Zeev Bonen.
Evolution of a sociotechnical system : a model and some implicatons online
. (page 1 of 2)
Online Library → Zeev Bonen → Evolution of a sociotechnical system : a model and some implicatons → online text (page 1 of 2)
Font size
'Xc^|i.CHrf.?^
WORKING PAPER
ALFRED P. SLOAN SCHOOL OF MANAGEMENT
Evolution of a Sociotechnical System
A Model and Some Implications
By
Dr. Zeev Bonen
WP 1054-79
March 1979
MASSACHUSETTS
INSTITUTE OF TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASSACHUSETTS 02139
Evolution of a Sociotechnical System
A Model and Some Implications
By
Dr. Zeev Bonen
WP 1054-79 March 1979
Evolution of a Sociotechnical System - A Model and Some Implication
Dr . Z. Bonen
Abstract
A complex sociotechnical system is composed of many people and many
types of equipment interacting together to perform seme set of tasks. A
model of the process of evolution of these systems is described, detailing
the search, decision and implanentation phases. Behavior in all stages and
the ensuing evolution path are strongly influenced by present system
properties, the decision process governing it and the quality of feedback.
A system acts as a filter that prefers those changes which cause minimun
disruption to its present state. Hence, radical disruptive change is
usually championed only by new comers and young fluid organizations;
whereas large mature systans would usually accept only incranental change
often characterized by technology lag and very slow adaptation of
structure. Caning papers will discuss various aspects of this process in
greater detail .
n ^^^"^"i
Introduction
A complex sociotechnical system (1 )(STS) is composed of many people
and many inanimate systans interacting together to perform a set of
tasks/missions. This definition encompasses a wide variety of systems i.e.
production lines and plants, canputeri zed business systems (air line
reservations system), rail network, R&D groups, military formations (tank
company, ground to air battery). More generally, a sociotechnical system
(STS ) may be as small as a small workshop with a few people or as large as
a big societal system (sociotechnical macrosystem) .
A complex STS is not created overnight or built in one piece. It
evolves gradually through an intricate process involving both technological
and social change. Tlius, a real issue posed by complex sociotechnical
systens to system theory is not how to design and optimize a large scale
system fVom scratch, but rather how to understand, facilitate and perhaps
improve the process of evolution.
Various authors have discussed factors influencing this process, each
enphasizing particular factors: econcmic , organizational, technological,
etc. Other authors have dealt iidth organizational tasks at various stages;
search, decision, implanentation. This paper attempts to present an
overall aggregate model and overview of this process, trying to include
most major stages and factors and their interaction. It deals specifically
with the evolution of a single STS which has some form of central
management .
Beginning with a description of a simplified model (sec. 2) later
sections detail the influence of the present STS throughout the process
(sec. 3). the search process and overall goals (sec. ^),'the decision
process (sec. 5), the implementation process (sec. 6) and finally a fuller
and more detailed overall model combining all the parts (sec. 7). The
implications of this model to the processes of evolution of various
sociotechnical systans are described in sec. 8. These conclusions are
supported by case studies in many fields.
In a specific case, the model could help in the analysis of the
problems of change, locate the major barriers to further evolution and aid
in designing a suitable change strategy.
I
I
2. Simplif ie d Model of the Evolution Process
A simplified model of the evolution of a STS is shown in fig. 1.
Opportunities and threats, perceived performance gaps and available
resources, all activate the search and decision process (2). The intensity
of search depends mostly on perceived performance gaps and on available
resources. In the decision process various alternatives for action,
including the status quo, are compared on a cost-benefit basis. Once a
change decision is taken, it will usually require a considerable time for
implementation, hopefully causing improved system performance. The quality
of performance feedback plays a major role in the unfolding of this
Simplified System Evolution Process
Opportunities
& Threats
System
Objectives
Available
Resources
Search &
De ci s i on
Process
Change
Decision
Perceived
Performance
Gap
Change
Implementation
Rate/Delay
System
Performance
Perceived
Performance
Feedback
Quality
Fig ire 1
process. Lack of feedback (e.g. military systens in peace time), distorted
and/ or delayed feedback, feedback magnifying small implementation problems,
these and other feedback aberrations lead to particular behavior.
This simplified picture omits the major role of the following factors
in determining the process of evolution:
1. Present STS state.
2. The decision structure governing the STS.
3. The type of change.
3. STS State Influence
STS state is defined by the following attributes:
1. Values and objectives.
2. Structure: the set of roles and relationships among STS
members .
3. Equipment and technological processes.
^. Personnel specific skills and training.
STS present state is due to past investment and efforts, that is sunk
costs. A change in any attribute involves costs. For example, changes in
structure, involving disruption of past roles and relationships, often
System Influence on Search, Decision & Impl emen tation
Search
Mechanism
Proposal
Decision
Process
De ci s i on
Outcome/
Delay
Implementation
Rate
System
State
Figure 2
cause uncertainty, apprehension and resistance anong members, making
sometimes the social cost of change much bigger than the monetary cost of
equipment changes. STa tends to resist costly changes, i.e. radical
changes which may cause large disruption in its present state (.3 pp. i^-^O,
4). Hence, a (large) mismatch between the present state and that required
by the (radical) change proposal will cause (.large) resistance to change in
both decision and implementation (fig. i?).
