Jay, Jason with George Roth, "From Continuous Improvement to Centralized Information: The Life and Times of a Systems Thinking Intervention", 2007 July 29-2007 August 2

Online content

From Continuous Improvement to Centralized
Information: The Life and Times of a Systems Thinking
Intervention

Jason Jay, M.Ed.
Organization Studies Group
MIT Sloan School of Management
50 Memorial Drive, E52-509
Cambridge, MA 02143

George Roth, Ph.D.

Lean Aerospace Initiative
MIT Sloan School of Management
77 Vassar Street, Building 41-205

Cambridge, MA 02139

Abstract

This multi-method retrospective case study examines the implementation of a systems
thinking intervention - The Manufacturing Game® - as part of a comprehensive
approach to “Proactive Manufacturing” and workforce involvement in continuous
improvement at an American oil refinery. Notable gains in local reliability metrics such
as equipment failure rate were achieved as a result of these efforts, which contributed to
impressive gains in financial, safety, and environmental performance. At the same time,
the overall contribution of the program to the refinery’s performance is ambiguous,
confounded by simultaneous investments in the physical capacity and technical
capabilities of the plant, along with increases in the crude-gasoline price spread that all
together generated significant value. This ambiguity created interpretive flexibility for
employees of the plant, who were able to both interpret and shape the continuous
improvement activities based on prior political divisions, mental models, and interpretive
biases. To make sense of these emergent and unintended consequences, a speculative
model of relational conflict that undermines workforce participation in continuous
improvement is proposed. 1

Keywords: Continuous improvement, workforce involvement, The Manufacturing Game,
systems thinking, organizational learning, organizational change, qualitative research.

Introduction

The translation of system dynamics models into board games has been a mainstay of
system dynamics theory and research, beginning with the Beer Distribution Game

' Note: This research is generously funded by the PIMO grant at the MIT Sloan School of Management.
Special thanks to John Sterman, Nelson Repenning, John Carroll, and members of the BPS student research
seminar for their contributions and input to the study.

Page 1 of 48
(Sterman, 2000; Sterman, 1989). Among those to make an appearance at the System
Dynamics Conference (e.g. Ledet, Monus, Cardella, & Burgess, 2005) is The
Manufacturing Game® (hereafter referred to as TMG or the Game). TMG focuses on the
relationship between equipment defects, maintenance activities, and operational
reliability and performance in a continuous manufacturing facility. Workshops
employing the Game have been conducted in a number of industries, with the majority
being in the oil and gas, chemical, and pharmaceutical industries (Ledet, 2007). The
Game provides an example of system dynamics being used to construct a social
technology (Jay, In Press) - a replicated, iteratively developed methodology for planned
organizational change.

This study examines the actual effects of this technology’s implementation in an oil
refinery, where the Game was used as part of a multi-year program of change labeled as
Proactive Manufacturing. Through our analysis, we seek to inform the underlying theory
of the Game and Proactive Manufacturing and their method of implementation in
organizations. We believe that our findings are generalizable to other programs and
technologies of organizational change, particularly those focused on workforce
involvement in continuous improvement.

Our research questions are simple, building off the basic claims of The Manufacturing
Game® seminars and design intent of Proactive Manufacturing as a social technology.
First, how successful were these efforts at their immediate goal, which is the
improvement of plant reliability through workforce involvement in cross-functional
teams? Second, what was the impact of any changes in reliability on overall financial,
safety, and environmental performance of the plant? Third, how can the answers to these
questions inform the underlying system model, the design of Proactive Manufacturing as
a whole, and continuous improvement activities more generally?

Before we answer these questions, we will give more detail about the program, the site of
implementation, and our research methodology. Then we will present our findings: that
the program was successful in improving aspects of plant reliability and likely
contributed to the refinery’s significant performance gains. We also find, however, that
political conflicts inside the refinery that pre-dated the program - left unaddressed -
continued to influence perceptions and undermined workforce engagement, thus limiting
the potential impact of Proactive Manufacturing. We will attempt to identify the causes
of these dynamics and use our research to augment some core assumptions of Proactive
Manufacturing and similar social technologies.

The Manufacturing Game

The Manufacturing Game® (TMG) was originally developed at Du Pont Corporation and
was built from a system dynamics model of chemical plant performance. Carroll,
Sterman, and Markus (1998) document the history of the Game, which emerged from a
benchmarking study of maintenance costs among Du Pont’s plants. One of the critical
features of the Game's history was that the system dynamics team originally tried to
convey insights from the model to chemical plant personnel through causal loop diagrams
and hands-on interaction with the Powersim-based computer model. These modalities
were not effective in stimulating interest or understanding, which led the team to design a

Page 2 of 48
board game and learning laboratory inspired by the Beer Distribution Game (Carroll et
al., 1998).

There are at least two lessons of the Game that were intended by its designers (Ledet &
Ledet, 2002). The first is similar to that of Repenning & Sterman’s (2002) work on
capability traps - that without investment in plant capability and a willingness to endure a
“worse before better” transition, it is nearly impossible to escape a reactive mode of
system behavior that further erodes capability. In the context of continuous processing,
this means being willing to stop the production process and forego revenues while
maintenance personnel work on the equipment to eliminate existing defects. The reward
for these investments is higher reliability, revenue, and safety and environmental
performance in the long run (sufficient accumulation of defects results in safety and
environmental incidents).

The second intended lesson is about the importance of cross-functional collaboration. In
order for the benefits of proactive maintenance to be realized, there must be coordinated
effort among the three roles in the Game - operations services that manages the parts
shop, maintenance planners who schedule and execute maintenance activities, and
operational managers responsible for meeting product demand and profitability goals.
These players must manage resources together to support preventative maintenance
activities. They can also allocate people from all departments to cross-functional teams
in order to conduct improvement activities and attain even higher performance.
Improvement activities serve to reduce the inflow of new defects, for example by
working with suppliers to reduce defect rates in new equipment arriving at the parts shop.

The system dynamics model used to develop the Game is rather complex and is in fact
proprietary to Ledet’s organization. For the purposes of this study, however, the core of
the model can be depicted as a simple co-flow.

Equipment ————=

Installation Decomissioning
Equipment
Defects
Defects from Defects removed
Design and ry through
- Suppliers decomissioning
+ Breakdowns
Defects Defect Elimination
—_—m through Use Through Maintenance
Cross-Functional - and Poor :
Improvement Teams Maintenance Capability
: ts)
Workforce Participation .
Proactive in Continuous Capability
Manufacturing ———” Improvement _ + Growth ¥
Program
8 Available Performance

Resources

Figure 1 - Core model of The Manufacturing Game®

Page 3 of 48
Here, equipment defects are seen as a property of equipment in the facility. Some arrive
with the equipment in the form of design flaws or poor quality from suppliers. Other
defects arise in the course of use or through improper or unskilled maintenance
procedures. Defect elimination activities can focus on preventative maintenance that
removes existing defects, or they can work proactively to slow the inflow of new defects
(e.g. through re-design, training of technicians, changing suppliers, etc.). If successful,
these defect elimination efforts result in fewer equipment breakdowns, improving the
capability and performance of the organization. These performance gains can generate a
self-reinforcing Capability Growth if the new resources are re-invested in further defect
elimination. The Manufacturing Game® gives players an opportunity to allocate people
among preventative maintenance and defect elimination activities and experience the
resulting improvements.

Proactive Manufacturing

Implementation of The Manufacturing Game® by Ledet and colleagues is not, however,
restricted to game play. As described in several publications about the Game’s use at
BP’s Lima, Ohio refinery (Griffith, Kuenzil, & Monus, 1999; Houshower, 1999; Ledet et
al., 2005), the Game has been incorporated into a system of activities designed to
reinforce and enact the insights of the game into an organization. Together these are
better conceptualized as a program or technology known as Proactive Manufacturing,
which has been compared to Total Productive Maintenance (Nakajima, 1988) by
proponents (Griffith et al., 1999). At Lima, the approach included the following
activities:

¢ Two-Day workshops in which The Manufacturing Game® was played by every

employee in the plant.

e Action teams, formed during the second day of the workshop, in which people
work together across functional lines to eliminate a defect in equipment,
procedures, or practices.

e A Continuous Improvement Forum (or “CI Forum”), meeting at regular intervals
and open to all plant personnel, in which people could present progress and
challenges in their defect elimination efforts. The CI Forum is also used to
deliver leadership and soft skills training to develop the capabilities of those
involved.

e A set of cultural artifacts to reinforce the principles of the game, including slogans
such as “Don’t Just Fix It, Improve It” and “Don’t Let the Bugs Bite” - including
cartoon images emblazoned on posters, banners, hard hat stickers, and
embroidered patches for shirts and coveralls.

