1. Effect of Technology on Sales Performance: Progressing from Technology Acceptance to
Technology Usage and Consequence
Author(s): Michael Ahearne, Narasimhan Srinivasan, Luke Weinstein
Source: The Journal of Personal Selling and Sales Management, Vol. 24, No. 4, Customer
Relationship Management: Strategy, Process, and Technology (Fall, 2004), pp. 297-310
Published by: M.E. Sharpe, Inc.
Stable URL: http://www.jstor.org/stable/40471971 .
Accessed: 22/10/2011 10:57
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
M.E. Sharpe, Inc. is collaborating with JSTOR to digitize, preserve and extend access to The Journal of
Personal Selling and Sales Management.
http://www.jstor.org
3. Selling& SalesManagement
298 Journal Personal
of
Figure 1 command-driven character-based interfaces. Therewerelim-
ExpectedPerformance PlottedAgainst itedapplications available. Applications weregenerally expen-
TechnologyUsage Time sive, in
and thecompetition anyapplication was
field relatively
limited. Indeed,manyoftheempirical studies used in testing
the earlyacceptancemodels measuredthe acceptanceand
usage of word-processing applications(Davis 1989; Davis,
Bagozzi, and Warshaw 1989, 1992). Today,the typewriter
has becomea relic, and there pervasive ofwordproces-
is use
sors,and thistoo has long ceased to be an innovation.
ci s' Over the past couple of decades,technology evolved has
and is significantly different than when acceptancemodels
werefirst introduced. Today,PCs have become pervasive in
I / business, withan installed base ofapproximately million201
(Juliussen 2002). Usersinteract withfriendly graphical inter-
facesthatare increasingly easierto use. Many vendorspro-
^-- > videa variety application
of and
software, prices keepdropping
Usage Key
dramatically. competitive factors a user'schoiceof an
in
-
applicationincludeproductfunctionality ease of use and
two keyantecedents TAMs. With increasing
of adoptionof
new technology just to staycompetitive, businesses have to
desirable behavior and impliesbetter performance" (2003, p. necessarily invest heavily in training their employees withnew
191). But this criticalassumptionthat usage is a proxyto technology (Marshallet al. 2000).
in
performance not been tested theliterature.
has Although havecertainly
we learned great
a deal from TAMs
We recognize thatthe focusof past research been the
has in theliterature, specific
the technologies examined havebeen
acceptance of technological innovations, and the objective evolving, of
and we can stillfindexamples technology imple-
was to accelerate acceptance technology. whathap-
the of But, mentationfailuresin recentCRM literature (e.g., Jones,
is
pensafter technology accepted?
any Would increased usage Sundaram,and Chin 2002; Riversand Dart 1999; Speier
always be desirable? Does technology alwaysimprove perfor- and Venkatesh 2002). Beyondacceptance, productivethe ef-
mance?Initially, thereare acquisition and setupcosts,train- fectof technology usage on performance a veryimportant
is
ingcosts,and maintenance costs.The positive returns would issue today.
start in
flowing onlyafter some time.How does technology A major contribution this studyis the criticalassess-
of
usage affectperformance overan extended period of timeaf- mentof theimplicit assumption a positive
of and linearrela-
terthe initialsetupperiod,training completed,
is and users tionship betweentechnology usage and performance. First,
are now usingthe system regularly? it possibleto evaluate
Is every salesperson knowsthattimeis a scarceresource be to
thereturns from specific
a technology? These important ques- judiciouslyallocatedamong varioustasks.Second, technol-
tionsarevery relevant managers.
to The purposeofthisstudy ogy cannot substitute severalnecessary tasks,such as travel
is to examinethe effect operational
of CRM technology us- and personal discussion. Third, can
a linearrelationship only
age on salespersons' performance after theyhave been using holdfor limited
a range, substitutability otherwise
as will elimi-
-
the technology at leastsix months.It is consistent
for with nate all expensive inputswith low marginal productivity
the research in
priorities sales strategy performance
and re- that is, everyresourcehas diminishing returns. Otherwise,
cently identified Leighand Marshall(2001): information technology be used increasingly, salespeoplecan be
by can and
technology (IT), sales/service/support systems, selection/ eliminated.
and However, thisis not likely happen.There is a
to
development of salespeople. limitto whattechnology do, and we recognize
can thattech-
nology is only a tool, and the substanceof personalselling
BACKGROUND stillresides withsalespeople.
