Multi-site Software Governance Impacts on Knowledge Management
1. THE
IMPACT
OF
MULTI-‐SITE
SOFTWARE
GOVERNANCE
ON
KNOWLEDGE
MANAGEMENT
Chris&na
Manteli,
Bart
van
den
Hooff,
Antony
Tang,
Hans
van
Vliet
VU
UNIVERSITY
AMSTERDAM
2. MULTI-‐SITE
SOFTWARE
DEVELOPMENT
GOVERNANCE
• SoHware
development
governance
ensures
that
processes
meet
the
requirements.
• A
governance
model
should
have1
– A
structural
perspec7ve:
“what
governance
looks
like”
– A
func7onal
perspec7ve:
“what
governance
does”
• Challenges
to
define
a
SoHware
Governance
Model
increase
when
development
ac&vi&es
are
distributed
among
remote
loca&ons.
• Challenges
on
KM
increase
when
development
goes
global
1.
P.L.
Bannerman,
“SoHware
development
governance:
A
meta-‐management
perspec&ve”,
in
Proceedings
of
the
2009
ICSE
Workshop
on
SoHware
Development
Governance,
ser.
SDG’09.
IEEE
Computer
Society,
2009,
pp.
3-‐8
3. RESEARCH
APPROACH
• A
structural
approach
to
a
Mul&-‐site
SoHware
Governance
Model
(SGM)
• Main
Knowledge
Management
(KM)
Challenges
in
Global
SoHware
Development.
Multi-site SGM
? KM Challenges
How
do
the
different
mul7-‐site
governance
structures
influence
the
knowledge
management
challenges?
4. MULTI-‐SITE
SOFTWARE
DEVELOPMENT
GOVERNANCE
• Business
Strategy:
the
outsourcing
strategy
with
the
remote
partners,
including
the
legal
implica&ons.
• Team
Structure
&
composiGon:
team
size,
role
descrip&ons
and
role
distribu&ons.
• Task
AllocaGon:
how
work
is
distributed
across
sites.
5. KNOWLEDGE
MANAGEMENT
CHALLENGES
• CommunicaGon:
communica&on
speed
and
frequency.
• Knowledge
creaGon
&
storage:
capture
informa&on,
record
it
in
a
medium,
transform
it
and
encode
it
as
knowledge.
• Knowledge
transfer:
knowledge
s7ckiness,
transac&ve
memory
systems
(TMS)
6. CASE
STUDY
OVERVIEW
• Océ
is
a
mul&na&onal
company
in
prin&ng
systems
(Canon
Group).
• Qualita&ve
data
analysis
– 20
interviews
• SoHware
Engineers,
Testers,
Site A
Architects,
Project
Managers.
– 3
loca&ons
– 1
project
Site B Site NL
7. CASE
STUDY
OVERVIEW
• Different
governance
structures
iden&fied:
Site NL-Site A Site NL-Site B
Business Different Companies: Legal Same Company: No legal barriers
Strategy barriers between sites between sites
• Site NL flat organization; Site A • Site NL flat organization; Site B
Team
hierarchically structured hierarchically structured
Structure
• Role descriptions differ between • Role descriptions differ between
&
sites sites
Composition • Unequal team sizes • Unequal team sizes
Co-development of distributed Independent development of
Task
components: Tightly coupled distributed components: Loosely
Allocation
activites coupled activities
8. THE
IMPACT
OF
BUSINESS
STRATEGY
ON
KM
• Knowledge
is
not
freely
shared
between
site
NL-‐
site
A
– Site
A
relies
only
on
the
available
knowledge
– Site
A
lacks
“system-‐generic”
knowledge
• Knowledge
shared
from
site
NL
to
site
A
needs
to
be
filtered
– More
&me
and
effort
spent
to
share
knowledge
9. THE
IMPACT
OF
TEAM
STRUCTURE
&
COMPOSITION
ON
KM
• Hierarchical
structures
create
boSlenecks
in
knowledge
sharing.
• Too
much
focus
on
agility
stresses
tacit
communica&on.
Documenta&on
remains
outdated.
• Different
role
descrip&ons
make
knowledge
difficult
to
locate.
• Knowledge
tends
to
s7ck
where
the
majority
of
the
teams
are,
or
where
the
larger
teams
are
located.
10. THE
IMPACT
OF
TASK
ALLOCATION
ON
KM
• Site
NL-‐Site
A:
Co-‐development
– Higher
communica&on
frequency.
– Increased
need
for
knowledge
sharing.
– Increased
need
for
codified
knowledge.
• Site
NL-‐Site
B:
Independent
development
– Communica&on
frequency
depends
more
on
the
development
phase.
– Knowledge
s&cks
the
more
‘independent’
teams.
11. COMMUNICATION
FREQUENCY
&
TASK
ALLOCATION
High
Site NL - Site A
Communication Frequency
medium
Site NL - Site B
Low
Requirements Architecture Design Coding Testing Integration Maintenance
12. SUMMARY
Multi-site SGM KM Challenges
• Business Strategy • Communication
• Team Structure • Knowledge creation
& Composition & storage
• Task Allocation • Knowledge transfer
13. LESSONS
LEARNED
• The
impact
of
mul&-‐site
governance
structures
on
knowledge
management:
– Legal
barriers
increase
the
effort
and
&me
spent
on
managing
the
crea&on,
storage
and
transfer
of
knowledge
– Unbalanced
team
structure
&
composi&on
impedes
smooth
flow
of
knowledge.
– Tightly-‐coupled
ac&vi&es
among
remote
teams
increases
communica&on
frequency
as
well
as
the
need
and
effort
spent
for
knowledge
sharing.
14. FUTURE
RESEARCH
• Expand
and
improve
the
3
structural
aspects
of
the
mul&-‐site
soHware
governance
model.
• Create
a
mul&-‐site
soHware
governance
framework:
– How
organiza&on
and
development
ac&vi&es
should
be
structured
to
best
align
business
and
development
goals.
15. THANK
YOU.
Chris&na
Manteli,
Bart
van
den
Hooff,
Antony
Tang,
Hans
van
Vliet
VU
UNIVERSITY
AMSTERDAM
16. Knowledge Management
Multi-site Software Governance
Challenges
• No direct documentation due to
Site NL-Site A:
information barriers.
• They are different companies • Information sent from Site NL to
and information barriers exist Site A needs to be filtered.
between the remote sites • Communication frequency is higher.
Business
Strategy
Site NL-Site B:
• They are the same company and
no information barriers exist
between the remote sites
Site NL-Site A:
• Site NL is a flat organization,
while Site A is hierarchically • Hierarchical structures create
structured. bottlenecks in knowledge sharing.
• Role descriptions differ between • Too much focus on agility stresses
sites. tacit communication and
Team Structure • Unbalanced team sizes. documentation remains outdated.
& Composition Site NL-Site B: • Different role descriptions makes
• Site NL is a flat organization, knowledge difficult to locate.
while Site B is hierarchically • Knowledge tends to stick where the
structured. majority of teams, or where the larger
• Role descriptions differ between teams are located.
sites.
• Unbalanced team sizes.
• Tightly coupled activities increase
Site NL-Site A: the need for knowledge sharing.
• They co-develop a function and • Co-development creates a greater
their activities are tightly coupled. need for codified knowledge.
• Communication frequency is high.
Task Allocation
Site NL-Site B: • Knowledge tends to stick to the
independent development teams.
• They develop independently and • Communication frequency depends
their activities are loosely on the release phase.
coupled.