Moreover, the search function is characterized by selective perception
(b, pp. 150-15b) confining the search mostly to present STS-matched
opportunities. Thus, STS tends to act as a filter that considers and
prefers those changes which cause minimum disruption to its present state
(fig. f).
Needs, H-oblems
Opportunities
'
Change Proposals
1
ik
t
Limits
to
Search
txisting System
Acts as a Kilter
i
Actual Process
of Change
Figure 3
1 Estimates of system- proposal mismatch and the generated resistance to
change due to attributes 1 and i? are appearently difficult. However, in
a specific case, one can estimate the magnitude of values and objectives
change, the number and influence of people who cannot or will not accept
these changes, the required indoctrination for other personnel etc.
Analysis of new versus old structure will reveal how many people would
suffer a major, minor or no change in role and relationships, what is
their influence and what kind of resistance they can generate. Though
these estimates look fuzzy. They are not worse than some cost estimates
for new high-risk technological equipment.
System state affects also performance feedback as described in the
following section.
4. The Search Process and Overall System Goals (Fig. 4)
Performance gaps may be caused by changes in the STS itself or by
changes in its environment including competition of other systems. Search,
however, is activated not by real performance gaps but by perceived
performance gaps. These depend on feedback quality which depends not only
on system type (e.g. military systems in peace time) but also on system
state. Danger signals are not perceived or ignored all-together due to
selective STS perception.
Search Process
Available
Resources
Overall
(}oals
-e|
Long
Term
Trends |
^^
â– " \
Active
Search
Opportunities
& Threats
X
Search
Mechanism
Perceived
Performance
Gap
Perceived
Performance
V
X-
Feedback
Quality
"T
i Filtering
i Effect
Ch ang e
Proposal
-â–ºFig. 5
System
State
.Fig. 8
Figure 4
Normal search activated by perceived performance gaps will usually be
limited to normal incremental change proposals (2) due to present STS state
filtering effect (unless the search mechanism is external to the systan,
sec. 3). Only when the performance gaps become large and really
threatening will the search mechanism change and start to look for radical
solutions .
This can be obviated by the existence of an active search policy regarding
the level and type of search as well as suitable search mechanisns (R&D
laboratory, independent market research, new-ventures organization)
encouraging
1. Search for new opportunities even when no performance gaps are
perceived.
2. Search for radical change opportunities.
Active search policy depends on overall system goals. Availability of
uncommitted resources will encourage active search for new uses for these
resources .
Overall system goals usually change very slowly responding to changes
in the envirorment as perceived by the search function. Goals' change
causes reformulation of system objectives and search of means to achieve
than. Fbwever, system goals are also affected by perceived performance.
Persistent success or failure in achieving certain objectives will lead,
after seme time, to reevaluation of system goals and changes in objectives.
5. The Decision Process
The decision process may vary all the way from a rational
analytic- synoptic one to an almost purely political bargaining process
depending on the following factors:
1. Decision Structure.
2. Change Proposal Properties.
3. STS Present State (sec. 3).
These are considerable differences in decision structures between
various types of organizations (e.g. business firms, goverrment agencies,
university departments) . Also the decision structure governing a specific
system may be, partly or even canpletely, outside it (i.e. the system has
little autonomy)
Change proposal properties include:
1. Size (Cost) and complexity.
2. Uncertainties in equipment technology, system structure and
external variables affecting change results.
3. Influence on system present state.
A specific change proposal, which is normal to one STS, i.e. requiring
little change is its present state may be quite radical with respect to
another STS requiring large change in its state and threatening severe
disruption .
An overall diagram of the decision process is shovn in Figure 5. The
"decision collective" includes all people involved in a specific decision,
formally or informally, inside as well as outside the formal boundaries of
the systan. Hence it depends on the decision structure, systan state, and
the specific change proposal considered. A large change proposal will
involve a larger part of the STS hence increasing the decision collective.
A large, radical change proposal threatening present structure will not only
increase the size of the collective, bringing in more and external
participants, thus making the decision process cumbersome and less
efficient, but also cause more goal heterogeniety anong its manbers. Thi©
may be sometimes counteracted by a common overall value/ goal system of all
or part of the decision collective. Decision collective effectiveness
depends on its size , efficiency and goal commonality.
Fig. 8
k
Change
Proposal
Required
System
State
Present
Systan
State
MisTiatch
Fig. n
Overall
Goals/Values
*
Decision
Structure
/
Common
Goals
i /
^
i
Decision
Collective
Decision
Collective
Effectiveness
Decision Process Properties
(Analytical Political)
1. Preferred Decisions
2. Decision Delay
Figure 5 Decision Process Determinants
There is a basic difference between operational and non-operational
goals. Following Simon's difinition (5, pp. 155-156) a goal is operational
only if there is an apparent way to connect goal satisfaction with the
consequences of proposed actions.
