Software Development Business Model for Top Level Process With
Competitive Bidding
Is Requirement Analysis the Unpaid Work?
Noriko Hanakawa
1
and Masaki Obana
2
1
Information management department, Hannan University, Matsubara, Osaka, Japan
2
Department of Information Science and Technolog, Osaka Institute of Technology, Hirakata, Osaka, Japan
hanakawa@hannan-u.ac.jp, obana@is.oit.ac.jp
Keywords: Software Process, Competitive Bidding, Top Level Process, Requirement Analysis, System Design.
Abstract: Recently, large-scale system development is ordered through competitive bidding in various computer
system. If competitive bidding runs in order to choose a best system proposal, development process is
greatly different from conventional software development process of the conventional software engineering.
Before developers and customers sign a contract document, requirement analysis and system design
activities are started. Of course, the activities are unpaid work because developers do not yet sign the
contract with customers. Therefore, we discussed software development business model with competitive
bidding. Problems in real competitive bidding are (1) unpaid work such as requirement analysis, (2) judging
a best system proposal by customers, (3) budget and schedule are determined before requirement analysis
activities. Based on the problems, we show a revised process of top level activities of planning phase and
competitive phase for developers’ profit and customers’ high satisfaction.
1 INTRODUCTION
Software engineering research field mainly focuses
on development techniques for large-scale software
such as requirement analysis, design software,
programming techniques, test techniques, and
operation and maintenance. Almost all the themes
are development techniques for developers. In
addition, because developers have to make high
quality software, software engineering researchers
frequently discuss the techniques without customers’
constraint such as budget and time.
However, large-scale computer system and
software consume much money. Of course money is
paid by customers. Customers are various
organization such as hospitals, banks, private
companies of public transportation, government and
local government, and educational organizations.
These customers have strong social responsibilities
in order to control modern societies’ system and
human life. Software and computer system should
be delivered by due day, and software and computer
system should be developed within their budget.
Therefore, we discuss “software development
business model”. Especially, we focus on top level
process with competitive bidding before making a
contract with customers. The top level process may
be beyond the scope of the conventional software
engineering research area. However, in current
industry, competitive bidding among system
development companies is popular and usual.
Especially, large-scale system should be bidded in
orde to keep low cost. Therefore, we propose a
software development busines model.
In this paper, we mainly discuss a top level
activities of software development process including
competitive bidding before making a contract. Atfter
that, an outline of the new software development
business model is shown. We show related work in
section2, in section3 we explain real competitive
bidding before signing a contract document. In
section 4, dilemma of customers and developers are
shown. We discuss software development business
model in section 5. Section 6 shows summary and
future researchers.
2 RELATED WORK
In conventional researches, Edward et al., shows a
173
Figure1: A software development top level process for large-scale infrastructural system in Japan.
model of the early estimating/planning stages of a
project (EEPS model) (Edwards JS., Moores TT.
1994). Because of unclear data in requirement
analysis, there were 30% budget error. Jamieson et
al. gave a model for pre- and post-contract phases in
agile development [12]. These researches mostly
cover our research topics. However, environment
surrounding development computer system
continuously change. We have to consider the
changes in the present day.
On the other hand, in recently, “Cho-joryu” for
software engineering has been proposed by
Muroya(IPA Software Engineering Center, 2012].
Muroya also provided a method of a contract for
software development , and important of customers’
activities in top level software development process.
Breiner et al. also discuss requirement engineering
in the bidding stage(K. Breiner, M. Gillmann, A.
Kalenborn, C. Müller, 2015). These researches claim
importance of competitive bidding. However,
research of competitive bidding process in software
engineering just have started. Concrete research
results are not described. In addtion, Takano et al.
show effective bidding strategy in a competitive
bidding simulation (Takano Y., Ishii N., Muraki M.,
2014). Pablo et al. propose effective competitive
bidding model in scoring and position probability
graph (Ballesteros-Pérez P., González-Cruz M. C.,
Cañavate-Grimal A., 2013). Also the other
researches discussed effective bidding and accuracy
of cost estimation. Management fields actively study
bidding way, bidding accuracy, and bidding
simulation. However, these researches discuss just
ways of bidding. Our research target is not bidding
system. Our research target is whole development
process including competitive bidding in software
engineering research field.
3 CURRENT PROCESS
3.1 Planning Phase
Fig.1 shows a typical top level software
development process for large-scale infrastructural
system in Japanese industry (Hanakawa N.,
Akazawa Y., Mori A., Maeda T., Inoue T., Tsutsui
S., 2006) (Hanakawa N., Obana M., 2010).
