software product development process model: case studies of mobile software companies

22
Software Product Development Process Model: Case Studies of Mobile Software Companies Tuure Tuunanen , Marianne Vainio Helsinki School of Economics, P.O. 1210, FIN-00101 Helsinki, Finland Abstract: We present a preliminary model for Customer-Oriented Software Product Development. We claim that there is a gap in the literature to integrate R&D and marketing in developing new software products. We reviewed the literature in new product development and information systems and find that a rich process discourse exists but lacks the answer to our problem. Hence, we conducted case studies of Finnish software companies to understand the development process. We found that by integrating the established StageGate model to a component based software development model the case companies were able to manage business and technology related risks associated with product development. Keywords: Software, New Product Development, Software Product Development, Case Study Corresponding author. Tel +358 40 544 5591, Fax +358 9 431 38700. Email addresses: [email protected] (T. Tuunanen), [email protected] (M.Vainio) Tel +358 50 5294723, Fax +358 9 431 38700. 1

Upload: others

Post on 12-Sep-2021

8 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Software Product Development Process Model: Case Studies of Mobile Software Companies

Software Product Development Process Model:

Case Studies of Mobile Software Companies

Tuure Tuunanen∗, Marianne Vainio†

Helsinki School of Economics, P.O. 1210, FIN-00101 Helsinki, Finland

Abstract: We present a preliminary model for Customer-Oriented Software Product

Development. We claim that there is a gap in the literature to integrate R&D and

marketing in developing new software products. We reviewed the literature in new

product development and information systems and find that a rich process discourse

exists but lacks the answer to our problem. Hence, we conducted case studies of Finnish

software companies to understand the development process. We found that by integrating

the established StageGate model to a component based software development model the

case companies were able to manage business and technology related risks associated

with product development.

Keywords: Software, New Product Development, Software Product Development, Case

Study

∗ Corresponding author. Tel +358 40 544 5591, Fax +358 9 431 38700. Email addresses: [email protected] (T. Tuunanen), [email protected] (M.Vainio) † Tel +358 50 5294723, Fax +358 9 431 38700.

1

Page 2: Software Product Development Process Model: Case Studies of Mobile Software Companies

1. Introduction The development of software to customers and or end-users outside of the organization

has been found to be a difficult task. Technological solutions do not usually limit the

R&D process, but usually it has been a case of market failure, when new software

products or services have been launched to the market place, like with the mobile Internet

in Europe in late 1990s (Peffers and Tuunanen, 2002) and later again with the Japanese

3rd generation mobile service FOMA (Sigvartsen, 2001).

Within marketing science, the dilemma has created a thriving discipline, new product

development (NPD), to focus on the problems related to the area. Researchers have

argued that integration of R&D and marketing functions is one of the key issues to

consider for having a successful product or service (Barczak, 1995; Gupta, 1985; Souder,

1988). However, the integration of functions has not been seen as the only required

solution, only a part of the whole. Other researchers have emphasized the relevance of

well-structured process (Cooper, 1990). Cooper (1990) has argued that by producing a

tightening pipeline of NPD with well-defined milestones, a firm is able to minimize the

market risk involved. In addition to ideas of having well-defined process, researchers

have argued for the importance of listening to the voice of the customer (Cooper, 2000;

Hauser, 1988).

In information systems (IS), researchers have struggled to solve the problems of

developing software to serve members of the organization. Lately, the attention has

turned to end users and other external end-users of information systems products or

services (Tuunanen, 2003). The research has focused on the areas of process

improvement, selection of right techniques, and during the last decade involving better

information management in the network that brings the voice of customers to the process.

Especially, an incremental approach of developing software and continuously involving

the market and customer information to the process has been argued to be the most

effective (Boehm, 1988).

We present a gap in the literature in integrating the R&D and marketing in software

development (SD). The gap is described in figure 1. We argue that even though the both

of the disciplines have emphasized the importance of well-structured process, we still

2

Page 3: Software Product Development Process Model: Case Studies of Mobile Software Companies

lack an integrated approach for software product development. Hence, we follow

Nambisan’s (2003) argumentation on using IS as a reference discipline for New Product

Development, and agree with Nambisan and Wilemon (2000) that a combination of

knowledge from the NPD and SD disciplines could provide a solution for the presented

gap.

”The GAP” Information New Software Systems Product Product

Development Development Development

Figure 1. Positioning the research.

Hence, we argue that by integrating the process thinking of new product development and

information systems development (ISD) we can fill the gap of the integration without

losing the voice of the customer (Griffin and Hauser, 1993). We base our argument on

two case studies of Finnish small and medium sized software companies. We present that

the marketing and R&D activities already exist in software organizations and our purpose

is to study how these activities are integrated from the process perspective. Based on

these results, we present a preliminary framework for customer-oriented software product

development.

