Cloud computing projects: critical success factors

Abstract Purpose The purpose of this article is to identify the critical success factors (CSFs) of information technology (IT) projects using cloud computing. CSFs are variables that can influence the success of projects and therefore need to be identified and managed appropriately. Design/methodology/approach This is an exploratory qualitative study with 23 experts in cloud computing projects through semi-structured interviews. The data was analyzed using content analysis. Findings The results present a list of CSFs for projects in a cloud computing environment. The study demonstrated that the CSF with greater relevance to cloud projects is the team’s technical capacity, followed by the support of senior management and the team’s soft skills. In addition, results demonstrated that contract item management is a limiting factor for cloud projects. Research limitations/implications The sample comprised only Brazilian experts, so it may not represent the same scenario as in other locations. The CSF ratio for cloud computing projects may vary depending on the company’s maturity in projects of this nature. Practical implications The CSF relationship can guide managers in properly conducting cloud computing projects, contributing to minimizing the risks and challenges that may interfere with the project. Social implications The relationship of the CSFs in cloud computing projects proposed fills a gap in studies specifically related to this context and tries to minimize project managers’ stress. Originality/value Contract items for the cloud context are added to the CSF literature in IT projects, which have not been addressed so far.


Introduction
Global business models have undergone profound changes related to the use of technology.In this context, cloud computing emerges as an evolution of information technology (IT), changing the way IT products (infrastructure, development platforms and software) are delivered to customers (Wang, Wood, Abdul-Rahman, & Lee, 2016).In this IT model, services (Mell & Grance, 2009).The cloud revolutionizes the traditional adoption of IT (Hsu, Ray, & Li-Hsieh, 2014), as it allows contracting infrastructure, platforms and software as services, and it is not necessary to make them an organizational asset (Sultan, 2011).This model increases the flexibility and scalability of the business, as resources can be contracted and immediately released when no longer needed (Zissis & Lekkas, 2012), and changes the way IT services are designed, developed, implemented, sized, updated, maintained and paid for (Avram, 2014).
In the on-premise model (Wang et al., 2016), applications, data, operating systems, servers, virtual machines, storage and network are maintained by the customers themselves, while in the cloud model, they are offered separately.As a result, incorporating IT through cloud services can shorten the schedule, optimize the scope, reduce the cost of IT projects (Wang et al., 2016) as well as minimize service provisioning issues, simplify systems and application management or reduce deployment costs.

Challenges of projects using cloud computing
Some features of the cloud imply challenges for organizations and PMs.Cloud environments are shared service environments where several clients can host their data on the same physical server or datacenter, increasing the security and control required in this environment (Google, 2020).In the traditional IT approach, complete control of installed servers and systems is under the responsibility of the IT department, while in the cloud, the customer has restricted access, which makes it difficult to measure performance and control the reliability of some services (Hofmann & Woods, 2010;Wang et al., 2016).Cloud services are acquired, managed and measured through contracts between customer and provider, making the proper management of contracts and changes fundamental factors for projects in this context (Wang et al., 2016).
In the cloud context, challenges for project management can be technical knowledge; decentralized virtual teams of the customer and the provider; and different cultures arising from the location of the customer and the provider (Sultan, 2011;Wang et al., 2016).Furthermore, considering that cloud services are contracted on demand (Armbrust et al., 2010), the set of the services influences their availability during the project.
IT service provider support and maintenance are based on the service level agreement (Hofmann & Woods, 2010).In this regard, the installation of services, the problem solution or the communication between the customer and the provider are given by the means and deadlines agreed to (Hofmann & Woods, 2010).Therefore, such aspects should be carefully considered in the project planning as these deadlines may not meet specific project needs that are subject to various risks and therefore need a good risk management plan.
Customers can receive constant updates in the Software as a Service (SaaS) cloud model.In this respect, the management of stakeholder expectations and the cultural and social expectations related to change are aspects to be considered in project management (Wang et al., 2016).Furthermore, about the transition to the cloud, Sheffield and Lem� etayer (2013) highlight interpersonal skills because the PM will have to integrate activities between different teams of the customer and the provider, in addition to managing project stakeholders that have different organizational cultures.