Change Action
Proposal
Consequences
-â–º _ , â–º-
O perational Goal
Goal
Satisfaction
According to Simon a collective decision process will be predominantly
analytical if decision collective members have common operational goals.
Otherwise decision will be reached by predominantly bargaining processes,
10
Radical or large change proposals involve substantial or even large
uncertainty. Therefore, their predicted consequences and contributions to
goal satifaction are vague and doubtful. Hence, even if the decision
collective has common goals, they can not be operational with respect to
radical change proposals.
To sum up, change proposal properties and decision collective
effectiveness togther determine the properties of the decision process for a
specific proposal. The following process attributes, paraphrasing fteck' s
attributes (6), will help in discerning decision process type:
1. Decision collective size, homogeneity, and efficiency.
2. Decision danands on rational capacity, objectivity and
perseverance.
3. Decision collective attitude toward and aptitude for radical
change .
4. Consequences of action, clear or vague.
5. Goals connection to action consequences, clear or vague.
6. One of a kind or statistical decision.
The interaction of these attributes to discover existence (or
non-existence) of common operational goals in shown in Figure 6. In cases
v*iere all the above attributes tend to be we 11- structured the process will
be mostly rational. The more any or all the attributes tend to be poorly
structured the more political the process will become.
Operationally, the decision process is characterized by the type of
decisions it prefers (small/large, normal/ radical) and by the time required
to arrive at a decision (Decision delay). The dependence of decision
outcome and decision delay on the decision process determinants is shown in
fig. 7.
11
Qu #6
Qu m
Qu #5
Situation and
Problem Dsfinition
Action Proposal
Predicted
Consequences
Operational Goal
Satisfaction
Qu #2
Common Operational Goals?
Analytical, political or
mixed decision process
Figure 6
cision [
lective I
Fig. 5
t
.^
Change
Proposal
Level of
Change
Uncertainty
Change
Predicted
Performance
Implementation
Problems
Feedback
1
Fig. 8
Present
Perceived
Performance
Fig. ^
Conservative
Bias
Discounted
[Change
[Performance
I
Perceived
- Fu tur e
.Benefit
Change
Cost
Fi
i;
Decision
Collective
Effectiveness
Cost
Resources
Ratio
Available
Resources
Cost-Benefit
Criterion
L
De ci s i on
- Delay
Decision
Outcome
(System
I Objectives
Figire 7 Decision Outcome and Delay
12
When the consequences of proposed changes are vague (Qu. #'4,5) due to
various uncertainties, a typical situation for an innovative radical change
proposal, the decision collective exhibits a conservative bias against it
(6, pp. 122-129). This bias increases with decrease in effectiveness of
the decision collective. The conservative bias causes a discount in the
value of the change proposal predicted performance, tipping the scale in
favor of the status quo or of lower uncertainty and wider concensus
proposals.
The less effective the decision collective is, the more it would tend
toward incremental, low risk (uncertainty) decisions; tending in the purely
political process toward disjointed incranental decisions (Lindblcm, 7).
The sane variables, decision collective effectiveness and change
uncertainty also influence decision delay. This delay increases with
increasing proposal cost/ available resources ratio.
6. The Implementation Process
The details of the implementation process, its stages and problems
vary very considerably between different systans. Here we shall describe a
simplified aggregate model, reserving detailed discussion for another
paper. This simplified model is similar in part to Pathak's model (8).
The implementation process model is shown in Figire 8. The nominal
Implatientation rate is determined by the required change (system -change
mismatch) and by the available resources. Problems which occur during
implanentation are of two types.
A disjointed incremental decision amounts to a small change dealing with
a small part/few dimensions of the system. This is mush easier to get
through due to smaller decision collective and consensus requiranents .
13
1. Normal debugging and learning problans associated vrith introducing
new equipment and procedures which increase with change uncertainty.
2. Resistance to change induced difficulties due to desired-actual
system state mismatch.
Hence, the rate of implementation problems depends both on
implementation rate and on syston state mismatch. As implementation
problems accumulate, more and more effort must be devoted to them,
decreasing the effective implanentation rate.
Implementation problems have two additional effects. First of all,
they reduce system efficiency and performance. Also feedback about
implementation problems, magnified and distorted by system members
resisting the change, leads to further discount of the impact of the change
on future performance.
Decision
Outcome
Required
Systan
State
r — "â–
Fig*! 7«» —
Systan
State
Mismatch
+ 1'
Available
Resources
zzznr:
Nominal
Implementation
Rate
Effective
Implementation
Rate
Present
System
State
Level of Change
Uncertainty
/
Resistance
A to Change
-f
Implementation
Problems
Feedback
Implementation
Problans
Baifi
Implanentation
Problems Level
Problans
Solution
Effort
Discounted
Change
Performance
Fig. 7
Figire 8
System
Per f o nn an ce
The Implementation Process
-^ ; —
Sometimes, one can avoid the resistance to change difficulties by
selecting (or creating) an implementing unit matched to the change tasks,
14
7. Overall Process Model
A combined and aggregated presentation of the overall process is shown
in Figure 9. It can be seen that while a STS evolves and adapts in
response to its changing environment, the path of evolution in all stages:
search, decision and implanentation, is constrained and to a large extent
determined by present systen properties, by the decision structure
governing it and by the quality of feedback.