Whatiftherelationship between technology usageand per-
When TAMs were first in
introduced the literature around formance curvilinear? Figure A meta-analysis time
is (See 2.) of
themid-1980s,the businessenvironment verydifferent and performance
was relationships showssimilar curvilinear pat-
thanit is today. The computer was
industry mainframe ori- terns (Sturman 2003). BefDre can testforpossiblemedia-
we
ented, withan installed base ofapproximately millionper-
19 torsand moderators affect relationship, mustfirst
that the we
sonalcomputers (PCs) in theUnitedStates(Juliussen 2002). understand how and whentheuse ofsalestechnology leadsto
Usersgenerally interacted with computers using unfriendly increased and
effectiveness efficiency (Tanneret al. 2005).
4. Fall2004 299
Letus consider operational
an -
CRM context a term coined Figure 2
by Tanneret al. (2005) to describe thatportionof CRM us- Proposed Curvilinear Relationship Performance
of and
age forsales processmanagement, such as lead generation, Technology UsageTime
contact, scheduling, performance tracking, and otherfunc-
Enabling
tionsin a salescontext. Intuitively, is reasonable expecta
it to f ,. ' ν
λ
a CRM system improve or her per-
to his j Disabling
salesperson using
formance effectively
by increasing or her usage of CRM
his
technology from, say,threeto fourhours per week. But if
ι ν '
thatsamesalesperson increased or herusageto 30 hoursa
his
week,thatsalesperson would be increasing or her CRM
his
usage at the expenseof otherworktasksthatmust be per-
formed. with any businessthatattempts optimizeits
As to
>
a
use of resources, salesperson needs to optimizehis or her
- -
allocation a keyresource time amongvarioustasks,
of in
orderto maximize or hersalesperformance.
his
It is natural thatthere in
would be heterogeneity technol- Usage
ogyusage across employees. Anecdotal evidence suggested
has
thatsalespeople amongthemosttechnophobic resis-
are and Note:Assumes is
technologyinstalled, training over, CRM is in
is and
tantof all white-collar workers(Mills 1995; Parthasarathy use.
and Sohli 1997), thoughthisshouldbe changing technol- as
ogy becomesmorepervasive. Some people takeenthusiasti-
callyto new technologies and are innovators. Othersprefer to better understand how technology acceptanceand usage
the old way of doing businessand are laggards. Hence, it is leadsto improved performance. TAMs provide witha
The us
quite that
conceivable salespeople may often or
overuse under- of
richfoundation theory and important constructs regard-
use CRM technology. Overusemay come at the expenseof ing usage,as outlinedin Table 1. The nextlogicalstepis to
other salestasks that mayhavea greater effect performance. studythe effect technology
on of usage varianceon actualper-
Underutilization would implythatthe salesperson less ef- is formance.
fective thanhe or shecould be, ifonlyhe or sheused thetools The major contribution this research to examineif
of is
of theCRM technology a greater
to degree. technology, such as capitaland labor,has diminishing mar-
If thereis a curvilinear relationship betweentechnology ginalproductivity in thecorporate production process.Time
usage and performance, has veryimportant
it implications of salespeople not unlimited
is and has to be allocatedacross
forthe effective of technologies personalselling.To
use in various tasks,withdiffering effects performance.
on Currently,
what degreeof usage is technology enabling? What are the thedifferential allocationof timeon technology voluntar-
is
factors thatcan help us predict when technology usage will ilydecidedbyeach salesperson, withno recommendations or
havea positive effect performance?
on Does technology have processtraining how to optimizetheallocations.
on Histori-
diminishing returns? What are the factors thatwill predict cally,time management salespeoplefocusedon the best
for
when usagewill have a disablingeffect performance?
on To way to rationa salespersons time among the accountsthat
answer suchquestions, researchers to movebeyondtech-
have makeup hisor herterritory Zoltners,
(e.g., Sinha,and Chong
nology acceptance and usagecontexts. need to study
We situ- as has
1979). Today, competition intensified, moreis expected
ationsafter particular
any technology is adoptedand installed of salespeople(e.g., Nonis and Sager2003), leadingto more
by an organization, salespeoplehave been trained ef-
and to in
stress salespeople. Time management now be seenas a
can
fectively thesystem.
use cluster behaviors
of thatare deemedto facilitate productivity
and alleviatestress(Lay and Schouwenburg1993). In this
study, deal onlywithtimespenton specific
we CRM tech-
THEORETICAL FOUNDATIONS AND
MODEL nology on
and its effect performance. This is not a broad
timemanagement study.
