Software Crowdsourcing Challenges in the Brazilian IT Industry

Leticia Machado

, Josiane Kroll

, Rafael Prikladnicki

, Cleidson R. B. de Souza

and Erran Carmel

Pontifical Catholic University of Rio Grande do Sul (PUCRS), Porto Alegre, Brazil

Department of Computing, (UFPA), Federal University of Pará, Belém, Brazil

School of Business, American University, Washington DC, U.S.A.

Keywords: Software Crowdsourcing, Software Dev

elopment, Software Engineering, Brazil, Challenges, IT Industry.

Abstract: Software crowdsourcing has been regarded as a new paradigm for the provision of crowd-labor in software

development tasks. Companies around the world adopt this paradigm to identify collective solutions to solve

problems, ways to accelerate time-to-market, increase the quality and reduce the software cost. Although

this paradigm is a trend in the software engineering area, several challenges are behind software

crowdsourcing. In this study, we explore how the software crowdsourcing has been developed in the

Brazilian IT industry. We have conducted 20 interviews with Brazilians practitioners in order to identify the

main challenges for software crowdsourcing in Brazil. Additionally, we identified and discussed enablers

and blockers’ factors, practice implications and directions for future research in the area. Our paper aims to

provide an overview of the software crowdsourcing in Brazil and motivation for researchers to better

understand challenges faced by the Brazilian IT industry.

1 INTRODUCTION

Crowdsourcing (CS) is defined as the act of an

organization to make its work available to an

undefined, potentially large networked of people – a

crowd - using an open call for participation (Howe,

2008). This concept has been adopted to disseminate

corporate tasks that were traditionally performed by

small groups of people.

CS has been adopted for several purposes such as

innovative design (Howe, 2008), information peer

production, knowledge and culture dissemination

data analysis (Brabham, 2008), and software

development (Lakhani et al., 2010; Wu et al., 2013).

CS in software development means to engage a

global pool of online workers that can be tapped on-

demand to provide software solutions or services

(Lakhani et al., 2010; Stol and Fitzgerald, 2014).

Many computational platforms were created to

handle the technical aspects of CS tasks, including

broadcasting of tasks, task assignment, and

submission and analysis of results (e.g. TopCoder,

Crowdtest and WeDoLogos).

In Brazil, software CS is in the early stages. We

have observed the lack of processes, models, and

practices to support the Brazilian community. This is

problematic because the adoption of CS in software

development activities can help to increase Brazilian

companies’ competitiveness in the global software

development market.

Our paper presents findings from a study

conducted in the Brazilian IT industry. This study

aimed to understand how software CS has been

adopted in the Brazilian IT industry and the main

challenges faced during its adoption. Additionally,

we identified and discussed enablers and blocking

factors for the software CS in Brazil, practice

implications and directions for future research in the

area.

Our study offers the following main

contributions:

 A Set of Challenges Concerning the

Adoption of Software CS by the Brazilian IT

Industry: since Brazil is in the early stages of

adoption of software CS, we explore the particular

challenges associated with it practice in Brazil.

 The Theoretical Foundation for Further

Research in the Area and Developing of Solutions

for the Brazilian Community: to identify and

understand which aspects are related to the adoption,

or not, of software CS in the Brazilian IT industry is

the first step towards integrating and facilitating the

CS model in other organizations.

482

Machado, L., Kroll, J., Prikladnicki, R., Souza, C. and Carmel, E.

Software Crowdsourcing Challenges in the Brazilian IT Industry.

In Proceedings of the 18th International Conference on Enterprise Information Systems (ICEIS 2016) - Volume 1, pages 482-489

ISBN: 978-989-758-187-8

2 SOFTWARE

CROWDSOURCING

CS is a hybrid model regarding to intrinsic and

extrinsic motivation (Mao et al., 2013). Motivation

can be driven by financial rewards, which are

extrinsic factors. On the other hand, crowd

participants are also interested in reputation that can

be earned through knowledge is shared – intrinsic

intentions (Olson and Rosacker, 2013).

Despite this outward similarity, characteristics of

service providers and suppliers are different in the

two models. While with traditional outsourcing an

entity subcontracts a handful of professional third-

party companies, CS model turns to scale via an

undefined, open, and heterogeneous online “crowd”

to source in these needs (Saxton et al., 2013).

