Abstract
In order to meet their strategic plans, companies have been investing in portfolios with a large number of projects; this phenomenon has been especially true for companies in the oil sector, whether because of a significant increase in society's demands for energy or the strong competition among these companies.
Within this context, the portfolio desired by companies is far beyond their capacities to carry them out, with the greatest restricting element being the lack of manual labor to conduct these projects. Accordingly, it is vitally important to develop an efficient methodology to give priority to projects and assist executives in choosing the most appropriate projects.
The objective of this article is to demonstrate the application of a methodology to give priority to projects using the analytic hierarchy process (AHP) technique in a large Brazilian company that transports and stores oil and oil products. The dynamic of the development of the methodology and the results achieved in the prioritization of the portfolio will be described.
Introduction
Project portfolio management has been a management alternative for companies that need to know which projects must be executed, due to a demand for projects that is greater than the capacity to carry them out and limited by resources. This discipline has generated a great deal of interest; many authors have studied it and have emphasized how essential it is in the conduct of organizational strategies.
Knowing which projects to invest in, and those that will promote a company's growth, is a challenge; on the other hand, investing in the wrong projects could lead to lost revenues, lost markets and, in extreme cases, make a company's existence unviable.
The organization studied, concerned with the way it assembled its project portfolio, resolved to perfect its methodology of giving priority to projects, which sparked the interest for this study. The objective of this study was to understand how to give priorities to a group of projects.
The work shows how part of the process of project portfolio management was implemented, utilizing the analytic hierarchy process (AHP) technique in an organization for the transport and storage of oil and oil products.
The details of this study begin in this section, which deals with the introductory aspects of the problem of giving priority to projects. In section two, the theoretical bases are presented, to support the case proposed together with the basis for understanding the problem. The principal authors who have dealt with the theme in question will be presented. In section three, the alternative methodology adopted and the procedures for the survey are presented. The results and their respective analyses will be presented in section four; finally, the conclusions are discussed.
Theoretical References
During the early part of this century, the topic of portfolio management assumed greater prominence in the literature of projects, partly due to the growing importance of project activity in organizations and partly due to problems arising from the coexistence of multiple and competing projects (Carvalho, & Rabechini Jr, 2006; Rabechini Jr., et al., 2005; Reyck, et al., 2005; Dye, & Pennypacker, 1999; Wheelwright, & Clark, 1992).
A system of portfolio management must provide elements so executives can decide which undertakings best translate into their corporate strategies; that is, “do the right thing,” (Drucker, 1963) which in the context of portfolio management, signifies mapping the most relevant opportunities and selecting the projects most aligned with the organizational strategy.
In this sense, the importance of the projects, must be established, taking into account the strategic aspects considered from the corporate environment point of view and also the real possibilities of the company being able to execute the undertakings in the portfolio.
The thematic influences presented as elements of the project portfolio management study go from the theory formulated by Markowitz (1952) to the conceptual approaches of Elkington (1997; 1998) on sustainability. This century, a new conceptual wave has arisen to interrelate and establish portfolio management as a discipline. It deals with portfolio models in the business and product development areas. The most noteworthy works on this topic were produced by Cooper et al (2001); for them, portfolio management signifies a process of dynamic decisions in which a list of active new product projects is constantly updated and revised. In this process, new product projects are evaluated, selected, and given a priority. Resources are then allocated and reallocated to the active projects.
The Project Management Institute's (PMI) (2006) vision of portfolio management shows the necessity of having the conditions to select the correct undertakings for the construction of the corporate strategies. For PMI (2006), the portfolio management signifies the centralization of one or more portfolios that include the identification, prioritization, authorization, management and control of projects, programs, and other related works, to achieve specific strategic business objectives.
The alignment between the business strategy of organizations and their portfolio of projects has been debated with interest by organizations and academics, and some models have arisen in both the corporate sphere and the academic milieu (Carvalho; Rabechini Jr., 2005 and 2006; Cooper et al, 1999 and 2001; Roussel et al., 1991).
The processes that give life to project portfolio management consist of mechanisms full of technical propositions, aimed at providing substance and consistency to the management itself. In this sense, the techniques of prioritization necessary to dealing with the multiple criteria of qualitative and, in some cases quantitative, evaluations are highlighted. One of the techniques presented for the process of project prioritization is the analytic hierarchy process (AHP). Conceived to provide priorities based on the comparison of similar pairs, taking into consideration criteria or properties in common, the technique conciliates the judgment of intensity of importance of one item over another (Saaty, 1980). The proposition of Saaty (1977) to measure the intensity of importance utilizing multiple criteria has its origin in a scale of nine levels. The first level represents the evaluation by one judge that two criteria have the same intensity of importance. The other levels represent increasing intensities of importance, always in the comparison of pairs of criteria.