Improving humanperformance organizations a primary
in is In theearly 1990s,Microsoft chairman Gatespredicted
Bill
goal fororganizations increase
to competitiveness (Marshall thatbusiness technology eventually would allowpeopleto do
et al. 2000). Organizationsspend significant amounts of a fullday'sworkin fouror fivehours,freeing themto spend
moneyon technology implementation look to get in-
and moretimeon leisure or and
pursuits withfriends family. In-
creased and
productivity performance from their investments stead,bythemid-1990s,theaverage was
workweek nearly an
(Marshall et al. 2000). To meetthesegoals,researchers need hourlongerthanin theearly1980s, and a survey foundthat
5. 300 Journal Personal
of Selling& SalesManagement
Table I
Review ofTechnology Acceptance Models in the Literature
Topic Reference Key Conclusions
AcceptanceModel (ΤΑΜ)
Technology Davis ( 1989); Usage can be predicted fromintentions.
Davis,Bagozzi, Warshaw( 1989);
and Keyantecedentsof usage are:
Taylor and Todd ( 1995) · perceivedease of use,
• perceivedusefulness.
Theoryof ReasonedAction(TRA) Davis,Bagozzi, Warshaw( 1989);
and Subjectivenormfrom TRA is examined, mixed
but
(an acceptancemodel) Jones,Sundaram, Chin (2002)
and resultson significance.
Triandis acceptance model of
(an Thompson,Higgins, Howell (1991)
and Key constructs thatwere significant
antecedents
PC utilization) of usage were:
• social factors(similar subjectivenorm),
to
• complexity use (similar ease of use),
of to
• job fit(similar perceivedusefulness),
to
• long-term consequences.
Innovation
Diffusion
Theory (IDT) Moore and Benbasat(1991) Looked onlyat adoptionversusnonadoptionand
(an acceptancemodel) added new constructsof
• observability,
• trialability,
• voluntariness.
MotivationalModel (MM) and Warshaw (1992)
Davis,Bagozzi, Enjoyment an antecedentof intention usage.
is and
(an acceptance model)
Theoryof PlannedBehavior(TPB) TaylorandTodd ( 1995) Subjectivenormsignificant,made up of
(an acceptancemodel) · peer influence,
• superior'sinfluence.
Adds perceivedbehavioral controlwithantecedents
of
• self-efficacy,
• facilitating
conditions.
Social Cognitive
Theory (SCT) Compeau and Higgins (1995); Self-efficacyan antecedentof usage,both directly
is
(an acceptancemodel) Compeau, Higgins, Huff
and (1999) and mediatedthrough anxietyand affect.
ΤΑΜ extendedwithTask-Technology
Fit Dishaw and Strong(1999); TTF does not directly usage but is mediated
affect
(TTF) Mathiesonand Keil (1998) through ease of use.
AcceptanceModel Extended
Technology Sundaram, Chin (2002);
Jones, and Adds to theΤΑΜ (Davis 1989) by including
(TAM2) Venkateshand Davis (2000) antecedentsof perceivedusefulness subjective
of
norm, image, relevance,
job outputquality, result
and
Moderatorsof intention use
demonstrability. to
are experienceand voluntariness.
Unified
TheoryofAcceptanceand Venkatesh al. (2003)
et Summarizes literature
the nicelyand proposes a
Use ofTechnology
(UTAUT) unifying framework. antecedentsof intention
Key
and thus usage are:
• performance expectancy(usefulness),
• effort expectancy (ease of use),
• social influence(subjectivenorms),
• facilitating
conditions.
Moderatorsare:
• gender,
•age,
• experience,
• voluntariness.
Addiction
Technology Brown( 199 1; 1993); Charlton(2002); Findsthe existenceof a formof behavioral
Griffiths
(1998) addictionin the use of computersand the Internet.
Addiction has six components: salience,mood
modification,tolerance, withdrawalsymptoms,
and
conflict, relapse.
LearningOrientationand Performance Ahearneet al. (2004); Shows the linkbetween learning
orientation
and
Orientation and Kumar(1994)
Sujan,Weitz, performance orientation the adoptionof
to
technologyand performance an SFA context.
in
Complexity Teo and Lim(1996) Studyof factorsassociated withusage.Adds
complexity in the contextof beingdifficult
but to
use,the opposite of ease of use.
