Importance-Performance Analysis Based SWOT Analysis

strengths and opportunities for achieving or maintaining competitive advantage.Quadrant 2 contains the attributes that are perceived as low importance tocustomers, but the organisation seems to provide high levels of performance. In this case, the organisation should reallocate resources committed to attributes in this quadrant to other quadrants in need of improvedperformance.Quadrant 3 contains the attributes with low importance and low performancewhich are referred to as the minor weaknesses. Thus attributes in thisquadrant do not require a great deal of priority for improvementQuadrant 4 contains attributes that are perceived to be very important tocustomers but performance levels are fairly low. These attributes arereferred to as the major weaknesses that require immediate attention forimprovement.Generally, data regarding customer perceptions toward a product or servicegathered via customer satisfaction surveys are examined for constructing anIPA matrix in which the main task is measuring attributes’ importance andattributes’ performance. Typically, method for measuring an attributes’ performance is well-established by using direct rating from customers survey inwhich the customers are asked to rate the performance of the attribute rangingfrom “very dissatisfied” to “very satisfied” in a 5-point or 7-point Likert scale.Whereas attributes’ importance can be measured either on a rating scale (selfstated importance) or estimated on the basis of performance (implicitly derivedimportance).The customers’ self-stated approach asks respondents to rate the importance of each product or service’s attribute and calculate attributes’ importancebased on customer preference. Although this is a commonly used approach, thismethod has some limitations. Firstly, adding questions for asking customers torate importance increases the survey length which might affect the response7

rates (Garver, 2003). Secondly, self-stated importance tends to have low discrimination power as customers tend to consider that all attributes are veryimportant (Gustafsson & Johnson, 2004). In addition, many researches arguedthat the customers’ self-stated importance is not an adequate method to measure importance (Matzler et al., 2003; Deng et al., 2008b) since this approachdepends on two assumptions1 that are erroneous in the real world.Consequently, the statistically inferred importance is introduced as a methodto implicitly derive attributes’ importance based on the relationships betweenattributes’ performances and overall performance measures such as overall satisfaction (Garver, 2003; Tontini & Silveira, 2007; Pezeshki et al., 2009). Thecommonly used statistical methods for deriving importance measures are multiple regression analysis (MLR) (Matzler & Sauerwein, 2002; Pezeshki et al.,2009; Ho et al., 2012) and partial correlation (Matzler et al., 2003; Deng et al.,2008b). Recently, Back-Propagation Neural Network (BPNN) has become analternative method for implicitly deriving importance (Deng et al., 2008a), dueto its ability to detect non-linear relationships between attributes’ performancesand overall satisfaction.3. IPA based SWOT frameworkThe main idea of this work is to use IPA to analyse survey data of theorganisation and its competitors, then the organisation’s SWOT factor is derivedfrom the IPA matrix as shown in Figure 2. The IPA based SWOT analysiscomprises of four steps:Step 1: Undertake a customer satisfaction survey. First and foremost, attributes of an organisation’s product or service are identified based on athorough literature review in each application area or by interviews (Martilla & James, 1977; Skok et al., 2001; Levenburg & Magal, 2005). Then, a1 TraditionalIPA has two assumptions which are (1) attributes’ importance is independentof attributes’ performance (2) there is a linear relationship between attributes’ performanceand overall performance (Matzler et al., 2004).8

Figure 2: Proposed framework for applying IPA to SWOT analysissurvey is developed regarding the identified attributes of an organisation’sproduct or service.Generally, a survey that is suitable for applying IPA consists of an assessment of respondents’ satisfaction for an organisation’s product or servicewhich is measured by a Likert scale with either five or seven levels (Lai &Hitchcock, 2015). In addition, the survey should contain an assessment ofoverall satisfaction on the Likert scale. Two surveys using the same set ofquestions are required in this study. The first survey focuses on the servicequality of the target organisation, while the second survey concentrateson the service quality of the organisation’s competitor.Step 2: Conduct an IPA on the customer survey. After the surveys are9

administered to customers of a target organisation and customers of anorganisation’s competitor, the customer survey data is processed to compute importance and performance for an individual attribute of an organisation’s product or service. Procedures associated to this step are shownin Figure 3, and discussed below.Figure 3: Steps for conducting IPA of customer survey data Calculate the attributes’ importance. Through this work, attributes’importance is derived from survey responses based on the relationships between attributes’ performance and overall satisfaction insteadof asking customers to rate the importance. Specifically, MLR is chosen to analyse the survey data and compute attribute’ importance asMLR is the best implicitly derived importance method with regardto an empirical comparison2 conducted by the authors.2 Theempirical study that investigates and compares three implicitly derived impor-tance measures including Multiple Linear Regression, Ordinal Logistic Regression and Back-10