The structure of the paper is following. We first go through a review of NPD literature

from the perspective of the development process. After this, we review how software

development literature has faced the process management challenges. Thirdly, we present

the limitations of the approaches and discuss the integration of the processes. Then, we

describe our methodology and describe the selected case companies. We present our

findings through a preliminary model for Customer-Oriented Software Product

3

Page 4: Software Product Development Process Model: Case Studies of Mobile Software Companies

Development and describe how the case companies have implemented it. After the

discussion section, conclusions and the future research topics are presented.

2. New Product Development Process The NPD literature has extensively argued for the integration of R&D functions and

marketing to have a successful new product development environment for a firm (Souder,

1988). Griffin and Hauser (1996) have found two distinguishing topics in the NPD

literature. The first is the physical integration of people by location and organizational

structure. The second is the process perspective. On the basis of the empirical studies, the

authors point out the importance of the well-structured process and list some good

examples, such as Cooper’s StageGate model (Cooper, 1990). Additionally, Voss (1985)

has studied the critical success factors for the development of software applications and

argued for good process management. In this study, we focus on the process aspect of

NPD.

As going through the literature, we found the StageGate model (Cooper, 1990) to be one

of the most coherent for examining the whole NPD process. The process is a conceptual

and operational road map for moving a project from idea to launch. The purpose is to

improve effectiveness and efficiency of the product. Even though, the StageGate model

may not provide solutions for every problem (Griffin and Hauser, 1996), the other

models, for example Quality Function Deployment (Haag et al., 1996) and the user

involvement in the process, concentrate only on some of the phases of the whole. We did

found other models within the NPD literature, such as Ambler and Styles’s (Ambler and

Styles, 1997), but we did not find these to be relevantly different than this widely used

StageGate model introduced by Cooper (1990).

Figure 2 illustrates the StageGate model. The process description is simplified to five

stages that present the major events in the NPD process (Cooper, 2000). Stages are cross-

functional, thus there is no R&D or marketing stage. Rather, each stage gradually drives

down the business risk by gathering vital information on technology, market, finance and

operations. The gates between stages serve as the quality-control checkpoints.

4

Page 5: Software Product Development Process Model: Case Studies of Mobile Software Companies

STAGE 1Customer value criteria

Preliminary market, technical& financial assessments

Action plan

STAGE 2User needs & wants study

Competitive analysisValue proposition defined

Technical feasibilityOperations assessment

Product definitionFinancial analysis

STAGE 3Preliminary market, technical

& financial assessmentsAction plan

STAGE 4Extended in-house testing

Customer field trialsTest market/trial sellFinalized launch and

operations plansPost launch & life cycle plans

STAGE 5Market launch & roll-out

Full production/operationsSelling begun

Results monitoringPost lauch & life cycle plans

under way

GATE 1Idea screen

GATE 2Second screen

GATE 3Decision to develop

GATE 4Decision to test

GATE 5Decision to launch

SCOPING BUILDING BUSINESSCASE DEVELOPMENT TESTING & VALIDATION LAUNCH

Figure 2. The Stage Gate model, modified from Cooper (2000).

Improved design and execution of NPD process minimizes the market risk by building in

the voice of the customer and focusing on the market performance of a product. For this

purpose, Cooper (Cooper, 2000) presents several key success factors that as incorporated

into the process minimize the product failure rate in the markets. Developing

differentiated, superior products means conducting a user needs-and-wants study early in

product development process in order to identify the components of a superior product. In

highly networked software industry, this might also mean tight cooperation with the

network parties who provide the knowledge of end-users. Furthermore, allocating more

time and resources on the activities preceding the design and development means

undertaking market and competitive analysis, concept testing and feasibility technical

assessments. Before the development stage, the StateGate process drives the focus on a

product definition which includes a target market definition, a product concept

comprising benefits to be delivered, the positioning strategy and more specific product

features and attributes. Additionally, the StageGate process gives attention early to the

market launch. One of the most important aspects of the StageGate process is that it

involves all process stakeholders, such as the top management, sales and engineering

functions, customers, suppliers, and partners in the product development.

In the following section, we examine how IS discipline has seen the development process

of new products and services.

3. Software Development Process In IS development discipline, the process development has been moving on the timeline

starting from disorganized ways of working and developing towards more organized

5

Page 6: Software Product Development Process Model: Case Studies of Mobile Software Companies

ways. Also the shift from sequential process models to evolutionary ones is evident. We

start our review with the oldest of development methods, code-and-fix, which includes

two steps in the process: 1) write code, and 2) test and fix it (Boehm, 1988). The resulting

product is reworked until it is in an acceptable stage (Boehm et al., 1984). The method

contains many problems with starting poorly understood requirements and problematical

structure of coding resulting high costs when re-work is needed later on (Boehm, 1988).

