Information & Management 48 (2011) 328–335

How organizations motivate users to participate in support upgrades ofcustomized packaged software

Huoy Min Khoo a,*, Cecil Eng Huang Chua b, Daniel Robey c

a Department of Information Systems and Technology Management at the University of Texas, San Antonio, USAb University of Auckland, New Zealandc Georgia State University, USA

A R T I C L E I N F O

Article history:

Received 5 May 2010

Received in revised form 14 March 2011

Accepted 4 August 2011

Available online 7 September 2011

Keywords:

Packaged software upgrade

Maintenance strategies

IT implementation

Framing

Enterprise resource planning

Enterprise systems

A B S T R A C T

Support upgrades are undertaken to correct errors, improve speed, and otherwise improve an existing

version of customized packaged software. Motivating such projects is especially challenging, because

users typically anticipate little benefit. We investigated ways of motivating user participation in

maintenance upgrading projects via an in-depth case study using the method of communicative framing.

This argues that (1) the positivity or negativity of a frame, and (2) the credibility, salience, and

consistency of the diagnostic, prognostic and motivational elements of the frame influence others’

willingness to believe and respond to a communication. Our case study explored user motivation and

participation in an upgrade of SAP software in an organization where no upgrade had been performed in

the past three years. We discovered that: (1) a negatively valenced communicative frame characterizing

an external party as a threat is most likely to motivate users, and (2) framing the support upgrade

simultaneously affected user motivation and reinforced the position of the IT support group.

Published by Elsevier B.V.

Contents lists available at SciVerse ScienceDirect

Information & Management

jo u rn al h om ep ag e: ww w.els evier .c o m/lo c ate / im

1. Introduction

Today, organizations buy or lease customized packagedsoftware in lieu of writing their own to improve the productivityof their business operations. Packaged software is developed by avendor and sold or leased to many users. Customized packaged

software allows adopting users to modify the base software to fittheir specific needs within certain limits imposed by the vendor. Incontrast, traditional packaged software allows little or no customi-zation; the user simply installs the software whose features havebeen completely determined by the vendor.

This distinction between these packaged software, of course,depends on the organizational context. For example, in someorganizations, Microsoft Office is treated as traditional packagedsoftware because the IT department only checks that new versionsof desktop tools work with the existing operating system.However, organizations that employ Excel-based macros considerthe same Office package as customized packaged software, becausethe installation of a new version would only be complete afterexisting macros had been tested and found to be compatible withthe new version of Microsoft Excel. Today, customized packagedsoftware is used in a wide variety of enterprise applications,

* Corresponding author. Fax: +1 210 458 6305.

E-mail address: huoy.minkhoo@utsa.edu (H.M. Khoo).

0378-7206/$ – see front matter . Published by Elsevier B.V.

doi:10.1016/j.im.2011.09.001

including accounting, demand forecasting, MRP, CRM, financialtrading (e.g., Murex), and decision support.

The maintenance of customized packaged software differssubstantially from that of custom software. It requires the initialinstallation of the software, followed by the installation ofupgrades or service packs provided by the vendor as well as themaintenance of custom code. Upgrading customized packagedsoftware is a non-trivial activity, often consuming a year or moreand expending resources equal to 20–30 percent of the cost of theoriginal software. Software upgrades may not work as planned andcan create serious organizational problems by interrupting servicefor a long time. For example, the Lockheed Martin Corp. had to shutdown its SAP servers for 71 h when upgrading from SAP 4.7 to 6.0,after which they experienced slowdowns in database queries anduser lockout problems. Software upgrade projects are thereforeexpensive, high risk propositions that must be carefully managed.

In many cases, upgrades are made though they have no directbusiness benefits. Indeed, the need to upgrade often arises becausesoftware vendors establish ‘‘sunset’’ dates beyond which they statethat they will not support earlier versions of packaged softwarewithout additional fees: thus firms are forced to upgrade, accrueadditional charges, or maintain the packaged software themselves.Because organizations often do not have access to the packagedsoftware source code, upgrading to a new version becomes themost realistic choice [9]. Such an upgrade can be considered asupport upgrade, which costs substantially more than an upgrade oftraditional package software. With customized packaged software,

H.M. Khoo et al. / Information & Management 48 (2011) 328–335 329

one must test and adapt the library of existing code andinfrastructure to the new version of the software.

Due to changes introduced in a new version of software,support upgrades may involve changes to business processes thatusers perform on a regular basis; thus users may resist theimposition of modifications that they did not request. Userreluctance thus creates an additional project risk.

The objective of our study was therefore to examine this issueand identify effective strategies for eliciting user participation in asupport upgrade. We report here on an in-depth analysis of a singlesupport upgrade case using the concept of communicative framing[3].