Opportunities
& Threats
System
Objectives
Overall
Goals
V A
Decision
Structure
Search
Process
Available
Resources
Decision
Outcome/
Delay
Inplementation
Rate/Delay
^—, -^
Resistance
to Change
Perceived
Performance
Gap
Chang e-sy Stan
Mismatch
System
State
Filtering
Effect
Perceived
Per f o rm an ce
Feedback
Quality
Environment
System
Performance
Figure 9 Systems Evolution Process Overall Diagram
Note that the search and decision bodies may be outside the STS
itself, in another unit or level of the organization or even without any
formal connection to it.
Hence, the selective perception filtering properties of the search,
decision and the implanentation stages (sec. 3) may be quite different.
15
For ex an pie, as described by Wilson (9) in a large diverse organization
there is a considerable variety of change propDsals including radical ones.
(Wide-band search filter), however its decision process is consensus
oriented and blocks major innovations (Narrow-band decision filter).
In other cases the decision process (decision collective, sec. 5) is
external and on top of the STS (strong corporate managanent) . It is not
affected (perhaps even affected positively) by the suggested radical change
(Wide-band decision filter) . It may well embark upon it, ignoring problems
in the STS itself which will surface during implanentation (Narrow-band
implementation filter) .
Thus, various combinations of all the above factors will lead to very
different evolutionary behaviors ranging from the normal evolution process
to the radical change process.
8. Processes of Evolution (10)
Normal evolution proceeds within a given framevork: Objectives,
internal structure, technology and external interfaces; hence, change-system
mismatch is small. Radical evolution on the other hand, involves far
reaching changes in the existing franework, a transition to a basically
different system and environment; hence, change-system mismatch is large.
In both cases, normal and radical, the implanentation of change is
increnental; one cannot change a complex STS overnight. However, the
results are very different.
As defined, normal change can usually lead to the following results:
1. Exploiting more efficiently present operating envirorment.
2. Obtaining better performance in a narrower operating environment.
3. Channeling into a dead end.
GE top managanent put transistors into its vacuum tube division. The
results were not satisfactory.
16
Pursuit of paths 1, 2, i.e. the path of adaptive speicali zation , can
lead to a dead end when system environment undergoes large changes. In
seme cases, where the envirorment changes slowly, the accunulation of
normal incremental changes over a long period of time may lead to a radical
system change .
Radical, qualitative change if successful, can lead to:
1. Large expansion of present operating envirorment.
2. TVansition to a different operating environment.
3. Creation of a new operating envirorment.
Radical change is required in order to cope with large changes in the
envirorment of the system.
Recalling STS behavior in response to change (sec. 3) it is evident
that mature inflexible sociotechnical systans, operating as narrow-band
filters will follow in almost all cases the normal path of evolution. This
tendency will be even stronger when the decision structure governing the
system is diffuse and heterogenious, causing decision collective
ineffectiveness. Radical change, triggered by technological opportunities,
changes in the environment or other factors, must therefore find other
routes. It will be chanpioned in many cases by newcomers creating new
uninhibited organizations penetrating, where possible via empty niches in
the envirorment. These, if successful in a relatively stable, possibly
new, environment, will in due course optimize their structure to that
envirorment; thus, beccming in their turn rigid and unable to accept
further radical change.
The evidence supporting these assertions is very impressive and comes
from many fields: business firms, military organizations and public
insitutions (3, 11, 12, 13, T*, 15).
17
Only a few exanples will be mentioned. The transistor revolution was
championed successfully by new companies. The attempts by the old vacuum
tube manufactures to move into transistors ended mostly in failure. The
diesel electric locomotive was introduced by General Motors, an outsider,
and not by the established locomotive manufacturers. The introduction of
major military innovations, i.e. ballistic missiles and naval nuclear
power, required the setting up of new organizations dedicated to these
missions.
A new technology which is applicable to various products and systems
will probably be applied, first of all, to apparently premising new
stand-alone products and small systems (digital watch, electric typewriter,
hand calculator) in many cases chatipioned as described above, by innovative
newcomers. These products and their champions serve as the vehicle for
transferring the new technology from the laboratory to the first generation
of actual use.
The application of new technology to a large mature SIS is a slow,
incranental, two-stage process
1. Substitution of equipment without changing the overall structure
of STS. (3. p. 10).
2. Re- structuring of the STS, including further substitution of old
equipment and introduction of new equipment. This process is
extremely slow and often not successful .