The first published model ‘waterfall model’ for software development was derived from

other engineering processes (Royce, 1970). It is a systematic, sequential approach, in

which each stage requires well-defined input, and results in well-defined outcomes. The

IS product is not delivered until the whole linear sequence has been completed. The

problems related to this linear model are stagnant requirements and badly structured

programming, which have been tried to avoid by including overlapping between stages

(Sommerville, 2001).

In contrast to the waterfall model, an evolutionary approach to software development is

often more effective in producing systems which meet the immediate need of customers

(Sommerville, 2001). In the evolutionary approach, the specification is developed

incrementally and reflecting users’ understanding of the problem (Sommerville, 2001).

Through this discourse, process models based on reuse-oriented development,

prototyping, increments, and spiral development have evolved. However, the

evolutionary approach has some problems. It is only an abstract representation, and thus

the process is not visible and systems often lack proper structure why special tools and

techniques may be required (Sommerville, 2001). Therefore, the evolutionary

development suites best small systems with a fairly short lifetime, as the problems

become acute in large systems.

In the evolutionary approach, reuse is often seen as essential for rapid system

development. In the past few years, an approach to component-based software

engineering has become increasingly popular. Also SEI has taken this into account and

has targeted the latest CMMI version 1.1 to integrated product and process development

(Software Engineering Institute, 2002). The engineering process areas of the CMMI

6

Page 7: Software Product Development Process Model: Case Studies of Mobile Software Companies

support component-based software development by emphasizing concurrent development

and focusing on all phases of the product’s life.

Furthermore, there is need to support process iteration where parts of the process are

repeated as system requirements evolve. The models for this purpose comprise

incremental development and spiral development. In incremental development software

is developed in small but usable pieces that can be delivered to a customer. Each

increment is an operative subset of the system and builds on the increments that have

already been developed (Pressman, 2000). Detailed design, coding, and testing occur

within these separate stages (McConnell, 1986). The process of development, validation

and integration continues until the delivered increments form a complete product.

In turn, the spiral development (Boehm, 1988) describes the software process as a spiral,

where each of the loops can be considered to represent, for example, one fundamental

step of the software process. Thus, the innermost loop might be concerned with

requirements engineering, the next with risk analysis that is followed by prototyping and

validating the found requirements. After one round the loop continues with a more

detailed abstraction level ending to the detailed design process, and finally

implementation. The spiral model assumes a risk-driven approach to the software

development rather than a primarily document-driven (waterfall) or code-driven

(prototyping) approach. Each cycle incrementally increases the degree of the system

definition and at the same time decreases the degree of technology risk (Boehm, 1988).

Next, we discuss the limitations of the NPD and IS development approaches.

4. Limitations of the Approaches Often the reason for the market failure of high technology products is the lack of

marketing knowledge and skills (Calantone and Cooper, 1981; Peffers and Tuunanen,

2002; Sigvartsen, 2001). Returning this, the StageGate model builds the success into the

process by designing a certain stage to gather information in order to lower the business

risk. Each stage cost more than the previous one, so that the model is based on

incremental commitments.

7

Page 8: Software Product Development Process Model: Case Studies of Mobile Software Companies

In IS field, the process management focuses on identifying risks and drawing up plans to

minimize their effect on the project. The IS literature specifies several main causes for

uncertainty in software development, for example people management (Sommerville,

2001), limited learning and knowledge management skills of software organization

(Mathiassen et al., 2001), failures in software cost estimation (Boehm et al., 1995), and

improper quality management. Generally, the recognized risk categories are project risks

that affect project schedule and resources, and product risks, that affect quality or

performance of the software being developed. Business risks affect the organization

developing the software (Sommerville, 2001), which indicates that business risk

management is not incorporated as such into software process models.

Presenting the limitations of the existing approaches, we assert that the integration of

both of the disciplines provides answers for creating customer-centric software products.

Because of the special nature of software product development for service providers of

the consumer markets, there is a need to justify the innovation process stage decision by

continuously evolving discourse by the involved parties. Then, gathered information

drives down both the technical and business risk.

5. Case Studies in Software Industry

Methodology We followed Klein and Myers (1999) in trying to understand the complex and fast

moving IS research topic. Hence, we chose a qualitative research approach. For the

purpose of developing the preliminary framework for customer oriented software product

development, we studied two new software product companies in Finland. We selected

the case study method because of its major strength in using many sources of evidence

and data collection methods (Eisenhardt, 1989; Wynn, 2001; Yin, 1994). Multiple

sources of evidence and multiple methods provide a better validity for findings. Primary

data collection method used in this study was theme interviews. We complemented this

by acquiring other written documents and information from the case companies. The

selected two cases are based on a convenience sample. However, we claim that they can

be used as illustrative examples of small and medium size companies developing mobile

software products.

