders,thesefirms and
many others founded within a few years of the unbundlingcontributedto an
expandingmarketfor insuranceapplicationsoftware.
While suchfirmsbeganto sproutimmediately,the life insuranceindustry
did not immediatelyrecognizethe importanceof the changestakingplace. The
IASA meetingsin 1970 showedthatinsurance
firmsandservicecompaniesserving
theinsurance
industrywerestruggling
to understand
theimpactthatunbundling
and
itsconsequent
highercostswouldhaveon their own operations[51; 18]. The one
conclusionthatseemedinescapable
wasthat"The new repricingof hardwareand
softwarewill makeall of usthinkquitedifferentlyin the future"[18, p. 92]. By
1971, however,theemphasishadshiftedto evaluatinginsurancesoftwaresystems
and servicesthat had proliferatedon the market[12, 37, 32]. At the 1971 IASA
conference,
onespeakergavehisassessment
of buyingpackagedsoftware:
It is usually,but not always,quickerto achieveresults,but for
this you mustpay moredollars. It may be moreerrortree, but
for thisyou haveto fit youroperationinto their package;it may
saveyousomeanalysisandprogramming
timeto implement,but
you maypay for thisby moretimeto adaptandmaintainthrough
itslifetime. Theseandmanyotheritemsareto be weighedin the
evaluationof a proposal[37, p. 47].
In spiteof thesetrade-offs,insurance
firmswereanxiousto takeadvantage
of the opportunities
createdby the proliferationof softwarefirms,products,and
servicesin the wakeof the unbundling.By 1972,responding
to "thevastnumber
of softwarepackages
beingdeveloped"for insurance,
to the "increasingutilization
of thesepackages
by thelife insurance
industry,"andto the manysoftware-related
inquiriesmadeto its offices,LOMA issuedthe first of its SoftwareCatalogs,a
seriesthat hascontinuedto the present[9]. The first editionwasactuallyin the
131
form of two Systemsand ProceduresReports: "EDP Software and Service
Companies,"and"EDP SoftwareCatalog"[23, 24]. The informationon software
and servicevendorscontainedin thesereportscame from a surveyof LOMA
memberlife insurance
firms[9, p. 23]. Thesevendorsweresubsequently
surveyed
to producedataon 81 vendorsandover275 softwarepackages.The catalogbroke
the applicationpackagesinto 25 categories,of which the mostnumerouswere
IndividualInsurance--Life,Actuarial,CFO SupportSystems,and ALIS support
systems.
A co-founder
of Cybertekexplainedlaterthat'"...nearlyall the[insuranceoriented]softwarecompaniesthat werefoundedright after IBM unbundledstarted
outby providingtheperipheralfunctions
a companyneededto make'62 CFO more
versatile'" [19, p. 28]. Even in the 1977 editionof the LOMA softwarecatalog,
mostof Cybertek'sproductsare still CFO-related,thoughrunningon the IBM 360
or 370 platform[22]. Many othercompanies
responded
to the marketopportunity
createdby ALIS's limitedsuccess
by developingalternativepackages
for moving
CFO usersupto IBM 360 hardware[19, p. 32]. Otherfirmsdevelopedapplication
packages for different hardware bases, such as Network Data Processing
Corporation'sLILA (Life InsuranceLogisticsAutomated)andrelatedprogramsfor
theBurroughs,Honeywell,UNIVAC, andNCR hardware[22, pp. 35-44; 6, pp. 7374]. But many of the firms listed in the LOMA cataloghad multiplepackages,
goingbeyondtheconsolidated
functionsoperationsintoa varietyof areasfrom file
managementsystemsto real estateand mutualfundsapplications.Thus software
packages
wereexpanding
fromtheconsolidated
functionsapplicationat thecenter
of life insurancedata processingto a variety of other applicationareasin the
insurance business.
1972wasan importantyearin theinsurance
industry's
recognition
of how
completelythe insurancesoftwarepicturehad changed. In additionto the first
LOMA softwarecatalog,a seriesof otherLOMA publicationsalso attemptedto
addressthe softwareand its evaluation[6, 25]. Similarly, the IASA conference
program for 1972 included several papers on evaluating software packages,
includingone by a representativefrom a firm that claimed it had evaluated20
softwarepackages/services
in the precedingfour years,andpurchased14 of them
[10]. These guides to evaluatingsoftware were designedto help member
companieschoosefrom the rapidly expandingarray of optionsin customand
packaged
softwareaswell ascomputerservices.In additionto suchguides,some
of the new softwarecompaniesthemselvespresentednew packagedsoftware
projects,suchas a packageto assistin underwriting,at the conference[33].