The SLtostitution of equipment in a large complex STS is a lengthy
process stretching over the many years required for development, production
and large scale incorporation. Because of this delay, as well as risk
aversion behavior common in a mature STS, the new equipment will usually be
WORKING PAPER
ALFRED P. SLOAN SCHOOL OF MANAGEMENT
Evolution of a Sociotechnical System
A Model and Some Implications
By
Dr. Zeev Bonen
WP 1054-79
March 1979
MASSACHUSETTS
INSTITUTE OF TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASSACHUSETTS 02139
Evolution of a Sociotechnical System
A Model and Some Implications
By
Dr. Zeev Bonen
WP 1054-79 March 1979
Evolution of a Sociotechnical System - A Model and Some Implication
Dr . Z. Bonen
Abstract
A complex sociotechnical system is composed of many people and many
types of equipment interacting together to perform seme set of tasks. A
model of the process of evolution of these systems is described, detailing
the search, decision and implanentation phases. Behavior in all stages and
the ensuing evolution path are strongly influenced by present system
properties, the decision process governing it and the quality of feedback.
A system acts as a filter that prefers those changes which cause minimun
disruption to its present state. Hence, radical disruptive change is
usually championed only by new comers and young fluid organizations;
whereas large mature systans would usually accept only incranental change
often characterized by technology lag and very slow adaptation of
structure. Caning papers will discuss various aspects of this process in
greater detail .
n ^^^"^"i
Introduction
A complex sociotechnical system (1 )(STS) is composed of many people
and many inanimate systans interacting together to perform a set of
tasks/missions. This definition encompasses a wide variety of systems i.e.
production lines and plants, canputeri zed business systems (air line
reservations system), rail network, R&D groups, military formations (tank
company, ground to air battery). More generally, a sociotechnical system
(STS ) may be as small as a small workshop with a few people or as large as
a big societal system (sociotechnical macrosystem) .
A complex STS is not created overnight or built in one piece. It
evolves gradually through an intricate process involving both technological
and social change. Tlius, a real issue posed by complex sociotechnical
systens to system theory is not how to design and optimize a large scale
system fVom scratch, but rather how to understand, facilitate and perhaps
improve the process of evolution.
Various authors have discussed factors influencing this process, each
enphasizing particular factors: econcmic , organizational, technological,
etc. Other authors have dealt iidth organizational tasks at various stages;
search, decision, implanentation. This paper attempts to present an
overall aggregate model and overview of this process, trying to include
most major stages and factors and their interaction. It deals specifically
with the evolution of a single STS which has some form of central
management .
Beginning with a description of a simplified model (sec. 2) later
sections detail the influence of the present STS throughout the process
(sec. 3). the search process and overall goals (sec. ^),'the decision
process (sec. 5), the implementation process (sec. 6) and finally a fuller
and more detailed overall model combining all the parts (sec. 7). The
implications of this model to the processes of evolution of various
sociotechnical systans are described in sec. 8. These conclusions are
supported by case studies in many fields.
In a specific case, the model could help in the analysis of the
problems of change, locate the major barriers to further evolution and aid
in designing a suitable change strategy.
I
I
2. Simplif ie d Model of the Evolution Process
A simplified model of the evolution of a STS is shown in fig. 1.
Opportunities and threats, perceived performance gaps and available
resources, all activate the search and decision process (2). The intensity
of search depends mostly on perceived performance gaps and on available
resources. In the decision process various alternatives for action,
including the status quo, are compared on a cost-benefit basis. Once a
change decision is taken, it will usually require a considerable time for
implementation, hopefully causing improved system performance. The quality
of performance feedback plays a major role in the unfolding of this
Simplified System Evolution Process
Opportunities
& Threats
System
Objectives
Available
Resources
Search &
De ci s i on
Process
Change
Decision
Perceived
Performance
Gap
Change
Implementation
Rate/Delay
System
Performance
Perceived
Performance
Feedback
Quality
Fig ire 1
process. Lack of feedback (e.g. military systens in peace time), distorted
and/ or delayed feedback, feedback magnifying small implementation problems,
these and other feedback aberrations lead to particular behavior.
This simplified picture omits the major role of the following factors
in determining the process of evolution:
1. Present STS state.
2. The decision structure governing the STS.
3. The type of change.
3. STS State Influence
STS state is defined by the following attributes:
1. Values and objectives.
2. Structure: the set of roles and relationships among STS
members .
3. Equipment and technological processes.
^. Personnel specific skills and training.
STS present state is due to past investment and efforts, that is sunk
costs. A change in any attribute involves costs. For example, changes in
structure, involving disruption of past roles and relationships, often
System Influence on Search, Decision & Impl emen tation
Search
Mechanism
Proposal
Decision
Process
De ci s i on
Outcome/
Delay
Implementation
Rate
System
State
Figure 2
cause uncertainty, apprehension and resistance anong members, making
sometimes the social cost of change much bigger than the monetary cost of
equipment changes. STa tends to resist costly changes, i.e. radical
changes which may cause large disruption in its present state (.3 pp. i^-^O,
4). Hence, a (large) mismatch between the present state and that required
by the (radical) change proposal will cause (.large) resistance to change in
both decision and implementation (fig. i?).