8

Page 9: Software Product Development Process Model: Case Studies of Mobile Software Companies

The data collection was conducted in spring and summer 2003. At first, we gathered

background information of the companies, such as financial information from 2000 to

2003, the future estimates for 2004, a strategy and an operating plans, an organization

structure, and product white papers, to better define the units of analysis. Then, we

conducted theme interviews to have an insight of business and technology management

in the companies. These resulted up to a hundred pages of research notes1.

We chose two types of participants for the theme interviews. The CEO or the head of

business development answered the questions which focused on management, business

development and marketing issues. The head of software development or the head of

technology gave us details concerning software development practices. Finally, we tried

to gain insight of the activities of NPD and software engineering on the basis of our

theoretical discussion.

The Selected Companies The selected companies operate in the mobile and wireless markets and the products are

used by wide audience end-users, and thus the cases are good and suitable examples to be

presented in this paper. Both companies have a strong growth initiative, which has also

been a reason for venture capitalists to make investments in the companies. The

companies differ from one another in terms of the development stage of a business; one

(Multimedia) is a start-up established 2001 that has just recently started to

commercialize the first version of a product, and the other (Messaging) has already made

four versions during six financial years. In 2003, the turnover of Multimedia was 2.5

million euros and it had 62 employees. The comparable figures of Messaging were 5

million euros and 65. By studying these two chosen companies, we were able to analyze

the integration of marketing and R&D in two slightly different business contexts. The

selected cases were named as Multimedia and Messaging after their product offerings.

Messaging has a middleware solution that enables mobile operators to increase the

control in wireless content business. Their product allows mobile operator and service

providers to manage the process of provisioning, delivering, and charging of a service

portfolio. It includes standard tools for integration to billing systems, subscriber, location, 1 The detailed themes of interviews are attached in Appendix 1

9

Page 10: Software Product Development Process Model: Case Studies of Mobile Software Companies

and service databases. Multimedia’s product, the Multimedia Framework, is a complete

system incorporating all technology required in implementing mobile video applications.

It consists of modules implementing the elementary functions such as video/audio coding

and protocols and of a virtual backplane providing the signaling and control for the

applications. The framework supports a variety hardware and software platforms and is a

suitable solution for mobile devices ranging from cellular phones to digital cameras.

Customer segments of the companies consist of component suppliers, device and phone

manufacturers, operators, and service providers. From the competitive perspective, both

companies try to operate as a pioneer in their specific field. Both of them also support

newest technological innovations for the sake of the future roadmap but as the chairman

of the board of Messaging summarizes: “We do not take risks as regards immature

technological trends.” Also the CEO of Multimedia concludes: “The failures in product

functionality would mean the loss of customers and the end of business, why we use only

proven technology.”

The business model2 of both companies is based on licenses, consultation and

implementation services, and product support fees. Messaging’s business is divided

equally into three parts, but Multimedia’s turnover mainly consists of license revenues

and product royalties. Multimedia’s business is only emerging, and for this reason, there

is yet no need to establish support services. Messaging has already three profitable sales

channels, but Multimedia has only closed the first deals with potential partners in this

year. Both of them have ten to twenty main competitors around the world.

6. Findings We conceptualized our findings in figure 3. The new product development process is

illustrated with the help of the StageGate model (Cooper, 1990) and it focuses on product

characteristics from the business and marketing perspective. The software development

process (Sommerville, 2001) focuses on the technology risk. In the chosen companies,

both of the processes are based on an evolutionary approach meaning that as product

2 The business model discussion is a thriving area in IS, but not in the scope of our paper. More information on business models can be attained: Timmers, P. (1999) Electronic commerce : strategies and models for business-to-business trading. Chichester ; New York: Wiley.

10

Page 11: Software Product Development Process Model: Case Studies of Mobile Software Companies

definitions become sharp and completed new software features are developed and

implemented by increments. In following we go through the presented model. We

illustrate the model by describing first how the NPD processes of the case companies

from marketing perspective. Then we continue with the software development processes

of the companies. Finally, we describe the integration of the StageGate (Cooper, 1990)

and software development processes (Sommerville, 2001) in the case companies.

Figure 3 Customer-Oriented Software Product Development Model

New Product Development From marketing perspective, the product innovation of the companies is controlled by a

NPD process. In this study it is conceptualized with the help of the StageGate model. In

both of the companies, the cross-functional and parallel stages are preceded by

Go/Kill/Hold/Recycle decision points, which support the evolutionary nature of the

process.

Scoping. Second screen. “The management group decides whether the idea merits any

work. There is no a certain procedure for an agenda, on the contrary, the meetings are

simple and relaxed. For example, last time we rented a cottage in Lapland for a weekend

and did not come back until we had finalized the plans,” said the CEO of Multimedia.