This approach was carried on to the Redberg Refinery” described in the present study,
where it was augmented with one additional element: the creation of Natural Work
Teams at each of the processing units in the plant. The purpose of these teams was to
serve as a unit-level analogue of the CI Forum. A coaching and training program

? The organization’ s identity and its members have been disguised for this study

Page 4 of 48
developed “first level learning leaders” for personnel at the manager, supervisor or hourly
levels to help facilitate the learning processes of these Natural Work Teams.

The intention of the consultants and plant managers involved with the Proactive
Manufacturing program at Redberg was to enhance the reliability of the plant through
engaging the workforce in defect elimination. Through these efforts, they hoped to
improve profitability, workplace safety, and environmental compliance of the plant in a
sustainable manner. They employed several threads of innovative organizational theory
and practice along the way:

e Practice fields and reinforcement of learning through action (Kim, 1989; Kim,
1995; Roth & Senge, 1996)

e Team-based organization and workforce participation in continuous
improvement, from studies of Japanese management practices (Womack,
Jones, & Roos, 1991)

e Insights about organizational culture and the importance of artifacts (the bug
imagery) and regular rhythms of events (the CI Forum) in embedding new
dimensions of organizational culture (Schein, 2004).

As we will see, however, in the complex social system of the refinery, these intentions
interacted with the history and social structure of the plant in unforeseen ways. Through
a detailed case history of Redberg Refinery, we hope to illuminate challenges that this
and other similar methodologies must confront.

Methods

In the tradition of Eisenhardt (1989), this multi-method case study is designed for
developing social theory and posing critical research questions. Strictly speaking, the
findings can not be generalized beyond the refinery under study, nor to an understanding
of the impact of other systems thinking methodologies, such as group model building.
We do pose questions, however, that may be relevant and worthwhile to consider in a
variety of other contexts.

Site - Redberg Refinery

Redberg Refinery is a mid-size oil refinery, processing over 300,000 barrels of oil per
day and producing jet fuel, diesel fuel, and gasoline for a variety of markets. The plant
employs around 800 people full time, 600 of whom are hourly workers organized by four
different unions. A few hundred contractors also work at the plant on any given day,
working on new construction or non-routine maintenance.

Over the past twenty years, the refinery has undergone several changes in ownership,
each of which has brought in a new plant manager. These changes were significant for
the refinery workforce, bringing with them changes in policies, compensation,
management philosophy, and job security. It was one of these plant managers, Eli
Larson’, who implemented the Proactive Manufacturing program, having worked with it

° This and all names in the document (other than Winston Ledet) are pseudonyms.

Page 5 of 48
in his previous position at another refinery. Eli retired when the Redberg refinery was
purchased in 2005, bringing in a new plant manager from elsewhere in the new owner's
corporation. The data collection period for this study was January-March, 2007, roughly
a year and a half after this transition.

For most of its history, Redberg synthesized its products from the industry standard W est
Texas Intermediate (WTI) crude oil. This changed, however, when the refinery’s owners
negotiated a contract for heavy, sour crude oil from a foreign supplier at a guaranteed
spread below WTI’s market price. Because this more difficult raw material requires
special equipment to process, the company secured financing for a significant $850
million capital expansion at the plant. As refinery capacity shortages nationwide
increased the market price of finished products, this guaranteed price spread was
exceeded and became an even greater profit margin. Proactive Manufacturing was
initiated during the capital expansion as a way of improving the reliability of existing
equipment and thereby optimizing the return on investment in new equipment. The
combination of technical capabilities from the plant expansion, contract for the sour
crude, market conditions, and reliability improvements worked synergistically to improve
the plant’s financial performance by an order of magnitude.

The downside of this synergy, however, is that the concurrence of management changes
and capital expansions confounds the effects of the organizational practices implemented
during that time. While multiple changes are a limitation for pure evaluation research,
they are also what make the present study more interesting. We found that people at the
refinery have as tough a time understanding what causes performance improvements as
we did. This uncertain causality has kept the sensemaking process active, and has
allowed divergent perspectives to endure about the value of Proactive Manufacturing
relative to other changes in the plant. We interviewed people at multiple organizational
levels, several years after Proactive Manufacturing was the focus for learning,
improvement and change, and found that these perspectives not only shaped people’s
feelings about the past but also explained their actions in the present. These perceptions
provided a window into the underlying political divisions of the plant and some of the
challenges that Proactive Manufacturing, or any continuous improvement efforts, would
have faced.

Interviews

The primary source of data for the study was a series of 66 interviews with 48 people
conducted by the two authors between December, 2006 and March, 2007. All but four of
the interviews were conducted face to face during visits to the Redberg refinery. Most
were conducted as one-on-one interviews behind closed doors, but there were five group
interviews with up to four people at once in a meeting room or operations control room.

Sampling of interview subjects began with Winston Ledet and former plant manager Eli
Larson, and then continued through a snowball process that was coordinated by the
manager of maintenance for the plant. Characteristics of interviewees are depicted in the
table below.

Page 6 of 48
Table 1- Number of interview subjects by job category and functional department

1" Line 2™ Line Grand

Function Hourly Supervisor Supervisor Manager Staff N/A | Total

Clerical 1 2
Engineering al iL. 2
Finance 1 1
HR 1 aT
HSE 1 1 2
IT 1 1
Maintenance 3 7 5 1 16
Management 2 2
Operations 9 2 1 1 13
Services 1 3 3
Training 3 3
Consultant 1 1
External 1 1
Grand Total 13 11 6 9 8 T 48

As a result of partnering with the maintenance manager as the primary contact inside the
plant, the sample of interviewees is by no means representative of the organizational
structure of the plant. It is heavily weighted towards salaried managers, supervisors, and
staff (n=34) with hourly workers making up a small part of the sample (n=13),
particularly given that of the roughly 800 employees on payroll, close to 600 of them are
hourly. The sample is also heavily weighted towards the Maintenance function (n=16),
whereas Operations is a far larger organization within the refinery.

With these limitations in mind, we did attempt to speak with structurally significant
members of the under-represented groups - we talked to the heads of the three largest
unions of the plant, including the committee chair of the local United Steel Workers
union that represents 500 of the plant’s hourly employees in operations and maintenance.
For the purposes of this research, which is to identify the political divisions and
narratives surrounding the Proactive Manufacturing program, this sample was adequate
to capture the local complexity.

Interviews lasted anywhere from 30-120 minutes, although most were 60 minutes long.
They were recorded on digital audio files that were compared to handwritten field notes
and transcribed for key quotes. Interviews consisted of three phases. In the first phase
the informant was asked for a brief career history, and we recorded the sequence of jobs
they had filled both prior to and since arriving at the refinery. In the second phase we
presented our informants with a timeline of events at the plant, divided into two columns:
one column for business events and results (e.g. the capital expansions and changes in

Page 7 of 48
ownership) and another column for improvement programs, including Proactive
Manufacturing. After presenting them with the timeline, we asked them to tell us stories
about events they found significant in the plant’ s history or about those they were directly
involved in. If we had omitted any events they thought significant, we encouraged them
to tell us, and we progressively edited the timeline with each interview. The third phase
of an interview involved asking specific questions, or clarification of what the respondent
said that might have differed from what we heard from others.

Observations

In addition to formal interviews, one of the authors conducted approximately eight hours
of ethnographic observations in control rooms, in morning meetings of maintenance
managers, at supervisors’ morning handoff meetings from the night shift, at a
maintenance planning meeting for a section of the plant, and at a root cause analysis
meeting for an equipment failure. During these observations, he took hand-written field
notes which he transcribed with additional observations and reflections each night, along
with notes and reflections from interviews. The other author was involved in a
consulting, training, and facilitation role for the first level learning leaders and with the
plant management team from 1998 to 2002. Records from that clinical engagement
served as a documentary data source to supplement his recollections.

Documentary sources

To supplement the interviews and observations, we collected a great deal of archival and
documentary material from the plant. This material included:

e Record keeping about The Manufacturing Game®, action teams, CI Forum, and
Natural Work Teams, including attendance

e Documentation of CI Forums, including video recordings of three sessions and
PowerPoint slides from almost everyone of the meetings

e Historical interviews of Redberg’s managers: One set were interviews with plant
management team members was conducted in 2000 by the consultant introducing
The Manufacturing Game®. Another set of interviews and notes were conducted
by the second author of this paper and provided as feedback in facilitating a series
of quarterly leadership team learning meetings from 2000 to 2002.

e Financial records, including variable costs for oil and energy purchase and
operational fixed costs including expenditures on labor and maintenance. These
records only date back to 1995; before that the records are the property of a
company with whom we did not have a research agreement

e Safety and environmental records, including OSHA and EPA recordables for the
years before, during, and after the Proactive Manufacturing Program

e Feed Rate and Reliability records for each of the major processing units in the
plant, from the period of 1991-2007

Page 8 of 48
Analysis

Analysis of the data was primarily qualitative, using the method of grounded theory
(Glaser & Strauss, 1967). Transcriptions of fieldnotes and interviews, along with other
archival documents, were imported into a database and analyzed using the Atlas.ti
qualitative data analysis software. Elements of these texts were coded by the first author
and reviewed by the second author using concepts emergent from the data and from the
academic literature on organizational leaning and change. Some concepts were linked
and grouped into themes in order to build the theoretical insights described in this paper.
Specialized queries and searches allowed some hypothesis testing and verification, along
with gathering of quotes cited as evidence below.