Information technology projects and critical success factors
CSFs are defined as the areas that, with satisfactory results, can bring competitive performance to the organization (Rockart, 1979).CSFs in projects were explored by Fortune and White (2006), who listed 27 factors for various types of projects.They proposed the formal system model that grouped the CSFs into goals and objectives, performance Cloud computing projects monitoring, decision-making, transformation, communication, environment, limits, resources and continuity.Several authors have studied which variables are essential for the success of IT projects, pointing out CSFs in this context.Some studies that stand out are Chow and Cao (2008), Nasir and Sahibuddin (2011); Sudhakar (2012); Imtiaz, Al-Mudhary, Mirhashemi, and Ibrahim (2013); Ahimbisibwe, Cavana, and Daellenbach (2015); Martins Muller and Dal Forno (2017); Fayaz, Kamal, Amin, and Khan (2017); and, finally, Stevenson and Starkweather (2017).Chow and Cao (2008) is the most cited among all these studies.From the analysis of 109 projects, the authors evaluated the CSFs of software development projects using agile methods and tested 12 CSFs extracted from a complete ratio of 36 CSFs, grouping them into dimensions.The authors related the CSFs to the perspectives of success considering quality, time, scope and cost.
When studying CSFs of software projects, Nasir and Sahibuddin (2011) suggest the following factors: clear and frozen requirements, a realistic estimate of schedule and budget, along with a competent PM.The study found that non-technical factors (94%) dominated over technical factors (6%).Sudhakar (2012) demonstrated that the project management dimension concentrates most of the CSFs of software development projects and that the main CSF is the support of senior management, followed by communication.Imtiaz et al. (2013) listed 15 CSFs in IT projects: top management support, leadership, work team, clear goals, team capacity, financial/budget support, effective communication, process quality, training, project progress monitoring, client/user involvement, risk management, effective monitoring and control, adequate requirements and correct team selection.Ahimbisibwe et al. (2015) study identified 37 CSFs for software development projects organized into four categories: organizational, team, clients and project.Finally, Martins Muller and Dal Forno (2017) studied the CSFs for software development projects considering the influence of the methodology used by the projects (agile, traditional or mixed), and pointed out that, for Brazilian organizations, there is a similarity in the contribution of the CSFs in software development projects that is independent of the methodology.
The study by Fayaz et al. (2017) suggests 15 CSFs: management support, budget support, effective communication, effective training, monitoring and control, leadership, clear goals, requirements specification, risk management, user engagement, project progress schedule, team capacity, right team, project duration and teamwork.Stevenson and Starkweather (2017) grouped the CSFs into five categories of analysis: communication, project, project management and team, organization and stakeholders.They point out that the main factor for the project to succeed is project management and team, especially the ability of the group to communicate at multiple levels.Table 1 shows the CSFs by the authors' approach.
It is possible to observe, in the cited works, that the CSFs have specific similarities and differences according to the focus of each study but are mainly based on variables found in the study of Fortune and White (2006) on CSFs in projects.However, observing more specific IT projects that deal with the implementation of enterprise resource planning (ERP) software, several authors such as Ehie and Madsen (2005), Finney and Corbett (2007), Françoise, Bourgault, and Pellerin (2009) and Beheshti, Blaylock, Henderson, and Lollar (2014) listed specific CSFs, which may mean that for specific types of projects, such as cloud projects, the CSFs may also be more specific.
Considering CSFs for project management in cloud computing, some studies address the theme in contexts of cloud model adoption in ERP projects (Gheller, Biancolino, Junior, & Giroletti, 2017), cloud in HR processes (Ziebell,

Cloud computing projects
Perello-Marin, 2019), cloud adoption in small and medium-sized companies (Hentschel, Leyh, & Baumhauer, 2019), cloud adoption in other industries such as construction (Oke, Kineber, Al-Bukhari, Famakin, & Kingsley, 2021) and cloud adoption in public sectors (Sallehudin et al., 2019;Mohammed, Ibrahim, & Ithnin, 2016), but studies do not address CSFs in cloud project management with an approach that considers the cloud as an IT project architecture, which is the focus of the present work.In a literature review by Hentschel et al. (2019) that analyzed 28 articles that deal with CSFs in the cloud, 26 of them dealt with CSFs of cloud implementation projects, confirming different objectives than the present article, which deals with the management of projects carried out in a cloud environment.Cloud-based CSFs can present a different degree of importance and relevance from other projects carried out in on-premise architecture environments.Therefore, this study presented and discussed a new CSF model that deals with the influence of the cloud contract on the project (analyzed with the interviewees), showing that the cloud contract items need to be known by the PM and may appear as limiting factors during the execution of the project.