Methodology
The methodological procedures of a qualitative character that guided this research were based on the case study method. According to Yin (1994), the essence of case study refers to the “attempt to illuminate a decision or set of decisions: why were they taken, how were they implemented, and with what result?” The preference for the use of the case study occurs when there are simultaneous events, in situations in which the relevant behaviors cannot be manipulated, but where it is possible to make direct observations and systematic interviews.
With these characteristics, the alternative to the case study methodology was considered as meeting the precepts of Yin (1994). This places the option for the only case when this may be considered revealing and when there is an opportunity to observe the process of development and, consequently, the learning of scientific knowledge.
The case study was based on interviews with members of the specialized teams, as well as the analyses of corporate documents from the company. Interviews structured by questionnaires were carried out with the employees, both in the engineering area and the three operational areas. Documents relating to the corresponding processes were analyzed.
A research question was drafted to guide the development of this study: how may we utilize the AHP technique to give priorities to engineering projects? Starting with this question, a process to prioritize the projects was studied.
The work was begun with the identification of those involved in the development of the methodology for prioritization, followed by the analysis of documentation and the interviews. Then, the information gathered (both from the analysis of the documents and the interviews) was analyzed and, finally, the conclusions of the study were reported.
The research flowchart is illustrated in Exhibit 1.
Exhibit 1 – Research Flowchart
Case Description
The organization that was the subject of this case study was a regional office situated in southern Brazil and belongs to one of the largest Brazilian companies in the transport and storage of oil and oil products.
The survey of data showed that the organization was prepared to execute approximately 50 projects per year at that time; although this appears to be a sizable number, the demand was, in fact, closer to 109 projects.
The process of giving priority to the prevailing projects at that time took into account the negotiating capacity of the representatives from the three operational departments. There was no structured process for the management of demand for projects, nor for the ratification of adequate criteria. The composition of the project portfolio was based on negotiations with the three operating areas, which were considered customers, utilizing the criteria of the potential risk to the company if the projects were not completed. The lack of any consistent management process and the exposure of the projects to negotiation have weakened the results obtained, as well as possibly generating an inappropriate portfolio.
It is possible to report that, perhaps due to the consequences of the above, the process of developing a system for the management of the project portfolio was initiated and aimed at improving the prioritization of the projects in the portfolio.
The establishment of a management process has the objective of identifying which projects are most important to the organization and at the same time which ones are capable of being carried out, even considering that the prioritization of the projects themselves was not the definitive element in defining the ideal portfolio.
The process of preparing a prioritization system began with the formation of an interdisciplinary team and the hiring of a specialist. Considering the information obtained in interviews, it was opted to utilize the multi-criteria analysis denominated analytic hierarchy process (AHP) for the definition of relative weightings for the criteria adopted.
Information regarding the opinions of those involved in the operational areas and concerning the adequacy of the unit's project portfolio management was gathered. Exhibit 2 shows this perception.
Exhibit 2 – Perception of the Portfolio Management
The process created took this important information into account and, according to the information from those interviewed, formed the construction and implementation of the model.
For the constitution of the management process, the method of multi-criteria analyses chosen was AHP, with seven criteria defined for the evaluation of projects. The criteria suggested by the workgroup and validated by the four regional managers are outlines in Exhibit 3.
Exhibit 3 – Criteria Chosen
With the criteria defined, their relative weightings were established. In the first attempt at establishing the weightings, a simplified form of parity comparison was utilized, in which mere preference for one criterion over another was used without the intensity of each preference being analyzed, as recommended by Saaty (1980).
The result of the analysis affected is illustrated in Exhibit.
Exhibit 4 – Parity Analysis of the Criteria
When the criterion of the line is preferable to the criterion of the column, number 1 is entered; otherwise, the number 0 is entered. For example, criterion A (operational safety) is preferable to all the other criteria, so number 1 is entered for the entire line of criterion A, and the relative weighting for A is equal to 25%. Another example is that criterion C (customer satisfaction) is only preferable to criterion G (complexity and multidisciplinary needs), and has a weighting equal to 7.14%.
The results of the analysis were submitted to the four regional areas and approved unanimously.
Although this evaluation is simple, it does have certain limitations. As can be seen, criterion A is preferable to all the other criteria, but its relative weighting is limited to 25%, a value that may not meet the concerns of the people involved.
The utilization of this type of matrix could lead to errors in evaluation, caused by the maximum limit in the weighting of the criteria, as seen in Exhibit, which was prepared taking into account a number of criteria, varying from two to seven. In this exhibit, the maximum weighting a criterion can have is shown and even preferable to all the others. For example, when six criteria are used, the maximum weighting is 29%, and when three criteria are used, the maximum weighting is 50%.
Exhibit 5 – Relative Maximum Weighting for a Criterion (Sensitivity Analysis)
After defining the criteria and their relative weightings, each project was evaluated with the assistance of a Likert scale, ranging from 1 to 5, according to the rules established; for example, those of the criteria “operational safety” shown in Appendix I and the information contained in the project file cards that had been completed by the operational areas. Excel spreadsheets were used to consolidate the data.