6. Fall2004 301
morethan68 percent respondents at leastsomewhat
of felt and used a CRM system 30-plus hoursper week clearly
for
more overwhelmed work today (Carey 1996). With the
at would have littletimeto make sales calls and close transac-
speedof technology, managers expectmoretasksto be com- tions. This would have a severe negative effecton the
pletedbyemployees thesameamountoftime.At themost
in salesperson's performance. thesame salesperson
Yet usingthe
basic level,employees have two resources, timeand energy, CRM technology say,three
for, hoursperweek,might an see
and theseresources undertheemployee control(Brown
are s increasein his or her productivity, per week statistics,
calls
and Leigh1996). So withincreased and
expectations demands, and salesas a percent quota performance.
of Maybethissame
salespeoplemust manage theseresources optimizetheir
to salesperson usingthe CRM forfourhoursper week would
effectiveness, with few,if any,tools to make objective
but see evenfurther in
increases theirperformance. theques-
So
observations decisionsin this resourcemanagement.
or In tion is, at whatpointof usage does performance to de-
start
the CRM world,othershave touchedon thisissue,such as cline? To capture the enabling and disabling aspects of
Engleand Barnes(2000), suggesting too much use and
that technology usage,we proposea quadraticmodel thatcan be
emphasis on theplanning function CRM systems
of maynot expressed as
be productive. * *
-
Performance (X + βλ Usage + β2 Usage2.
Similarly, self-regulatory capabilities critical success
are for
in boundary-spanning such as personalselling(Brown,
roles The above model is a generalized way of capturing the
and
Leigh, Jones 2005). Self-regulatory effectiveness is defined enabling of if
effect technology β{ is positive and significant,
as theability to monitor and control ones own goal-directed is and
and if/?2 negative significant, similar theBass (1969)
to
and
behaviors performance (Bagozziand Dholakia 1999). In in
model. It is powerful thatit captures neteffect vari-
the of
complex taskenvironments such as sales, individualsmust ous factors thatcan influence and ßr
βχ
pursuemultiple goalssimultaneously (Austinand Vancouver There would be a minimallevelof sales performance ex-
1996; Bagozziand Dholakia 1999). Attention effort and can pectedforeach salesperson whenthere little
is use ofthespe-
be channeled from mostproductive
the tasks thosethatare
to cific technology under investigation. The coefficient is βχ
for
less instrumental attaining centralgoals (Brown,Leigh, to be positive, theenabling -
effect thatis,
expected denoting
and Jones2005). While research shownthatself-regula-
has initially using technology improves performance. However,
tory skillscan be enhancedthrough training areassuch as
in technology expectedto have diminishing
is returns. Thus,
timemanagement (Leach 1999), how do we train salespeople thecoefficient is expected be negative,
β2 to denoting thedis-
- point,moreusage
withself- regulatory on theusageofCRM systems,
skills when ablingeffect thatis, beyonda particular
we do notunderstand effect suchusage
the of on their -
would detractfromperformance because otheressential
perfor-
mance? taskswith greater marginalproductivity would have to be
of
We claimthatone possiblebenefit CRM implementa- neglected or sacrificed.High levels convenience
of usingtech-
tionswould be more effective of the salesperson's
use time nology or highlevelsof "relative enjoyment" to
(relative other
(Tanneret al. 2005), butwe do not have sufficient empirical tasksperformed salespeople)using technology,
by self-effi-
data to support thisclaim,or how it maybe moderated a by are
cacy,ease of use, usefulness, social factors additional
or
salespersons technology expertise. Absenttoolsto objectively facilitating factors thatmay increaseusage of technology in
optimize their management timeand effort
of resources, sales- sales situations.
persons will makesubjective to
estimates prioritize opti-and Certaintasksare moreconvenient perform to thanother
mize usage of these resources.We theorizethat affective tasks.For example,a salesperson might findit moreconve-
orientations tasksfrom
to whichsalespeopleget internal re- nientto workwithhis or her operational CRM technology
wards couldhaveenabling in accepting usingtech-
effects and insteadofgetting intoa carand driving visita customer.
to It
nology, whereasresistance avoidingtasksthatare not as
to could be even moreconvenient the salesperson tiredor
if is
enjoyable others
as could contribute a disablingeffect,
to as expects to encounter objections from prospective
a customer.
would the limitation time.This theory only being of-
of is This increases positiveeffect working
the of withthe CRM
feredas a way of understanding why different individuals technology, taskthatan employer
a would typically re- still
might havedifferential of
utilization technology. Furthermore, quire the salesperson perform, allowsthe salesperson
to and
thisexplanation offered
is only as a background under-
for to partially avoid an important alternate taskwith possible
standing affinity to
or disaffinity technology and is not the negative effect.