At first, customers make a plan of a new system.
Customers make a plan in order to take budget in an
organization. Managers decide to make a new
system according to planning documents. In
addition, managers decide budget and schedule. The
budget is decided by managers according to
strategies of the organization. The budget usually is
built during the previous year. Because budget for
new system is a part of whole budget of the
organization, it is difficult to revise the budget. In
addition, the plan includes concrete due date and
schedule.
3.2 Competitive Bidding Phase
Competitive bidding phase includes the following
activities.
(1) C5: Making a system proposal request
Customers make a system proposal request
document. The document includes as follows;
x System purpose
x Target users, and the round number of users
x Images of several use scenes
x Vague requests such as “improving our
performance” or “our workload reduces”.
x Rough schedule and concrete due date.
x Rough images of functions of the system
Seventh International Symposium on Business Modeling and Software Design
174
In addition, various constraints are shown in the
proposal request documents. For example, system
administrators consume workload as little as
possible. These constraints are not described
concrete quantitative values of performance
requirements. Although concrete requests are not
described, many images of customers’ request are
enumerated as sentences and simple figures.
Developers and sales engineers should satisfy all
requests even if the requests are vague and not clear.
(2) C6: Delivering the system proposal request
The system proposal request documents are
opened to development companies in industry. The
system proposal request are uploaded a Web site that
everyone can access.
(3) D1:Understand system proposal requests
Salesmen, sales engineers, and system engineers
read the system proposal request document. Often,
the number of page of a system proposal request
document is 100 pages more over. The reading and
understanding workload is also huge.
(4) D2: Making a system proposal
Salesmen and sales engineers and system
engineers discuss the new system. Examples of
workload for making a system proposal in our
university’s case are as follows;
x Reading and understanding the system proposal
request documents (1man-month)
x Investigating new technologies (3 man-months)
x Investigating similar system and trend of the
similar system in industry (1 man-month)
x Interviewing customers and analysing customers’
problems (3 man-months)
x Estimating workload and cost (2 man-months)
x Making a system proposal document including
the billed amount (4 man-months)
Total workload is 14 man-months.
(5) D3: Submitting system proposal
The completed system proposal is submitted to
customers by due day. Of course, customers receive
several system proposals from several companies.
(6) D4: Presentation of proposals by companies
The companies present their proposals. This
presentation is most important in a competitive
bidding phase. Developers may consume several
man-months’ workload. The presentation strongly
influences results of the competitive bidding.
In our university’s case, one company’s
presentation time is 2 hours, and question and
answer time is 1 hour. 6 companies submitted the
system proposals. So, it took 18 hours. It takes three
days for finishing all presentations.
(7) C7: Judging the proposal documents
Figure 2: Similar proposals for similar systems.
Customers determine a best proposal. Customers
check whether the proposals satisfy requests. Next,
proposals are evaluated by various points;
completeness, innovativeness, technical quality, and
consistency of system design. That is, the customers
evaluate essences of system proposal. If evaluation
is low, the system proposal will be rejected.
After that, the full amounts are compared. A
system proposal with smallest full amount is
selected. Competitive bidding is finished. A
development company are determined.
3.3 Contract and Development Phases
A development company that won the competitive
bidding makes a contract with customer. The due
day and the full payment amount are written on the
contract document. Of course, because details of
requirement analysis have not yet determined,
developers do not almost know what they should
develop. Customers clearly know only payment
amount and due day. Next phase is development of
new system. Activities run according to the
conventional software engineering.
4 PROBLEMS IN BIDDING
4.1 Is Analysis the Unpaid Work?
Developers consume large efforts to make a system
proposal. At first, developers should be clear vague
requests. Developers often visit customers, and have
many meetings in order to clarify their requests.
Sometimes, meetings take several hours. Meetings
are just like requirement analysis activities in
development phase. To achieve more detailed
system design and more complete system proposal,
developers have to do long interview even if the
interview activities are the unpaid work.
Customers felt enthusiasm in the long interviews.
Because customers often judge a best proposal
Software Development Business Model for Top Level Process With Competitive Bidding - Is Requirement Analysis the
Unpaid Work?
175
subjectively, the enthusiasm of developers is
important. Almost requirement analysis finishes in
competitive bidding phase. Of course, these
activities are the unpaid work. On the other hand,
developers have to describe full amount in a system
proposal (Lopez-Martin C., Isaza C., Chavoya A.
2012). To estimate cost, developers need more
detailed requests. The full amount is calculated by
workload, computer and software cost.