2. Related literature

2.1. Packaged software implementation and upgrades

Organizations choose to use packaged rather than customsoftware to reduce development costs, shorten implementationtimes, acquire state-of-the-art best practice, reduce maintenance,and obtain extended functionality [11]. However, softwareproblems may arise after packaged software is installed. Oneongoing challenge in its management is in coping with itsevolution, as vendors often change software capabilities whenintroducing new versions and terminating support for older ones[10]. Although client organizations may influence vendors, theylack direct control over software support. Upgrades initiated byvendors can introduce unwanted additional functionality oreliminate existing desired functions. In addition, upgrades arecostly. An ERP upgrade may cost the implementing firm a millionUS dollars or more.

Academic research on packaged software issues has predomi-nantly focused on initial implementation (e.g. [2]). Of the fewstudies on customized packaged software maintenance, themajority address descriptive research questions instead of buildinga theory to explain how or why organizations cope with packagedsoftware maintenance [13]. The research that has attempted tobuild a theory focused on the decision to upgrade rather than theprocess of upgrading.

To our knowledge, only Nah and Delgado [12] have explicitlystudied the upgrade process by comparing the CSFs for packageimplementation with those for package upgrades. They reportedthat CSFs related to change management, such as user training andcoordination across all affected parties, are important for success.Users are important to an upgrade, because new versions aretypically designed to suit a generic business, thus users must eitheradapt to these generic prescriptions or participate in thecustomization and configuration of the package to fit with theirbusiness processes. Users also create idiosyncratic adaptations andworkarounds to overcome limitations in any customized software.For example, Excel spreadsheets are often employed in place ofdatabase systems to allow users more control over their data. Suchunintended work practices, while critical to actual businesspractice, are often unknown to a maintenance team and/or thecustomized package software vendor.

Furthermore, lack of user participation can create difficulties.For example, users may refuse to relinquish their IT equipment toallow the support upgrade. Similarly, users may complain aboutthe upgrade or otherwise disrupt work on the support upgrade.

2.2. Communicative framing

Communicative framing theory explains how messages can bestructured to motivate other users to action. As such, it makes itpossible to understand the mobilization of users to participate insupport upgrades of customized packaged software.

Communicative frames have been used in the marketingliterature to show how binary characteristics of advertisementsinfluence targeted recipients. For example, researchers havestudied the impact of a promotion window as restrictive orexpansive in its effect on consumer purchasing decisions [5], andthe impact of stating a discount either as a percentage or as a dollarvalue [6].

Our purpose, however, was not to understand how messagescould mobilize purchases, but how they can motivate organiza-tional stakeholders. We therefore turn to communicative framingresearch in the communications and social movement literature.The literature in these fields revolves around two questions: (1)how does positive and negative framing influence effort? and (2)how does social context influence the construction of communi-cative frames?

2.2.1. Negative/positive framing

Most research on negative and positive framing stems fromprospect theory, the basis of the work that earned Kahneman the2002 Nobel Prize in Economics. According to prospect theory, aproject situation framed in terms of its likelihood of loss may elicita different effort from project members than one framed in termsof potential gain: the effort is likely to be asymmetric. Also,Kahneman and Tversky had come to prospect theory by noting thatpeople preferred a focus on a positive outcome (you have a 50%probability of surviving) to the equivalent negative statement (youhave a 50% chance of dying). Some research has suggested,however, that context determines whether a frame should beconstructed positively or negatively; thus there may be no need toframe a customized package support upgrade positively, e.g., bysaying it results in a better life: it could be re-framed as necessaryto avoid a worse life. When maintenance is performed well, usersmay fail to see that the effort improves their lives.

IS projects create disruptions to the social order of a firm, andthis can create anxiety that can cause performance to fall [7].Therefore, the change management literature would suggest thatpositively framed messages should be employed during acustomized package support upgrade.

2.2.2. Communication context

Other characteristics of a communicative frame can influenceits effectiveness. However, most research has found that theseother characteristics depend strongly on the specific communica-tive frame being examined (e.g. [8]). For example, for discounts, thequestion arises: Should the discount be stated in absolutemonetary terms or as a percentage of the cost? Such questionsare irrelevant to the framing of restrictions to civil liberties duringa war [4]. More relevant questions include whether suchrestrictions should be framed in an objective or advocacy tone[1], or whether heroic images should be used.

Research on communicative frames has identified two generalfactors that influence a frame’s resonance with recipients: framecredibility and frame salience. Frame credibility is the recipient’swillingness to accept the message as true. Frame salience is theimportance of the message to its recipients.

Frame credibility can be considered from two dimensions:empirical credibility (the believability of evidence presented) andframe consistency (the continuity of the message as it is createdand disseminated) across three core framing tasks: diagnostic,prognostic, and motivational. Diagnostic frames describe aparticular problem and suggest blame or causality resulting fromit. Prognostic frames propose a solution to the problem and alsoidentify strategies that can be used to solve it. Motivational framesdevelop rationales for action that encourage groups of stake-holders to act. Effective frames must be consistent across theirdiagnostic, prognostic, and motivational forms.

Diagnostic

Prognostic

Motivational

IT Group'sFraming

Users' Receptionof Frame Project

Outcome

Credibility-Empirical Credibility-Frame ConsistencyandSalience User Participation

Fig. 1. Process model of support upgrade.