focusoftheresearch. Our current study dealswiththeexplo- The userwilltendto spendmoretimeon thosetasks from
rationof thediminishing of
returns technology usage. whichhe or she getsmoreenjoyment, insteadof beingever
To assumethatperformance would continueto increase vigilant aboutwhichtasks willcontribute mostto perfor-
the
linearly with technology usage is intuitively appealing.
not for
mance. Because timeis a scarceresource anysalesperson,
a
To illustrate, successful outside salesperson who adopted a salesperson spending lotoftimeon an enjoyable would
a task
7. 302 Journal Personal
of Selling& SalesManagement
have less time to spend on tasksthatmay have a relatively technology usage and its relationship performance,
to and
greatereffect performance.
on Hence, thereexistsa need to greaterhomogeneity withinthe sample would ensurethat
balance use of any resource with its marginal
productivity, extraneousfactors not confound results.
do the Therefore,in
thatis, use technology long as it has a positiveeffect
as on thisstudy's
context, having a singlefirm an advantage.
is
performance, but not beyondthat.For thisto happen,one
has to first a
establish curvilinear then a feed-
relationship, Measures
backsystem, a
and, finally, correctivemechanism thatguides
a salesperson optimally any technology.
to use The testing the proposedrelationship
of betweenIT usage
and salesperformance depends on twocritical variables: tech-
nologyusage In
and performance. past research technol-on
METHODOLOGY
ogy usage, researchers have relied almost exclusivelyon
self-
reported of
measures technology usage(Jones, Sundaram,
Sample and Site Selection
and Chin 2002; Speierand Venkatesh 2002). In ourstudy, we
We conducted our research withinthefemale health-care di- sought an objective measure salesrepresentative
of technology
visionof a mid-sized pharmaceutical firm. This divisionof usage.In an effort accurately
to track technology
the usageof
thecompany a SiebelCRM system placefor
had in morethan salesrepresentatives, worked
we witha company thatspecial-
one year to
prior theimplementation our research.
of Siebelis izesin Siebeltechnology in
implementation thepharmaceuti-
widely considered be themarket
to leaderin CRM systems in cal industry. In cooperationwith the IT specialistsat the
the pharmaceutical industry,holdinga dominating shareof mid-sizedpharmaceutical company, software
a productwas
themarket. configuration theSiebelsystem this
The of in study installedthatruns alongsidethe Siebel CRM software that
is typical the pharmaceutical
for industry. The Siebel system tracks usage of technology sales representatives. soft-
by The
included salestoolsdesigned facilitate planning,
to call postcall wareprogram placesa timestampwhena salesrepresentative
recordkeeping, information gathering, intrafirm communi- beginsusingthe Siebel software a timestampwhen the
and
cation,and customer accountmanagement. salesrepresentative completes his or hersession. The software
Our analysis sample consists 131 sales representatives. also timestamps
of and track exactscreen usagewithin 90-plus
54 of
Approximately percent themare female, and the me- screens availableto thesalesrepresentative. Every timea sales
dian age is in the rangeof 26-35 years. The averageexperi- representative synchronizes with the companysystem re- to
ence in a salesjob is 8.5 years(s.d. = 5.7), theaverage tenure port their activity to downloadnew customer
call or data,the
withinthe companyis 5.8 years(s.d. = 4.4), and the sales- software transfers salesrepresentative's informa-
also the usage
peopleworked their in territory average 2 A years(s.d. =
an of tionto thecompany's database. The usagefigures not in-
did
1.4). All salesrepresentatives collegegraduates,
were withap- clude tracking timespentin administrative activities(suchas
proximately 15 percent having advanceddegrees. the upload/download functions), but focusedon screens for
We considered two majorconditions necessary our
as for territory analysis activities,
precall planning and preparing ac-
research: theuse of technologies voluntary
(1) was such that tivities,and postcallplanningand reporting activities. The
in
variance IT usageamongsalesrepresentatives existed, and CRM system had been in place formorethanone year, and
(2) thetechnology been in place long enoughto provide
had the data capturesystem had been in place formorethansix
sufficient familiarity establish
and stable of
patterns useamong months without any majorproblems. onlyused data for
We
thesalesforce. These werenecessary becauseour primary in- salespeople who had been usingtheCRM system at least for
vestigation was to measurethe consequenceof technology sixmonths aftertraining to eliminate confounding the
any on
usage on sales performance. wanted to ensurethat the
We data frominitialtraining. noted earlier,
As system usage is
was
technology adopted,salespeoplehad been trained, and voluntary, usageis not evenrequired minimal
and for report-
therewas heterogeneity usage among the salespeople.