A significant distinction between software

development strategies is the duplication of work. In

CS, activities are performed in parallel, distributed

in many chunks instead of single projects. The main

differences among Innersourcing, Outsourcing,

software CS, and Open Source Software (OSS) are

presented in Table 1. The payment characteristics of

software development strategies also are different. In

software CS payments are based on reward per tasks

(Âgerfalk et al., 2015). In OSS, knowledge is for

sharing, with the focus on the development of better

software and little if any attention given to

profitability. Software CS is an application of the

OSS principles to other industries. However, it

receives an open and unidentified group that

competes to solve a problem (Olson and Rosacker,

2013).

For each CS area three main elements are

adopted as shown in Figure 1. The first component

is the CS platform, which acts as the intermediates

between the two other components and consolidates

the tasks outcomes. The second component is the

Crowd, which is globally dispersed. The third

component is the Requesters. They are the

companies or the individuals whom demands the

work (tasks) (Prikladnicki et al., 2014).

Figure 1: Basic crowdsourcing model.

Many authors argue that CS promotes creativity

and problem solving (Kittur et al., 2013). However,

software CS has many issues and unique features.

Some of them still need for support (Wu et al, 2013):

complex and heterogeneous tasks, interdependent

tasks, several types of expertise, and activities of

collective control.

These issues in software CS include quality, cost,

diversity of solutions, delivery speed and,

competitive scenarios. Furthermore, studies discuss

many challenges and opportunities for better

understand, evaluate and support the CS influence

the software industry (Huhns et al., 2013).

2.1 The Brazilian IT Industry

Brazil is one of the largest economies in the world.

Brazil has unique characteristics. Population of 200

million people; large and expanding domestic

market; single language; sophisticated financial

market; industry and business knowledge; qualified

human resources; infrastructure; governmental

support; favorable economic, political, and legal

environment (Prikladnicki and Carmel, 2014). Thus,

Brazil plays an important role in the global

economy.

We decide to investigate software CS in the

Brazilian IT industry because Brazil has the biggest

and most diversified science, technology and

innovation system of Latin America. According to

Forbes magazine (2014), Brazil’s economy

outweighs that of all other South American

countries, and Brazil is expanding its presence in

world markets. We expect having emerging and

large economies like Brazil taking place in the

software CS market with more platforms, requesters

Table 1: Software Development Strategies.

Innersourcing Outsourcing CS OSS

Concept

Traditional business

model

Traditional global

business model

Web business model Online Community

Nature of workforce

Specific group Specific group

Open and undefined

group

Open group

Incentives

Extrinsic Extrinsic Extrinsic/Intrinsic Intrinsic

Intellectual Propriety

Enterprise Enterprise

Winning solutions CS

enterprise

Members’ license

Payment

On payroll Contract Pay per tasks Often unpaid

Software Crowdsourcing Challenges in the Brazilian IT Industry

483

and large crowds.

The Brazilian IT industry does not take a risk

before something (technology, model or process) is

been widely known in other international markets.

The Brazilian market has a more pre-cultural attitude

in order to preserve conditions, principles, and

existing processes. This market keeps a traditional

way to carry out business. Few companies engage in

innovative projects and pilot projects with a degree

of uncertainty (Prikladnicki et al., 2014).

3 RELATED WORK

Whereas CS has been discussed in a wide variable

domain, still have a small number of studies

discussing software CS initiatives in different

countries.

The adoption of CS platforms in South Africa is

discussed by Chuene and Mtsweni (2015). This

study was performed in order to understand how it is

used and by whom. The authors report the

embracing of CS initiatives has been slow,

especially amongst public organizations, due to

various reasons, such as lack of awareness. Some

local CS platforms such as “Txteagle” launched in

Kenya and Rwanda enables citizens to earn few

dollars by completing micro-tasks on their mobile

phones. A South African platform presented is

focused on crowd-funding, social, and government

crowdsourcing aspects. This study reveals a lack of

information pertaining to the status and number of

users benefiting from the adopted and/or deployed

platforms.

In a recent study performed by To and Lai

(2015), the authors discuss the latest developments

in crowdsourcing in China. They describe CS

scenarios in terms of concerns and opportunities.