Propositions
When analyzing the studies on CSFs in IT projects, it was not possible to observe, in the set of CSF established as vital, references to some specific characteristics of cloud projects, such as items related to contract management (Wang et al., 2016), change management (Wang et al., 2016), communication management (Wang et al., 2016;Sheffield & Lem� etayer, 2013)  Capacity and competence of the team; experience of the team; commitment and motivation of the team; experience and competence of the project manager; relationship team/user; specification of requirements; definition of objectives and goals; schedule planning; project planning; risk management (prior analysis and treatment); training and learning of users/clients; development methodology; team size; team composition; team experience with the methodology used; appropriate process and methodology; infrastructure and support tools; support from higher management; commitment of higher management; project leadership; adequate resources (human, financial and material); organizational culture (of the client); user/customer participation; user/customer commitment; variable factors; realistic expectations of the user/ client; internal communication of the project; estimates of previous realized costs; realistic budget; project complexity; project size; monitoring and effective control of the project; quality control; process tests Stevenson and Starkweather (2017) Ability to communicate at multiple levels; define the project appropriately; customer acceptance; ability to coordinate; stakeholders have a common understanding of success criteria before the project starts; clearly articulation of what should be done; competence of the PM; communication and collaboration between PM and project owners; listening; commitment; meeting user requirements; top/senior management support; meeting stated objectives; ability to deal with ambiguity and changes; keeping the team moving toward a common goal Source: Prepared by the authors Table 1.
RAUSP 58,1 how the characteristics of the teams impact the success of the projects (Wang et al., 2016).This suggests the need for a more cloud-driven CSF study.
We listed the propositions below from the theoretical framework analyzed to verify the CSFs for projects using cloud technology (Table 2).The Organizational, People, Processes, Technique and Project dimensions emerge from the study by Chow and Cao (2008), who listed the CSFs for agile projects.They were selected for their relevance and as the result of grouping several CSFs listed in other studies.The dimensions Contract management, Change and risk management and Communication management are proposed based on the literature on cloud projects.

Method
This research was based on the following phases: study for the theoretical framework; preparation of the interview protocol; face validation; interviews and objective questions with experts; transcription, codification and CA of the interviews; and presentation of results, analyses and conclusions.
A qualitative and exploratory approach was adopted, with data collection through semistructured interviews (Bardin, 1977) with cloud experts.A script guided the interviews on the CSFs in IT projects to analyze their application to cloud projects.The script contained questions concerning the characterization of the respondent and CSFs, based on the propositions presented in Table 2.This instrument was tested with one respondent, and we certified that it was suitable for use in the other interviews.However, to minimize any bias of the researcher, a new instrument was created with multiple-choice questions addressing the CSFs based on the 20 propositions for the interviewees to evaluate the degree of influence each CSFs had on the success of cloud projects (five-point Likert scale, ranging from 1 -no influence to 5 -total influence).This new instrument was also validated with the same respondent.
For this research's unity of analysis, IT projects that used cloud computing technology as an IT architecture were considered, either as Infrastructure as a service, Platform as a service or SaaS.The projects' financial contribution was not relevant to the research, nor was the size of the organization to which the projects belonged.The region in which the company was located was not relevant, either.Respondents, or research subjects, were experts in cloud computing projects with experience in managing cloud projects, being leaders, PMs or directors directly involved in the project.There was no minimum amount of experience in managing projects of this type, but it was necessary that interviewees be involved with project management and not with the direct programming of IT resources in cloud environments.
We interviewed 23 specialists working in Brazil, totaling 23 h and 45 min of video recordings.The interviews were conducted through videoconferencing, using Skype and Google Meeting tools.In the sample of the interviewees, 2 were women and 21 were men.The participants ranged from 25 to 68 years (an average of 44).The professionals' experience in IT ranged between 7 and 40 years (an average of 20 years), while the time of experience in cloud projects was between 6 months and 12 years (an average of 4 years).
Of the interviewees, 5 worked for national companies and 18 for multinationals.Companies were divided into sectors: technology services (13), business process outsourcing (BPO) (4), transport (3), telecommunications (2) and insurance (1).Most of these companies had more than 2,000 employees.The experience of organizations whose interviewees provide services to (customer companies) is relatively new to cloud projects, ranging from two to five years (an average of four years).Throughout this document, interviewees (characterized in Table 3) are called I01 to I23 for confidentiality reasons.The interviews were transcribed for later analysis, which was carried out through CA (Bardin, 1977;Silva & Russo, 2019) by one of the researchers.The phases of CA are divided into pre-analysis, exploration of the material and analysis and treatment of the results.In the pre-analysis phase, the corpus of the research is constituted.In this work, it consisted of the transcription of 23 interviews with 24 questions per interviewee, totaling 552 answers.In the exploration phase, the recording and context units are constituted.In the case of this article, 18 CSFs and their strengths (positive view from the interviewee) and weaknesses (negative/doubtful view from the interviewee) were explored.In the last phase of CA, data were categorized and analyzed with inferences and interpretation.