11

Page 12: Software Product Development Process Model: Case Studies of Mobile Software Companies

“Information for idea screen and scoping is collected by communicating with our

network partners such as component suppliers, phone manufacturers, operators and

technology partners,” he continued.

Quite on the opposite, Messaging screens various sources, such as customers, partners

and competitive environment through competitor offers and market follow-up. Customer

feedback is collected through a digital channel and stored into a database. The responses

are given priorities according to three levels and the results are reviewed regularly by the

management group and the product steering group. The groups also deal with discussions

with customers and partners. During recent years, the company has also conducted a

market research for their main customers. “We wanted to study the current market

situation, the future trends, and timing of those trends,” said the CTO of Messaging, “On

the basis of the results, we adjusted our plans. We will repeat the research yearly.”

Whether the idea seemed to justify more extensive investigation, the management group

decided to explore the business case. In both of the companies, there was no a clear

boundary between the first and the second stage.

Building business case, Decision to develop. A common method to complete a product

definition in both companies was to review plans with customers and network partners.

“We throw the ball to our customers and partners and refine the idea with them. Actually,

we have formed few partnerships for the sake of getting a better access to this market

knowledge,” said the CEO of Multimedia. Also salesmen of Messaging asked customers:

“Would you like this kind of feature..?”

“We make these normal analyses such as product life cycle analysis, cost estimations,

market research, and sales estimates,” stated the CEO of Multimedia. However, one of

the most significant outcomes of both companies was a product positioning that related

the value added to the ones offered by partner network. “That way, we don’t step on

anyone’s toes and strengthen our position compared to competitors’,” said the CEO of

Multimedia.

Overall, there was no significant variation between the explicit results of this stage in the

companies. Only existing difference was the amount of information processed, on which

the decision to develop was based. In other words, Messaging examined the business case

12

Page 13: Software Product Development Process Model: Case Studies of Mobile Software Companies

from several angles and with numerous iterations. The management group went through

the feasibility analysis several times. The business case was verified sound if the project

aligned with the business’s strategy and the product provided distinct value to customers.

Development, Decision to test. After examining the technological feasibility of the

solution, companies started to the develop process, which is described in detail in the next

section. Hence, we move to the next phase: decision to test.

There was no a clear point when to make a decision to test, rather the procedure was

evolutionary. As product components were completed, they were gradually transferred to

in-house testing and customer trials. And, as both the CTO of Messaging and the CEO of

Multimedia noted: “When development begins, the salesmen start to sell the product.”

Testing and validation, Decision to launch. In both of the companies this stage of NPD

process was highly overlapping with the previous one. The responsibility of software

testing was carried by the software engineers. However, after verifying technological

feasibility, the companies tested the usability, appropriateness and efficiency of product

features by delivering demo versions to certain customers. “Yet, installing demos require

the same amount of work than normal product implementation, so it is not very easy way

of doing customer trials,” the CEO of Multimedia pointed out.

“Development and validation activities are carried out concurrently with rapid feedback

across these activities,” described the head of software engineering of Messaging, “And

finally me make the decision to start selling the update.” The extent of testing and

validation of product innovation was larger in Messaging because they had the possibility

to exploit more parties in feedback gathering. “Our customer base is small, so we deliver

only two or three demos. But it is enough for us,” ascertained the CEO of Multimedia.

Launch, Post-launch review. “There is no such a thigh as a product launch,” said

respondents from both of the companies. However, “We try to set the time of the market

launch with the most significant industry fairs, such as GSM World or Cannes,”

explained the CTO of Messaging. This was also an effort in Multimedia. In other

respects, salesmen just begin to sell a new update or a product version. However,

delivering updates was performed carefully. “Customers can’t digest too heavy updates,”

13

Page 14: Software Product Development Process Model: Case Studies of Mobile Software Companies

concluded the CTO of Messaging. Finally, the results of the launch were monitored and

used for the future developments.

Software development In both of the companies, the software development was based on reusable components.

“This is incremental development,” the head of technology of Messaging said, “When

there is a new feature, we complete a new increment iteratively. An each increment

provides a subset of the total system and integration with existing increments.” In both

companies, reuse was considered essential for rapid system development. “Mainly, our

software development is just making updates of existing components,” described the head

of technology of Messaging, “The market develops so fast that for example capacity

requirements have multiplied. Actually, we have developed the central parts of our SMS

gateway tens of times.” Sometimes these components are systems in their own right that

provide specific functionality such as picture formatting, message sending, or traffic

monitoring.

The component based development influenced the software development process models

of the companies. The CMMI (Software Engineering Institute, 2002) software

development model is used to illustrate this in Figure 3. It includes five iterative and

looping phases: 1) requirements development, 2) technical solution, 3) product

integration, 4) verification, and 5) validation. Due component based development, the

finalization level of a product was high in both companies. Generally, the final products

consist of two to ten application modules and from twenty to thirty platform components.

