Professional Documents
Culture Documents
net/publication/224687175
CITATIONS READS
11 1,343
3 authors:
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Antonio Carlos Tonini on 20 November 2014.
ABSTRACT
Six Sigma increases quality by reducing process variability and aligning customer’s
expectations, providing high financial returns. First applied in manufacturing
companies, it also becomes very effective on finance, management and service in
general, including software development.
DMAIC (Define-Measure-Analyze-Improve-Control) has been widely used as the
method for Six Sigma implementation projects in manufacturing, once its procedures
are based on the well known PDCA (Plan-Do-Check-Act) principles. However, it does
not fulfill appropriately all software development requirements, which leads to
researches for introducing new practices in the DMAIC method or developing other
methods and techniques.
Based on a research performed through a multiple cases study, it is possible to suggest
improvements in the DMAIC method. Some of the considered issues are: cost
evaluation, time and customer satisfaction impact on business; the improvement
processes verification; organizational standardization and learning achieved from Six
Sigma projects.
The main result is a proposal of a specific roadmap for Six Sigma projects application in
software development process improvement, which includes the treatment of the aspects
found.
INTRODUCTION
Software development companies have looked for excellence in their product quality
due to growing pressure of their customers. So, they have developed strategies that
make them unique in comparison to their competitors. One strategy is continuous
improvement of software development process.
Six Sigma is a methodology that increases quality by reducing process variability and
aligning customer’s expectations, providing high payback. First applied in
manufacturing companies, it also has shown to be very effective on other businesses,
such as finance and management, mainly where customer service presence is relevant.
DMAIC (Define-Measure-Analyze-Improve-Control) has been the most used method
for Six Sigma projects. Nevertheless, it was designed for industrial environment, where
production processes are much more predictable than service. For that reason, the
method needs some alterations, which represents an opportunity to develop and create
new methods.
Based on a research performed through a multiple cases study, this paper suggests some
improvements on the DMAIC method and proposes a specific roadmap for Six Sigma
projects on software development process improvement.
SIX SIGMA
Motorola developed the first Six Sigma implementation method in order to solve some
already known quality problems. The method was developed by Dr. Mikel Harry and
was called MAIC (acronym of Measure, Analyze, Improve and Control). It was
constituted by four stages [11]:
• Measurement: data gathering regarding the as-is picture of the process, including
input variables, output and defects (out of compliance);
• Analysis: understanding of current process performance causes;
• Improvement: elaboration of improvement alternatives for process performance;
• Control: procedures and rules to maintain the obtained improvements, making them
maintainable and durable.
GE, unlike Motorola, didn't exactly know its quality problems and, therefore, first
needed to map the real quality problems in order to solve them. Besides that, GE
intended to use Six Sigma for new problems, as soon as they appeared, so the method
should also contain the problem identification task. Dr. Mikel Harry took advantage of
the acquired experience in Motorola and elaborated a new method that got known as
DMAIC (acronym of Define, Measure, Analyze, Improve and Control). Nowadays, it is
the most acquainted and used Six Sigma implementation method [11]. This method is
composed by five stages that should be accomplished in a sequential way for any Six
Sigma project:
• Definition: identification of problems and situations to improve, remarking what is
critical to quality (CTQ);
• Measurement: selection of critical characteristics to quality (CTQ) through
necessary measures in the process;
• Analysis: identification of the root-causes that are responsible for the identified
problems;
• Improvement: specification of process characteristics to be improved;
• Control: documentation and follow-up of the new process conditions.
The increase of Six Sigma use unchained two new implementation method needs: the
attention to Six Sigma philosophy diffusion into the organizations that were adopting it
[13] and incorporation of management tasks on its own model [16].
Thus, in consequence, several methods appeared, such as Breakthrough Strategy,
Roadmap, New Six Sigma, Eckes method, Six Sigma Roadmap, DMADV, IDOV and
DMEDI.
This method, proposed by Dr. Mikel Harry and Dr. Richard Schroeder in 2000, is a
problem driven method in three organizational basic levels: business, operation and
process. It obliges organization to maintain constant innovation strategy (breakthrough
strategy driven) [12].
At business level, it is guided to increase profitability and keep the organization running
in the long term; in the operation level, it is driven to increase process efficiency,
eliminate non quality costs and reduce waste; in process level, it is used to reduce
defects and variation, increasing process capacity and keeping operational and strategic
goals aligned.