of we
ing purposes.However, eliminated sevenusersfromour
Choosingtherespondents from single
a firm opposedto a
(as samplewithfewer than 50 hitson the system the three-
in
cross-sectional studyusing variousfirms) an advantagein
is monthperiodmeasured, orderto eliminate
in users who only
the present as
setting, it controlsforconfounding external accessedthe system minimally who chose not to use the
or
effects to the variability market
due in contexts (e.g., com- system all.Adoptionofthetechnology a precursor our
at is to
petitive situations) and organizational factors (e.g., informa- studyabout usage,and thisfitsour requirement. This lefta
tionsystems salesmanagement
and practices). The limitation sampleof 131 salespeople conducting analysis.
for our
of investigating salespeoplefromany singlefirmcan, how- Sales representatives unawarethatthe parentcom-
were
ever, lead to a questionof representativenessthe firm
of and pany was tracking, had the abilityto track,theirCRM
or
thegeneralizability results.
of However, what is beingtested usage patterns. This was the companypolicy.It was quite
in our modelis thetheoretical relationship between degree of clear,based on discussions withseniorsalesmanagement and
8. Fall2004 303
salesrepresentatives, thesalespeople notknowabout
that did ure 2, two regressions wererun. First, simplelinearregres-
a
theusagetracking. sion withtechnology usage as the independent variableand
Data from system
this werecollected a three-month
for pe- job performance as the dependent variable was run.Second,
riodfor salesrepresentatives.
all A measure thetotalnumber
of a multiple regression as
withboth usageand usage1 indepen-
of system (or totalscreens
hits used in a three-month period) dentvariables and job performance thedependant
as variable
wasusedas a measure salesperson
of technology utilization(av- was run. (See Table 2 fortheregression results.)
eragehits= 283; s.d. = 228). We used hitsrather thantotal First, observethatthe linearrelationship
we betweenIT
timebecause, withtotaltimeon thesystem, thereis a possi- usageand job performance not significant > 0.05). This
is (p
bility of sales representatives leaving a screenrunning with- clearly calls into questionthe assumptionmade in past re-
out usingthesoftware. One shouldnote,however, these that searchthat the relationship linearand providesevidence
is
measures hitsand timeare highly
of correlated (0.98). Per- forits negation.Second, supportis foundfora curvilinear
haps we did not have to worry thataccount.An added
on relationship. effect IT usage, is positive signifi-
The of βχ, and
advantageto using hits is that it capturesthe wider usage cant (B = 0.64, ρ < 0.05), and theeffect oiusage2^^ is nega-
acrossmultiple screens and, also, multiple accessto thesame tive and significant = -0.55, ρ < 0.05), supporting
(B our
functional screens, it was seen to be beneficial.
if Further, we hypothesis. including quadraticterm, fortechnol-
By the /?2,
ran our analysis using hitsand separately using timeas our ogy usage, the varianceexplainedin salesperson perfor- job
measure usage.Our results
of weresignificant irrespective of mancesignificantly increased < 0.05) from1.2 percent
(p to
whether measured
we usagewithusinghitsor time. 4.2 percent. Although the amount of varianceexplainedin
We obtainedthesalespersons performance from company job performance (percent quota) is rather low,the following
records. Salesperson performance was operationalized using points should be considered when interpreting relation-
this
thepercent quota persalesrepresentative on achieved
of based ship.First, theeffect "robust,"
is becauseitlinkstwodifferent
saleslevelsoverthethree-month periodcorresponding the to data sources(both of whichare objective).Second, this4.2
technology usage data collection. did not attempt con-
We to of in
percent variation salesperson is
performance explained
trolforsalesperformance priorto theimplementation the of only by IT usage.Third, thisis the incremental increase in
CRM system fortworeasons. First, sincetheSiebelCRM sys- performance beyond whatwas accomplished theolderSFA
by
temwas implemented, there havebeen territory realignments technology thatwas replacedwith the current technology.
along with addition many
the of newsalespeople. Second,there Fourth, amountofvariance
the explained in salesperson per-
wasa different CRM system placeprior theSiebelsystem
in to formance compares favorably theindividual
to contributions
and thuspriorperformance could only allow us to look at made by othervariablesin previoussales studies.The vari-
incremental ofthenewsystem
gain versus old system.
the How- is
ance explained consistent withprior literature findings that
ever, Churchill al. (1985) noted,percent quota is a
as et of "no singledeterminant explaina very
can largeproportion of
strong measure performance, it controls externalities
of as for thevariancein objective salesperformance" (Churchill et al.