China offers CS advantages, it because includes the

online population’s size people’s and their

willingness to participate of the CS activities. Other

opportunity was the Zhubajie’s CS Chinese

platform, where it services aren’t closed off to

English speakers. Chinese freelancers also

participate in other markets. CS barriers reported in

this study are the language between client and CS

partners. Another concern was the censorship and

government control, and intellectual property rights.

China and India are successful in attracting

global outsourcing industry (Perera and Perera,

2014). China has a large educated workforce, high

quality infrastructure, government keenness, etc.

While India has lower labor cost, familiarity with

western business practices, positive time zone

difference, Indian owned global delivery centers and

strong private-public partnership. In China,

crowdsourcing platforms such as Zhubajie

(zhubajie.com), TaskCN (taskcn.com) and K68

(k68.com) have attracted a lot of innovative talent

and solution seekers, which greatly enhance business

operations (Shao et al., 2012).

4 RESEARCH METHODOLOGY

An initial ad hoc literature review was carried out

with the purpose of sharing the basic CS concepts

with the research team and identifying the

challenges to be addressed. Semi-structured

interviews were conducted iteratively with

Brazilians practitioners from different Brazilian

companies. Our goal here is better understand how

the software crowdsourcing has been adopted in the

Brazilian IT industry. The interviews focused on

both industry and academic perspective. Each

interview lasted between 30 and 60 minutes.

We created an interview protocol with open-

ended questions focusing on the discussions of both

perspectives: industry - organizational motivations

for leveraging the crowd, specific tasks to be

completed and perceived impacts on the

organization; and academic – sharing characteristics

with related research areas like software testing,

collaborative software engineering, distributed

software development and distributed collaborative

programming.

4.1 Participants

We conducted a total of 20 interviews, in which the

majority of interviewees were males (18) and the

others females (2). Thirteen participants were from

the first group (industry) and 7 participants were

from the second group (academic). These

participants are mainly from the south of Brazil.

Participants have 3 years of working experience in

average.

The industry participants are classified under

three different CS perspectives: buyer, platform, and

crowd. We interviewed participants from two

pioneer Brazilian CS platforms – Crowdtest and

WeDoLogos. These companies are the two largest

crowd testers and crowd designers in Latin America.

We interviewed academic participants during

The Brazilian Conference on Software: Theory and

Practice (CBSoft 2014). The CBSoft is one of the

largest events held by the Brazilian Computer

Society, with the goal of promoting and encouraging

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

484

Table 2: Participants’ information.

Participant # Job Title

Type of experience

(Academic or Industry or Both)

Element

P1..P6 Tester Industry and Academic Crowd

P7, P8 Developer Industry and Both Crowd

P9..P11 Assistant Professor (System analyst) Academic Crowd

P12 Developer Both Requester

P13 Manager (IT Consultant) Industry Requester

P14 IT Manager (Host Service Company) Industry Requester

P15 IT Manager (E-Commerce Company) Industry Requester

P16 Manager (IT Company) Industry Requester

P17 Manager (Media Company) Industry Requester

P18 CEO (Innovation Consultant) Industry Requester

P19 CEO (WeDoLogos) Industry Platform

P20 CEO (Crowdtest.me) Industry Platform

the exchange of experiences between the scientific,

academic and professional communities in Software

Engineering (SE). The participants’ details are

presented in Table 2. The last column called Element

presents the three basic elements of the CS model, in

which each participant was classified (see Figure 1).

4.2 Procedure

We conducted the interviews face-to-face, by voice

or video conference call, and by email. Some

conversations were not audio recorded because of

companies’ confidentiality issues. The interviewees

were asked to report their experiences in software

CS under six aspects: (1) CS initiative, (2) CS

platforms, (3) CS tasks and projects, (4) CS

payment, (5) business impact, and (6) the future. We

present the interview questionnaire in Table 3.

Table 3: Interview Questionnaire.

Aspects Questions

Initiatives

Do you know CS?

Tell us about your CS experience?

Are you doing micro-tasking specifically?

Or are you doing macro task projects?

Platforms

Which Platforms (middlemen) have been

used?

What is the number of workers in the

crowd?

CS Tasks

and

Projects

What have you done to achieve and inspect

for quality? What has worked best?

How do you manage day-to-day tasks?

Payment

Is the enterprise encouraging / discouraging

the use of paid CS?

Business

impact

By what measure was it successful? What

has made this challenge a success?

How did you measure success?

Future

What are your plans for CS?

What is the Brazilian CS scenario for the

next three years?