MLR is applied to the survey data to create a model for discoveringthe relationship between the attribute performance of the productor services and overall satisfaction which reveals the attributes thatinfluence the overall satisfaction. All attributes’ performance are setas independent variables and overall satisfaction is set as dependentvariable.The regression coefficients obtained from the MLR model can be referred to as implicit importance since the regression coefficient generally indicates how much a one unit increase in the independentvariable results in an increase or decrease in the dependent variablewith all other variables held constant (Nathans et al., 2012). Calculate the attributes’ performance. The performance for eachattribute of organisation’s products or services is computed by averaging performance ratings from all respondents to the questionnairewhich is called “actual performance”. Construct the IPA Matrix. The grand mean of attributes’ importance and grand mean of attributes’ performance are calculated, andthen used to divide the IPA matrix into four quadrants. Finally, allattributes’ importance and attributes’ performance calculated in previous procedures are plotted on the x-axis and y-axis of IPA matrixrespectively.Step 3: Identify strengths and weaknesses through the IPA matrix. Withregard to the IPA matrix produced in Step 2, an organisation’s attributeslocated in Quadrant 1 and Quadrant 2 are identified as strengths as theyare having high performance. Whereas an organisation’s attributes located in Quadrant 3 and Quadrant 4 are identified as weaknesses as theyPropagation Neural Network to one self-stated importance measure using direct-rating scales,across three datasets. The evaluation metrics are predictive validity, discriminating powerand diagnosticity power.11

are having low performance. Based on the same principle, strengths andweaknesses of the organisation’s competitor are identified from the IPAmatrix of competitor.Step 4: Identify opportunities and threats through IPA matrix. By comparing the attributes of an organisation and its competitor that were previously labelled as strength and weakness, opportunities and threats ofthe organisation can be identified based on the ideas of Pai et al. Paiet al. (2013) which stated that “the strengths of competitor become thethreats of the organisation and the weaknesses of competitors can becomethe opportunities of the organisation”.A summary of the identification for all aspects of SWOT and their managerial implication is presented in Table 1 (Matzler et al., 2003; Lee et al.,2010) and described as follows:Strength (S). Attribute is labelled as an organisation’s strength, since itis identified as a strength of an organisation and its competitor. Thismeans both a target organisation and its competitor are performingwell at providing this attribute. The organisation should maintainthe performance of this attribute to ensure that the attribute is notturned into a threat when its performance is lower than that of thecompetitor.Weakness (W). Attribute is labelled as an organisation’s weakness, sinceit is identified as a weakness of an organisation and its competitor.This means both a target organisation and its competitor are notperforming well at providing this attribute. The organisation shouldimprove the performance of this attribute in order to obtain a competitive advantage in the target market over the competitor.Opportunity (O). Attribute is labelled as an organisation’s opportunity, since it is identified as strength of an organisation but it isidentified as a weakness of a competitor. This means the competitor12

is not performing as well as the organisation at providing this attribute, implying that the organisation has a competitive advantageover the competitor. The organisation should maintain or leveragethe performance of this attribute to stay competitive.Threat (T). Attribute is labelled as an organisation’s threat, since it isidentified as a weakness of an organisation but it is identified as astrength of a competitor. This means the organisation is not performing as well as the competitor at providing this attribute, implying that the organisation has a competitive disadvantage to thecompetitor. Hence, an organisation should be aware of it and takean immediate action to improve the performance of this attribute inorder to prevent a potential loss of profit.Table 1: SWOT identification tableStrength - WeaknessSWOT head competitionWOCompetitive advantageSTCompetitive disadvantageWWNeglected opportunitiesWAdditionally, each SWOT factor is weighted as the product of importanceand performance. Specifically, a positive value of performance is assignedto the strength and opportunity factors since these factors have performance higher than or equal to the means of overall performance. On theother hand a negative value of performance is assigned to the weaknessand threat factors since these factors have performance less than means ofoverall performance. This weighting scheme enables factors in each SWOTaspect to be prioritized regarding the magnitude of the weight for whicha factor with high magnitude that has higher priority in maintaining orimproving than the lower one.13

4. Case study of Higher Education Institutions in ThailandTo evaluate the proficiency of the IPA based SWOT analysis in the real-worldsituation, the case study of Higher Education Institutions (HEIs) in Thailandwas conducted. Specifically, two Computer Engineering departments (henceforth, department A and B) of Thailand’s leading university were selected. Department A is referred to as the target organisation while Department B isreferred to as the competitor of the target organisation.Department A, the selected target organisation, was established in 2006 andlocated in the second campus of the selected university. Department A was chosen as a target organisation because this department needed an improvementto be in the Thailand top 10 for computer engineering. However, DepartmentA still has no concrete future plan since the department’s future plan was formulated based on an imprecise SWOT lists in which some ideas were raisedfrom personal attitudes in the brainstorming with no supporting evidence ordocuments.Department B, the selected competitor, was established in 1989 and locatedon the main campus of the selected university in the state capital. Department Bwas chosen as a competitor of Department A because it has academic strengthsand expertise in computer engineering which is evidenced by both national andinternational awards. It is also ranked in the Thailand top three for computerengineering with regard to central university admissions test of Thailand in2014.The datasets used as a case study, an implementation of IPA based SWOTanalysis and result evaluation are described in the following sub-sections.4.1. Data collectionRegarding the first step of an IPA based SWOT framework, input data forconducting IPA was collected in the classroom via a questionnaire responded byundergraduate students of Department A and Department B. A questionnaireusing closed-response questions on a five-point Likert scale (1 Very Dissatisfied to 5 Very Satisfied) was designed and developed. The questionnaire14