Table 1Summary of interviews.

Stakeholders # Respondents # Interviews

SAP Support Group 10 18

Users 14 17

Total 24 35

H.M. Khoo et al. / Information & Management 48 (2011) 328–335330

An important element of frame consistency is its valence. Adiagnostic frame that is positively valenced, but which leads to anegatively valenced prognostic frame would be inconsistent.However, two positively valenced frames can remain inconsistentif other elements of the frames do not align.

Frame salience refers to the importance of the message to itsrecipients. A frame can be credible, but if the message is viewed asunimportant, the recipients will not be motivated to act.

The concepts of frame credibility and frame salience areelements of our process model (Fig. 1). IT project leaders arepositioned in our model as framers of messages that seek toencourage user participation in a customized package supportupgrade. Message framing involves diagnostic, prognostic, andmotivational frames disseminated to users, and the credibility andsalience of the frame influences user participation. A frame that isneither credible nor salient is less likely to be received well, thusaffecting user participation and, ultimately, the project outcome.

3. Method

Given the lack of research on customized packaged softwareupgrades, we employed an exploratory, inductive, qualitative casestudy approach. This approach allowed us to collect in-depth dataon a project in its natural environment. Our approach wasinterpretive because we relied on subjective reports fromorganizational stakeholders and used communicative framingtheory when we interpreted the outcomes. It was inductive, as theresults and theory were developed during analysis. In line withinterpretive case studies, our aim was to generate novel localinsights into the research phenomenon rather than to generate ortest general theoretical propositions.

3.1. Case selection

We selected a software package upgrade in a large organizationthat used many software packages and that had conducted manyupgrade projects. The first author requested and obtained access tothe Consumer Market (CM) division of a Fortune 500 companyheadquartered in the southeastern United States. This companyhad two large divisions, the CM and an industrial market division.CM reported over $3 billion in sales revenue in 2000 and had over120 manufacturing facilities in North America.

An SAP enterprise system package upgrade was selected frommany possible projects for two reasons:

1. It was large and complex. Other projects involved very smallproject teams consisting of a few IT personnel and users.

2. It had recently been completed, and so the experience was freshin the minds of the participants.

The SAP case involved an upgrade from SAP 3.0f to 4.6c, a leap of13 versions. The upgrade took six months beginning in July 2000

and ending in December 2000, 15 months before our study began.For the purpose of our study, there were two principal stakeholdergroups involved in the project: the SAP Support Group (approxi-mately 90 people including 65 full-time employees and contrac-tors directly responsible for maintaining the SAP software), as wellas affiliated IT staff in charge of maintaining the UNIX and databaseinfrastructure); and users, who belonged not only to the companyheadquarters, but also to 44 of the 120 field offices andmanufacturing facilities in North America.

3.2. Data collection

The head of the SAP upgrade team was the field researcher’sprimary point of contact. This person introduced the researcher totwo other department heads responsible for the Database andUNIX Operating Systems groups. The researcher set up an initialinterview with each department head. Snowball samplingtechniques were then applied to obtain additional interviewsubjects.

Data were collected in two rounds. The first began in April 2002and lasted nine months. The second was conducted in May 2004 tofollow up on contextual details. Data collection in this roundincluded only interviews; this round took only two days.

Although the second round occurred a substantial time afterinitial data collection, this did not compromise our results becausethe second round focused on the collection of factual data onsituational nuances. Most of our questions dealt with divisions oflabor. For example, we needed to ascertain who was initiallytasked with testing the accuracy of a particular SAP report. Also, noother significant software upgrade occurred during the timebetween our rounds, so interviewees could not confuse theirresponses due to incidents occurring in other upgrades. Finally,this was a significant upgrade, and thus certain details were etchedin respondents’ minds.

Interview techniques followed accepted case study bestpractice [14]. A total of 35 interviews were conducted with 24respondents. Table 1 summarizes the number of respondents andtheir classification as either IT staff or users. All interviews lastedbetween 60 and 90 min except for a few follow-up interviews thatwere shorter.

Initial interviews began with questions about the subjects’backgrounds: their positions and roles, and how they perceivedtheir roles to fit the overall organizational context. We then asked

H.M. Khoo et al. / Information & Management 48 (2011) 328–335 331

general questions about subjects’ perceptions of the organization’scustomized package upgrade policies: including their opinion onthe frequency of upgrade, the decision processes involved inupgrading, who was involved in such decisions, and whatinfluenced upgrade decisions.

We then moved to specific questions about the upgrade project,such as the factors influencing the upgrade project, who influenceddecision making, and how communication about decisions wasdisseminated.

Follow up interviews were scheduled under three circum-stances: (1) when allocated time was insufficient for theinterviewer to ask all questions, (2) when new questions aroseafter reviewing transcripts or speaking with other subjects, and (3)when other respondents provided information that contradictedthat from a respondent. With permission from respondents, allinterviews were audio-recorded and transcribed.