Cloud computing projects
In this study, the categories were separated by CSF.After that, strengths/weaknesses in the interviewees' statements that characterized the relevance of the HR were evaluated.Thus, each HR was interpreted individually.
No specific software was used to support analysis; instead, a Word document was produced observing each CSF for each of the interviewees and determining the importance and prominence of the CSF in their responses.

Analysis of critical success factors in cloud projects
According to the interviewees, several CSFs mentioned have some or total influence on the success of cloud projects, whereas others do not.The CSFs of this study are addressed

Summary of critical success factor analysis in cloud projects
In addition to the CA of the questions addressed in the interviews, the objective questions sought to consider the degree of influence that each CSF considered in this study has on the success of the cloud project (on a scale of 1-5, ranging from no influence to total influence, respectively).These questions allowed an overview of this matter, in addition to minimizing any bias in the inference of the qualitative analyzes summarized in Table 4.Although the number of responses is too small to make statistical inferences or reflect the reality of this article's analysis effectively, it corroborates the interviews' CA.It is worth noting that the objective of these closed questions is to compare the CSFs based on the interviewees' responses and corroborate with the analyses and inferences previously made.Table 4 shows the results obtained from the objective questions.
It is noteworthy that several CSFs had a median of 5.0, which means that among the interviewees, more than half of them attributed these CSFs as having total influence on the success of cloud projects.
It is also noteworthy that the contract items factor, suggested in this work for the context of cloud IT projects, seems to have become more relevant over the years as the company becomes more experienced in using cloud technology [1].From this, there is space for future research to address the influence of cloud maturity on the CSFs of projects conducted in such an environment.

General analysis of propositions and discussions
Based on the analysis of the collected data, it is possible to present the confirmation of the propositions (Table 5).In addition, the most relevant points analyzed in the CSF proposition stand out.In the table, "yes" means that the proposition was confirmed because responses that attest to the relevance of the CSF (positive mentions, agreement) were found in most of the interviews.In addition, when observing the closed answers that deal with the CSFs, the average of the interviewees corroborates this inference.On the other hand, "Not necessarily" represents that, despite being mentioned by the interviewees, no great relevance was found on the proposition (doubt, partial agreement), in addition to not being confirmed by the interviewees in the closed questions whose average was equal to or below 4.0.
Propositions P1 and P2 were not confirmed as CSFs for agile projects, according to Chow and Cao (2008); however, they are essential for projects conducted in the cloud, according to the interviewees of this study.Propositions P3, P4, P5, P6 and P8 were confirmed by Chow and Cao (2008) and by our interviews.On the other hand, Propositions P9 and P10 were confirmed only by Chow and Cao (2008), while Propositions P11, P12 and P13 were neither confirmed in this study nor by Chow and Cao (2008).Propositions P7, P14, P16, P17 and P18 that emerge from the theoretical framework for the cloud were confirmed, but Proposition P15 was not confirmed.
We noticed that, in the cloud environment, a new CSF dealing with contract items has now arisen after not being previously evidenced.It is suggested that this factor arises as a result of the management of the cloud environment through contracts (Wang et al., 2016).On the other hand, the CSFs change management (P17), risk management (P16), conflict of interest management (P7) and communication (P18) are not evidenced in the work of Chow and Cao (2008) but are present in the CSF literature in Fortune and White (2006), Martins Muller and Dal Forno (2017) and Stevenson and Starkweather (2017).It is also noteworthy

Not necessarily
The type of project is linked to the drive (prioritization, time, visibility) that the project will have within the company P13) The project schedule contributes to the project's success Not necessarily Controls and directs the project, manages expectations, but needs constant review and may not adapt to the reality of the project P14) The items negotiated in the contract between customer and provider (agreed service level, access restriction, cost, prioritization of services) contribute to the project's success Yes.
Govern how cloud services will be made available and consumed; need to be reviewed so as not to interfere with the project's progress P15) The knowledge of the PM in contract management contributes to the project's success

Not necessarily
It helps to understand the deadlines and constraints that can influence the project's progress; several organizations have a specific area that assists the PM with contractual ones (continued)  Chow and Cao (2008)], whereas the dimensions team management, organizational and people stand out.The dimensions contract, changes, risks and communication are not addressed by Chow and Cao (2008) but are evidenced as significant for cloud projects.
Based on the general analysis of the research, we developed a framework with the results and relationships between the CSFs in cloud computing projects, as shown in Figure 1.
Thus, it is possible to note the relevance of the People and Organizational dimensions, which have a more significant set of CSFs with a strong relationship, and the low relevance of the technical and project dimensions that have a more significant

Relationship between
CSFs in cloud projects and their dimensions Table 5.