Table 1 – Breakthrough Strategy structure
Stages Activities Description
Identification Recognition Six Sigma benefits understanding
Definition Definition of projects to be executed
Characterization Measurement Process as-is picture
Analysis Root-cause problems identification
Implantation Improvement Identification of variables that improve the
process capacity
Control Improvement effectiveness verification
Institutionalization Standardization Alignment of all improvement projects with
learning replication
Integration Diffusion through the company of the good
practices learned
The method consists of eight activities grouped in four stages, as shown on Table 1.
Roadmap method
This method, developed by John M. Gross in 2001 is only guided to Six Sigma
launching, and helps to estimate efforts and costs on Six Sigma projects and inform
them to the whole enterprise [10].
The stages of this method are the following:
• Management training on objectives establishment, possible benefits mapping and
the necessary resources availability;
• Identification of a support consultancy to prepare an internal specialist team (black
belts and green belts) for selection and operation of Six Sigma projects;
• Definition of a top executive to be the Six Sigma program champion and establish
goals and implementation plans;
• Establishment of quantified goals so that Six Sigma implementation can be
compared to something at the end of the process.
This method was developed by Matt Barney and Tom McCarty in 2003 and it is based
on the learning at Motorola University, also aimed to launch Six Sigma philosophy in
the organization. It assumes that the organization integrates Six Sigma with their main
business metrics (scorecards) [01].
This method proposes three stages:
• Evaluation of current organizational performance through the SWOT (Strengths,
Weakness, Opportunities and Threats) technique, involving analysis of enterprise
image to customers and the establishment of objectives and of performance driver
metrics;
• Training concerning Six Sigma for the responsible teams and main sponsor
executives;
• Project management with constant revision by senior leadership.
ECKES method
This method was proposed by George Eckes in 2001 and is also oriented for Six Sigma
launching. It was built with business process concepts and takes the following actions
[09]:
• Business objectives determination and agreement among all personal involved;
• Essential process creation, Six Sigma projects targets, detailing of other key-process
levels and process-drivers identification;
• Process leaders and owners identification;
• Process measurement creation and validation and publication in collective panels;
• Six Sigma projects selection criteria definition;
• Six Sigma projects selection;
• Continuous process management.
This method was developed by Peter Pande, Robert Neuman and Roland Cavanagh in
2001 and enables its adoption through three strategies, individually or concomitantly
used: current process improvement, process reengineering and process management
[18].
It is a quite complete method that includes the Six Sigma implementation and
management. It is composed by five stages:
• Identification of the main customer-key processes, which allow to create a customer
vs. process table;
• Customer needs definition through direct customer research (VOC), benchmarks
establishment, requirements description that allow to analyze and prioritize business
needs and strategies;
• Current performance measurement, to know the current process capacity and the
defective points;
• Prioritizing, analyzing and implementation of improvements, in a way that
maximizes returns;
• System expansion and integration, through actions to maintain improvements, define
responsibilities and control improvement process.
DMADV method
This method was developed by Thomas Pyzdekis and is a DMAIC variation that has
been applied to developing new products. According to this method, when activities are
unique (specific projects), there is no way of improving or controlling. For this reason,
the DMAIC improvement and control stages were, respectively, substituted by the
design and verification stages. The method was known as DMADV (acronym of
Define, Measure, Analyze, Design and Verify). The stages are the following [20]:
• Definition: customer's needs definition on order to set project scope and plan;
• Measurement: customer’s needs understanding and CTQ definition;
• Analysis: development of high level design concepts;
• Design: detailed project, control and test plans development;
• Verification: design tests and implementation of the processes in operational scale.
IDOV method
The method IDOV (acronym of Identify, Design, Optimize and Validate) was created
by David Woodford in 2003 for the Institute iSixSigma. This method is oriented to the
Six Sigma projects regarding new products release. It contains the following stages
[22]:
• Identification: customer's requirements obtaining and CTQ establishment;
• Design: transformation of the customer's requirements in functional specifications
and solution alternatives;
• Optimization (improvement): product quality and process performance prediction
through use of statistical projection tools;
• Validation: checking that all requirements were attended.
DMEDI method
This method (acronym of Define, Measure, Explore, Develop and Implement) is a tool
kit to guarantee the necessary strictness to release a new product or service. This
method was elaborated by Kimberly Watson-Hemphill for Pricewaterhouse & Coopers
and it considers the following stages [21]:
• Definition: identification of the true problem and its unfolding;
• Measurement: knowledge of each detail of the new product;
• Exploration (analysis): evaluation of all possible development alternatives for the
new product;
• Development (improvement): development of the new product through the most
appropriate solution;
• Implantation: execution of the chosen solution.