Conclusion
The explosivegrowthof insurance
application
softwarethatbeganin 1969
wasnotcausedsolelyby IBM's unbundling.As computertechnology
becamemore
powerful, especiallywith IBM's System/360line, more functionscould be
computerized;at the sametime, the enormityof the programmingtaskalsogrew,
makingpackaged
softwareincreasingly
attractivefinancially[8, p. 323]. But until
theunbundling
move,the largelife insurancefirmsgenerallydevelopedtheirown
software,andthe smallto mediumsizedonestendedto adoptIBM's bundledCFO
or, to a lesserextent,ALIS. Over time,the lowerlevelof popularityof ALIS might
have encouragedthe growth of the small independentsoftwareindustry. By
132
separating
the softwarefrom the hardwaredecisionandby increasingandmaking
visiblethecostof IBM software,however,theunbundling
decisiongaveimpetus
to theemerging
insurance-focused
application
softwareindustry.Moreover,from
SPAN to ISA andEquimatics,the insurance
industryalsocontinuedits traditionof
shaping,aswell asbeingshapedby, theinformation
processing
industry.
References
1.
2.
Robert E. Baines,"1410 ComputerApplicationsat RepublicNationalLife InsuranceCompany,"
Proceedingsre' IASA , 1964, pp. 35-40.
CharlesJ. Bashe,Lyle R. Johnson,
JohnH. Palmer,andEmersonW. Pugh,IBM's Early Computers
(Cambridge,MA, 1986).
3.
Walt Bohne,"SoftwareTravelingSpecializationRoad," ElectronicNews,(May 4, 1970), section
2, p. 55.
4. KennethR. Brown."The AdvancedLife InformationSystem- Its Value to You," Proceedit•gsqf
IASA, 1969, pp. 167-168.
5. MartinCampbell-Kelly,
"Development
andStructureof the International
SoftwareIndustry,19501990,"Bttsiness
and Ecomm•icHisrosy.24 (Winter 1995), forthcoming.
6. CharlesH. CissleyandJeanBarnes,EDP &,stemsandApl)licationsin Li/k btsurance(NY. 1972).
7. RaymondA. Duncan,"FacilitiesManagement:
Why Wouldan Established
Life Insurance
Company
Contractfor ComputerServiceswith a FacilitiesManagementCompany?"Proceedings•/' the
LOMA SystemsFot'ltm,1971, pp. D4-1--D4-3.
8. FranklinM. Fisher,JamesW. McKie, andRichardB. Mancke.IBM at•d the U.S. Data Processi•lg
hlduslrwAll Ec'OiIOiiI•('
Hi,•tot)'(New York, 1983).
9. JamesF. Foley,Jr., "LOMA DevelopsEDP SoftwareServicesCatalog,"Proceedingsqf IASA,
1972. pp. 23-26.
10. RayfordA. Freemort."FourteenSoftwarePackages
in Fore' Years"Proceedb•gs
of IASA. 1972.
pp. 27-28.
1I. GordonGoodwin, "TechmcalDevelopmentsin EDP - ProgrammingMethods,"Proceedingsof
IASA. 1960, pp. 234-236.
12. H. GordonGoodwin,"Evaluationof ComputerSoftwareSystems,"Proceedings
of IASA, 197I. pp.
49-55.
13. JerryGrisham,•'EDSWill Offer Life Insurance
Mgmt.Package,"
Electro•icNews,(April 19, 1971),
p. 58.
14. W.K. Headley.•'TheCFO Packageas a Guide," Proceeditlgso! IASA, 1964, pp. 13-16.
15. KennethM. Hills, blsura•lceDaterProcesstrig:Property--Liabili(v--Lil•k
(Philadelphia,1967).
16. AllenM. Johnson.
"ALIS andPALlS for HealthProcessing,"
Prm'eedil•gs
ef IASA, 1967,pp. 123126.
17. LeithG. Johnson,
comp.anded.."FromPunched
Cm'dstoSatellites,andBeyond:A historyof data
processingat AetnaLife & Casualtypreparedfor an exhibitheld in the AetnaInstituteGallery,"
April 5. 1984 (typescriptfrom Data Processing,
AetnaLife & CasualtyCompanyArchives).
18. JackK. Jones."Unbundle?!"Proceedit•gsq! IASA, 1970, pp. 93-94.
19. GregoryJordahl,"SoftwareExplosion:
The Choiceis Yours,"ln.•ttrro•ce
& Technology,18 (August
1993), 28-32.
20. Elmer C. Kubie, "Reflectionsof the First SoftwareCompany,"IEEE Amtalsq['the HistotT ql•'
Computitlg,16 (1994), pp. 65-7 I.
21. Doron P. Levin. Irreconcilable Diffkrem'es:RossPerot ue•xusGeneral Motors (Boston, 1989).
22. Life Office ManagementOrganization(LOMA), Operationsand SystemsReportNo. 38, "EDP
SoftwareCatalog"(NY, 1977).
133
23. __,
LOMA SystemsandProcedures
Report#14, "EDP SoftwareandServiceCompanies"
(NY, 1972).
24. __,
25.
LOMA SystemsandProcedures
Report#15, "EDP SoftwareCatalog"(NY, 1972).
, LOMA SystemsandProcedures
Report#16, "Evaluationof SoftwarePackages"
(NY,
1972).
26. Richard
W. Longing,
"A LookatC.O.S.ForC.F.O.on360,"Pr(;ceedmgx
o/' IASA,1967.pp.327329.