Where appropriate with the quantitative time series data, simple regression analyses were
conducted to detect linear and quadratic trends over time.

It is worth noting that clear analysis of the interview data in this case is by no means
straight forward. Perceptions of the value, the meaning, even the sequence of events
involved in Proactive Manufacturing differ sharply depending on whom we asked in the
refinery. Some were clearly advocates of the program, saw it as an approach and
philosophy of work, and believed it had been critical to the operational and financial
success of the plant. Others thought the program was a fine idea but irrelevant and
poorly implemented, paling in comparison to the capital expansions occurring in parallel
or the changes in management before and after the program. Still others, particularly in
the hourly ranks, were opponents of the program, who thought of it as an attempt by
management to get them to work harder, and that the net impact on the plant has been
deleterious.

For many of the respondents, however, their views were mixed. Often they would talk
about the value of the program and its early successes, and point to ways that current
practice reflects the ideals they learned. Then they would add additional stories about
failures to invest in necessary maintenance improvements or about subsequent
management behaviors and programs that undermined the spirit of cross-functional
collaboration. Given this complex picture, it would be more accurate to organize the data
in terms of narratives rather than opinions or attitudes (Czarniawska-Joerges, 1997;
Ewick & Silbey, 1995). Instead of forcing the data to fit into an assumption that people
hold unitary views, a narrative mode allows us to describe the raw material and process
of sensemaking (Weick, 1995).

Results

Timeline of events

The most basic result of the research, which we developed iteratively in conducting our
interviews, was the development of a timeline of the Redberg Refinery’s history. While
our informants described a variety of events, a few were repeatedly cited as critical to the
plant’s culture and performance and the careers of the people who work there. They also
provide an important backdrop and context for the implementation of Proactive
Manufacturing, which helped generate its successes and challenges.

Page 9 of 48
Table 2 - Timeline of events at Redberg Refinery

Year Management Event
regime

Late 1980’s | OldCo TQM implementation - early attempt at workforce
involvement in continuous improvement

1992 OldCo OldCo reduces dual train refinery to single train
refinery, massively cutting personnel from 3000 to 800
and decomissioning equipment to shave operating costs
in preparation for sale

1995 SmallCo - SmallCo purchases Redberg for $89 Million

1995-1998 Sam Taylor Efforts by new managers to reach out to hourly
workers lead to conflict with supervisors.
Early investments in reliability improvements

1998 Announcement of deal for heavy, sour crude and
Heavy Oil Upgrade Project (HOUP)

1998 SmallCo - Arrival of Eli Larson as plant manager

1999 Eli Larson —_>yfanufacturing Came and action teams begin
“Boris the Bug” imagery and slogans posted

2000 Continuous Improvement Forum established

2000 HOUP construction begins, financed by $650M in
loans.

2001 SmallRefCo - | SmallCo sells off retail, becomes refining-only

Eli Larson company SmallRefCo

2001 Natural Work Teams established, Productive
Conversations coaching commences

2002 HOUP construction finished; heavy, sour crude
processing begins

2002 IPO of SmallRefCo - hourly workers excluded

2002 25% reduction in salaried workforce for cost control

2003 Lubrication program begins - offshoot of Natural
Work Team

2005 MidCo MidCo purchases SmallRefCo for $8 Billion. Nearly
half of the company’s capacity is in Redberg and more
than half of total margin, placing its evaluation at close
to $4 Billion. Eli Larson retires.

2007 Research for this paper conducted

Page 10 of 48

In this timeline, the grayed cells highlight the implementation of elements of Proactive
Manufacturing. A few things are critical to note about events before, during, and after
the implementation of this program.

e Before Eli Larson’s arrival, the plant management team led by Sam Taylor
implemented two key policies that would impact later events: beginning to invest
in maintenance and reliability improvements; and reaching out to hourly workers
through a Union Management Team and open-door policy. The former is critical
in people’s sensemaking about plant performance; the latter shaped the political
landscape into which Proactive manufacturing arrived.

e Simultaneous with Proactive Manufacturing, as discussed earlier, the Heavy Oil
Upgrade Project significantly changed the supply and product mix. This
substantially boosted the financial performance of the refinery. With increases in
the market price of fuels, these organizational, technological, and strategic
improvements increased the annual profit margin of the refinery by 3,150%
between 1999 and 2006.

e These capital investments created a temporary situation of limited cash and
extensive debt. In this context, Proactive Manufacturing came with substantial
cost and risk - the initiators of Proactive Manufacturing repeatedly emphasized
that the program was initiated by Redberg itself, without the backing of
SmallCo’s senior management team. It also required an investment of more than
$1 Million in consulting fees, licensing, and personnel time away from regular
work. The successes achieved were therefore highly significant to leaders and
advocates of the program.

e After Proactive Manufacturing efforts began, two events occurred that affected
morale, sensemaking, and participation in continuous improvement activities: an
IPO of the company that brought financial gains to managers while excluding
hourly workers; and a reduction in salaried workforce that eliminated numerous
staff and supervisory positions.

With this context in mind, we can begin to analyze the effects of Proactive
Manufacturing on equipment reliability, the influence on overall plant performance, and
the political dynamics arising in the course of its implementation.

Impact on reliability efforts and outcomes

To the extent that the program was successful in improving reliability of the equipment,
we would expect to see improvements in mechanical availability, as well as qualitative
evidence that the program left its mark. While not perfectly conclusive, the available
data are indicative of a successful program that paid for itself many times over.

Action teams and maintenance efforts

Action teams formed after The Manufacturing Game® workshops are the route of most
direct influence for such improvements. On the second day of the Game workshop,
participants are asked to form small teams composed of people from multiple levels and
functions of the plant. These teams were assigned problems to solve in the plant, framed
as defects or “bugs” in the plant. If the team knew of a more pressing issue to tackle then

Page 11 of 48
it was their prerogative, but the majority seem to have taken on the project assigned. A
spreadsheet used to track action team status from 1999-2000 includes 88 action teams
formed around pre-existing issues and 27 “proactive” teams who developed their own
project. Success stories about these teams abound, and managers credit these teams with
financial benefits that paid for the Proactive Manufacturing program several times over.

At the close of the second day, at the end of the day, we say “ we’ ve got these
projects and think you would work good on this team.” They split up into teams
to work on these projects. Some of those projects have actually been successful,
they were successful and we still talk about some of them today. One of them was
one at 43 they had a seal that kept going out on a certain pump... Well,
[engineer] started asking questions about this pump and why, you know, the next
thing that come up was, “why do we run that pump?” We run it because of this.
Next thing that come up was, “Fix it? We can shut the whole system down.” We
didn’t really need it. They were just kinda runnin’ this system. And they shut it
down and saved so many dollars because they didn’t have to fix the problem, 1.
And 2, it was a bunch of hardware you didn’t have to upkeep. It was a major
accomplishment.

We had refractory [insulation] in the regenerator at the FCC to keep the metal
from burning up... Refractory would fall out of the overhead of that regenerator
and then you had to shut the cat cracker down because you had an unexpected
fall down. Well, when the Manufacturing Game came in, that just happened to
happen sometime during that time frame. Here was a chance to put their money
where their mouth was. They decided that they were going to do it right [the
management team]. They said it might take us a few more days to stay down, and
may cost us some extra dollars, but they had proven to those who worked in that
area that they really meant what they were saying... They didn’t just patch it,
they cleaned it up and put it in right the way it should be. It cost several thousand
dollars to do that, but the end result is they didn’t have a fall down again. Matter
of fact, I don’t know that to this day, it has fallen down since.

There was a team put together to do one thing, but they said we’ve got a problem
over here. One of them was at the old coker. They had a problem with the air
system - it was always dirty. Water and stuff in it. Their team came up with a
design to dry the air out, have cleaner air for their tuggers - they use air tuggers
in the drilling process... And we published that for many years, for a long time,
and presented it at the Manufacturing Game.