4.3 Data Analysis

Our data analysis was guided by techniques

associated with less procedural versions of the

grounded theory (GT). Specifically, we applied the

techniques of coding and constant comparison as

recommended by Corbin and Strauss (2008). These

techniques helped us to elicit emergent themes in the

Brazilian IT industry, to identify concepts in the

collected data and to link these concepts to higher-

level categories.

5 FINDINGS

Our findings show that both academic and industry

participants have different experiences using CS for

micro and complex tasks. They have adopted CS for

software and other domains. They have performed

tasks such as, testing service and image recognition.

We also identified collaborative tools adopted by

Brazilians to improve software development. These

tools are based on crowd knowledge such as GitHub

and Stackoverflow.

Table 4: CS research in Brazil.

Research Area Crowdsourcing Topics

Experimental SE

Software Engineering

Barriers to contribute to the

open sourcing process

Software testing

Crowd testing model of the

enterprise

Collaborative Software

Engineering

Collaborative tools – CS

Platform

Crowd knowledge

Software Ecosystems

Platform

Distributed software

development and open

participation (crowd)

Distributed collaborative

programming

Motivation, coordination

and sharing knowledge

Software Crowdsourcing Challenges in the Brazilian IT Industry

485

Researchers are investigating Crowd testing in

two contexts – distributed and traditional testing

software. We observed from these findings research

opportunities in other software ecosystems that share

the same characteristics with other research areas.

Table 4 presents the research areas and topics in

software CS.

5.1 CS Elements Perspective

Our findings show that Brazil has a few CS buyer’s

initiatives and platforms. On the other hand,

Brazilian active members’ are emerging both on

national and international CS platforms.

We describe our findings based on three

perspectives: Crowd, Requesters, and Platform.

These results can be categorized in enablers and

blockers’ factors. Table 5 shows these factors.

Table 5: Enablers and Blockers’ factors in Software CS.

Elements Enablers Blockers

Crowd

Extra money Poor feedback

Shared

knowledge

Few collaboration

Curiosity

Scarce context

project information

Free time

Unavailability of

documentation

Requesters

Scalability

Specific business

knowledge

Save money and

time

Low quality of

services

Creativity

Maturity of

suppliers (crowd)

New ways to do

the same

Identifying a

specific process

Platform

Fast delivery Data confidentiality

Reduced cost

Very specific

business rules

Diverse types of

testing

Laws and taxes

involved

An enabler factor means a characteristic that

promotes or motivates the CS practice. On the other

hand, a blocker factor means a characteristic that

inhibits or limits the CS practice.

As enablers, there is the collective intelligence of

the software engineering industry with more

diversity, creativity and knowledge sharing.

Scalability, cost reduction and time-to-market are

also important enablers for platforms and requesters.

As blockers, the requesters pointed out the low

quality of services, the difficulty in identifying a

specific process to distribute tasks to the crowd and

the maturity and adoption of CS in Brazil.

5.2 Challenges

We identified challenges related to three areas:

Tasks, Processes and People. Each area can describe

one or more challenges.

 Tasks – Lack of Quality

Challenges related to tasks area includes the lack the

quality of platforms and micro-tasks. The platform

should provide clear information and support

documentation to the crowd, appropriate structure

for the submission of solutions, and feedback to

submitted tasks.

Micro-tasks refer to the personal demand created

by requesters. The configuration of the micro-task

request is performed through CS parameters such as

specific subject, constraints, quality issues, monetary

reward, and target worker. Participants reported the

unavailability of documentation on tasks

requirements, specific business rules and scarce

context about the tasks. In some projects,

participants report the need for more information to

complete and deliver distributed tasks. They also

need to achieve the requesters’ expectations. Stol

and Fitzgerald (2014) also describe the task quality

assurance challenge under the requester perspective.

Testers describe the lack of information on

reported bugs. A tester participant, who has worked

on the Crowdtest Brazilian platform, describes how

errors are reported.

“Everyone reports errors in the same place and

the tool does not return with the reason why the

reported error was not considered as an error to the

platform (client), and we were no guarantee"

(Participant 5).

On the other hand, the requester describes the

quality of crowd deliveries. In his opinion, the lack

of qualified works is the main problem.

"Sometimes there is a lack of professionalism in

this environment. Most of the time filters are not

performed by the platforms to allocate the profile of

qualified members” (Participant 14).