The maximum score that can be attributed to a project is 5, and in the first evaluation, the highest score obtained was 3.43, and the fiftieth project obtained a score of 2.21.
The result of the prioritization was presented to the operational areas, which requested revisions of the evaluations effected and the relative weightings of the criteria.
A new evaluation of the relative weightings to the criteria was effected, this time considering the intensity of the preferences in the parity evaluation according to the basic Saaty scale, which had been presented to and approved by the operational areas.
The result of the new evaluation of the relative weightings may be seen in Exhibit 6 where it may be observed that the criterion with the highest relative weighting (operational safety) had its weighting increased from 25% to 43.43%.
Exhibit 6– Relative Weighting of the Criteria
With the new analysis, the priorities were altered. Of the 50 principal projects evaluated in the second evaluation, only 40 had appeared in the first evaluation, indicating clearly that the weightings utilized in the first evaluation did not meet the necessities of the organization. In the second analysis, the project with the highest score received 4.30 and the fiftieth project obtained 2.66.
Exhibit 7 – Evaluation of the Projects
The scores obtained by the projects in the two evaluations may be seen in Exhibit 7, which was prepared taking into account the 109 projects in the “X” axis and the scores from the two evaluations of each project in the “Y” axis. The chart was placed in order from the second evaluation, with the first project being that which obtained the highest score in the second evaluation, and the last project being that which obtained the least score. For example, project number 1 obtained a score of 4.32 in the second evaluation and 3 in the first, whereas project number 109 obtained a score of 1.11 in the second evaluation and 1.18 in the first. The chart also shows the limit of the 50 projects that obtained the highest scores in the two evaluations.
In order to compare the prioritizations, it was necessary to make the two lists of projects adequate. In the first evaluation, there were 125 projects, with 112 evaluated; the second list had 125 projects, with 117 evaluated, but only the 109 utilized in this work had both evaluations and prioritizations.
The work undertaken by the prioritization methodology development team was considered satisfactory by the organization. In addition, because the work was undertaken by company staff with the guidance of an external specialist and, allied to the active participation of the operational areas involved throughout the process, this contributed significantly to the results achieved.
Conclusion
This work provides an understanding of how to implement a system of project prioritization in an organization of transport and storage of oil and oil products.
The principal points considered were (1) the possibility of using the analytic hierarchy process AHP technique, (2) the advantages of prioritizing projects, (3) a comparison between two forms of evaluating criteria.
When analyzing sets formed by many items, the number of comparisons creates a supermatrix unviable enough to
The AHP technique proposed by Saaty (1980) presupposes parity comparisons between the criteria to identify the relative importance of each one, and also parity comparisons of the alternatives to identify the degree to which they are preferable in relation to each criterion, admitting that the maximum number of alternatives is nine.
be applied (Dyer, 1990, p 255). To resolve this problem, the practice of evaluating each item as a function of its contribution to each criterion was adopted. In the case of the organization under study, the quantity of comparisons effected was equal to 109 items times 7 criteria, which equal 763 comparisons.
Certain care must be taken in relation to the scales utilized. Some authors, such as Stevens (1946), Forman (2001), and Kendrick and Saaty (2007) do not recommend the utilization of the Likert scale but prefer using rating scales, because, unlike the Likert scale, this scale allows the zero point. It is also an absolute scale for each criterion, because some problems could arise when the Likert scale is used in this type of evaluation.
For example, in a Likert scale of five levels, like the one that was utilized by the organization, the possible evaluations can only be 1, 2, 3, 4, or 5, and because there is no zero score, when an item does not impact a given criterion, it still receives a score of 1, which does not reflect reality. Another type of problem is the utilization of the same scale to measure different criteria. For example, with the criteria “legal and regulatory” and “financial,” it is difficult to represent them correctly using the same scale.
The organization studied utilized the concepts of the AHP methodology by initially adopting a simplification: a parity evaluation of the criteria without considering the intensity of importance. This procedure, although much simpler to implement, does not find support in scientific publications, and the results obtained did not fully meet the expectations of those involved. The work was subsequently redone, this time considering the intensity of importance by adopting the basic Saaty scale. Starting from the new relative weightings of the criteria and calculating the relative importance of the criteria according to the AHP methodology, it was possible to reach a satisfactory result.
It was not possible to evaluate what the impact would have been on the prioritization result affected if the rating scales had been used instead of the Likert scale, because in this situation, representative scales would have to be prepared for each criterion and every project would need to be reevaluated as it applies to each criterion.
The study was limited to approaching the theme of prioritization, without approaching the management of the portfolio as a whole; this theme will be studied in another forum. The results observed here are strictly related to the unit researched and cannot be generalized.