such as differences acrossterritories, including territory size, 1985, p. 116). Furthermore, Churchill al. foundthat"the
et
market potential, and economicconditions, thepercent-
and to
abilityof individualdeterminants predictperformance
to-quota measure in
should, part, for
control prior performance. seems rather unimpressive" (1985, p. 110). Our studyfo-
Sales quota was calculatedbased on the volume of product cused on the incremental difference performance to
in due
sold(prescriptions) customers
by in
(physicians) a salesperson's theupdating theCRM technology. does notcapture
of It the
territory (a company-defined of physicians), compared
set as entire effect technology onlythecapacity theupdate
of but of
to a target quota thatis setat thebeginning theyearbyan
of (in the CRM technology studied)to improve performance.
external organization in
specializing salesforce compensation. The 4.2 percent variance explained compares wellto theother
Because prescription in
information the pharmaceutical in- findingsin the literature,ranging from 3 to 9 percent
dustry is accurately tracked thephysician
to level(due in large (Churchill al. 1985; Vinchuret al. 1998).
et
partto the factthatthe industry heavily is regulated the
by By examining plot of the effects IT usageon sales-
the of
Food and Drug Administration), withthevastmajority all of personjob performance in Figure3a, one can see thatthe
pharmacies reporting customer-prescribing to twomajor
data idealsystem usage(488 hits), thepointatwhichusagemaxi-
or
secondary data houses (NDC and IMS), this represents an is
mizesjob performance, higher thantheaverage current sys-
accurate picture of a salesrepresentative's performance. tem usage (283 hits).A histogram IT usage is shownin
of
Figure One shouldnotethatabout20 percent thesales-
3b. of
people in the sampleuse the system above the ideal levelof
Results
488 hits. One has to be cautiouswith using "hits,"as it is
In orderto testforthepresence a curvilinear
of relationship possible,but not probable,thatif the sales training effec- is
IT as in
between usageand job performance, represented Fig- tive,salespeople having more"hits" navigating lot before
are a
9. 304
s s s?
/NI Ο O*
■D
8
Ή îg gpSpqp
Ö Ö Ö Ö Ö
M *C
ΰ
(Λ
κ
υ υ - ή
**■ rs - t oo η
- 00 ΙΛ On ^
+, η M '0 Ν Ov
<N - - <N -
m cm |
g - <N
φο
(Λ
II ^ 3 S3
ι
1 «8
Sie, ik{ ^ "ü
*£+ "g
0>ft> *λ - Oro - Ο
SS
Dr)
'S
« cfc^pppp
loI^TJ Cjujrooi^OO
Si 8S 8ΰ *λ
^^ ^ ö ö c
:g φ
+ + < ΐ
II II g îg ^J
ÎÎ „ ϊ?
κ
Su
3
si
ÎÎ
CC
§§ öS«,-!
flyON^<Nt*vO»^>
I
<u<u Loproppio
II **
°* £
Ί-Ι- '
- ^
2î ^ y
ΣΣ -g -SoSoSS-p
LUli: Σ - (Ν Σ - CM Τ
10. Fall2004 305
Figure3a
PlotofEffects IT Usageon Salesperson Performance
of Job
use
Avg.
Usage(4g8^
Qptjma|
/
-ι
150 (
y-
140 -/-
130 jf-
I 120 ^^^^ »^^
0 110 -^S~
£ 100 -N^
Ü x
c 90
0 80
i™
0
n
60
50 J 1 1 h 1 1 1 1 1 1 1 1 «
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Total overThree-Month
Usage(Hits Period)
Figure3b
Histogram IT Usage
of
401
Iliiiii-
' '%%%
%%% _
%%%o
' ' ' ' ' ' ' ' ' "·%
XVVVVVVVW
ITUsage
11. 306 Journal Personal
of Selling& SalesManagement
finding whatthey want,and there stilllearning-curve
are ef- Alternately, underusers be identified
can and sent in for
fectsthathave not stabilized. as has
However, training been additionaltraining thattheyare able to get to the opti-
so
provided, and a yearhas passed,we do not expectthispos- mum point faster. Overusers can also be identified sent
and
viewpoint be likely.