A buyer, who has bought services on

WeDoLogos platform since 2010, says that in this

field there are few professionals with specific skills.

For him, the majority of workers are not skilled.

Although the quality of the service is considered

good for him, he receives few proposals to perform

the work (approximately 10 proposals per project).

On the other hand, another participant describe a

project in which he received more than 200

proposals. However, the quality of the delivery by

the crowd was very low. Only 10% of solutions

received from the crowd could be take. For some

projects, he has considered to return to the

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

486

traditional model of hiring service to meet his needs.

"I believe that for small businesses the

crowdsourcing model can work. The quality of

service is low and any delivery is part of the crowd"

(Participant 12).

 Processes – Lack of CS Processes

Since the crowd and requesters from Brazil have

little experience in software CS, we observed

challenges related to the immature adoption of CS

processes. Users are not familiar themselves with the

CS processes.

Participants from the buyer’s group report the

importance of having a process to support the

adoption of CS initiatives. Participants emphasize

limitations to adopt CS models in business. These

limitations are related to CS management processes

in terms of collectively coordination,

communication and collaboration.

“Brazil is a very conservative country and it

needs to prepare people to work with

crowdsourcing. It is necessary to have a strong

process behind the platform and people to support

business” (Participant 13).

“Crowdsourcing is difficult in our company

because it is necessary to have a visibility of tasks

(progress activities, for instance “to do”, “doing”

and “done”). We have a strong work process

orientated on quality and productivity.

Crowdsourcing could be a new direction to the

future but it requires a maturity level and another

mindset” (Participant 16).

A participant reports the use of CS model for

open innovation domain in Brazil. This project

started in 2006. Recently, this project has a

partnership with a CS innovation platform called

Innocentive. The requester describes how the CS

processes work in this project.

“I believe in the crowdsourcing model for my

business and my clients. The process created by

them has been used with good results for the clients.

The success in crowdsourcing projects depends on

some factors, such as number of projects,

participant engagement, ideas proposed by

participants, deployment of solutions, number of

participants, and concept of phases to

crowdsourcing projects (dependent on the

complexity and investment by client)” (Participant

18).

The participant also describes tools to support

the process. These tools are only adopted as an

interface between customers and network (crowd).

For him. It is more important to focus on how the

project design is planned and executed by the

workers.

 People – Cultural Barriers

Brazil has particular laws and legislation. The

country is very concern on labor law and

bureaucratic issues.

Brazil is a conservative country in terms of

distributed software development. Usually, Brazil

receives a lot of outsourcing demands but it is not

used to outsource. The most of the time, Brazil is

much more a supplier than a consumer.

The country is a special case in terms of software

development. Brazilians companies and labor

participate in global CS marketplaces, but it also

“plays in their own sandbox”. It may happen

because of language issues. Portuguese is the official

language in Brazil and the majority of the population

do not speaks another language.

Another fact is that Brazilians prefer to have a

permanent job besides to be to have freelancer job.

In addition, participants report do not trust in having

a virtual contract of work. They are very

conservative people.

One of our findings shows a contradiction

between the collaborative culture of Brazil and the

competitive environment in the economy. Brazil still

trying to introduce a new model of work.

“Companies control logic and create an

antagonism in relation to digital networks.

Companies have difficulty in changing their business

to work in a complex world. Companies want the

control of their idea” (Participant 8).

Brazilians companies do not understand the

distributed workforce within a collaborative system.

They have concerns about intellectual property and

business rules. According to the participant 18, CS is

much more talked about than consumed in Brazil.

6 DISCUSSION

CS is an emerging topic in software industry. It

provides a new approach for companies involve their

workers with innovation activities. However, despite

the positive effects, many challenges are identified

for CS practice. The Brazilian IT industry has

specific challenges that make this country different

from others.

We found that the main challenges faced by

Brazilians practitioners are concentrate in three areas

Tasks, Processes, and People. We also found eleven

enablers and blockers factors for the CS practice in

Brazil.

Tasks are difficult to manage in CS

environments. Requesters expect to receive a task

with certain level of quality. However, in some cases

Software Crowdsourcing Challenges in the Brazilian IT Industry

487

the delivery do not attend the expectations of the

requester. According to (Li et al., 2013), one of the

most problems in CS is quality control to ensure the

quality of results. The factors of quality for CS tasks

are the number of participants, tasks assignment to

workers according to their individual expertise, and

the reward amount. The inappropriate worker-task

matching may harm the quality of the software

deliverables (Mao et al., 2015).