I was on the one that won the President's Award. We had a lot of leaks in the
sewer out here, and we were trying to identify where these streams were coming
from. There is a company that came in and helped us do a fingerprinting process.
They run tests on samples and they could tell exactly what stream was going into
the sewers and we could backtrack and fix the problem. We saved several million
dollars that year. Everybody knew that, you know, at the waste water plant we
were killing bugs out there [specialized algae and bacteria used to clean waste
water]. We were killing off the bugs, and it can be pretty expensive. And then we
get cited also along with that. So we knew we had to address it.

Page 12 of 48
They were not big things - they were small things. Nothing wrong with that.
Enough small improvements make a big change in the bottom line. One thing is
that they fixed the blowers at 1241. The blowers, that’s something that brought in
combustion air for the regeneration process. They were falling down all the time.
Every time one of them would fall down, they would cut throughput for the units
[which cost $80K/day]. They put together a team of operations, maintenance,
everybody. One thing that’s nice, you do have cross-functional teams. They
looked at equipment and found it was an alignment problem between the driver
and the blower... They went from 6 months or less on a run to 2 or 3 years
without falling down... Savings were $700K/year. See that is significant with the
operators too because they don’t have the blowers falling down at 2 in the
morning when it’s 30 degrees outside.

It’s only mechanical about 20% of the time. That's where Eli Larson, why we
have so much faith in this thing, even though you spend half a million dollars [for
the program], that sounds like a lot of money. No matter what plant you go into,
they are wasting money hand over fist... it’s easy to find something that will pay
for that investment. We found something within the first few months that paid for
that three times over. It had to do with a flaring process we were doing here. We
were importing gas from [supplier]. The procedure simply called for us to import
it at a certain rate, which was more than we could handle. All we had to do was
slow that down. Because we were bringing it in at such a high rate that we ended
up flaring a lot of it. Simply by changing the procedure and getting with the
supplier to say, hey, we need to take this at a little bit slower rate. They said, OK,
and it was no big deal. All of a sudden we stopped flaring about a million dollars
worth of gas a year... They worked all that out, and found out hey, this is a
procedural thing. You know how much that costs? It costs ZERO.

We got together as a group of admins and stenos. We said what are you doing,
how are you ordering, where are you getting this from? How can we make this a
group effort where we get the most for our bucks? Best quality at best price.
That fell into food service as well as office supplies... After we looked at, OK, this
is what we are doing, this is what is resulting, how can we make this work to
where we can get the most for our buck? We got a team together and evaluated
the office supply situation, compared different office vendors, supply vendors and
looked at maybe a new contract would be good if we could get corporate wide.
That’s what we did. SmallRefC o-wide contract for office supplies. It worked out
to where we saved at least $50,000 over a period of time for that year, because of
using a source that was common to everybody as opposed to a source here or
there.

In these stories, we have numerous examples of Proactive Manufacturing directly
affecting behavior and plant performance. While detailed impact analyses of these teams
were not always conducted, they were repeatedly cited by informants as having led to
improvements in reliability and reductions in operating costs. Informants also
emphasized that these early action teams set an example that is now followed as part of
the regular, everyday activity at the plant - cross-functional teams for defect elimination
are now commonplace.

Page 13 of 48
As further evidence of the plant’s commitment to reliability improvements, we can see a
surge in maintenance expenditures from 1999-2001, the period during which Proactive
Manufacturing was initiated. These costs then decreased from 2001 to 2003, which may
be evidence of the “worse before better” transition (Repenning et al., 2002) predicted by
The Manufacturing Game®. Once preventative and proactive maintenance begin to
improve reliability, the costs due to unplanned maintenance may have decreased.

60,000

50,000 /
40,000 IN

/ A

/ [-*= Maintenance Costs

30,000

20,000 7

10,000

$ (Thousands)

1994 1996 1998 2000 2002 2004 2006 2008
Year

Figure 2 - Maintenance costs at Redberg Refinery, 1995-2006

Following this period, maintenance costs then began increasing again from 2003-2006.
There are three possible reasons for this increase: it may be indicative of further
preventative and proactive maintenance investment; it may be a result of the organization
backsliding into reactive maintenance; or it may simply reflect the costs associated with
integrating and maintaining new equipment brought online with the Heavy Oil Upgrade
Project in 2002. This is an example of how plant-level indicators are difficult to interpret
during the post- 1999 period because of the simultaneity of multiple processes in the
plant. To get a clearer picture, it is critical to examine more of the quantitative and
qualitative data.

Rotating equipment reliability

At the most micro level of rotating equipment (pumps, motors, compressors, etc), internal
data show significant improvements in reliability following Proactive Manufacturing.
Figure 3 below, extracted from a Computerized Maintenance Management System
(CMMS), shows a steady decrease in failure rates from 2001 to 2004. Ideally we would
observe data of this type from before 2001 in order to see the longer term trend, but there
are two notable features of this data display. First, the reason the graph begins in 2001 is

Page 14 of 48
that this is the year when the CMMS was implemented, which itself was a part of
reliability investments at the plant. Second, 2001 is a year after The Manufacturing
Game® was completed and the CI Forum began. This evidence is suggestive (but my no
means conclusive) that Proactive Manufacturing efforts contributed to reliability
improvements.

Figure 3 - Internally circulated data on Failure Rate

Rotating Equipment F ailure Count by Equipment Types
for 2001, 2002, 2003, 2004

cent pone mTuReNe——_uoron TE CENT_COMPRLUBEIRECIP_PUNP
mew Fan [BBRECI COMP mmICRANE mm GEARS ano ciiss

;
a

a a i

No. Of Failures

Lubrication Program

One compelling manifestation of this change in culture has been the development of
additional, more focused improvement programs by participants in Proactive
Manufacturing. In the context of a Natural Work Team, one hourly operator with a
passion for machines became curious about high rates of motor breakdown and its
relationship to lubrication practices. The company funded him to attend a conference on
lubrication, and then helped him attend a four-day course to be certified in lubrication
science and technology. He used this education to begin asking questions about
lubrication use at Redberg. His inquiry, in partnership with a maintenance technician in a
refinery laboratory, discovered mismatches between lubrication and metallurgy and
problems with contaminated lubricants related to operations practices. Beginning with
experiments in his own area of the refinery, he discovered the increase in MTBF and
decrease in maintenance costs associated with proper lubrication.

This local inquiry then led to a systematic overhaul of lubrication procedures and a round
of training for operators and maintenance personnel about the science of lubrication,
which seems to have increased MTBF on critical equipment. Advocates of continuous
improvement efforts know about the lubrication program and cite it as the cutting edge of
the program. And it is an example of the success that comes from connecting a
philosophical approach to cultural artifacts and a willingness to invest in following
through.

Page 15 of 48
Along with the Natural Work Teams, came the bug thing - don’t just fix it.
Something like this allowed me to stand up and say, “ hey, guys you're preaching
this, but all we did was repair this machine again. We didn’t find out what's
happening to it. We’re just fixing it all the time.” After I had a little education,
knew what I was talking about... Well, it kinda put the company on the spot.
We're going to preach this, then we gotta practice it. More than once [I pointed
to the bug]. No doubt, it gets people's attention. And it did. With everybody
involved, and different ideas, between machinists, like I said, as a group making
decisions, instead of just one person... Started using 1 micron oil everywhere in
the plant... delivered straight from the truck - that’s a standard. There’s not
many suppliers who can do it for you...

We started our own training program out here in 2005... We trained, we brought
all the operators and all the machinists to the training course... It took about a
year to get everybody through it. And it’s not easy - you got... you walk in and
try to make a lot of changes to people my age in this kind of service and they're
going hell, we’ ve been doing it this way for 30 years, what's the matter? You still
gotta get the operators to focus. I mean... keepin’ things clean would be the
biggest thing. Not checking a lot of things they could be checking.

From the perspective of the Proactive Manufacturing program architects, this story is
exemplary. It is not, however, without its challenges. The lubrication program requires
cooperation among operations and maintenance, both of whom must utilize the right
lubricants and monitor their effects.

At the management level, the relationship between operations and maintenance that the
Proactive Manufacturing program tried to institute appears to be well modeled. At the
union level, however, cross-functional collaboration is not as simple. Operators and
many maintenance crafts people are organized by the United Steel Workers. This is by
far the largest union, representing 500 hourly workers. Electricians (who do
instrumentation) are organized by the IBEW, Machinists by the IAMAW, and clerks by a
division of the USW, all of which add up to fewer than 100 people. Structurally this
disparity in size poses a problem. Conflicts over turf can be poignant, and can impede
cross functional efforts like the Lubrication Program. A fter the more positive part of the
head operator's story, we heard the following:

This has been my struggle. Lubrication belongs to machinists. I have been
allowed to practice this in my area... It’s kind of like I was invading their
territory. I had one of them almost choke me for it. The only reason I’ve been
allowed to practice is that their work has gotten so busy in repairs that their
preventive maintenance work is not getting done. Some of them are glad - I didn’t
want to mess with those filters anyway. Someone like them sees... I get my hand
slapped. So it’s really been a struggle with two different unions involved.

