Transcript Key issues in Managing a Global Software Outsourcing relationship

Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Key issues in Managing a Global
Software Outsourcing
relationship between a Norwegian
and Russian firm: Some Practical
Implications
Vegar Imsland, Sundeep Sahay and Yvonne Wartiainen
April 9th 2003
1
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Agenda
• The context of the research
• Methodology
• Discuss 3 key issues that influenced the
relationship, and suggest some points of action:
• Knowledge sharing
• Estimation
• Project Management
• Communication
• Discussion
2
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
The context of the research
•
•
•
•
NorSys: The Norwegian customer
RussCo: The Russian developer
Relationship started in 1999
Possible future projects
3
The context of the research – compared
with IN364 issues
• Theoretical foundations
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
• Research methodology
• Illustrating complexity
• Structuration theory: Understanding culture
• Organizational complexity
•
•
•
•
Geographical distance
Context specific knowledge
Language
GSO vs. Participatory Design
4
Methodology
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Data collection:
• 12 interviews in Norway and 14 interviews at RussCo in St.
Petersburg.
• Reading project documents.
• Analyzing tools used in the project.
• Discussions.
Other:
• Feedback to the Norwegian project staff – used by the projectmembers to justify their actions.
• We felt moral obligations to give something back to the
companies.
• The Norwegian company sponsored our trip to Russia.
• …but never instructed us what to write.
• Our engagement in the project has had practical implications which
we consider positive.
• We believe we could not have collected our data being less involved
in the project.
• We were asked for advice during the project.
5
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Knowledge sharing
• Definition: Concerns the various processes through which
information and knowledge is acquired, shared and
integrated.
• Various forms of knowledge: the Norwegian tax- and salary
rules, about developing estimates, about SalarySystem,
TestTool, and Delphi.
• Various mechanisms: face-to-face meetings, written
documents, e-mails, telephone-conversations and ICQ-chats,
project-reports and bug-reports.
• Local knowledge: Knowledge which could not be transferred,
which was in the heads of people and embedded in work
practices.
6
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Knowledge sharing
• Problem 1: Lack of domain knowledge
• Problem 2: Inadequate efforts to
transfer knowledge
• Estimation
7
Knowledge sharing
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Lack of domain knowledge
• NorSys was supposed to have complete
knowledge of SalarySystem
• SalarySystem is very context specific
• Little written documentation
• RussCo was believed to have established
procedures and understanding of how
outsourcing should be done
• Redesign projects are different from ”new”
projects
8
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Knowledge sharing
Inadequate efforts to transfer
knowledge
• Knowledgeable RussCo people spent time
on other projects
• Language problems
• Positive: TestTool and the test cases
improved the knowledge sharing
9
Knowledge sharing
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Estimation
• Technical focus
• Top-down approach
• RussCo wanted the contract → made low
estimates
• No negotiation process
• No overall evaluation; interconnected
modules
10
Knowledge sharing
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Suggested points of action:
•
•
•
•
•
Earlier and regular face-to-face meetings
Encourage more communication channels to be used
Videoconferencing
Closer control over movement of personnel
Language and cultural training
Estimation:
• Take the GSO nature of the project into consideration
• Do not make estimates unless the understanding of the system
is appropriate
• Emphasize the need to develop “correct” rather than “low”
estimates
11
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Project management
Within the issue of project management, we
include the formal (and also informal)
routines and procedures for application of
knowledge, management of skills, use of
tools and techniques for project activities in
order to meet the common aim of the
partners to successfully complete the
project.
We first discuss some project management
related problems and then discuss some
suggested points of action.
12
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Project Management
• Problem 1: The role and movement of
key people
• Problem 2: Project Reporting
13
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
The role and movement of key people
Key people: those people that because of their
role, knowledge, and work-tasks are
important to the project
Organizational structure: the informationprocessing network in the organization
including points of team members’ power
14
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
The role and movement of key people
• No formal demand related to the team
members skills and experience was made.
• The challenge of language led to bigger
workload on two people (the Russian
speaking developer on the Norwegian side,
and the Russian project leader on the
Russian side) than originally intended.
