Towards Building Knowledge on Causes of Critical Requirements Engineering Problems Marcos Kalinowski UFF Niterói, Brazil kalinowski@ic.uff.br Rodrigo Oliveira Spínola UNIFACS/Fraunhofer Salvador, Brazil rodrigo.spinola@pro.unifacs.br Tayana Conte UFAM Manaus, Brazil tayana@icomp.ufam.edu.br Rafael Prikladnicki PUC-RS Porto Alegre, Brazil rafael.prikladnicki@pucrs.br Daniel Méndez Fernández Technische Universität München München, Germany daniel.mendez@tum.de Stefan Wagner University of Stuttgart Stuttgart, Germany stefan.wagner@informatik.uni- stuttgart.de Abstract—[Context] Many software projects fail due to problems in requirements engineering (RE). [Objective] The goal of this paper is to gather information on relevant RE problems and to represent knowledge on their most common causes. [Method] We replicated a global family of RE surveys in Brazil and used the data to identify critical RE problems and to build probabilistic cause- effect diagrams to represent knowledge on their common causes. [Results] The survey was answered by 74 different organizations, including small, medium and very large sized companies, conducting both, plan-driven and agile development. The most critical RE problems, according to those organizations, are related to communication and to incomplete or underspecified requirements. We provide the full probabilistic cause-effect diagrams with knowledge on common causes of the most critical identified RE problems online. [Conclusion] We believe that the knowledge presented in the diagrams can be helpful to support organizations in conducting causal analysis sessions by providing an initial understanding on what usually causes critical RE problems. Keywords—Survey; NaPiRE; Knowledge Building; Requirements Engineering; Problems; Causes; Causal Analysis. I. INTRODUCTION The importance of high-quality requirements engineering (RE) has been widely accepted and well documented. Pfleeger [1] states that efficient RE is one of the main factors to avoid software project failure. RE constitutes a holistic key to successful development projects [2]. However, industry is still struggling to apply high-quality RE practices [3] and getting a further understanding on common RE problems and their causes is of great interest to both industry and academy. Therefore, many researchers have addressed identifying and analyzing RE problems faced by industry [4][5]. More recently, a project called NaPiRE (Naming the Pain in Requirements Engineering) comprises the design of a family of surveys on RE practice and problems, and it is conducted in joint collaboration with various researchers from different countries [6][7]. The goal of this project is to lay an empirical foundation about the state of the practice in RE to allow steering future research in a problem-driven manner [6]. Currently, the NaPiRE survey is being conducted in several countries around the globe. Conducting causal analysis sessions [8] is an efficient means for organizations to improve their practice to overcome problems faced during software development. In these sessions, the causes of problems are identified and addressed to prevent their recurrence in future projects. Experience reports on conducting causal analysis sessions on RE problems can be found in [9], [10] and [11]. One of the main difficulties reported during those sessions concerns the absence of a starting point for identifying potential causes. An initial solution concept to address this problem has been proposed in [12], where an approach for integrating knowledge of successive causal analysis sessions is described. This approach introduced the concept of a probabilistic cause-effect diagram, and of using such diagrams to present accumulated knowledge on the probabilities of causes based on the organization’s prior causal analysis experiences on similar problems. However, although this approach and the probabilistic cause- effect diagram showed to be useful to support causal analysis sessions in a proof of concept [13], an experimental study [14] and an industrial experience [9], the knowledge depicted in the diagram has to be generated based on intra-company data from previous causal analysis sessions. Thus, it has to be built gradually and from scratch for each context, as there is no general documented and empirically grounded knowledge causes of critical problems that could be used as a starting point. In this paper, we aim at gathering information on relevant RE problems and to represent knowledge on their most common causes as reported by the industry. Therefore, we replicated the NaPiRE survey in Brazil. We got answers from 74 different Brazilian organizations, spread across the country. We then used the data to identify the reportedly most critical RE problems and organized knowledge on their common causes by building