Moreover, the search function is characterized by selective perception
(b, pp. 150-15b) confining the search mostly to present STS-matched
opportunities. Thus, STS tends to act as a filter that considers and
prefers those changes which cause minimum disruption to its present state
(fig. f).
Needs, H-oblems
Opportunities
'
Change Proposals
1
ik
t
Limits
to
Search
txisting System
Acts as a Kilter
i
Actual Process
of Change
Figure 3
1 Estimates of system- proposal mismatch and the generated resistance to
change due to attributes 1 and i? are appearently difficult. However, in
a specific case, one can estimate the magnitude of values and objectives
change, the number and influence of people who cannot or will not accept
these changes, the required indoctrination for other personnel etc.
Analysis of new versus old structure will reveal how many people would
suffer a major, minor or no change in role and relationships, what is
their influence and what kind of resistance they can generate. Though
these estimates look fuzzy. They are not worse than some cost estimates
for new high-risk technological equipment.
System state affects also performance feedback as described in the
following section.
4. The Search Process and Overall System Goals (Fig. 4)
Performance gaps may be caused by changes in the STS itself or by
changes in its environment including competition of other systems. Search,
however, is activated not by real performance gaps but by perceived
performance gaps. These depend on feedback quality which depends not only
on system type (e.g. military systems in peace time) but also on system
state. Danger signals are not perceived or ignored all-together due to
selective STS perception.
Search Process
Available
Resources
Overall
(}oals
-e|
Long
Term
Trends |
^^
â– " \
Active
Search
Opportunities
& Threats
X
Search
Mechanism
Perceived
Performance
Gap
Perceived
Performance
V
X-
Feedback
Quality
"T
i Filtering
i Effect
Ch ang e
Proposal
-â–ºFig. 5
System
State
.Fig. 8
Figure 4
Normal search activated by perceived performance gaps will usually be
limited to normal incremental change proposals (2) due to present STS state
filtering effect (unless the search mechanism is external to the systan,
sec. 3). Only when the performance gaps become large and really
threatening will the search mechanism change and start to look for radical
solutions .
This can be obviated by the existence of an active search policy regarding
the level and type of search as well as suitable search mechanisns (R&D
laboratory, independent market research, new-ventures organization)
encouraging
1. Search for new opportunities even when no performance gaps are
perceived.
2. Search for radical change opportunities.
Active search policy depends on overall system goals. Availability of
uncommitted resources will encourage active search for new uses for these
resources .
Overall system goals usually change very slowly responding to changes
in the envirorment as perceived by the search function. Goals' change
causes reformulation of system objectives and search of means to achieve
than. Fbwever, system goals are also affected by perceived performance.
Persistent success or failure in achieving certain objectives will lead,
after seme time, to reevaluation of system goals and changes in objectives.
5. The Decision Process
The decision process may vary all the way from a rational
analytic- synoptic one to an almost purely political bargaining process
depending on the following factors:
1. Decision Structure.
2. Change Proposal Properties.
3. STS Present State (sec. 3).
These are considerable differences in decision structures between
various types of organizations (e.g. business firms, goverrment agencies,
university departments) . Also the decision structure governing a specific
system may be, partly or even canpletely, outside it (i.e. the system has
little autonomy)
Change proposal properties include:
1. Size (Cost) and complexity.
2. Uncertainties in equipment technology, system structure and
external variables affecting change results.
3. Influence on system present state.
A specific change proposal, which is normal to one STS, i.e. requiring
little change is its present state may be quite radical with respect to
another STS requiring large change in its state and threatening severe
disruption .
An overall diagram of the decision process is shovn in Figure 5. The
"decision collective" includes all people involved in a specific decision,
formally or informally, inside as well as outside the formal boundaries of
the systan. Hence it depends on the decision structure, systan state, and
the specific change proposal considered. A large change proposal will
involve a larger part of the STS hence increasing the decision collective.
A large, radical change proposal threatening present structure will not only
increase the size of the collective, bringing in more and external
participants, thus making the decision process cumbersome and less
efficient, but also cause more goal heterogeniety anong its manbers. Thi©
may be sometimes counteracted by a common overall value/ goal system of all
or part of the decision collective. Decision collective effectiveness
depends on its size , efficiency and goal commonality.
Fig. 8
k
Change
Proposal
Required
System
State
Present
Systan
State
MisTiatch
Fig. n
Overall
Goals/Values
*
Decision
Structure
/
Common
Goals
i /
^
i
Decision
Collective
Decision
Collective
Effectiveness
Decision Process Properties
(Analytical Political)
1. Preferred Decisions
2. Decision Delay
Figure 5 Decision Process Determinants
There is a basic difference between operational and non-operational
goals. Following Simon's difinition (5, pp. 155-156) a goal is operational
only if there is an apparent way to connect goal satisfaction with the
consequences of proposed actions.
Change Action
Proposal
Consequences
-â–º _ , â–º-
O perational Goal
Goal
Satisfaction
According to Simon a collective decision process will be predominantly
analytical if decision collective members have common operational goals.