The limitation of our retrospective methodology is that it is unclear whether these and
other conflicts existed before the inception of the program, undermining cross-functional
work throughout its history, or if this is a more recent phenomenon. Regardless, while
the Lubrication program has on the whole successful, it highlighted the importance of
political dynamics, which we will discuss in more depth in a later section.

Page 16 of 48
South side green-field culture

Changes at Redberg associated with Proactive Manufacturing are particularly noticeable
on the south side of the plant where capital expansions associated with the Heavy Oil
Upgrade Project occurred. To facilitate financial management of the loans needed to
build the new equipment, that part of the plant started as a separate subsidiary company at
the start, and workers were employees of that subsidiary. Although personnel and
managers were drawn from the existing organization to work on the new equipment, the
south side has the quality of a partially “green field” area. A single control room filled
with dozens of computer screens controls multiple processing units; this affords different
operational procedures than the older sections of the plant where small, unit-specific
control rooms are the norm. Other innovations, for example the installation of a
simulator for training of operators, helped create the idea that the south side would be a
site of innovation and high performance. The manager in charge of setting up south side
operations was also a particularly strong advocate for Proactive Manufacturing, carefully
chose the workers for the new plant, and sought to infuse the new culture with this
philosophy.

In our interviews with south side supervisors and hourly workers, we noticed a level of
both consensus and enthusiasm that differed from the other section of the plant. One
head operator put it as follows.

[On the south side] management and operations actually work together. They
don’t fight each other as much. It’s not an us-against-them thing most of the time.
I’m sure on the other side of the yard you hear that a lot... Philosophy is different
from what I was used to on the other side of the yard. This was the first time I
heard proactive. I didn’t have a clue what it was. They actually try to be
proactive here. Try to do routine preventative maintenance stuff. Actually do it,
not just have it on paper. Try to head something off before there’s a problem.
Most of the time. It’s totally different from the north side... First day I was over
here, [manager] said this is how it’s going to be, and it actually has been. Before
people wouldn't talk to somebody that high. They never came around. Now you
see them all the time. You can talk to these people just like you and I are sitting
here talking. To me there’s not that much animosity between the jobs... It was
the purpose at the time, because if that project didn’t go through, the whole plant
would have been sunk.

This quote indicates both the increased capability of the south side and the way that
capability derived from being a site of novelty and innovation. Another head operator
and the maintenance and operations supervisors were all highly optimistic about the
culture and performance of their units. Some of the common worker-supervisor conflict
and discontent we documented (described below) did not seem as active among south
side personnel.

We also received reports from an outsider in the engineering team that meetings and
cross-functional coordination practices were particularly effective. This led the first

Page 17 of 48
author to observe two meetings - one planning meeting focused on maintenance
activities, and one root cause analysis team inquiring into the repeated failure of a
ventilation door on one of the units. Both meetings convened supervisors and hourly
personnel from maintenance and operations as well as some engineering staff. Both were
conducted with a high level of energy and precision as compared to the plant-wide
moming handoff meetings we observed. And both meetings involved moments where
the principles of The Manufacturing Game® were clearly and assertively brought into
play.

At one point in the planning meeting, an operations supervisor named Max Neville
paused the group to emphasize the importance of coordinating schedules among
operators and maintenance to facilitate preventative maintenance activities. Then, a
moment later, when a maintenance supervisor described a successful repair completion,
Max again paused the conversation to emphasize the cost savings involved in that
activity. This use of “teachable moments” to encourage a proactive and collaborative
approach characterized Max’s participation in the Root Cause Analysis meeting as well.
When it seemed that the group was converging on a quick-fix solution to the failed
ventilation door, Max emphasized the importance to endure extra down-time and
operational cost to “fix it right the first time,” using almost verbatim one of the slogans
from Proactive Manufacturing that hangs on walls throughout the refinery.

These observations highlighted both the embedding of Proactive Manufacturing
philosophy into the culture and practice of the refinery and the need to continually
maintain it. Max’s assertions of proactive approach are pedagogical - an attempt to
change or maintain proactive principles. Later in an interview he emphasized the
importance of ongoing work to stay in a proactive mode. A strong advocate of the
Natural Work Teams (NWT) approach and host to what seems to be the longest running
NWT, Max recounted his recent need to re-emphasize principles of disciplined operation
with the team because safety incidents had begun to occur under his watch.

More generally, our interaction with Max also highlighted the importance of supervisors
in maintaining shifts in organizational culture. Although his recent NWT meeting got

labeled as the “ass-eatin’ meetin’,” it was clear that Max maintains a style of empowering
management compatible with a workforce engagement approach.

If you tell people what to do, and give them the tools, most of the time they will
produce it for you... that’s why I’m talking to you relaxed because I know they’re
on the job. I have time to look down the road a bit.

He encourages Natural Work Teams to form action teams to work on specific
improvement efforts, for example on tolerance alarms or maintenance procedures. But
these projects don’t occur until a worker volunteers and educates themselves to become a
“guru” in that area.

The difficulties that Max does face highlight the importance of the distinctive south-side
culture. When he has had workers transfer from the older section of the plant into his
unit, they have had trouble adapting to an atmosphere where they are expected to take
charge and engage in continuous improvement. Four of the six workers who transferred
there from a utility plant ended up returning to their old unit after the “ass eatin’

Page 18 of 48
meetin’.” Although the two that remained fit in better with the south side culture, one of
Max’s head operators expressed similar concems.

It's slowly changing back the other way... That sense of being needed or
whatever, don’t know if that makes sense. It has gotten to the point where okay,
you're established, we got what we wanted, now we're going to go back the other
way. More friction, us against them. It’s headed that way... Just a feeling. A
sense... [When things were starting up] everyone was on the same page, we was
all together, one big happy family. Now we're getting’ new people over here,
they didn’t go through that. Sometimes the buy-in and stuff is not what it should
be, in both directions. Both being management and labor or whatever.

As we will discuss further below, the new intended behaviors and social structures of
Proactive Manufacturing and continuous improvement are always subject to buffeting
and retrenchment into pre-existing ways of operating.

General positive accounts

Beyond the specific stories and localized efforts, we heard general accounts that were
consonant with the intended outcomes of Proactive Manufacturing and describe changes
in the organizational culture.

One major thing I think Proactive Manufacturing did, and the idea of a cross-
functional team did... prior to that, there was finger-pointing. There was the
tendency to lay blame [among operations, maintenance, and engineering]. Lots
of finger pointing and wanting to assign accountability to another organization.
The one thing that Proactive Manufacturing has done by developing the cross-
functional team approach... when we got an issue to be resolved, there is
engineering, there is operations, there’s maintenance, there may be
enviornmental health and safety... whoever the right folks are, they get together
and talk about it, and they solve the problem. And many times it goes
undocumented...

When you make the transition out of the finger pointing mode, and people work
together on some of these things, when something doesn’t go right, then if it’s an
operations issue then operations is much more willing to put their hand up and
say “this is our fault. We let that bearing run dry. Or we let that pump run dry.”
Rather than try to cover it up and try to bring the maintenance guys or
engineering guys to try to solve a problem that didn’t exist in the first place.

-Member of Management Team

We ran this refinery for years as reactive. Every time something broke, run out
there and fix it. We try to be very very proactive... Let me put it this way - we
have to he, to survive, to be proactive. I have 8 guys, and several thousand pieces
of rotating equipment. You do the math. There is no way you can repair all this if
it’s falling down all the time with 8 guys. We have to work smarter. The repairs
we make have to last longer. The troubleshooting has to be better. The material
we go to has to last longer. The conversations you have with operators - what

Page 19 of 48
actually is running in this? With the engineers - are we seeing H2S in this? What
is the limit on the amine here? We have to have better conversations, we have to
talk better, we have to be more clear so that we can provide better equipment so
we get longer runs on our equipment.

But the thing about Proactive Manufacturing, what that does is give you the big
picture on how one piece of the puzzle affects everything else. That’s what
Proactive Manufacturing is about - it’s seeing the big picture. You have to start
there. You can’t have people in their own little world, just working on this here.
They have to understand this pump affects this unit, this unit affects this unit, this
unit affects the other side of the refinery, which affects the whole refinery.
Proactive manufacutring is to put in the minds of the people an understanding of
how what we do - what every individual person does - how important it is to the
whole organization.

-South Side Maintenance Supervisor

It seems to me that things have taken time but things have changed. It has taken a
long time but has had some effect. They stayed with the program so long that it’s
now part of the culture. But there are still those who don’t buy in... They don’t
believe in it... Some of them, they are just old school, old paradigms, where it’s
“once again, here’s another program.” Apathetic look - like “that won't do
nothing for us.” But if you look at it in long run, I think it really has improved
things... The improvement in the attention to the equipment, and the seriousness
of keeping equipment in good shape.