On the other hand, workers report the lack of

information that can result in a task delivered with

low quality. When workers understand what

information is need for the task specification, it will

be possible to provide solutions to problems that

meet customers’ needs. However, to Wu et al.

(2013) the vendor selection has a direct correlation

with the quality of an outcome. Workers are

attracted by an open call format rather than being

selected. It encourage the non-skilled workers to

participate. The list of countries with higher level of

active members shows Brazil on 14º ranking

position between 50 countries. Country rankings are

based on an aggregation of the TopCoder members

within a particular country that have competed

within the last 180 days

(https://community.topcoder.com/).

The lack of processes definition is another

challenge faced by the Brazilian IT industry. To take

advantage the power of software CS, it is important

to define the proprieties, elements, responsibilities

and interaction flows of software CS as a new

software development process (Kittur et al., 2013).

While other countries like United States adopt and

invest in crowdsourced development processes,

Brazil adopts a timid posture regards to it. Brazil has

only two CS platforms to support software activities.

We believe that online markets for software CS

tasks such as software project development

activities, still have not received attention from

companies and workers. Currently, Brazilian

platforms do not meet the requesters’ expectations.

According to our findings, Brazilian platforms

support only few types of activities.

Portuguese is the official language in Brazil.

English is the global language of business. The

majority of Brazilians speak only Portuguese. Thus,

the Brazilian community face difficulties to use

international CS platforms due to the language.

The intellectual propriety in software CS is a

world polemic question. In Brazil, this question is

amplified because the Brazilian laws and legislation

have characteristics of trade protectionism.

In literature, few authors explore region-specific

practices in CS software project. Europe and United

States are well populated with CS participants, but

that still does not say much about potential

differences in acceptance of CS across the globe.

In our study, some cultural aspects in CS are

presented. Brazilians are highly creative in their own

way, but the country is still underdeveloped in terms

of software CS. Cross-cultural differences in the

adoption of CS and open approaches to business are

still under-explored.

CS is a business concept that focus on the use of

intelligence, collectivity and volunteer knowledge to

solve problems, improve or develop new products,

technologies and services (Brabham, 2008).

Nevertheless, CS still not clearly understood by

many companies that can take advantage of this

concept. Brazilians are highly creative in their own

way, but the country is still underdeveloped in terms

of software CS. Cross-cultural differences in the

adoption of CS and open approaches to business are

still under-explored.

Under the Brazilian perspective, there are many

issues regards to CS elements’ (requester, platform

and crowd). To Carmel and Eisenberg (2006),

Brazilian national software builds pride inside the

Brazilian software community to develop software

under conditions of hardship. For these authors,

Brazilian software companies do not believe in its

capacity to create and offer job for other workers

and other markets.

Every country is unique and has its own specific

challenges when it comes to change the way of

work, like implementing software CS. This study

gives a starting point on region-specific practices in

crowdsourced software development.

6.1 Limitations of this Study

We are aware of the limitations of this study, since

our study does not seek to establish any causal

relationships, we do not discuss threats to internal

validity.

The qualitative analysis of the interviews was

performed by the authors together, which limited the

effects of possible researcher bias in the analysis.

We also adopted grounded theory to analyze

collected data using descriptive statistics and

techniques (Corbin and Strauss, 2008).

We interviewed Brazilians practitioners with

different experience levels. The imbalance

experience could have influenced positively or

negatively in our findings. Unfortunately, the

identified findings are not exhaustive. They only

represent those that have been experienced and

observed by our participants. We have carefully

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

488

selected the participants in this study.

7 CONCLUSIONS AND FUTURE

WORK

In this study, we investigate how the software CS has

been developed in the Brazilian IT industry. We

found that Brazil is very conservative and moderate in

terms of adoption of software CS in IT scenario.

Brazil has a few requesters’ demand, both in national

and international CS platforms. Also, we found few

CS platforms to support the Brazilian market.

Although Brazil is an important software market

and one of the most important emerging economies,

we are not surprised with our findings. Brazil has a

weak participation in software CS initiatives.