In addition to interviews, the field researcher attended onemeeting between the SAP group and its affiliates, participated in auser training session, and sat with CM employees to observe orderentry using SAP. The field researcher also toured the data centerand sat in technology governance committee meetings to observediscussions on upcoming projects. In addition, the field researchercollected documents from interview respondents, the organiza-tion’s intranet, and the Internet. These included SAP screen shots,the SAP training manual, CM’s organizational chart, and annualsales figures.

3.3. Data analysis

Our analysis involved a differentiated role strategy. The firstauthor assumed an embedded role by acculturating within theorganization and associating with the individuals who participatedin the project. This embedded researcher developed a rich,descriptive case study report. Other authors played the role ofdispassionate outsiders who reviewed the case to suggest conceptsand theories of relevance to researchers. A discussion ensuedbetween authors as theories were identified, adopted, or discarded.Authors played devil’s advocate roles to each others’ analyses andoffered complementary insights. The idea of applying communi-cative framing as an analytical lens emerged from this discourse.While communicative framing was not the initial idea that guideddata collection, it was nevertheless possible to apply it, as one ofthe principal thrusts of data analysis during the project.

Within our process model, seven constructs were important:(1) diagnostic, (2) prognostic, and (3) motivational frames, the (4)credibility and (5) salience of the frames as a whole, (6) userreception of the frames, and (7) user behavior resulting fromframing. We reviewed the transcripts and coded speech fragmentsand decided which of them belonged to which of the abovecategories; of course, a speech fragment could be categorized assupporting more than one construct.

A speech fragment was considered supportive of:

� A diagnostic frame if it identified a problem, made an attribution,or assigned blame.� A prognostic frame if it outlined a solution to a problem or

articulated an upgrade strategy. Generally, speech fragmentsabout prognostic frames included a diagnostic frame component.� A motivational frame if it identified a message from the IT project

group to be articulated to a specific stakeholder group or if itinvolved asking a stakeholder group to do something.

Credibility and salience were not directly extracted from speechfragments. Instead, they were assessed via triangulation. To assessthem, we examined subjects’ accounts to determine whether sub

dimensions of credibility and salience were present. For credibility,we compared interviews to determine if there was a fit betweenframing and events. Also, we assessed the congruency of beliefs,claims, and actions between IT and users. Similarly, to assessrelative salience, we judged whether the framed message wasessential to the targeted population, and resonated with targets’personal experiences. User behavior resulting from framing wasobtained from users’ responses to questions about tasks theyperformed during the upgrade.

4. Results

The CM division elected to implement SAP in 1996. A decisionwas made to implement a custom version of SAP because thestandard version did not support many of the organization’sspecific, critical business processes. Most customizations weremade at user exits.

We seldom change the standard SAP. . .. But they have places called

user exits where you can put in your functionality so it will go out to

your stuff and then back into theirs. If you use it properly, you can

customize it to fit your business models. So, we do use that a lot.

. . .Ideally, we won’t ever use them but we can’t help it. In our

industry, it’s so specific that we have to use them. – SAP SupportGroup

However, there were some business processes that could onlybe implemented by modifying the core SAP code. As a result, theversion of SAP implemented incorporated customized code thathandled the organization’s processes. Upgrading SAP thus carriedconsiderable risks for users, as functionality custom-built into theexisting SAP system could be overwritten and lost when a new SAPsystem was installed.

The negative thing about upgrading is that when you upgrade, SAP

creates different codes potentially. And guess what? That core

modification no longer is in that program. . . .That means that we

have to put it in all over again and we have to hope that it works.

It’s very, very difficult to upgrade systems when you’ve added

additional code that’s not part of standard SAP programming. –SAP Support Group

In addition to lost functionality, an SAP upgrade couldpotentially make data difficult to retrieve, as the underlyingdatabase was restructured in the upgrade.

Unfortunately SAP is not really consistent in every area as to how

they do it. In some cases, you can still get to your old data over

here. . .and in some instances they will migrate that old data to the

new space. So then you have to figure out how to get to all of it. So

they aren’t necessarily consistent across the entire board. – SAPSupport Group

Thus, users understandably generally resisted upgrading SAP.

But typically, the business doesn’t like upgrade. It’s. . .IT saying, we

need to upgrade. We need to be more current. Because they don’t

want to go through it. They don’t want to be re-trained. They don’t

want to have the downtime. – SAP Support Group.

The IT Policy of the SAP Support Group was to upgrade SAPevery one and a half years. However, due to business changesthat demanded the SAP Support Group’s attention, the supportgroup had not conducted any upgrades for more than threeyears.

H.M. Khoo et al. / Information & Management 48 (2011) 328–335332

4.1. Framing by project team

4.1.1. Diagnostic frame

The decision to upgrade to SAP 4.6c from 3.0f was made by theSAP Support Group for three reasons:

1. The company’s IT policy was that all package software must besupported by the vendor.

2. SAP had announced that version 3.0f was no longer going to besupported. This would have a potentially deleterious effect onorganizational performance, because SAP would no longer fixknown bugs in version 3.0 software.