When the customer places an order, the product is already 90% ready-made. 10% of the

work consists of integrating the components into a final product. There were no major

differences in the technological assembly between the companies. However, there was a

difference in how they have implemented CMMI model.

The major difference was in the requirements development phase that was closely linked

to scoping and business case building in NPD process. Messaging’s requirements

development approach was clearly more structured. In turn, Multimedia invested more in

exploring product requirements and understanding the customers’ needs. Messaging’s

requirements development (e.g. specification) was comparable with other common

14

Page 15: Software Product Development Process Model: Case Studies of Mobile Software Companies

processes models, like waterfall. However, the intermediate stages include also

component analysis and integration to the system originating from CMMI. “The process

reflects the incremental structure of the product, “said the head of technology. In

addition, Messaging had divided the development organization into five teams that focus

on certain components or stages of the process.

On the contrary, software engineers of Multimedia worked more with customers and

partners to explore the feasibility of requirements and to deliver final system. Their

process model is flexible and has only abstract stages of design, development,

verification, integration, and validation. “Small chaos is necessary”, said the CEO of

Multimedia, “But documentation and version management are performed very properly.”

Both companies produced a new version of the product every year.

Messaging systematically collected information from several sources, such as customers

and partners, that made the process more structured. “Every requirement comes from the

customers,” described the CTO of Messaging. Unlike Messaging, Multimedia’s offering

is based on technology push because of the emerging market. The company had no

systematic information management to forecast end-users’ needs and wants, and as the

CEO of Multimedia summarized: “The product definition is based on intuition of the

management group and some discussions with network partners. We do not have any link

to end-users.”

Rapid prototypes (McConnell, 1986) were common in both companies. “We develop a

prototype, but actually it serves more salesmen than software engineers. It is not

productive to present a piece of code to anyone, but we have to conceptualize the product

in order to collect the initial customer feedback. Otherwise, it is difficult to transmit an

image of the product concept,” said the CTO of Messaging, “We experiment the

prototypes especially when users’ needs and wants are not well understood.”

The Integration of Software Development and Marketing – Requirements Management We found that the integrative thing between software development and marketing (NPD

process) was requirements management. We consider it best describes the

transformation of customer needs to technical specifications that was occurring in the

15

Page 16: Software Product Development Process Model: Case Studies of Mobile Software Companies

case companies. Requirements management was the glue that attached together the two

processes and the people. The customer focus was derived from NPD work, whereas

software development provided the actual means to build the product.

The integration of the two processes was characteristic in two parts of the model. Figure

3 describes these as double arrows. First integration was evident during scoping and

business case building and requirements development. In both cases this loop, or even a

pump, provided the needed information for NPD and software development processes to

continue forward. Management tried to listen to customers and, in turn, software

engineers start to developing technical requirements as stated by product definitions.

“All of the product requirements come from the ones who talk with customers, we

(software developers) just concentrates on developing those features and

producing quality software,” stated the head of software engineering of

Messaging.

The second integrative loop was with testing & validation and launch and product

integration. The testing & validation were heavily overlapped with software

development’s product integration, verification and validation. The participants were not

able to separate these two in practice, but they saw these as one. Similarly, the launch of

the product was tightly integrated to the previous, which, of course, is hardly a surprise.

In following we describe how the organizations dealt these two loops, or pumps, in

practice.

In managerial level, there was a difference between the amounts of integrative functions

in the companies. Whereas Messaging had clearly more managed functions, which are

instantiated by an individual project, group, or organizational function, Multimedia had

only one. The function existing in both companies was the management group. “The new

product development is purely a synonym to the management group work,” was the case

in both companies. The management groups of the companies consist of delegates from

all the main functions.

Messaging had specific product development functions, such as a product steering group

that consists eight to ten members from sales and engineering, and with which Messaging

16

Page 17: Software Product Development Process Model: Case Studies of Mobile Software Companies

manages the co-operation between sales and engineering. The head of business

development and the leader of the group concluded that

“The product steering group has regular meetings and we gather information from

various sources. The sources are for example a yearly customer survey and

customer feedback collected by sales team. Finally, we analyze the data and

present the results to the management group,”

In Messaging, customers were served by account managers from both sales and

technology. There were also two product managers led by a product director operating

between engineering and sales and providing resources to development teams.