15
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
The role and movement of key people
• Too much information is routed through one
person, increasing the risk for overload – and
“person dependent”. The result of overload is often
selection, which again might lead to ignorance of
important issues.
• The organizational structure of the Russian
company prevents the project leader from taking
important decisions based on the knowledge
he/she’s gained – “responsibility without
authority”.
16
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Project reporting
• Testers are pressurized due to lack of time
to close and report a build as verified
(successfully fixed), before they are satisfied
with the corrections made.
• Both the companies lacked the appropriate
level of understanding about the project
status in the beginning of the project due to
general reports, and that no checks were
made to verify the status.
17
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Project reporting
• The programmer on the Norwegian side has
no formal responsibility to report his
progress on the calculation engine,
complicating the parallel programming
and integration process.
• Detailed reports are good for the project
members closely involved in the project, but
more high-level and general reports should
be made for the upper-management.
18
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Suggested points of action:
• Make it a contractual obligation for the
vendor to provide the buyer with CVs
from project members, and report shifts of
people and get approvals from NorSys.
• Allow more direct communication lines to:
1. spread knowledge among people.
2. reduce risk of overload on one person.
3. make cross-site problem solving more efficient.
19
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Suggested points of action:
• Decrease the consequences of shift in
personnel by introducing
shadowing/buffering – let one extra person
follow the project so that if one leaves, this
one can step in on short notice.
• CMM : establish routines, procedures, and
make documentation about the project
more available.
20
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Suggested points of action:
• Reexamining the relationship between
testers and developers to find answers to
whether closing builds premature creates
new bugs and will result in further delays in
the long run. Release build every second
week instead of every week? When 80% of
bugs are fixed.
21
Project Management
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Suggested points of action:
• Generate more general and less-detailed
report for the overall management of the
project – for top management.
• Introduce another level of management that
reports to the project leader from the
testers and programmers.
22
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Communication
Definition: describes the transmission
and receiving of information, and the
feelings and attitudes of the people
involved with the overall purpose of
having it understood and producing a
response.
23
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Communication
• Problem 1: Content of
communication.
• Problem 2: Channels of
communication.
24
Communication
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Content of communication
• Issues of high and low context
communication: Norwegian expect direct
communication, while Russians has a more
indirect communication style.
• Concepts are not universal so that even if
you could translate the word
“bilgodtgjørelse” to English or Russian, you
need to fill in with an explanation of the
actual meaning.
25
Communication
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Content of communication
• Due to English language limitations
Russian team members seems
reluctant to write to the e-mail list
because it is to serve as formal
documentation. Explaining technical
issues, and concepts that are local to
Norway was further complicated by
the limitation of language.
• Black-box communication.
26
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Communication
27
Communication
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Channels of communication
Different channels have different characteristics
which makes their use more, or less, critical at
different stages in the project:
• Face-to-face: in this project these meetings have
been too few, to late, and with not enough people
attending.
• E-mail: result of the e-mail list: team members
reluctant to use it due to visibility to management.
Led to use of informal channels and more work for
the Russian project leader.
28
Communication
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Channels of communication
• ICQ: has encouraged its own use because it allowed for
informal, invisible communication in Russian between the
two sites. Was removed for security reasons.
• TestTool: made monitoring of bugs efficiently, made customer
input on bugs possible, increase the customers sense of
control.
• Telephone: due to costs and lack of shared language this has
not been used, but for those speaking Russian it would have
been preferred.
• Video conferencing: provide similar opportunities to face-toface meetings, but is not used in this project.
• Fax: not used in this project.
29
Communication
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Suggested points of action:
• Explore additional channels
• Increase amount of face-to-face meetings
• Consider the use of videoconference as a
complement to face-to-face meetings
• Sort out the technical problems with NetMeeting
because it will provide the capabilities of ICQ and
the visual dimension
• Encourage more direct channels of communication
• Introduce language and cultural training
30
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Communication
31
Vegar Imsland, Sundeep Sahay, and Yvonne Wartiainen
Department of informatics, University of Oslo
April 9th 2003
Discussion
•
•
•
•
Methodology
Problems
Suggested action
Questions
32