-Hourly Operator

For the maintenance managers, Proactive Manufacturing has been a powerful influence.
As Carroll, Sterman, and Markus (1998) point out, maintenance is not part of the “image”
of a plant, usually being considered a service or cost center. The Manufacturing Game®
attempts to shift that discourse and thereby legitimate the role that maintenance plays in
plant performance. Proactive Manufacturing is, for them, an ideology and a way of life -
in essence, the right way to do the job. Their narrative of success is woven with a
narrative of the “good old days” and of a lost hero in the form of the plant manager Eli
Larson. In a group interview with the second-line maintenance supervisors:

S - Eli Larson brought Proactive Manufacturing, Productive Conversations.
People thought it was flavor of the month. But driven by plant manager who
walked the talk. Operations and maintenance was us and them. We got more
trust and respect - issues easy to solve. Eli set expectations, you knew where you
stood.

W - Eli preached it every day - talked about it every day. Pounded in idea of
reliability. Insisted that we fix things right - long term solutions. Eli knew
everyone.

S - Knew weaknesses and strengths. Empowered you to do work. Best manager
I’ve ever seen.

Page 20 of 48
Other directors in the management team - engineering, health/safety/environment - take
the stance of favorable bystanders who reaped ancillary benefits from the program.

The action teams that were spawned out of the Proactive Manufacturing
workshop concepts, and the notion of employee involvement and getting
employees directed towards solving problems was a way to real solutions... I love
that concept. That was something that OldCo really tried to get going through a
quality movement in the late 1980's with some limited success. It really took off
here, and helped in the environmental area because a lot of the problems -
equipment reliability problems - one of the consequences of those problems are
environmental emissions or excursions. So people working on those problems
were solving my problems. And also there were several teams... in fact 1 counted
at one time there were at least 23 teams - Proactive Manufacturing teams - that
were working on solutions to environmental problems.

[I ask him to compare TQM and Proactive Manufacturing]

I would have to say... I would have to give the edge slightly to the Proactive
Manufacturing movement here. I think it was more effective. I think it won over
more people than the quality movement. I think people attended a lot of the
quality meetings and training classes because they were obligated to do it but in
my recollection there wasn’t that level of volunteerism and commitment to the
movement that we had in Proactive Manufacturing. PM had the edge. I also
have to say that Eli as a leader of it, and face and spokesman for it... he was
much more effective than his counterparts in the quality movement... He was
determined to win over the hearts and minds... he was really committed and
really believed in it, and he was determined to persuade as many people as he
could at Port Arthur that it was really the way to success and sustained success,
through reliability. Reliability in our equipment, our procedures, and even our
people. We had to eliminate breakdowns and failures, and wherever we did that,
success would follow in all these other areas. It was a very compelling argument.

-EHS Manager

The current plant manager, who assumed his position when Redberg was acquired in
2005, is intrigued by what occurred, and is interested in the findings and
recommendations in our case study as part of his thinking for how to go forward.

Other supporters include supervisors and superintendents from both maintenance and
operations. Typical benefits cited include:

Improvement of Mean Time Between Failure and other measures of reliability
Erosion of “us and them” between maintenance and operations

Fixing it right the first time - going an extra step in maintenance activities,
realizing that doing so leads to reduced workload in the long run - “If you can get
to a place where mechanics can do preventive maintenance and not fight fires all
the time.”

Ease of pulling together teams to work on necessary repairs and improvements

Page 21 of 48
Plant-level performance

As we can see from these responses, Proactive Manufacturing seems to have had a
powerful effect on aspects of plant reliability and on the organizational culture
surrounding maintenance and reliability improvements. The contrary views highlight
some challenges the program has faced, and the need for ongoing reinforcement of the
ideas, but the picture is on the whole positive.

When we examine the unit-level and plant-level reliability data, that picture is somewhat
less clear as we will see below. To the extent that the program was successful in
improving reliability of the equipment, we would expect to see improvements in
mechanical availability and uptime, with concomitant increases in profitability. We
might also expect to see a decrease in safety and environmental incident rates due to
equipment failures. While substantial performance gains are evident, their timing relative
to the simultaneous influence of the Heavy Oil Upgrade Project (HOUP) and Proactive
Manufacturing make causality difficult to determine. What we found was that this
ambiguity kept sensemaking active and allowed for multiple interpretations both for us as
researchers and more importantly for people working in the refinery.

Plant-level up-time and reliability

Reliability of the plant is measured by counting the number of hours during which
processing units are taken offline, and comparing it to the ideal case of full operation - 24
hours per day, 365 days per year. The difficulty in accounting for these “down hours” is
that there are several reasons why a unit might be taken offline:

1. Unplanned maintenance, as during an equipment failure

2. Planned maintenance or regulator inspections, including turnarounds that occur at
regular intervals

3. Economic down time, in which one unit is slowed or shut down because its product
is less profitable than others’ in the plant

4. Shutdown of equipment upstream or downstream of the unit that require the unit’s
shutdown in order to halt the flow of product

In order to examine the impacts of a reliability program, it is best to look at operational
availability, which is based on categories 1 and 2. Category 3 is omitted in order to
exclude the influence of external price fluctuations that yield economic downtime.
Category 4 is omitted in order to avoid double-counting outages due to mechanical
failure (although doing so creates a conservative measure of the impact of operational
down time).

The graphs below present operational downtime data for 11 of the plant’s 19 units. These
11 were selected because reliability data was available for the period 1994-2006, with
some data extending back to 1991. A simple graph is presented as Figure 4.

Page 22 of 48
Figure 4 - Operational Down Hours for selected units

Operational Down Hours (Including Turnaround)

2500

2000

[-e= Crude Unit
|-=—Vacuum Tower

Reformer 1344
1500

The challenge of making sense of this graph puts readers in the shoes of people inside the
plant. First, no time series graph of this sort had been produced as part of the archival
materials we saw. Plant level data about reliability, combining data from multiple units
around the refinery, is available primarily to personnel in the operations services group.
Quantitative data is not, of course, the only raw material for sensemaking - direct
experience of working as an operator or maintenance technician certainly grants a sense
of trends over time, and discontent with poor availability of equipment can get conveyed
through stories and other kinds of talk. But a systematic analysis of trends over time, let
alone their connection with interventions like Proactive Manufacturing, is beyond the
likely scope of such conversations. Reliability data, whether presented quantitatively or
through conversation, must make sense of complex stochastic processes involved in the
failure of equipment and the need for maintenance. These ongoing change processes in
the equipment itself create ambiguity about the influence of systematic, deliberate
intervention.

For the purposes of our analysis here, however, we were more precise with the data
available. Following the recommendation of the plant's reliability engineer, this
somewhat chaotic data can be smoothed out by normalizing the effect of planned
tumarounds. For each of the units, the number of down hours due to tumaround was
summed and averaged over the period of observation, yielding a measure of Adjusted
Operational Down Hours. This measure can serve as a dependent variable in regression
analysis, with measurement year as a predictor, in order to analyze trends in operational
downtime.

Page 23 of 48
Figure 5 - Operational Down Time, analyzed across 11 refining units.
Planned turnarounds have been normalized across the observation period.

Operational Down Time

8
3

Quadratic fit
e—S 95% Cl
———- Median spline

Hours Per Year

8
sj es & wow ew See ww ww
PLP LS LS SP SS SS SS SS 5

Year

Here in Figure 5 we can see that there is a general downward trend in operational down
hours in the plant, and thus an improvement in operational reliability. The polynomial
regression has a somewhat better fit than a linear regression, however (Linear: F=6.75,
p=.01, R2adj=.0335; Polynomial: F=5.38, p=.006, R2adj=.0501). This indicates an
increase in operational down hours, peaking in 1996, followed by a decline. This is not
the pattern one would expect if Proactive Manufacturing (which began in earnest in
1999) had been the most significant influence on reliability. Instead we see improvement
beginning in 1995, which is consistent with the idea that pre- Proactive Manufacturing
efforts to improve reliability may have had an important influence.

Still, the data does not contradict the idea that Proactive Manufacturing helped. The
median spline in Figure 5 shows rough behavior over time for operational downtime at
the refinery. In particular, from 1999 to 2003 we can see a pattern that could be
interpreted as a “worse before better” transition - an increase in operational down hours
for proactive maintenance in 2000 followed by an improvement in plant reliability over
the subsequent three years. Following this same logic, however, we could say that there
is a larger worse-before-better pattern coinciding with the overall tenure of SmallCo from
1995 to 2005, beginning before Proactive Manufacturing.

