McCall’s model is one of the most commonly used software quality models (Panovski, 2008). This model provides a framework to assess the software quality through three levels. The highest level consists eleven quality factors that represent the external view of the software (customers’ view), while the middle level provides twenty three quality criteria for the quality factors. Such criteria represent the internal view of the software (developers’ view). Finally, on the lowest level, a set of matrices is provided to measure the quality criteria [8] . The contribution of the McCall Model is assessing the relationships between external quality factors and product quality criteria [9] . However, the disadvantages of this model are the functionality of a software product is not present and not all matrices are objectives, many of them are subjective [10] .

In order to evaluate the quality of software products, Boehm proposed quality model based on the McCall’s model. The proposed model has presented hierarchical structure similar to the McCall’s model [11] . Many advantages are provided by the Boehm’s model, namely taking the utility of a system into account and extending the McCall model by adding characteristics to explain the maintainability factor of software products [9] . However, it does not present an approach to assess its quality characteristics [12] .

The FURPS model was introduced by Robert Grady in 1992 [13] . It’s worth mentioning that, the name of this model comes from five quality characteristics, including Functionality, Usability, Reliability, Performance and Supportability. These quality characteristics have been decomposed into two categories: functional and nonfunctional requirements [13] . The functional requirements defined by inputs and expected outputs (functionality), while nonfunctional requirement composes reliability, performance, usability and supportability. However, the one disadvantage of this model is the software portability has not been considered [14] .

Dromey’s model extended the ISO 9126: 1991 by adding two high-level quality characteristics to introduce a framework for evaluating the quality of software products. Therefore, this model comprehends eight high-level characteristics. Such characteristics are organized into three quality models including requirement quality model, design quality model and implementation quality model [15] . According to Behkamal et al. [10] , the main idea behind Dromey’s model reveals that, formulating a quality model that is broad enough for different systems and assessing the relationships between characteristics and sub-characteristics of software product quality.

The One disadvantage of Dromey’s model is the reliability and maintainability characteristics could not be judged before a product actually implemented [9] .

ISO 9126 is an international standard for software quality evaluation. It was originally presented in 1991; then it had been extended in 2004. The ISO 9126 quality model presents three aspects of software quality which address the internal quality, external quality and quality in use [16] . Therefore, this model evaluates the quality of software in term the external and internal software quality and their connection to quality attributes. In this respect, the model presents such quality attributes as a hierarchical structure of characteristics and sub-characteristics. The highest level composes six characteristics that are further divided into twenty one sub-characteristics of the lowest level. The main advantage of this model is the model could be applied to the quality of any software product [9] .

Because of the ISO/IEC 9126 provides quality characteristics and sub-characteristics that are general and common for evaluating the quality of every type of software products, the recently proposed quality models have been derived from ISO/IEC 9126. For example, Kumar et al. [6] have proposed quality model depend on ISO/IEC 9126 in order to evaluate the quality of aspect-oriented software. Adnan et al. [17] have also proposed model to evaluate the quality of COST systems. Additionally, Bertoa et al. [18] have adapted the ISO/IEC 9126 to establish a component-based systems quality model. Therefore, the ISO/IEC 9126 is adapted in this work to propose the ERP software quality model.

The quality characteristics, functionality, reliability, usability, efficiency, maintainability, and portability have commonly been proposed in most quality models. However, scholars have different opinions while choosing subcharacteristics of these characteristics. This research concentration is on the product quality rather than on quality in use. Therefore, this section defines the various characteristics and its sub-characteristics in term of ERP systems.

The Functionality has been defined by ISO [21] as the capability of the software to provide functions which meet the stated and implied needs of users under specified conditions of usage. In order to evaluate such characteristic, it has been divided into four sub-characteristics, namely accuracy, suitability, interoperability, and security [6] . Adapting the functionality of the ERP systems reveals that the systems software should provide its functions, namely financial process, human resource management, supply chain process, manufacturing process and/ or customer service process as per the requirements when it is used under specific conditions. Therefore, as previously mentioned additional two sub-characteristics have significantly been proposed under this quality characteristic including modularity and compatibility.

The reliability is the capability of the software to maintain its level of performance under stated conditions for a stated period of time. Reliability has three sub-characteristics consist maturity, fault tolerance, and recoverability [9] . In terms of ERP systems, the reliability refers to the capability of the systems to maintain its service provision under specific conditions for a specific period of time. In other words, the probability of the ERP system fails in a problem within a given period of time.