As a result, a company wins the competitive
bidding. A company can receive money for work of
the following development phase. However,
companies that did not win the competitive bidding
cannot receive any money
4.2 Comparing System Proposals
On the other hand, customers also have to make a
difficult decision in competitive bidding. Although
customers have only vague requests, customers have
to assess submitted system proposals. It is difficult
to judge whether proposals satisfy their requests.
Therefore, easiest judgement is to check
completeness of system proposals. For example,
customers check how a proposed system design is
complete. More detailed design and deeper
investigation are assessed at good proposals. That is,
if developers consume more workload for making a
system proposal, customers feel better because
developers sincerely make a system proposal.
Judgement is subjective rather than objective.
Customers do not usually have sufficient knowledge
in order to judge objectively system proposals.
5 TWO DILEMMAS
5.1 Making Proposals with Profit
Companies that challenge competitive bidding have
to keep profit. Even if a company did not win a
competitive bidding, a company can keep profit.
Fig.2 shows how companies make system proposals
with keeping profit. Developers submit a system
proposal for similar system. Almost all parts of
system proposals are re-used in similar system
proposal bidding. That is, developers consume little
workload for making system proposals. System
proposals describe similar system design and similar
software design. These designs are already proposed
to the other customers. However, a part of the design
is original design. For example, although software
design consists of several package software, original
software design is combined package software
(Kataoka N., et al. 1998).
Because workload for making a system proposal
is a little, companies keep profit even if companies
did not win competitive bidding. However, if
developers submit same system proposals without
any revisions, the system proposals will be rejected.
Customers can easily know that developers skimp
their effort to make a system proposals. Original
requests are not satisfied in the proposal. Such
system proposals will not win competitive betting.
Developers struggle with a dilemma between
unpaid work and winning of bidding. If developers
want to win bidding, developers should consume
unpaid work. If developers want to avoid unpaid
work, developers could not win competitive bidding.
5.2 Fixing Budget and a Delivery Day
Business of company should be improved more
efficient through information technology. However,
they do not have sufficient knowledge of
information technology. Therefore, managers cannot
design computer system. They have only rough
images of several use scenes, then, they do not have
sufficient skill for making concrete system requests.
In addition, development of computer system
needs much budget. New system greatly influences
business process and business concept. Sometimes
managers employ several new faces, and new
sections for new business process. Therefore, budget
and schedule of the new system is not only elements
of the new system but also elements of the company
strategies. Therefore, budget and schedule are
usually determined as an annual plan of the previous
year. Budget and schedule for new system are one of
important strategies of a company. Budget and
schedule cannot be easily revised. Even if new
functions or new system design are proposed by
developers at requirement analysis activity,
customers cannot pay additional fee for new
functions. Moreover, time for developing additional
functions is also unacceptable.
On the other hand, when customers judge a best
system proposal in bidding, customers want more
detailed system design and system proposals and
investigation. Customers want to choose a best
system proposal and a best developers’ team.
Customers’ dilemmas are (1) budget and
schedule are fixed as an annual plan although
development has not yet started, (2) it is difficult to
explain detail requests, (3) customers require high
completeness of system proposal in order to judge a
best proposal against vague customers’ requests.
Seventh International Symposium on Business Modeling and Software Design
176
Figure 3: Gaps of between conventional development process and real process including bidding.
6 AN OUTLINE OF SOFTWARE
DEVELOPMENT BUSINESS
MODEL
Fig. 3 shows gaps between development process of
conventional software engineering and real
development process including competitive bidding.
The following sub-sections explain an outline of
software development business model with the gaps.
6.1 Development Process of Model
Development process reflects real process including
competitive bidding in Fig.3. Main process is as
follows;
(1) Determine budget and schedule in an annual
plan in customers’ company
(2) Making system proposal request by customers
(3) Making system proposal including system
design, requirement analysis, investigation, and
budget and schedule by developers
(4) Competitive bidding and making a contract
(5) Development phase including requirement
analysis, design, implement, and test.
Basically, unpaid works for developers are the
above process (3), the above process (4). In the
above process (5), developers can earn money from
customers. Moreover, of course, customers do not
pay any money before making a contract.
6.2 To Avoid Unpaid Work
Various products are clearly divided to two steps;
the first step is unpaid work, the second step is paid
work. In the first step, developers make a system
proposal as unpaid work. In the second step,
developers make a system design as paid work.
Fig.4 shows outline of the two steps.