Health, Safety, and Environmental Results

As the refinery’s manager of health, safety, and environment (HSE) indicated, many of
the Proactive Manufacturing efforts were focused on safety and environmental problems

Page 24 of 48
in the plant. If successful, the program should have had an effect on the plant’s
performance in these domains.

In the safety realm, data is readily available in the plant due to reporting requirements by
the United States Occupational Health and Safety Administration (OSHA). Job-related
injuries, whether by falling off a ladder, being burned by hydroflouric acid, or getting a
paper cut in an office, must be recorded and reported to OSHA. The refinery had records
dating back to 1993. As a benchmark for these data we used the petroleum refining
industry average (USDOL-BLS, 2007). Trends in the refinery and industry data are
plotted in Figure 6.

OSHA Recordable Injuries

—e— Refinery data
—?— Industry average

Injury rate

SNE Ys

so # # « ¢ sw ww se Se
ES vs 9
LPP LS PP SP PS FS LS LS LS LS LS

Year

Figure 6 - Safety performance: OSHA Recordable injuries per 100 full-time equivalent workers.
Petroleum refining industry average included as benchmark

Here we can see a general downward trend in the injury rate, with the most significant
improvements occurring between 1993 and 1998. These improvements moved
Redberg’s injury rate from high above to somewhat below the industry average. Since
then, the refinery has won more than thirty industry awards for safety performance, and
achieved milestones such as 1 million man-hours without a serious injury (i.e. one
resulting in lost work time). These performance gains were likely due to targeted
behavior change programs initiated by the HSE department, including significant rewards
for safety performance. One such program included a pickup truck raffled to one of the
workers, contingent on the plant’s meeting its safety goals.

Notably, however, these improvements occurred before the implementation of Proactive
Manufacturing. If we examine the period after 1998, when Eli Larson arrived and
initiated the program, we see a pattern of bumpy but flat performance in the OSHA
recordable rate. A linear regression on the post-1998 data reveals no significant change

Page 25 of 48
over time (t=.25; p=.814). This pattern implies that either Proactive Manufacturing had
no effect on this metric, or that its effect was eclipsed by simultaneous increases in
recordable injuries, such as might occur in the south side construction activities.

On the environmental side, we can examine the two most common types of
environmental pollution - oil spills and airborne emissions in excess of the refinery’s
permits, both of which are tracked and reported to the US Environmental Protection
Agency. Unfortunately the same industry-level data were not available from the EPA for
comparison, so we simply include the trends in reportable spills and emissions in Figure
7 below.

Air Emissions

Oil and liquid spills

Number of reportable incidents

Year

Figure 7 - Environmental performance: Reportable air emissions and liquid spills, for years with
available data

In the case of these environmental excursions, we see a similar improvement in plant
performance over time, at least for the period with available data (1995-2006). Much of
the improvement in environmental performance, however, occurred prior to 1998 and the
initiation of Proactive Manufacturing. In fact both graphs show increases in the number
of environmental excursions around that time, with improvements that follow.

We can examine whether Proactive Manufacturing coincided with improvements in these
metrics by examining data from 1999 onwards, when the program was in effect. For
liquid spills, there is no significant improvement during the 1999-2006 period (t=1.48,

Page 26 of 48
p=.19). For air emissions, there is small but significant improvement over time (t=2.87,
p=.028). These latter improvements may be due to the Proactive Manufacturing efforts,
but as with the safety data, any effects are confounded by simultaneous programs that
took environmental performance as their primary focus. In 2000, for example, an
environmental awards program was instituted that offers a cash bonus to all workers in
the plant if environmental targets are met. The sharp decline in liquid spills following the
initiation of that program may be indicative of its success, as well as longer term changes.
But such causation is impossible to infer from data available either to us as researchers or
to the plant personnel.

Financial Performance

Archival sources from the plant show that during the period under study, Redberg
Refinery enjoyed substantial gains in financial performance. The following graph of the
plant’s gross margin and EBITDA depicts this trend.

Figure 8 - Financial performance of Redberg refinery

2500

2000

1500
—¢— Gross Margin
—s— EBITDA

$ (Millions)
~
Ss
Ss

1995 1997 1999 2001 2003 2005
Year

From refinery earings in 1999 of $50 million, we see an increase by a factor of 31.5 to
$1,580 million in 2006. By using measures of gross margin and net margin per barrel of
production, we can compare these performance measures to those of the refining industry
as a whole. Doing so allows us to account for the effects of fluctuations in crude oil and
refinery product prices. Industrial comparisons for financial data are taken from the U.S.
government's Financial Reporting System Survey (USDOE-EIA, 2007). Here we can see
that the refinery began to exceed industry performance in 2001, and has continued a trend
of increasingly outperforming the industry.

Page 27 of 48
Figure 9 - Redberg financial performance compared to industry average data from USDOE.
Gross and Net Margin measured in US$/Bbl of refinery production

€
i 2

=
: He, 2 - @ -- Redberg Gross Margin/Barrel
i 10 - &- Industry Gross Margin/Barrel
q ? %. Ja —+#— Redberg Net Margin/Barrel
ge - — - - : Industry Net Margin/Barrel
a | "3g et fe

6 os :

=a as

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Year

Much of the performance gain visible in these graphs appears due to strategic business
decisions made by SmallCo. When SmallCo purchased the refinery, they did so in
partnership with a private equity group who invested upwards of $850 million in
alterations to the plant. The financiers did so because SmallCo had brokered a ten-year
deal with an foreign oil company, allowing Redberg to purchase heavy, sour crude oil for
an attractive price spread below the market rate for the more standard West Texas
Intermediate crude oil. The investments in the plant were necessary to process this new
supply - a decoker and hydrotreater to handle the heavy molecules and an extensive
sulfur processing facility to handle the “sour” sulfur content. The contractual price
guarantee was needed to ensure that the substantial capital improvement loans could be
repaid.

These expansions all occurred on the south side of the plant, divided by a road from most
of the older equipment. To assist monitoring and operation of this new equipment, a state
of the art centralized control room was built - two rooms full of computer screens,
complete with simulation capability for training new operators. Collectively, these
expansions were known as the Heavy Oil Upgrade Project or HOUP.

People in the plant are unanimous that the guaranteed margin associated with the heavy
sour crude deal was a boon to the plant’s financial success. The financial data for the
plant show a significant improvement in 2000 associated with this change. The later
refining market conditions made this investment even more profitable than had been
originally forecasted.

Page 28 of 48
Some informants, particularly in the operations services group, also emphasize the
importance of the plant's flexibility in offering a dynamic product mix to the market.
This flexibility of production if afforded by three elements of plant capability: the ability
to handle multiple kinds of crude oil, including heavy sour crudes; sophisticated
monitoring and control systems allowing fast changes in operational plans; and cross-
functional skill and collaboration necessary to handle changes in supply and product mix.

Given the significant increases in financial performance, members of the refinery and
consultants rightly regard Redberg to be a breakthrough success story. When we asked
why some of the consulting activities around Proactive Manufacturing had been
discontinued, one manager said, “what do we need consultants for, to help us count our
money?”

Interpretative F lexibility

The trouble, as mentioned above, is that the simultaneity of multiple change processes
produces interpretative flexibility about the cause of Redberg’s success, whether in the
financial, safety, or environmental realm. Faced with ambiguous internal data,
employees and managers face the same challenges we do as researchers. Is the plant
successful because of Proactive Manufacturing and a highly committed workforce,
because of the technical capabilities enabled by capital expansions, or because of
business environment changes and improved market prices? Teasing apart these three
effects would require a prodigious accounting analysis that the plant’s financial team has
not conducted. Differentiating the effects of Proactive Manufacturing from those of
focused safety and environmental programs on HSE performance would be even more
difficult. Making sense of the plant’ s history therefore falls to the individuals and
political factions involved.

For the designers of Proactive Manufacturing, the intent was to create deliberate synergy.

The investments that had been made over the last four years or so was just a very
bare maintenance but probably not even called a sustaining maintenance level.
So some of the facilities had regressed and degraded, and reliability, you know,
wasn’t good. A lot of room for improvement. And of course then when I got
there, it was very obvious to me anyway that we had a huge project that could be
very very economically viable and yet it was all dependent upon good safe and
highly reliable operation of the old plant. It was all fully integrated together -
you could not run one without the other. So then that’s when we decided to take
the same approach - pretty much the same - as was done at Heartland, in terms
of using the Manufacturing Game, bringing in all the people, putting them
through the two day workshops, making them become familiar with the theory
behind it. Forming action teams to go off and make improvements. And starting
a CI Forum... with pretty much the same representation as at Heartland.

Eli Larson, former plant manager

For some people within the plant, however, their own sensemaking has led them to a
more binary conclusion.