Cloud computing projects
set of CSFs with a weak relationship in the success of cloud computing project management.

Final considerations
The list of CSFs in cloud projects proposed in this study fills a gap in related studies because this work deals with CSFs in the context of cloud as an architecture of IT projects.In addition, the contract items for the cloud context, which have not been addressed until now, are added to the CSF literature in IT projects, suggesting the relevance/originality of the theme.Contract items need to be part of the PM's planning review in the early phases of the project as they can be limiting.Another relevant factor is that cloud projects need technical teams with experience, something not so relevant in the CSF analysis of other IT projects.
Thus, this study's theoretical contribution stands out for presenting significant variables for the success of projects in cloud architecture environments and expanding the theoretical framework on the subject.As a social implication, it is possible to minimize the PMs' personal stress regarding the risk of failure of the projects they manage.
In terms of contributions to the practice of IT cloud project management, the set of CSFs proposed in this study can be helpful in guiding PMs and IT on aspects that need to be appropriately managed, so that cloud projects have a greater chance of success.Furthermore, attention to the proposed CSFs can contribute to the realization of diagnosis and improvement plans in the context of cloud projects.
This study has some limitations.First, the sample included only Brazilian experts, so the findings may differ from other populations.Second, the CSF ratio for cloud computing projects may vary depending on the company's maturity in projects of this nature.Finally, the selected interviewees were directly managing the cloud projects, so experts from the project teams were not heard, which may represent a variation in the CSFs.These limitations may also suggest that future studies investigate each of these aspects.
Future studies can be conducted to evaluate CSFs quantitatively in a sample of professionals and companies working with cloud projects.Quantitatively evaluating the propositions of this work can confirm the CSFs found for the cloud and confirm the cloud CSF's influence when compared to on-premise projects.A quantitative study may also cover a more extensive sample regarding location and diversity.In addition, there is room for elaborating a specific CSF scale for cloud computing projects.

Figure 1 .
Figure 1.Relationship betweenCSFs in cloud projects and their dimensions

Table 1 .
Albors-Garrigos, Schultz, Schoeneberg, & Synthesis of CSFs by authors with technology/development methodology; proper planning; appropriate development processes/methodologies (process); updated progress reports; effective monitoring and control; adequate resources; good leadership; risk management; complexity, project size, duration, number of organizations involved; effective change and configuration management; support tools and good infrastructure; committed and motivated staff; good quality management; clear assignment of roles and responsibilities; good performance by suppliers/ contractors/consultants; provision of end-user training Sudhakar (2012) Communication in the project; leadership; relationship between users and IT team; reduction of ambiguity; maximization of stability; technical tasks; problem-solving; technical uncertainties; technical implementation problems; system integration; top management support; realistic expectations; organizational policy; financial support; strength; user engagement; customer engagement; partnership with suppliers; events in the external environment; customer acceptance; output accuracy; output reliability; output opportunity; quality control; system documentation and procedures; team capacity/skills; teamwork; correct selection of the project team; team coordination; task orientation; project plan; control mechanisms; project schedule; management skills; clear goals for the project Imtiaz et al. (2013) Top management support; leadership; work team; clear goals; team capacity; financial support/budget; effective communication; process quality; training; monitoring project progress; customer/user involvement; risk management; effective monitoring and control; proper requirements; correct team selection Ahimbisibwe et al. (2015) Higher level management support; organizational culture; project planning level; leadership; vision and mission; monitoring and controlling; change management skills; team commitment; internal project communication; team training; team composition; the project team's experience with the tasks; general knowledge of the project team; lack of development team's skill; the project team's experience with software development methodologies; user participation; user support; training and customer education; customer experience; lack of end-user experience; technological uncertainty; (continued) or

Table 3 .
Compiled by the authors Cloud computing projects according to their dimensions, namely, Organizational, People, Processes, Technique, Project, Contract management, Change and risk management and Communication management, as available in a complementary file (Supplementary materials).
Notes: IT time: professional IT experience time; Cloud time: professional cloud project experience time; Enterprise cloud time: enterprise cloud project experience time Source:

Table 5 .
Analysis of the research propositions that the dimension that encompasses CSFs related to projects loses importance for cloud projects [which differs from what was found by