comprised of questions asking students about their level of satisfaction (performance) toward six attributes of department and one overall student satisfaction(see Appendix A). These attributes were selected based on a list of attributesdefined in previous studies of student satisfaction (Siskos & Grigoroudis, 2002;Silva & Fernandes, 2011; Grebennikov & Shah, 2013)The questionnaire was piloted among 32 undergraduate volunteers of Department A and Department B, and the reliability of the questionnaire was assessedby Cronbach’s alpha as one of the most frequently used methods for calculatinginternal consistency (Saunders et al., 2009). As a rule of thumb Cronbach’salpha value greater than 0.7 is considered reliable (De Vaus, 2002). The Cronbach’s alpha for each attribute of the department ranged from 0.86 to 0.97 asshown in Table 2. The Cronbach’s alpha of all groups was greater than 0.7 thusit can be concluded that the questionnaire has good internal consistency.The questionnaire was distributed online to undergraduate students (fourconsecutive years) of the two departments during April and May of 2014, inthe second semester of 2014/2015 academic year. A total of 155 and 43 validquestionnaires were collected for analysis from Department A and DepartmentB respectively. Note that the sample size of both department were greater than30 which is the minimum required sample size for data analysis using MLRcomputed by G*Power (see Appendix B). Thus, it can be assured that theMLR was properly conducted which results the reliable regression coefficientsobtained from both dataset.4.2. Implementation of IPA based SWOTAfter the questionnaires which were developed in the first step of IPA basedSWOT framework were administered to students of the two departments, thestudent survey data was then processed following step 2 - step 4 of IPA basedSWOT framework as described in Section 3 in order to generate SWOT ofDepartment A.To create the IPA matrix of the two departments, importance and performance for individual attribute of the department were computed. For each15

Table 2: List of attributes used in the case study and Cronbach’s alphaAttributeQuestion/FactorAcademicTeaching ability of teaching staffPersonalSubject expertise of teaching staff(Teacher1-5)Cronbach’s alpha0.86Friendliness of teaching staffAvailability of teaching staffAdvice and support in learningTeachingLecture materials0.92and Learninge-learning resources(Teaching1-6)Assessments (clarity and timely feedback)Class sizeAccurate and up-to-date unit contentTeaching facilities and classroom conditionAdministration(Admin1-5)Knowledge of rules and procedures of staff0.97Knowledge of the information about courses, exams, activities of staffInterest in solving student’s problems by staffFriendliness of staffAbility of staff to provide services in a timely mannerComputerQuality of computer facilitiesFacilitiesAvailability of computer facilities(CompFac1-4)0.95Availability of internet accessAvailability of printing and photocopying Cultural exchange programs with foreign country0.97Field tripsMoral development activitiesHealth development activitiesInterpersonal skills development activitiesPersonal learning and thinking skills development activitiesSocial volunteer activitiesAdditional Services(AddService1-4)Financial aid for studentsMedical support to studentsDepartment websiteLibrary160.96

attribute, importance was obtained as a regression coefficient by regressing theperformance of questions related to the attribute (independent variables) withthe overall student satisfaction (dependent variables) using SPSS.To make the importance measured as regression coefficient from data ofDepartment A and Department B comparable, all importance was expressedas a percentage contribution of factor, with the negative importance set tozero. This approach has been used in some previous comparative studies ofimportance measures (Gustafsson & Johnson, 2004; Pokryshevskaya & Antipov,2014).For each factor, the performance of Department A and Department B iscalculated as a mean of satisfaction. Finally, a mean of all performance andmean of all importance were calculated. Then, the grand mean of performanceand importance were calculated as the average of these means. Subsequently,the grand means were used as the intersection to create the IPA matrix ofthe Department A and Department B. The use of grand means provides a faircomparison between the IPA matrix of both departments.The IPA result including importance (I), performance (P) and IPA quadrantof Department A and Department B is presented in Table 3 in which the firstcolumn is a short name of factors and their description is provided in Table 2.These short names of factors are also used in the other tables of this paper.Table 3: Importance-Performance of the two departmentsDepartment AFactorIPDepartment er51.9344.23220.0003.7912Continued on next page17

Table 3 – Continued from previous pageDepartment AFactorIPDepartment BIPAIPQuadrantIPAQuadrantTeach

involved with quantitative SWOT analysis and customer oriented SWOT are investigated. An overview of IPA is provided where the main focus is approaches that have been used for measuring attribute importance. 2.1. SWOT analysis SWOT analysis