3. The maintenance team would be better able to handle userrequests for new features, because many user requestedfeatures were already incorporated into SAP 4.6c.

Overall, the anticipated benefits of upgrading SAP were over-whelming in comparison to not doing so. The diagnostic frame thusleft little room for argument: the planned upgrade was necessary.

4.1.2. Prognostic frame

Given existing problems with SAP 3.0f, the SAP support divisionreviewed the situation and decided that it was important toupgrade to 4.6c. However, the SAP Support Group realized thatconvincing users to upgrade would be difficult. They would have to‘‘sell’’ the upgrade to users.

There’s a lot of salesmanship that goes into it. You can’t just walk

into a big meeting once and say we’re going to upgrade, because

they won’t go for it. – SAP Support Group

The SAP Support Group decided that the appropriate way to ‘‘sell’’this support upgrade was the need to maintain vendor support.

The key reason we upgraded then was that we were losing support

from SAP on our current release. They only support it for 3

years. . .you can receive support from them but you have to pay for

it and they also treat you with a lower priority than customers who

have a support agreement. – SAP Support Group

This message about upgrading to maintain SAP support wasconsistent with the decision to withhold new functionality.Computer processes in the upgrade would be kept as close aspossible to the existing computer processes of SAP 3.0f.

In an upgrade, we never introduce new functionality. . . .But our

approach on upgrades [is to] get our current functionality working.

– SAP Support Group

The third user problem: that new SAP features should beincorporated to address user requests, was not the focus of theupgrade project but was to be addressed in subsequent projects.

This was what we called a technical-only upgrade. In other words,

we were not updating the software and saying: here are all of the

areas where SAP is giving you improved functionality and let’s take

advantage of it all right now. – SAP Support Group

4.1.3. Motivational frame

Because SAP upgrades usually involved downtime and training,business users normally preferred to defer an upgrade as long aspossible.

They are so comfortable with the way it works now, and you have

to go train three or four thousand people every time you change it,

they just want no part of it. – SAP Support Group

To motivate user participation, IT informed users that the entireorganization had no choice in the upgrade. The vendor was forcingthe organization to upgrade, and the entire organization would justhave to accept this.

We typically sell this project as a ‘‘must have,’’ that is, we need

to stay current, we need to get support from the vendor. We can

improve the business although we can’t always really prove it.

So, we don’t typically have to do return on investment and there

is no capital available, usually. It’s just our time. – SAP SupportGroup

However, the SAP Support Group promised to do its best toensure that there would be minimal disruption to user activities;for example, by trying to avoid changes in SAP functionality duringthe upgrade.

While the upgrade problem was framed in terms of a lack ofchoice, the organization had a real choice to refuse to upgrade. Theorganization had not upgraded for three years. The version of SAPthat was being employed was stable, and serious bugs had long agobeen identified and either resolved or worked around.

But honestly we have been live on that for several years so we

didn’t need a great deal of help from SAP at that point. . . .The longer

you’ve been on a release, your reliance on the vendor becomes less

so your incentive for an upgrade actually becomes less. . .We have

the choice too. – SAP Support Group

Thus, while the impact of failing to upgrade SAP would havebeen serious to the organization, it would not have beencataclysmic. Not upgrading SAP was thus a viable (if expensive)choice for the organization, and the framing of communications toconvince users to accept the upgrade was necessary.

4.2. Credibility, salience, and users’ reception of frame

4.2.1. Credibility

The message sent by the SAP Support Group was consistent andcredible. The problem identified in the diagnostic frame was thesame as that conveyed to users, who completely accepted the ideathat they had no choice but to upgrade due to the vendor’s decisionto cease support.

I just know we needed to; SAP was not going to support the upgrade

that we had, so that was the big part of it. So they came to us. . .and

said, by the end of 2003, we are not going to support, you know, in

case if there is a major breakdown,. . .they won’t support us. – User

In addition, the behavior of the SAP upgrade team wasconsistent with this message. The team’s actions demonstratedthat the upgrade would avoid new functionality and preserveexisting functionality. For example, to ensure that SAP 4.6c wasworking, the SAP upgrade group employed the same scripts asthose created to test the 3.0f version.

We used the existing test plan. . .as [if]to say ‘‘here’s what I’m trying

to do and here are my expected results’’. . . .we wanted the same

results they were already getting. – SAP Support Group

Of course, some of the SAP software had changed beyond thecontrol of the SAP Support Group.

The screen looks different. . .Here’s how you change your password

now. But. . .we did not change the functionality on them. – SAPSupport Group

Thus, the diagnostic, prognostic, and motivational framesaligned with one another, enhancing the credibility of the SAP

H.M. Khoo et al. / Information & Management 48 (2011) 328–335 333

Support Group’s framing of the upgrade strategy. Users felt that themessage was credible because their observations were consistentwith the message of the SAP Support Group.