Otherwise decision will be reached by predominantly bargaining processes,
10
Radical or large change proposals involve substantial or even large
uncertainty. Therefore, their predicted consequences and contributions to
goal satifaction are vague and doubtful. Hence, even if the decision
collective has common goals, they can not be operational with respect to
radical change proposals.
To sum up, change proposal properties and decision collective
effectiveness togther determine the properties of the decision process for a
specific proposal. The following process attributes, paraphrasing fteck' s
attributes (6), will help in discerning decision process type:
1. Decision collective size, homogeneity, and efficiency.
2. Decision danands on rational capacity, objectivity and
perseverance.
3. Decision collective attitude toward and aptitude for radical
change .
4. Consequences of action, clear or vague.
5. Goals connection to action consequences, clear or vague.
6. One of a kind or statistical decision.
The interaction of these attributes to discover existence (or
non-existence) of common operational goals in shown in Figure 6. In cases
v*iere all the above attributes tend to be we 11- structured the process will
be mostly rational. The more any or all the attributes tend to be poorly
structured the more political the process will become.
Operationally, the decision process is characterized by the type of
decisions it prefers (small/large, normal/ radical) and by the time required
to arrive at a decision (Decision delay). The dependence of decision
outcome and decision delay on the decision process determinants is shown in
fig. 7.
11
Qu #6
Qu m
Qu #5
Situation and
Problem Dsfinition
Action Proposal
Predicted
Consequences
Operational Goal
Satisfaction
Qu #2
Common Operational Goals?
Analytical, political or
mixed decision process
Figure 6
cision [
lective I
Fig. 5
t
.^
Change
Proposal
Level of
Change
Uncertainty
Change
Predicted
Performance
Implementation
Problems
Feedback
1
Fig. 8
Present
Perceived
Performance
Fig. ^
Conservative
Bias
Discounted
[Change
[Performance
I
Perceived
- Fu tur e
.Benefit
Change
Cost
Fi
i;
Decision
Collective
Effectiveness
Cost
Resources
Ratio
Available
Resources
Cost-Benefit
Criterion
L
De ci s i on
- Delay
Decision
Outcome
(System
I Objectives
Figire 7 Decision Outcome and Delay
12
When the consequences of proposed changes are vague (Qu. #'4,5) due to
various uncertainties, a typical situation for an innovative radical change
proposal, the decision collective exhibits a conservative bias against it
(6, pp. 122-129). This bias increases with decrease in effectiveness of
the decision collective. The conservative bias causes a discount in the
value of the change proposal predicted performance, tipping the scale in
favor of the status quo or of lower uncertainty and wider concensus
proposals.
The less effective the decision collective is, the more it would tend
toward incremental, low risk (uncertainty) decisions; tending in the purely
political process toward disjointed incranental decisions (Lindblcm, 7).
The sane variables, decision collective effectiveness and change
uncertainty also influence decision delay. This delay increases with
increasing proposal cost/ available resources ratio.
6. The Implementation Process
The details of the implementation process, its stages and problems
vary very considerably between different systans. Here we shall describe a
simplified aggregate model, reserving detailed discussion for another
paper. This simplified model is similar in part to Pathak's model (8).
The implementation process model is shown in Figire 8. The nominal
Implatientation rate is determined by the required change (system -change
mismatch) and by the available resources. Problems which occur during
implanentation are of two types.
A disjointed incremental decision amounts to a small change dealing with
a small part/few dimensions of the system. This is mush easier to get
through due to smaller decision collective and consensus requiranents .
13
1. Normal debugging and learning problans associated vrith introducing
new equipment and procedures which increase with change uncertainty.
2. Resistance to change induced difficulties due to desired-actual
system state mismatch.
Hence, the rate of implementation problems depends both on
implementation rate and on syston state mismatch. As implementation
problems accumulate, more and more effort must be devoted to them,
decreasing the effective implanentation rate.
Implementation problems have two additional effects. First of all,
they reduce system efficiency and performance. Also feedback about
implementation problems, magnified and distorted by system members
resisting the change, leads to further discount of the impact of the change
on future performance.
Decision
Outcome
Required
Systan
State
r — "â–
Fig*! 7«» —
Systan
State
Mismatch
+ 1'
Available
Resources
zzznr:
Nominal
Implementation
Rate
Effective
Implementation
Rate
Present
System
State
Level of Change
Uncertainty
/
Resistance
A to Change
-f
Implementation
Problems
Feedback
Implementation
Problans
Baifi
Implanentation
Problems Level
Problans
Solution
Effort
Discounted
Change
Performance
Fig. 7
Figire 8
System
Per f o nn an ce
The Implementation Process
-^ ; —
Sometimes, one can avoid the resistance to change difficulties by
selecting (or creating) an implementing unit matched to the change tasks,
14
7. Overall Process Model
A combined and aggregated presentation of the overall process is shown
in Figure 9. It can be seen that while a STS evolves and adapts in
response to its changing environment, the path of evolution in all stages:
search, decision and implanentation, is constrained and to a large extent
determined by present systen properties, by the decision structure
governing it and by the quality of feedback.
