Using Welphi to build a multicriteria model for the selction of software in the oil&gas industry

Wondering if Welphi is the tool to help your decision problems while meeting corporate environments? We can help answer that question for you: definitely yes! Read below to find out how Welphi was successfully used to solve a real decision problem of the leading oil & gas company in Portugal regarding the selection and integration of a much needed software: a data integration platform.

Welphi was used in this project as a tool to collect information to latter inform the building of a multicriteria model meant to assess vendor’s proposals regarding the evaluation and selection of an Enterprise Management System, i.e. a data integration platform.  

Adopting a framework combining concepts of multicriteria value measurement with participatory processes, two Welphi processes where developed and implemented, which where both technically sound in MACBETH, in order to make clear the value system of the actors engaging in the processes these Welphi processes where used to either,

• Determine the added value from a set of performance levels in their respective categories, i.e. within each criterion (criteria value functions).
• Determine the partial value of each criterion considered in the model (weighting criteria).

 |Aim

•  Delphi for Criteria value functions: Collect qualitative pairwise comparison judgments between performance scale levels on each one of multiple evaluation criteria.
•  Delphi for weighting criteria: Collect qualitative judgments of importance of swinging between least and most preferred performance levels on the criteria.

|Objects of study: A previously selected set of evaluation criteria, determined by the company, considering both the functional and technical stands of software solutions applicable in the context.

|Panel: A total of 73 experts and future users of the platform belonging to district working areas within the company, distributed by two panels: Functional and Technical, according to their area of knowledge and expertise. A common design for the Welphi value function process was implemented simultaneously with the two panels.

|1^st round:

• Delphi for criteria value functions: For each of the considered criteria the respective performance intervals were displayed, each one representing an increase in performance. The participants were asked to answer the following question:

“With regard to this criterion, which do you consider to be the increase in preference between each two levels of performance?”

Answers were provided according to the MACBETH qualitative judgment scale (from “no increase” to “extreme increase”). The sequence of the participants’ answers enabled to extract their implicit main concerns for each criterion.

• Delphi for weighting criteria: The participants were asked to answer the following question:

“Regarding the selection of the proposal for a data integration platform that best meets the company’s needs, suppose there is a proposal with neutral performances in all criteria. What would be the importance of improving it from neutral to good on each of the criteria?”

Answers were provided with the MACBETH qualitative judgment scale (from “no importance” to “extreme importance”).

In both Welphi processes a “don’t know/don’t want to answer” option was also available for selection and comments could be provided.

|2^nd and 3^rd rounds: The extracted results of the 1^st round, participant’s percentage of answers and comments, were presented at the beginning of the 2^nd round, allowing panelists to maintain or revise their answers. Specifically, for the Welphi for criteria value functions, participants were presented with information about the percentage of participants corresponding with each criterion and each one of the extracted implicit main concern. At the beginning of the 3^rd round, panelists received again information synthesizing the answers of the panel after the 2^nd round. Accordingly, these had the opportunity to maintain or revise their answers.

|Results: Resulting outcomes of this processes served as feed in information for the company to construct the aimed multicriteria evaluation model, having a strategic group make final decisions based in the provided compilation of judgments form the participants in the Delphi processes.