Did we get any more functionality? No, that wasn’t the intent. The

upgrade was primarily technical. – User

4.2.2. Salience

Users also felt the upgrade was important to them and that afailure of the upgrade would impact the entire organization andtherefore affect their future.

It really impacted not only us but our customers. . . .you can’t just

look internally about what’s going to happen to you, you have to

decide, what and how you are going to communicate to the

customer. And how you are going to mitigate that risk? . . .Can you

ship on time? Can you invoice on time? Can you take their orders?

Are you responsive? – User

4.2.3. Users’ reception of frame

Because users felt that the upgrade was credible and salient,their response to the project was generally positive. Users felt andresponded to its urgency.

We needed to get to 4.6 promptly knowing SAP was no longer going

to support us on 3.1. – User

Users also believed that upgrading to 4.6 was the only solutionto the problem of lack of future support. They did not resist theproject: they felt they had no other option.

For me, as far as the upgrade. . .we have no choices. . . .it’s up to the

business. – User

4.3. User participation

Users’ positive reception of the frame translated into useraction. Users followed the SAP upgrade team’s instructions and didwhat they were told to do. For example, users were receptive totraining sessions and attended them to learn how their interfacehad changed. Training required time commitment – as much as40 h worth, but users were willing to participate in order to beready for the upgrade.

There was extensive training for us to understand how to use the

new system, and how to know where to look for the different fields,

and what fields we had to populate, and just in general how to

process an order. – User

Users also committed time to test the new version whenrequested to do so.

Usually. . .they. . .ask us to help run some test, to make sure,

because we are the one who is using it. – User

Beyond the training and testing requested by the SAP SupportGroup, users proactively performed other tasks that helpedidentify issues with the upgrade. They committed personal timeto practice on the new SAP interface.

We make sure we have a lab where they can [check] the system.. . .

– User

Even individuals who only used SAP sporadically helped makesure the upgrade went smoothly.

For the most part I had a good team. So I really didn’t have to sit

down and [demonstrate the upgrade]. I did do some. . .because I

wanted to be familiar with the system. . .. But on any given day

things would come up and I might sit there and key a couple of

orders in. . .to help them out. – User

Users also formed communities to help each other learn thenew version of SAP. Those who mastered the SAP material betterhelped those who had difficulties.

We. . .established the mentor group, they were quick to learn and

able to help when we weren’t right. – User

Users also spent time gathering information about specificareas where they needed help from the upgrade team. Both usersand upgrade team members spent time interacting directly withworkers.

You become that floater on the floor, helping people when they’re

stuck. . . .That’s why we picked 6 people and that was basically one

person from every team so that each team has someone sitting near

them. . .. If you’re sitting with them you can hear them slamming

the keyboard and go over there and say ‘‘excuse me, what are you

trying to do? Let me help you’’. – User

4.4. Project outcome

The director of the SAP Support Group and the users agreed thatthe upgrade project was completed successfully. There were nomajor problems experienced by users, and no drop in productivity.Individuals involved in the project attributed the successfulupgrade to all participants’ effort.

It’s because we spent so much time on the upgrade. – SAP SupportGroup

Consistent with the SAP Support Group’s framing of theupgrade, neither the project team nor users detected any changein productivity. The framing of the upgrade focused on maintainingthe status quo.

No, there wasn’t any impact on the productivity negatively or

positively. . . .it was just a different way of doing things. – User

Despite the apparent neutrality of this result, the upgrade wassuccessful because of the consistent set of diagnostic, prognostic,and motivational communicative frames used to ‘‘sell’’ the upgradeproject to the users.

5. Discussion

Our research illuminates an unusual problem – how to mobilizeusers to participate in a support upgrade project that providesminimal benefits. Normally, one might expect users to bedisinterested, or resistant to supporting maintenance with nobenefits. Yet, support upgrades are bound to become morefrequent and potentially disruptive to users in the future. Ouranalysis of this problem suggests that IT groups can gain userparticipation and commitment through their framing of theupgrade.

The Support Group’s message centered on overcoming theexternal threat created by SAP’s decision to ‘‘sunset’’ the legacyversion of SAP software and downplayed the new featuresavailable in the new version of SAP. In response, users not onlycooperated with the support upgrade, but also held back on newfeature requests to reduce the complexity of the upgrade. As a

Table 2Case summary.

Research model constructs Case evidence

Diagnostic frame

Identification of concern (1) Adhere to corporate guidelines.

(2) Maintain support with SAP.

(3) Add new features requested by users.

Highlighting of concern Issue of SAP sunset highlighted.

Attribution of blame/causality SAP refuses to support the old version.

Prognostic frame Preserve existing operations. New features to be explored later. Primarily, maintain existing features.

Users and maintenance team must cooperate.

Motivational frame Maintain organizational operations during upgrade. Everyone in the same situation.

SAP Support Group to make sure that the upgrade is minimally disruptive to operations.

Credibility Users believed the upgrade was necessary to obtain continued support from SAP.

Salience Users believed the upgrade was important to them, as a system failure could cause disruption

of business operations.

User reception Aligned with communicative frames.