In the first step, developers make a system
proposal as typical system design and software
design. Typical system design means normal system
design without original design and original
customization. Therefore, system designs of
proposal are same as system design of the other
customers. Of course, system design does not
require results of requirement analysis. Likewise,
software design consists of commercial packages
without original software. That is, developers
prepare system proposal packages for competitive
bidding. Of course, correct total cost and time are
not clear in the first step. Therefore, for example,
budget is 30 % increase money of the typical design,
and schedule is 30 % long of typical design. The
additional money and the extended time are rough
estimation for original system design and original
software in the second step.
After developers win competitive bidding,
developers start development phase including
requirement analysis, system design, and
investigation of new technologies according to
customers’ requests. Results of requirement analysis
reflect to original system design and original
software. These design and software are customized
the system proposal. In this way, developers earn
proper money as requirement analysis, system
design, and investigation of new technologies.
Software Development Business Model for Top Level Process With Competitive Bidding - Is Requirement Analysis the
Unpaid Work?
177
Figure 4: Products between unpaid and paid work.
6.3 To Achieve Effective Bidding
To achieve effective bidding, customers check gaps
between their requests and system proposals.
Customers do not evaluate satisfaction of their all
request in system proposals at competitive bidding.
Customers understand how system proposal is
different from their requests. Customers know what
requests are not satisfied in system proposals. Most
important customers' action is to understand gaps
between customers’ requests and package software
and normal system design. Of course, the
understanding of gaps does not need deep
knowledge of information technologies. Customers
only know whether packages support their request
functions. Customers do not need to understand
mechanisms of software packages. If
customersasked question whether packages support
their request functions at competitive bidding,
developers would easily answer the questions.
Customers make a list of non-supported
functions and unsatisfied performance on each
system proposal. Customers should select a system
proposal that includes fewest non-supported
functions and unsatisfied performance. If customers
have difficulty of making the list, customers should
employ engineers who can make the list of non-
supported functions and unsatisfied performance.
The list of non-supported functions and unsatisfied
performance are implemented as “original system
design” and “original software” in Fig.4.
7 SUMMARY
We showed real top level software development
process including competitive bidding. Problems of
top level process are (1) unpaid work such as
requirement analysis, (2) judging a best system
proposal by customers, (3) budget and schedule are
determined before requirement analysis activities.
Therefore, we discuss an outline of software
development business model for developers’ profit
and customers’ satisfaction. Limitation of system
proposals submitted to competitive bidding are
clarified. We suggested a way of select a best
proposal with a list that described non-support
functions and unsatisfied performance.
In future, reconstruction of competitive bidding
for large-scale computer system will start little by
little in Japan.
ACKNOWLEDGEMENTS
This work was partially supported by JSPS
KAKENHI Grant Number JP26330093.
REFERENCES
IPA Software Engineering Center, 2012. A "Cho-joryu"
Approach. In The Seventeen Principles for System
Development, Japan.
K. Breiner, M. Gillmann, A. Kalenborn, C. Müller, 2015.
Requirements Engineering in the Bidding Stage of
Software Projects – A Research Preview. In REFSQ
2015 21st International Working Conference, pp.270-
276.
Takano Y., Ishii N., Muraki M., 2014. A sequential
competitive bidding strategy considering inaccurate
cost estimates. Omega, vol. 42, issue 1, pp.132-140.
Ballesteros-Pérez P., González-Cruz M. C., Cañavate-
Grimal A.,2013. On competitive bidding: scoring and
position probability graphs. International Journal of
Project Management, vol. 31, issue 3, pp.434-448.
Hanakawa N., Akazawa Y., Mori A., Maeda T., Inoue T.,
Tsutsui S., 2006. A Web-based integrated education
system for a seamless environment among teachers,
students, and administrators. International Journal of
System
Computer in Japan, Vol.37, Issue 5, pp.14-
24.
Hanakawa N., Obana M., 2010. Mobile game terminal
based interactive environment for large-scale lectures.
In WBE2010 Eighth IASTED International Conference
on Web-based Education.
Kataoka N., Koizumi H., Takasaki K., Shiratori, N., 1998.
Remote joint application design process using package
Seventh International Symposium on Business Modeling and Software Design
178
software. In ICOIN-12 Twelfth International
Conference on Information Networking, pp.495-500.
Lopez-Martin C., Isaza C., Chavoya A. 2012. Software
development effort prediction of industrial projects
applying a general regression neural network.
International journal Empirical Software Engineering,
vol.17, num.6, pp.738-756.
Edwards J.S., Moores T.T. 1994. A conflict between the
use of estimating and planning tools in the
management of information systems. European
Journal of Information Systems 3(2), pp.139-147.
Software Development Business Model for Top Level Process With Competitive Bidding - Is Requirement Analysis the
Unpaid Work?
179