In contrast to Messaging’s product-lead organization, Multimedia’s organizational

structure is divided simply into the management, sales, and engineering. About 50 out of

62 employees belonged to engineering. “We are just ramping up the sales for our

product,” said the CEO of Multimedia, “And now we try to hire a marketing director and

few more salesmen.” The management group that is the only integrative function

between sales and engineering, had meetings twice in a month. “There are also meetings

in every three months that serve as summary points,” added the CEO. “In the

management group, there are all important people of the company making decisions,” he

concluded, “Such as the CEO, the head of sales and marketing, the CTO, the head of

software engineering, and the head of business development. But being a small firm, we

communicate every day around the coffee table.” However, this tacit knowledge is

incomplete to manage integrative activities as being only partially performed and not

planned.

7. Discussion We argue that our preliminary model can provide a solution to the gap of integrating

R&D and marketing, particularly in software companies operating in mobile and

wireless. We found that the reviewed literature in NPD and in IS development gave us

good grounds to form a preliminary understanding of the process model for Customer-

Oriented Software Product Development Model.

17

Page 18: Software Product Development Process Model: Case Studies of Mobile Software Companies

We found that that there is no a single process for developing new software products, but

instead, the NPD process is tightly integrated to software development process. The

findings present how the framework is distinctly divided into parts accordingly how it

deals with project risks. The evolutionary NPD stream, based on the StageGate (Cooper,

1990), is argued to manage the business risks, whereas the incremental software

development process considers the technology risks. We also found that the NPD process

is mainly driven by the management group, and that the software development is

controlled through it. The product definition is developed in NDP process and transmitted

to software development, which focuses on requirements development, technological

solution, verification, component integration, and validation.

Even though the stages of both processes exist in both companies, we found that as a

company and the market matures, the process framework becomes more evident and

structured. Additionally, not so start-up companies can invest more resources in

completing a product definition, which is seen in Messaging’s case. Furthermore,

listening to the voice of the customer (Griffin and Hauser, 1993) became more systematic

in NPD in forms of a product steering group, product managers, customer surveys,

gathering feedback from lead users, and working with network partners. An additional

finding was that both companies considered that a small chaos in the development is

needed to create innovations. Although both companies drove towards organized

processes, it is necessary not to kill all ideas by structuring the processes too much.

8. Conclusion We found that our findings support the previous claims asserted within NPD literature

(Barczak, 1995; Gupta, 1985; Souder, 1988). The findings clearly showed the need of

integrating R&D and marketing to software development in order to decrease the

business and technical risks in NPD. However, we claim that no model has been

presented for managing this integration.

To fill this gap, we used Nambisan’s (2003) idea of using IS as a reference discipline to

extend the already created understanding in NPD. We found that the process literature for

developing information systems would add a contribution to the area. By doing case

studies we found that the integration of the process models in both NPD and software

18

Page 19: Software Product Development Process Model: Case Studies of Mobile Software Companies

development is essential. With this paper, we present a preliminary model of Customer-

Oriented Software Product Development that fills the presented gap.

We argue that our model brings contribution to researchers and practitioners alike. For

academia, we present a way to bring closer NPD and IS research communities and thus

create new knowledge in managing software product development. Secondly, the

proposed model can decrease the risk of development failure for IS products. For

practitioners, our research brings more understanding of how the product development

can be organized to decrease the development risk for software products. Our model

presents a way to manage the NPD process and how integrate it to the software

development. Furthermore, we claim that our model also gives a recommendation of how

to organize the NPD of a small and medium sized software company.

In the future, we seek to study more deeply how the integration phases of the model work

and how this relates to research in requirements engineering discipline. We also see that

more research should be done in validating our model. Our case studies show only

preliminary results. We plan to continue our research with a quantitative approach.

References Ambler, T. and C. Styles (1997) "Brand development versus new product development: toward a process model of

extension decisions," Journal of Product & Brand Management (6) 4, pp. 222-234.

Barczak, G. (1995) "New Product Strategy, Structure, Process, and Performance in the Telecommunications Industr," Journal of Product Innovation Management (12) 3, pp. 224.

Boehm, B. (1988) "A Spiral model of software development and enhancement," IEEE Computer (21) 5, pp. 61-72.

Boehm, B., T. E. Gray, and T. Seewaldt. (1984) “Prototyping vs. Specifying: A Multi-Project Experiment.” 7th International Conference on Software Engineering, 1984, pp. 473-485.

Boehm, B. W., B. Clark, and e. al. (1995) "Cost model for future life cycle processes: COCOMO 2," Annals of Software Engineering (1pp. 57-94.

Calantone, R. and R. G. Cooper (1981) "New Product Scenarios: Prospects for Success," Journal of Marketing (45) Spring 1981, pp. 48-60.

Cooper, R. (2000) "Winning with New Products Doing it Right," IVEY Business Journal (July-August 2000pp. 54-60.

Cooper, R. G. (1990) "Stage-Gate Systems - a New Tool for Managing New Products," Business Horizons (33) 3, pp. 44-54.

