Selection of software agile practices using Analytic hierarchy process.

Context: The increasing use of software and information technology in modern society requires that the deployment of IT solutions should be more efficient and controlled. In this sense, agile methodologies are essential to achieve this goal, but a relevant question arises: Which is the right methodology for reaching that goal? Objective: The purpose of the research was to study the prioritization of agile practices and improvement objectives in the context of software development, using the Analytic Hierarchy Process (AHP) method. For this purpose, 42 agile practices, 16 improvement objectives, 4 organizations dedicated to software development in Colombia and 40 professionals in the area of information and communication technologies in the same country were considered. Method: ology: The methodological approach used was mixed. On the one hand, we applied a quantitative approach for the treatment of data with the AHP method and, on the other hand, a qualitative analysis by consulting experts through a digital survey to validate the prioritization of the improvement objectives. Questionpro was used as a support tool for multicriteria comparison. Results: The main results show that applying AHP allowed us to prioritize 6 improvement objectives and 5 agile practices, where it is highlighted that the prioritization does not discard the other objects of comparison but allows us to put into practice the prioritized elements to favor their progressive implementation. The non-prioritized elements could be part of future iterations of multicriteria comparison that go hand in hand with capacity development and organizational maturity models in the context of global software development. Conclusion: It is highlighted that, although there is evidence of the application of multi-criteria comparison methods in the software development industry, this work applies such methods in the context of four different agile methodologies, which translates into a precedent for the conformation of hybrid methodological approaches.