27. Malcolm MacKinnon, "A Brief History of the Developmentand Installationof Prudential's
AdvancedOrdinarySystem( 1963-73)," stenciledinternalPrudentialdocumentwith handnotation
of date andauthor,"presented
at lunchto commemorate
completionof AOS on Dec. 1l. 1973."
Documentobtainedfi'omMalcohn MacKinnonandconfmnedby him in interview.
28. __,
interviewconductedby RobertHancke,(Sunapee,NH, July 21, 1994).
29. ThomasP. Maher,"The IBM AdvancedLife lnfommtionSystem,"Proceedblgs
o/' IASA. 1968, pp.
118-122.
30. BraceK. McBeath,"Up-to-Date-Appraisal
of '62 C.F.O. Package."Proceedings
•/' IASA, 1964,
pp. 16-18.
31. RobertW. McComb,Jr., "TheIBM ALIS Package--Modification,
Conversion,
andImplementation
Problems,"Proceedings
q/'theLOMA AutomatiottFortim, 1968,pp. 189-191.
32. EarleA. McKever,"Experiencewith andevaluationof computersoftwareservices,"Proceedmg•
q/' IASA, 1971,pp. 53-55.
33. VaughnW. Morgan,"ComputerAssistedUnderwriting,"Proceedings
q/' IASA, 1972,pp. 48- 56.
34. NathanO'Neil, SystemsSupervisor,
menloto Irwin J. Sitkin,Asst.Secretary,AccountsDept., May
9, 1962.O'Neil DP Records,Chronological
File, 1959-1963,AetnaLife & CasualtyCompany
Archives.
35. __,
Administrator,
ComputerSe•Mces,memoto H.G. Goodwin,Asst.Secretary,Computer
Services,"Progress
ReportJanuary13to January26, 1967,"January27, 1967,O'Neil DP Records,
Chronological
File, 1967-68,AetnaLife & CasualtyCompanyArchives.
36. Dale H. Pearson,"An Evaluationof IBM's ALIS," Proceedi•gsq/' IASA, 1967,pp. 319-321.
37. RobertL Plageman,
"Experience
withandevaluation
of computer
softwarepackages,"
Proceeding
t?/'IASA, 1971,pp. 46-49.
38. EmersonW. Pugh, Lyle R. Johnson,and John H. Pahner.IBM'x 360 and EarIx' 370
(Cambridge,MA, 1991).
39. JamesR. Robinson,"A & S ALIS at Occidental,"Proceedings
o/' IASA, 1967,pp. 126-133.
40. RalphRoseman,"ALIS Implementation--One
Company'sExperience,"Proceedingxq/'the LOMA
AutoulatiotlFornm, 1968, pp. 192-196.
41. Norbert W. Schultz,"ConversionProblemsDuring the Initial ConversionTo ElectronicData
Processing,"
Proceedings
tt/'IASA,1962,pp. 593-599.
42. RobertShafto,CEO, The New England,formerlyof EDS, interviewconductedby JoAnneYates
andHansGodfrey(Boston,September1, 1993).
43. Douglas
L. Smith,"Experiences
of a Functioning
CFO '62Operation,"
Proceedmgx
q/' IASA, 1965,
pp. 13-17.
44.
Robert Sobel, IBM.' Coto.v.vus
i• Transition (NY, 1983).
45. SPAN Data Processing
Center, Inc., November16, 1964 report,"Developments-- June 1960 to
Date,"in Oversized
Documents:
SPANGuestRegister;
AetnaLife & CasualtyCompanyArchives.
46. D.W. Stinson,"Long-Range
ComputerSystemsPlanning- Software,"Proceedingxo/' IASA, 1972,
pp. 140-I41.
47. J.H. Talbot,Jr., "The Mark IV Programming
SystemUsedby the PrudentialInsuranceCompany
of America,"Proceedins,,x
o/' IASA, 1969,pp. '7-9.
134
48. GeorgeR. Tindall,Jr.. '"62 CFO For HealthPolicies,a Historyand Evaluationof One Installation,"
ProceediJ•gx
•i" IASA, 1966, pp. 225-227.
49. GeorgeR. Van Wyck, "Consolidated
Functions
- SmallScaleComputers,"Pr•;ceedt'm,,x
•;/the IASA,
1963, pp 80-82.
50. Gilbert M. Whitfill, "ConsolidatedFunctions- Small ScaleComputers,"Prt•ceedi•tg.v
t•l'the IASA,
1963, pp 82-83.
51. DonovanL. Wilson,"ComputerRelatedServices: Unbundling,"Pt't•ceedmgs
•[ IASA, 1970, pp.
92-93.
52. JoAnneYates,"Co-evolution
of InformationProcessing
TechnologyandUse:Interactionbetween
the Life InsuranceandTabulatingIndustries,"Btt.vine.v.v
HistoryReviews'.
67 (Spring 1993), 1-51.
53. __,
"FromTabulatorsto Early Computersin the U.S. Life InsuranceIndustry:Continuities
and Influences,"Sloan WP 3618-93 (October 1993), revised,September1994,