Reaction Users cooperated with IT to reduce impact of change.

Users formed communities to facilitate learning.

Project outcome Successful upgrade with little benefit to users.

H.M. Khoo et al. / Information & Management 48 (2011) 328–335334

consequence, although successful, respondents agreed it did notcontribute to organizational productivity.

Table 2 shows how each element of our model is supported bycase evidence. The diagnostic, prognostic. and motivational frameswere credible and salient, leading users to accept them and to aligntheir efforts to implement the upgrade.

5.1. The negatively valenced external threat

Our case illustrates one of the more important insights fromcommunicative framing theory: the benefit of a negativelyvalenced message to galvanize users to engage in a supportupgrade project. The IT support group recognized framing wasnecessary for this project; its members consistently used terms like‘‘sell’’ or ‘‘salesman’’, suggesting that they recognized that usershad a choice to opt out of the upgrade.

Their framing was so successful that users accepted the sunsetdate argument and agreed to upgrade. Because SAP was runningcritical business operations, users not only proactively facilitatedproject success, but also feared project failure. The case thusillustrated how worry of the effect of an external party canfacilitate cooperation across diverse stakeholder groups. Theportrayal of SAP as having the power to end support for SAP3.0f was instrumental in aligning users and the SAP Support Groupagainst a common threat.

5.2. Framing’s effect on IT support group behavior

As predicted by communicative framing theory, a credible andsalient frame influences the recipient’s behavior. Users respondedpositively to the SAP Support Group’s frame and contributedproactively to help provide project success. Consistency ofmessages convinced users that the SAP Support Group was ontheir side, thus increasing credibility and reducing project risk.Notably, the SAP Support Group elected to focus on the goal ofmaintaining SAP support and chose not to implement new featuresinto the upgrade. Inconsistencies between the old and newversions of the software were addressed by ensuring that the newversion behaved just like the old one.

These actions provide strong evidence that the SAP SupportGroup’s framing influenced their own conduct during the upgrade:their efforts to gain user participation were sustained by theconsistency of their behavior. Change agents may encounter risks iftheir use of frames is viewed as manipulative or exploitive. Clearly,IT staff must proceed with a strong purpose, but their interests aremore likely to be served if their own actions are seen as consistentwith their framing of the problem and its solution.

5.3. Anxiety and change management

Our research also identified a practical implication formanaging organizational change. Change agents may introducespecial measures to reduce anxiety so that workers can contributeto the change. Tvedt et al. [15] stressed the need for managers to beavailable, to embrace diversity, and to listen to the concerns ofaffected workers. They argued that managers should clarifychanges in job roles quickly and accept resistance.

However, the SAP Support Group used anxiety to make theircommunicative frames salient. Users were told that a failure wouldpotentially disrupt work, thereby compromising company opera-tions and users’ jobs. Furthermore, the Support Group suggestedthat increased participation would increase the probability ofimplementation success, thereby reducing user anxiety.

Instead of trying to convince workers that they will remainvaluable and retain status in an organization after an organiza-tional change, our case suggested that it would be more effective toconvince workers that they would lose value and status if theorganizational change failed. With no reason to contribute effort tomaintain the status quo, participants need to become anxious thattheir failure to participate may lower their value and status.

6. Conclusion

We addressed the question of how one should encourageparticipation by users in a customized package support upgradeproject where packaged software is replaced with a new version ofthe same software from the same vendor. The upgrade is primarilyintended to benefit the IT group supporting the upgrade. The issueof gaining voluntary user participation in such projects is critical,since such projects bring little or no obvious benefit to their users.Through our analysis of a case study, we obtained insights into howusers can be convinced to participate.

From a theory perspective, we discovered that one effectiveframing for customized package support upgrades portrays usersand the maintenance team as working together to address anexternal threat. A successful framing encouraged users to respondto the threat to maintain the status quo. We also discovered that theframing of the message to users motivated the IT group to actconsistently with its own frames.

Although our research provides an initial investigation into thephenomenon of support upgrades, the empirical support for ourfindings were limited to a single upgrade case.

Furthermore, our data were collected during the mid 2000s atwhich time the support upgrade decision was principally restrictedto either upgrading the customized packaged software, or leaving

H.M. Khoo et al. / Information & Management 48 (2011) 328–335 335

the software at the same version number. However, the risingpopularity of customized packaged software has meant that otherplayers have entered the market. A third option, to outsourcemaintenance of legacy customized packages to support vendors, isnow viable.

Communicative framing can help IS researchers and ITpractitioners understand the communication process involved inmobilizing stakeholders to participate in a project. To date, most ISresearch has acknowledged the importance of user participation inproject success. Studies have not focused on the issue of motivatinguser participation. The use of communicative framing can help ISresearchers and practitioners to understand how to mobilize usersto participate in IT projects that promise no tangible benefits tothem.