• Verification: design tests and processes implementation in operational scale.
A field research was accomplished with the intention of knowing and understanding
some aspects related to the Six Sigma implementation in software organizations, as well
as process improvement projects implementation. The research was conducted by the
Multiple Case Study method, with an empiric, qualitative and explanatory character,
emphasizing the individual's perspective and examining the phenomenon in its natural
environment through multiple evidence sources [23].
In order to doing so, the research protocol was elaborated containing the general view of
the study, the selection of information sources, the analysis units adopted criteria, the
subjects, the propositions, the analysis variables and the field procedures. [23].
Three analysis units were chosen, constituted by software companies that use Six Sigma
in their software process improvement. A support questionnaire was elaborated for the
interviews, containing subjects related to Six Sigma implementation in the organization
and the methods used in the execution of specific projects. Regarding the data
gathering, at least three workers were interviewed in each organization, one of them
from the process improvement department, one of the software development area and an
executive. The data gathering was completed after semi-structured interviews and
observation of registrations and documents. The results, as expected, were affected
intensely by the organization context, highlighting its influence.
The following cases report the analyses related to each one of the three analysis units,
identified in this work as unit 1, unit 2 and unit 3.
Case 1
Case 2
The unit 3 is a great Brazilian insurance company that develops software for its own
use. It doesn't show any quality certification, neither maturity in its software
development area.
The adoption of Six Sigma for software process improvement happened after positive
experiences in the use of this methodology in the business processes.
The Six Sigma projects execution is conditioned to improvements that result in return
on investments regarding the consumer market. The first projects happened in existent
work processes and were related to the administrative, managerial and technical
structure of the software development department.
With the acquired experience, the company understands that the new Six Sigma projects
should be related to the launching of new products that involve information technology
(IT), and result in new technologies adoption. In these projects the participation of third
parties are relevant, a fact that, besides Six Sigma, has led the company to implement
ITIL.
Table 4 - Six Sigma implementation method at unit 3
Stage Process Details
Initiation Definition Project needs definition
Recruiting Suppliers recruiting and training
Execution Measurement Understanding of customer needs together with
CTQ definition
Analysis Conceptual development of the project and the
standards of the new technology
Project Detailed project development, control planning
and test elaboration
Validation Coverage of all project details
Verification Test of the design and implementation of the
processes in operational scale
Finalization Project delivery, training and implantation
Closing Documentation Acquired knowledge registration
The method adopted in Six Sigma implementation was inspired in the method
DMADV. The processes are listed in the table 4.
CONCLUSIONS
DMADV
ECKES
DMAIC
DMEDI
IDOV
1 2 3
BS
NSS
SSR
Analysis X X X X X X X
Design X X X
Improvement X X X X X
Implementation
Validation X X X X
Verification X X X X X
Finalization X X X
Control X X X X
Registration and X
Documentation
Standardization X X
Integration X X X X
Mngt Continuous X X
management
The table 5 compares the pointed processes in the methods and those practiced by the
researched organizations.
Roadmap, New Six Sigma and ECKES methods are based on a group of
recommendations for the adoption of Six Sigma philosophy. DMADV, IDOV and
IMEDI methods show punctual variations to DMAIC method, especially when referring
to processes related to new products and larger attention on the tests of implemented
improvements: validation (if all the requirements were considered) and verification (if
each one of the requirements is in accordance with what was requested).
The Six Sigma Roadmap method adds the Customer's dimension to the improvement
process, as well as the prioritization for improvement projects and integration of the
improvement results (synergic effect).
The Breakthrough Strategy method is really an evolution of DMAIC, since it separates
Six Sigma philosophy launching activities and the processes from each improvement
project.
At the three units studied, Six Sigma philosophy was implemented either externally
(units 1 and 2) or in function of other processes (unit 3).
Nevertheless, summed up to the implementation methods are the processes related to
financial feasibility evaluation, under the business (cost) and customer (satisfaction)
points of view, and also studies on process capacity (technological condition, schedule
and readiness).
They also emphasize the importance of some subjects that are not listed in the methods:
• Tests of suggested improvements (validation and verification) to better know
problems and root causes
• Implementation team training
• Lessons learned spread in the enterprise and close to the third party teams that might
come along in the software production
• Standardization of improvement processes.
As for the Six Sigma philosophy implementation, there is no evidence that can justify a
specific method for software companies.