The usability is the capability of the software to be understood learned, used, and attractive by the users, when used under specified conditions. The usability has set of sub-characteristics, including understandability, learn ability, and operability [22] . This characteristic is employed in this study to suggest that the ERP systems should be understood, learned, used and executed under specific conditions. Thus, the complexity has been proposed as additional sub-characteristic under this quality characteristic.

The efficiency refers to the capability of a system to provide performance relative to the amount of the used resources, under stated conditions. To be measured, it has also been divided into three sub-characteristics, namely time behaviour, resource utilization an efficiency compliance [21] . Adapting this characteristic to the ERP systems suggests that the systems should be concerned with the used software and hardware resources when providing the ERP systems’ functions.

The maintainability is the capability of the software to be modified. The maintainability consists five sub-characteristics, including analyzability, changeability, stability, and testability [14] [16] . In this research, any feature or part of the ERP system should be modifiable. As well as identifying a feature or part to be modified, modifying, diagnosing causes of failures, and validating the modified ERP system should not require much effort. Thus, reusability has been proposed as sub-characteristics under this quality characteristic.

Finally, the portability of software refers to the capability of the software to be transferred from one environment to one another [21] . Therefore, the ERP system should be applied using different operating systems; be applied at different organizations or departments; and be applied using a variety of hardware. Similar to the previous quality characteristics, the portability has set of sub-characteristics, namely adaptability, installability, coexistence, and replace ability [9] .

Table 1. ERP Systems Quality Model (ERPSQM).                                                        

In this study, survey on twenty experts has been conducted, in order to assign pair wise relative weights to ERPSQM characteristics and sub-characteristics. Out of these experts, eight professionals are working at software industry and experts of ERP systems development. Remaining twelve are academicians who have whether good knowledge of the ERP systems or are doing their research in such area. Only fifteen of the experts had filled the survey forms and successfully sent it back. The survey form involves seven tables for filling the pair wise relative weight values of ERSQM characteristics and sub-characteristics. First table is for filling the pair wise relative weight values of the model characteristics, namely functionality, reliability, usability, efficiency, maintainability and portability. The remaining tables are for filling the pair wise relative weight values of sub-characteristics of the ERSQM’s characteristics. The means of collected data on pair wise weight values of characteristics and sub-characteristics are filled in the square matrices, to apply the PHP process and then calculate the Eigenvectors and Eigenvalues (ℷ ). For instance, A [aij], in Equation (4) represents the square matrix of six main the ERSQM’s characteristics.

The next step is to determine the eigenvectors. In this respect, there are many ways to calculate the priority vector, which is the Eigenvector. Multiplying the entries of each row in the matrix and then calculate the 6^th roots of a product. The 6th roots are summed and that sum used to normalize the eigenvector values, since the sum of its column is one. Later, the6th root of each column is divided by the sum of the 6th roots column.

micians’ perspectives functionality of the ERP systems is the most important one over other characteristics, followed by reliability, usability, efficiency, maintainability, and the portability of the ERP systems is less important. These results provide more than ranking. In fact, the relative weights are a ratio scale could be divided among such quality characteristics.

The next step is to check the consistency of participants’ answers by obtaining consistency index (CI) and consistency ratio (CR) from Equations (2) and (3) to do that, the eigenvalueis required. Eigenvalue could be calculated from the Formula (1). All the values which obtained were merged with the values in the Table 2. These results reveal that ℷ valueshave satisfied the condition of ℷ  > n, and the mean of ℷ  values is 6.388 > 6.

By applying Equations (2) and (3), CI=ℷmaxnn-1=6.388-65=0.078 , so CR = 0.063. Therefore, the estimation is acceptable because CR < 0.1.

Similarly, AHP process has been applied to calculate eigenvectors and eigenvalue of sub-characteristics of quality characteristics of functionality through portability. the results come as the following: the eigenvector values (0.382, 0.252, 0.081, 0.119, 0.90, 0.048 and 0.029) are for sub-characteristics of functionality, (0.687, 0.226 and 0.087) are for sub-characteristics of reliability, (0.608, 0.217, 0.086 and 0.089) for sub-characteristics of usability, (0.505,0.196, 0.140, 0.063 and 0.096) for maintainability sub-characteristics, (0.536, 0.241, 0.144 and 0.078) for efficiency sub-characteristics, and (0.705, 223 and 0.073) for sub-characteristics of portability. Regarding the consistency test, all the estimates are acceptable, since all the CR values of the sub-characteristics are less than 0.1.

Therefore, this empirical study provides companies and organizations with the quality characteristics and their importance that should be taken into account in developing and implementing the ERP systems. In which it can be assured that the ERP systems will be successfully implemented.