Our results show that the main challenges in

software CS in Brazil are related to Tasks,

Processes, and People. We believe that for the

software CS to work effectively, it is important to

better understand the issues related to the three CS

elements (Crowd, Requesters, Platform). Given this

perspective, the research we present here is of value

to both to academic and industrial communities. We

also believe that these findings are particularly

important from the Brazilian perspective; however,

they also help add to the body of evidence in the

field of software engineering.

In spite of the challenges, we believe that CS will

get new labor markets in future. Markets that are

disrupted, like the Brazilian software market, shall

see changes in the types of tasks that are currently

being performed. Also, software development

though CS may it will help to alleviate the Brazilian

cultural limitations mentioned in this study.

More empirical research is needed on how to

develop software CS in Brazil. We plan to follow

our case organizations to see how they minimize the

identified challenges, collecting more detailed data

about their software practices, as well as by

additional interviews.

ACKNOWLEDGMENTS

This work is partially funded by CNPq (312127/2015-4).

REFERENCES

Âgerfalk, P. J., Fitzgerald, B., and Stol, K. J., 2015.

Software Sourcing in the Age of Open: Leveraging the

Unknown Workfoce. Springer.

Brabham, D. C., 2008. Crowdsourcing as a Model for

Problem Solving: An Introduction and Cases. In

Convergence, 14, 1.

Carmel, E., and Eisenberg, J., 2006. Narratives that

software nations tell themselves: An exploration and

taxonomy. Communications of the Association for

Information Systems, 17(1), 39.

Chuene, D.; Mtsweni, J., 2015. "The adoption of

crowdsourcing platforms in South Africa," in IST-

Africa Conference, 2015, pp.1-9.

Corbin, J. M. and Strauss, A. L., 2008. Basics of

qualitative research: Techniques and procedures for

developing grounded theory, 3rd edition. Sage.

Forbes Magazine, Brazil: Profile. Available at

http://www.forbes.com/places/brazil/. Access on

November 2015.

Howe, J., 2008. Crowdsourcing: Why the Power of the

Crowd Is Driving the Future of Business. In Crown

Business.

Kittur, A., Nickerson, J. V., Bernstein, M. S., Gerber, E.

M., Shaw, A., Zimmerman, J., Lease, M. and Horton,

J. J., 2013. The Future of Crowd Work. In Proc.

CSCW. ACM.

Lakhani, Karim, David A. Garvin, and Eric Lonstein,

2010. TopCoder (A): Developing software through

crowdsourcing. Harvard Business School General

Management Unit Case 610-032.

LaToza, T.D., Towne, W.B., van der Hoek, A. and

Herbsleb, J.D., 2013. Crowd Development. Proc. 6th

CHASE Workshop. San Francisco, CA.

Li, K., Xiao, J., Wang, Y., and Wang, Q., 2013. Analysis

of the key factors for software quality in

crowdsourcing development: an empirical study on

TopCoder.com. In 37th Annual Computer Software

and Applications Conference, pp. 812-817.

Mao, Ke, et al., 2015. A Survey of the Use of

Crowdsourcing in Software Engineering. RN 15:01.

Olson, D. L., and Rosacker, K., 2013. Crowdsourcing and

open source software participation. Service

Business, 7(4), 499-511.

Prikladnicki, R., Machado, L., Carmel, E., & de Souza, C.

R., 2014. Brazil software crowdsourcing: a first step in

a multi-year study. In 1st International Workshop on

Crowdsourcing in Software Engineering(pp. 1-4).

Saxton, G., Oh, O., Kishore, R., 2013. Rules of

Crowdsourcing: Models, Issues, and Systems of

Control. Information Systems Management 30, 1–13.

Stol, K. J., and Fitzgerald, B., 2014. Two's company,

three's a crowd: a case study of crowdsourcing

software development. In Proceedings of the 36th

International Conference on Software

Engineering (pp. 187-198). ACM.

To, Wai-Ming, and Linda SL Lai., 2015. Crowdsourcing

in China: Opportunities and Concerns." IT

Professional 17.3: 53-59.

Wu, Wenjun, Wei-Tek Tsai, and Wei Li., 2013. An

evaluation framework for software crowdsourcing.

Frontiers of Computer Science Journal 7, no. 5: 694-

709. Publisher Springer Berlin Heidelberg.

Software Crowdsourcing Challenges in the Brazilian IT Industry

489