Eisenhardt, K. (1989) "Building Theory from Case Study Research," Academy of Management Review (14) 4, pp. 532-550.

Griffin, A. and J. R. Hauser (1993) “The Voice of the Customer,” in Marketing Science, vol. 12, pp. 1-27.

Griffin, A. and J. R. Hauser (1996) "Integrating R&D and marketing: A review and analysis of the literature," Journal of Product Innovation Management (13) 3, pp. 191-215.

Gupta, A. (1985) "The R&D-Marketing Interface in High-Technology Firms," Journal of Product Innovation Management (2) 1, pp. 12.

19

Page 20: Software Product Development Process Model: Case Studies of Mobile Software Companies

Haag, S., M. K. Raja, and L. L. Schkade (1996) "Quality Function Deployment Usage in Software Development," Communications of the ACM (39) 1.

Hauser, J. R. (1988) "The House of Quality," Harvard Business Review (66) 3, pp. 63-73.

Klein, H. K. and M. D. Myers (1999) "A Set of Principles for Conducting and Evaluating Interpretive Field Studies in Information Systems," MIS Quarterly (23) 1, pp. 67-93.

Mathiassen, L., J. Pries-Heje, and O. Ngwenyama (2001) Improving Software Organizations: From Principles to Practice. U.S.: Addison-Wesley.

McConnell, S. (1986) “Rapid Development-Taming wild software schedules,”. Redmond (WA): Microsoft Press.

Nambisan, S. (2003) "Information Systems as a Reference Discipine for New Product Development," Mis Quarterly (27) 1, pp. 1-18.

Nambisan, S. and D. Wilemon (2000) "Software Development and New Product Development: Potentials for Cross-Domain Knowledge Sharing," Ieee Transactions on Engineering Management (47) 2, pp. 211-220.

Peffers, K. and T. Tuunanen. (2002) “Using Rich Information to Plan Mobile Financial Services Applications with Maximum Positive Impact: a Case Study.” Mobile Round Table, Tokyo, 2002, pp. 31.

Pressman, R. (2000) Software engineering -a practitioners approach: McGraw-Hill.

Royce, W. W. (1970) "Managing the development of large software systems: concepts and techniques," Proc. IEEE WESTCON pp. (Ch. 3).

Sigvartsen, A. L. (2001) "NTT DoCoMo faces FOMA failure," InfoSatellite.com, http://www.infosatellite.com/news/2001/10/a151001foma_failure.html (14.10., 2003).

Software Engineering Institute, C. M. U. (2002) Capability Maturity Model Integration (CMMI), Version 1.1, Staged Representation. Pittsburgh: Addison-Wesley.

Sommerville, I. (2001) “Software Engineering,”, 6th edition: Addison-Wesley.

Souder, W. E. (1988) "Managing Relations Between R&D and Marketing in New Product Development," Journal of Product Innovation Management (5pp. 6-19.

Timmers, P. (1999) Electronic commerce : strategies and models for business-to-business trading. Chichester ; New York: Wiley.

Tuunanen, T. (2003) "A New Perspective on Requirements Elicitation Methods," JITTA : Journal of Information Technology Theory & Application (5) 3.

Voss, C. A. (1985) "Determinants of Success in the development of Applications Software," Journal of Product Innovation Management (2pp. 122-129.

Wynn, E. (2001) “Qualitative Research in IS: Issues and Trends,” in T. E. M. (Ed.), Hershey, PA: Idea Group Publishing, pp. 306.

Yin, R. K. (1994) Case Study Research: Design and Methods, Second edition. Vol. 5. Thousand Oaks: Sage Publications, Inc.

20

Page 21: Software Product Development Process Model: Case Studies of Mobile Software Companies

Appendix 1

New products development strategy:

How would you characterize your new product strategy?

How many competitors you have? Who are they?

How often you launch new products or product versions?

New product development process:

How do you generate ideas for new products?

How do you screen ideas?

How do you test the feasibility of the requirements?

How do you perform concept definition and testing?

How do you perform business analysis?

Do you develop in-house prototype?

How do you perform market testing?

How do you perform market introduction?

Who are involved in the NPD activities in each phase?

What is the structure for the new product development in your company?

What are the roles of the parties involved in the process?

Who is in charge managing the NPD process in your company?

How is the process organized?

21

Page 22: Software Product Development Process Model: Case Studies of Mobile Software Companies

How do you combine R&D with marketing?

How do the development team and marketing people cooperate?

Software development process:

What software development approaches you use in your company?

Do you have specific methods?

What phases the process includes?

Who are involved in the process?

How do you organize software development?

Would you describe your software development more like a chaotic process or very

organized one? Why?

Do you have quality systems for software development? If so, what specifications are

used?

How do you manage requirements?

How do perform requirements development?

How do you document the requirements and system specification?

How do you perform verification and validation?

22