Acknowledgements

We would like to thank the participants at the IMARC ResearchWorkshop, Nanyang Technological University, Singapore, andmembers of the SIGITPM Workshop 2009 for helpful commentsand insights. We would like to especially thank Wee Kiat Lim andAdrian Yeow for conversations about Communicative Framingtheory. We would also like to thank Dr. Edgar H. Sibley and threeanonymous reviewers for their valuable feedback and suggestions.

References

[1] S. Aday, The framesetting effects of news: an experimental test of advocacy versusobjectivist frames, Journalism and Mass Communication Quarterly 83, 2006, pp.767–784.

[2] D. Aloini, R. Dulmin, V. Mininno, Risk management in ERP project introduction:review of the literature, Information and Management 44, 2007, pp. 547–567.

[3] R.D. Benford, D.A. Snow, Framing processes and social movements: an overviewand assessment, Annual Review of Sociology 26, 2000, pp. 611–639.

[4] M.P. Boyle, M. Schmierbach, C.L. Armstrong, J. Cho, M. McCluskey, D.M. McLeod,D.V. Shah, Expressive responses to news stories about extremist groups: a framingexperiment, Journal of Communication 56, 2006, pp. 271–288.

[5] A. Cheema, V.M. Patrick, Anytime versus only: mind-sets moderate the effect ofexpansive versus restrictive frames on promotion evaluation, Journal of Market-ing Research 45, 2008, pp. 462–472.

[6] D. DelVecchio, H.S. Krishnan, D.C. Smith, Cents or percent? The effects of promo-tion framing on price expectations and choice, Journal of Marketing 71, 2007, pp.158–170.

[7] M. Fugate, A.J. Kinicki, G.E. Prussia, Employee coping with organizational change:an examination of alternative theoretical perspectives and models, PersonnelPsychology 61, 2008, pp. 1–36.

[8] L.H. Hoffman, M.D. Slater, Evaluating public discourse in newspaper opinionarticles: values-framing and integrative complexity in substance and healthpolicy issues, Journalism and Mass Communication Quarterly 84, 2007, pp.58–74.

[9] H.M. Khoo, D. Robey, Deciding to upgrade package software: a comparative casestudy of motives, contingencies and dependencies, European Journal of Informa-tion Systems 16, 2007, pp. 555–567.

[10] R.J. Kohl, Requirements engineering changes for COTS-intensive systems, IEEESoftware 22, 2005, pp. 63–64.

[11] B. Light, S. Sawyer, Locating packaged software in information systems research,European Journal of Information Systems 16, 2007, pp. 527–530.

[12] F. Nah, S. Delgado, Critical success factors for enterprise resource planningimplementation and upgrade, Journal of Computer Information Systems 46,2006, pp. 99–113.

[13] F.F.-H. Nah, S. Faja, T. Cata, Characteristics of ERP software maintenance: amultiple case study, Journal of Software Maintenance and Evolution: Researchand Practice 13, 2001, pp. 399–414.

[14] C. Standing, I. Sims, P. Love, IT non-conformity in institutional environments: e-marketplace adoption in the government sector, Information and Management46, 2009, pp. 138–149.

[15] S.D. Tvedt, P.Ø. Saksvik, K. Nytrø, Does change process healthiness reduce thenegative effects of organizational change on the psychosocial work environment?Work and Stress 23, 2009, pp. 80–98.

Huoy Min Khoo is a senior lecturer in the Department

of Information Systems and Technology Management

at the University of Texas, San Antonio. She received her

doctoral degree in Business Administration from

Georgia State University. Her research interests include

understanding technological impacts and coping strat-

egies from the perspectives of stakeholders, decision-

making in information systems, and organizational and

technological issues related to Net-enabled organiza-

tions. She has published in European Journal of

Information Systems, Journal of Information Technolo-

gy, Journal of Electronic Commerce, IEEE Transactions

on Professional Communication, and Communications

of the AIS.

Cecil Eng Huang Chua is an associate professor at the

University of Auckland. He received a PhD in Informa-

tion Systems from Georgia State University, a Masters

of Business by Research from Nanyang Technological

University and both a Bachelor of Business Adminis-

tration in Computer Information Systems and Econom-

ics and a Masters Certificate in Telecommunications

Management from the University of Miami. Cecil’s main

research interest is in how technological artifacts shape

and are shaped by human governance and social

structures. Cecil has published widely in such journals

as Decision Support Systems, Journal of the AIS, MIS

Quarterly and the VLDB Journal.

Daniel Robey is an Emeritus Professor of Information

Systems at Georgia State University. He received his

Doctorate in Business from Kent State University in

Ohio, USA. His research interests include empirical

examinations of the effects of a wide range of

technologies on organizational structure and work

practices. It also includes the development of theoreti-

cal approaches to explaining the consequences of

information technology in organizations. His work

has appeared in many journals in information systems

and management. He received the 2009 LEO award

bestowed by the Association for Information Systems

and the International Conference on Information

Systems for a lifetime of distinguished contribution.

He also received the Lifetime Service Award from the

Organizational Communication and Information Systems Division of the Academy

of Management in 2010. He is Editor-in-Chief of Information and Organization.