siblealternative to fortraining reducetheirtechnology
to usage, and evaluate
whether thereare sufficient productivity gainsfromthe ex-
MANAGERIALIMPLICATIONS tentof technology A feedback
use. mechanism a guiding
and
system arefuture developments thatcould vastly assist im-
in
of
Most providers CRM software havetheir rootsin SFA, and provingoverallperformance. recognizethat individual
We
so "thebulk of CRM functionality designedto enhance
is firms will have theirown policies,both in the opennessof
salesand salesmanagement functionality" (Shoemaker 2001, monitoring, well as in human resources
as training and de-
p. 178). Our research showsthatthere a positive
is payoff to velopment. Also,sometrainees be quicklearners com-
will and
SFA/CRM, but thatthe enablingeffect tapersoff,and the fortable withtechnology, whereas someothers maybe slower
relationship betweentechnology usage and performance is in acquiringand usingtechnology improvetheirperfor-
to
curvilinear.This is a powerfulimplication. For in our par- mance.Would theoptimum pointbe reached quicker the
for
ticularsample, average, appearsthatincreasing usage
on it the fast learners withgreater technology expertise? Would there-
will
oftechnology havea positive effect performance.
on Tech- be
lationship different different
for salespeople basedon their
usageshouldbe encouraged the
until optimum point- technology expertise? Only a close monitoring technology
of
nology
is, -
that thepointofinflection where positive
the effecttapers and itseffect performance yieldusableresults
on will to
usage
offto zero. How can we identify optimumpoint?We
this and
managers, periodicupdating thestudy of suggested here
needto keeptrack technology
of usageand updatethesystem would be extremely useful. address issueoftheeffects
We the
to incorporate changes to the expert system thatmay be in of heterogeneous technology expertise acrossthe sales force
use, orto provide valuable information themanager guide
to to in
further thenextsection.
salespeople about their technology usage. To achievethe above recommendation identifying
of the
Tracking technology usageposesan interesting conflictfor optimumpoint of usage,it is conceivablethatduringperi-
managers between confidentialmonitoring of technology use odic sales meetings, salespeopleat or nearthe optimum
the
versusprivacy concerns theirusers.Althoughoutsidethe
of be provideda forumto discusstheiruse of technology, so
scope of thepresent we
study, recognize thatit is an impor- thatthey becomeexemplars theothers emulate.
for to Behav-
tantissue,and firms will need to decide how to respondto ioralmodification through such recognition be quite ef-
can
this The
conflict. salesmanagement literature underscored
has fective. anecdote,instances wheretechnology
By of playeda
theimportance trust
of amongsalespersons salesmanag-
and positiverole can be identified used in training
and sessions.
ers (e.g., Flaherty and Pappas 2000; McNeilly and Lawson Each firm may use itsown antecedents technology that
of use
1999; Rich 1997). If a firm had an open policyof tracking it observes itssalespeople.
in Equallyimportant is theidenti-
technology use, the salesperson would knowwhen the opti- of
fication factors thatcontribute the disablingeffect,
to so
mum point is reached,but thereis also the possibility that thatpeople are persuadednot to overusetechnology when
-
thesituation be "gamed" thatis, ifone is rewarded
can for productivity declines.
marginal
logging on a certain numberof times, keepingthe screen
or Although we havea situation whichthebenefits tech-
in of
on for certain
a amountoftime, these evenbe automated.
can nologyusage are capturedthrough performance, Rivers and
The potential salespersons
for manipulating use of tech-
the Dart (1999) findthatformal quantification ofcostsand ben-
nology, when thereis the open policyof tracking such use, is
efits uncommon.Erffmeyer Johnson and (2001) statethat
assumesan adversarial relationship betweenemployees and systematic formal planning and evaluation practices werenot
management. Similarly, advisingsalespersons increaseor
to widelyadopted. Positiveoutcomeswere improved commu-
decreasetheir of technology
use surrenders confidential-
the nication withclients and accessto information. Negativeex-
ityof monitoring such use and mayadd to the lack of trust periences included underestimation cost and amount of
of
between and
salespersons salesmanagers. training required.In thisstudy, thoughwe did not specifi-
If a firm wishesto keep the monitoring confidential, the cally look at quantification costs and benefits opera-
of of
managercan use the information periodic discussions
for tional CRM, a long-termstrategicpayoffis improved
about how particular salespeoplecan improvetheirperfor- performance, such returns recurring.
and are
manceby usingtechnology a greater lesser
to or degree.This Productivity as measured payback
in periodofoperational
way the overusers also be identified,
can and, in theircase, CRM (SFA) can be found in the literature. Gillan (1992)
themanager could counselthemto perhapsspend less time a
reported paybackperiodof 18 monthsfora $1.5 million
withtechnology to improve some otheraspectof their
or on SFA; Taylor(1993) foundabout two-thirds respondents of
job requirements. in a survey reported a payback period lessthaneightmonths.