Some of the considered issues are: costs evaluation, time and customer satisfaction
impact on business; improvement processes verification; organizational standardization
and learning achieved from Six Sigma projects.
Table 6 –Six Sigma implementation suggested roadmap
REFERENCES
[01] Barney, M. and T. McCarty. The new six sigma: a leader’s guide to achieve
rapid business improvement. Englewood Cliffs: Prentice Hall, 2003.
[02] Basili, V. R., G. Caldiera and G. A. Cantone. “Reference archiecture for the
component factory”. ACM Transaction on Software Engineering and
Methodology. Vol 1. nº 1, January 1992. pp 53-80.
[03] Biehl, R. E. “Six Sigma for software”. Quality time. IEEE Software. V. 21. n.2.
mar-apr, 2004. p-68-70.
[04] Blakeslee Jr., J. A. “Achieving Quantum Leaps in Quality and Competitiveness:
Implementing the Six Sigma Solution in your Company” in: Proceedings of the
53th Annual Quality Congress. American Society for Quality, Anaheim:
Califórnia, pp. 486–496, May 1999.
[05] Card, D. N. “Managing Software Quality with defects”. 26th COMPSAC:
Computer Software and Applications Conference. Proceedings. IEEE Computer
Society. Aug, 2002. p.472-474.
[06] Carmel, E. “Offshoring, Software Exports and the Place of Brazil”. I DO 2004 –
Desafios e Oportunidade. São Paulo. Dez.2004.
[07] Carvalho, M. M. “Medindo o Sigma do Processo” in: Seis Sigma: Estratégia
Gerencial para a Melhoria de Processos, Produtos e Serviços. Roberto G.
Rotondaro (Coord). São Paulo: Atlas, 2002.
[08] Cusumano, M. A. “Software Factory: a historical interpretation”. IEEE
Software, Washington, p. 999, março 1992.
[09] Eckes, G. A revolução Seis Sigma. Rio de Janeiro, Campus, 2001.
[10] Gross, J.M. “A road map to Six Sigma quality”. Quality Progress, Milwaukee,
v.34, n.11, p.24-29, Nov.2001.
[11] Harry, M, J. Six Sigma: “A Breakthrough Strategy for Profitability”. Quality
Progress. v.31, n 5, May 1998.
[12] Harry, M. J. and R. Schroeder. Six Sigma: the breakthrough management
strategy revolutionising the world’s top corporations. London: Currency
Publishers, 2000.
[13] Hoerl, R.W. “Six Sigma and the Future of the Quality Profession”. Quality
Progress. v.31, n 6, June 1998.
[14] Hunter, D. and B. Schmidt. “Six sigma: benefits and approaches”. Chemical
Week, v.161, n.37, p.35-36, out, 1999.
[15] Infantini, S., F. J. B. Laurindo and M. S. P. Pessôa. “Six sigma and information
technology: a case study in an insurance company” in: EUROMA-POMS
JOINT INTERNATIONAL CONFERENCE, 1., Como, Italy, 2003.
Proceedings. Padova: SGEditoriali, 2003.. v.3, p. 591-600.
[16] Marash, S. A. “Six Sigma: Business Results Though Innovation” in:
Proceedings of the 54th Annual Quality Congress of the American Society for
Quality. Indianapolis: Indiana, pp. 627–630, May 2000.
[17] Murugappan, M. and G. Keeni. “Quality Improvement: The Six Sigma Way”.
Proceedings Asia Pacific Conference Quality Software, IEEE CS Press, Los
Alamitos, California, 2000, pp. 248–257.
[18] Pande, P.S., R. P. Neuman and R. R. Cavanagh. Estratégia Seis Sigma. Rio de
Janeiro: Qualitymark, 2001.
[19] Parasuramam, A., V. A. Zeithmal and L. L. Berry. Delivering service quality:
balancing customer perceptions. New York: The Free Press, 1990.
[20] Simon, K. “DMAIC versus DMADV”. Retrieved 08/21/2005,
http://www.isixsigma.com/library/content/c001211a.asp.
[21] Watson-Hemphill, K. “Design financial services with DMEDI”. Retrieved
08/21/2005, http://www.isixsigma.com/library/content/c040101b.asp.
[22] Woodford, D. “Design for Six Sigma: IDOV Methodology”. Retrieved
09/05/2005, http://www.isixsigma.com/library/content/c020819a.asp.
[23] Yin, R.K. Estudo de caso: planejamento e métodos. São Paulo: Bookman, 2001.