Opportunities
& Threats
System
Objectives
Overall
Goals
V A
Decision
Structure
Search
Process
Available
Resources
Decision
Outcome/
Delay
Inplementation
Rate/Delay
^—, -^
Resistance
to Change
Perceived
Performance
Gap
Chang e-sy Stan
Mismatch
System
State
Filtering
Effect
Perceived
Per f o rm an ce
Feedback
Quality
Environment
System
Performance
Figure 9 Systems Evolution Process Overall Diagram
Note that the search and decision bodies may be outside the STS
itself, in another unit or level of the organization or even without any
formal connection to it.
Hence, the selective perception filtering properties of the search,
decision and the implanentation stages (sec. 3) may be quite different.
15
For ex an pie, as described by Wilson (9) in a large diverse organization
there is a considerable variety of change propDsals including radical ones.
(Wide-band search filter), however its decision process is consensus
oriented and blocks major innovations (Narrow-band decision filter).
In other cases the decision process (decision collective, sec. 5) is
external and on top of the STS (strong corporate managanent) . It is not
affected (perhaps even affected positively) by the suggested radical change
(Wide-band decision filter) . It may well embark upon it, ignoring problems
in the STS itself which will surface during implanentation (Narrow-band
implementation filter) .
Thus, various combinations of all the above factors will lead to very
different evolutionary behaviors ranging from the normal evolution process
to the radical change process.
8. Processes of Evolution (10)
Normal evolution proceeds within a given framevork: Objectives,
internal structure, technology and external interfaces; hence, change-system
mismatch is small. Radical evolution on the other hand, involves far
reaching changes in the existing franework, a transition to a basically
different system and environment; hence, change-system mismatch is large.
In both cases, normal and radical, the implanentation of change is
increnental; one cannot change a complex STS overnight. However, the
results are very different.
As defined, normal change can usually lead to the following results:
1. Exploiting more efficiently present operating envirorment.
2. Obtaining better performance in a narrower operating environment.
3. Channeling into a dead end.
GE top managanent put transistors into its vacuum tube division. The
results were not satisfactory.
16
Pursuit of paths 1, 2, i.e. the path of adaptive speicali zation , can
lead to a dead end when system environment undergoes large changes. In
seme cases, where the envirorment changes slowly, the accunulation of
normal incremental changes over a long period of time may lead to a radical
system change .
Radical, qualitative change if successful, can lead to:
1. Large expansion of present operating envirorment.
2. TVansition to a different operating environment.
3. Creation of a new operating envirorment.
Radical change is required in order to cope with large changes in the
envirorment of the system.
Recalling STS behavior in response to change (sec. 3) it is evident
that mature inflexible sociotechnical systans, operating as narrow-band
filters will follow in almost all cases the normal path of evolution. This
tendency will be even stronger when the decision structure governing the
system is diffuse and heterogenious, causing decision collective
ineffectiveness. Radical change, triggered by technological opportunities,
changes in the environment or other factors, must therefore find other
routes. It will be chanpioned in many cases by newcomers creating new
uninhibited organizations penetrating, where possible via empty niches in
the envirorment. These, if successful in a relatively stable, possibly
new, environment, will in due course optimize their structure to that
envirorment; thus, beccming in their turn rigid and unable to accept
further radical change.
The evidence supporting these assertions is very impressive and comes
from many fields: business firms, military organizations and public
insitutions (3, 11, 12, 13, T*, 15).
17
Only a few exanples will be mentioned. The transistor revolution was
championed successfully by new companies. The attempts by the old vacuum
tube manufactures to move into transistors ended mostly in failure. The
diesel electric locomotive was introduced by General Motors, an outsider,
and not by the established locomotive manufacturers. The introduction of
major military innovations, i.e. ballistic missiles and naval nuclear
power, required the setting up of new organizations dedicated to these
missions.
A new technology which is applicable to various products and systems
will probably be applied, first of all, to apparently premising new
stand-alone products and small systems (digital watch, electric typewriter,
hand calculator) in many cases chatipioned as described above, by innovative
newcomers. These products and their champions serve as the vehicle for
transferring the new technology from the laboratory to the first generation
of actual use.
The application of new technology to a large mature SIS is a slow,
incranental, two-stage process
1. Substitution of equipment without changing the overall structure
of STS. (3. p. 10).
2. Re- structuring of the STS, including further substitution of old
equipment and introduction of new equipment. This process is
extremely slow and often not successful .
The SLtostitution of equipment in a large complex STS is a lengthy
process stretching over the many years required for development, production
and large scale incorporation. Because of this delay, as well as risk
aversion behavior common in a mature STS, the new equipment will usually be
1 2
Online Library → Zeev Bonen → Evolution of a sociotechnical system : a model and some implicatons → online text (page 1 of 2)
Using the text of ebook Evolution of a sociotechnical system : a model and some implicatons by Zeev Bonen active link like:
read the ebook Evolution of a sociotechnical system : a model and some implicatons
read the ebook Evolution of a sociotechnical system : a model and some implicatons