Page 29 of 48
I can tell you right now, Sam Taylor was probably the biggest factor in getting the
[heavy sour crude] deal and making Redberg what it is today... Between Taylor
and the deal and the plans for the decoker and then [SmallCo CEO] going on
through with it and everything else, by the time Eli Larson came in here, he was
just basically a figure head. Everything had been set up as far as moving
forward. He came in as a plant manager, nothing else. His deal was the
Proactive Manufacturing and the CI Forum - that was his big push for anything,
so... And really what put us in the place of where we are now is Taylor, the CEO,
and, you know, the money... they were fixin to spend a billion dollars on this
place and it wasn’t worth at the time... in fact they were still tied up in debt with
OldCo. So it was a good PR program, it was a good group that went out there
and did the financing. Like I said, Larson came in... his part in it was a whole lot
less than the rest of them, almost non-existent... He stayed for, you know, quite
awhile and ended up smelling like a rose. He is one of those guys who landed in
the right spot at the right time.

This quote comes from an hourly maintenance worker who is the head of a union in
Redberg, and presents a contrary view - that success derived from strategic changes and
capital expansions, and Proactive Manufacturing was irrelevant. This quote indicates
how political contests within the refinery - such as those between union, supervisors and
management - can shape the data people look for and the conclusions they draw. Given
the interpretive flexibility of the data available, there is no objective arbiter among these
perspectives, which have real consequences in the way events play out as will emerge
below. People held closely to their perspectives and accumulated supportive experiences,
thereby reducing any need for, or influence of, data that would confirm or refute their
personally held views, or of those views of the people with whom they identified
themselves.

Conflict and C ontroversy

In the course of our interviews, a number of people expressed negative views of the
program and the plant as a whole. Although their causal reasoning differs, there are two
things that most of these attributions and narratives hold as their focus. The first is the
degree to which a “proactive” approach to maintenance and equipment repair has actually
been implemented at the plant, as part of Proactive Manufacturing or otherwise. The
second is the failure of the program, particularly the Continuous Improvement Forum, to
gain significant participation from the hourly workforce. This non-participation, and the
change in CI Forum over time, illuminates dynamics worth considering in the
implementation of similar programs in other contexts.

Perceived limitations to Proactive Manufacturing

Most of those who described limits to the Proactive Manufacturing approach said that the
plant had failed to implement one of the key lessons of The Manufacturing Game®: that
you have to spend money on maintenance to get out of a reactive mode. Some attributed
this to the fact that SmallCo, the owner of the plant when the program was implemented,
was short on funds and was unable to invest in plant upkeep. Others felt that managers
simply hadn’t learned the lessons of the Game. Nearly every operator we spoke to, and

Page 30 of 48
several supervisors, could name a few pieces of equipment that had been awaiting repairs
for what they thought was an excessive period of time.

SmallRefCo was tight on cash and lots of stuff wasn’t implemented because they
didn’t have means. They fell by wayside. For example, there is one thing where
we could switch a turbine drive with a motor drive to get better steam balance on
one of the units. We could use less 15 lb steam. Typically we vent and lose
energy. They knew about that for years but never did it because of money. Even
now hasn’t been done.

-Maintenance planner

Across the street they always had two big huge wet gas compressors. You'd run
one one month and then you'd go over there, and the other one would be slow
rollin’ and you'd switch over and run it, and do any maintenance you had to do
on that other one. Sometime about 5-6 years ago, they took this equipment and
they just decided that, well, we'll repair it and we won't run it. So what it did is it
set over there and it had some problems with it... the refinery fuel with H2S in it,
it’s got corrosives in it, well, it eat up the shaft in there. So they had to pull the
shaft out and replace it, and worked on it. But I’ll bet you that dang compressor
hasn’t run in over three years. Runnin’ wet gas through it. And you would never
see that before, you'd always see them maintain the equipment and swap from one
to the other. Now basically you've got one compressor over there and if it goes
out, you're going to the flare. And you don’t have a backup to do it. So it’s
gonna be a loss of product, it’s gonna be a loss of, you know, I guess you could
say an environmental excursion. It’s gonna be all kinds of stuff because it wasn’t
deemed necessary to maintain the equipment... The guys over here would speak
up about it. They would say, “this is a situation where you don’t want to get
into.” And they'd hear, “ well, you know, we’re not gonna spend the money, or
you know it costs too much,” and I think that’s kind of like that BP [Texas City
refinery disaster] deal - you get to the point where these guys in those downtimes
look for so many different ways to cut cost, that you know it’s to the detriment of
the people, the jobs, and the actual bottom line of the corporation. It’s an
immediate impact. I guess from workin’ out here through OldCo, SmallCo,
SmallRefCo, NewCo... that it’s a short term fix for a lot of things. They don’t do
a long term... because I guess you could say maybe they won't be around. Why
worry about it if we’re not gonna be around?

-Maintenance Worker

Others acknowledged investments in reliability, but emphasized that these had occurred
during Sam Taylor’s management tenure in the early days of SmallCo’s regime. Given
the curve in reliability data that begins declining in the mid-1990's, this is a plausible
interpretation, one that downplays the role of the action team projects described above.

The critical thing we noticed, however, is that these views were much more likely to be
expressed by informants who had other concerns related to political conflicts within the
plant. Given ambiguity in the data, people interpreted the success or failure of the
program through the lens they wore. To understand the sources of these political

Page 31 of 48
conflicts, it is necessary to look at other stories told about the history of the plant that
emphasize separation between management, supervisory, and hourly worker levels.

Policies differentially affecting hourly workers, supervisors, and managers

While some degree of conflict among hourly workers, supervisors, and managers is
expected in any manufacturing facility, particular events in Redberg’s history helped
shape the norms and interpretive schemes that interact with those of Proactive
Manufacturing. One of the most salient processes that people discussed was the change
in management when SmallCo first purchased Redberg Refinery in 1995. The new plant
manager Sam Taylor and his management team, along with the investments in the
HOUP, established a relationship with the union leadership in the plant that had not
existed under OldCo. A Union-Management team was formed, which was designed to
address consistent grievances that union membership had and to engage hourly workers’
help in improving the plant.

As part of these activities, union leaders were given an opportunity to name supervisors
that they wanted fired because of conflict with hourly workers. One story about this
process, told in hushed tones, was that the HR manager and union leader would sit ina
room and invite one supervisor after another, at which time the HR manager would ask,
“Should we keep him?” Along with the firings, the plant management instituted an
“open door policy” allowing workers to come to managers with complaints about their
supervisors; managers responded by overturning some of the supervisors’ decisions. One
story to emerge along these lines was about an overtime lunch policy, in which a
supervisor tried to follow a plant policy about the provision of lunch for workers on
overtime.

I didn’t feel that SmallCo held honesty and integrity very high on their priority
list... They would try to promote an initiative behind closed doors but when we
tried to manage it in the plant, we got no support... Things as simple as an
overtime lunch - this is a small example but indicative of the philosophy. We met
with all the supervisors and they said, you will administer the overtime lunch
policy. If you work overtime, from the graveyard over to today, it’s an
unexpected overtime so you are allowed to order breakfast. In some cases you
can order breakfast and lunch. Supervisors were trying to administer that policy.
It wasn’t long before one of the hourly workers called the managers and said, “I
think I deserve a lunch.” The operator thought it was unfair. We administered
the policy as it was written. Instead of getting two overtime lunches he got one.
Or might have been the dollar amount. But the plant manager overrode and
completely redid, and I guess gave in to the guy. We felt like he cut that
supervisor off at the knees. Eliminated any authority or credibility that
supervision had in the plant. From that point forward it became common practice
- word gets around - that any time an hourly worker didn’t like something, they
went around the supervisor and the area supervisor directly to the plant manager.
In all the years I’ve been in the plant, the SmallCo senior management was the
weakest we've ever had. Weak in leadership. I think we developed some bad
habits. We developed some philosophies that we're having to change now and
that we started to change when Eli Larson came on board.

Page 32 of 48
-Member of plant management team

While these moves intended to show management's seriousness in engaging the hourly
workforce, the unintended consequence was to alienate and frighten supervisors. The
phrase “cut off at the knees” emerged in several interviews with people who had been
supervisors in this time period. In response, the supervisors banded together to organize
a facilitated session in which the “97 Issues of the Supervisors” were elicited. These 97
Issues served similar function to union grievances, most of which asked for clarification
about compensation and policies for supervisors and their workers.

It was in this context that Eli Larson arrived and brought Proactive Manufacturing. On
the one hand, the intention of the program was to unite people across levels and functions
in the organization, through mechanisms like action teams. On the other hand, the
program itself became an object of interpretation within the political landscape of the
plant. The same supervisor who had helped organize the 97 Issues and “almost got
myself fired” showed great passion and interest for Proactive Manufacturing and was
promoted to operations manager in part for this reason. Because the earlier moves had
made the distinction between supervisors and hourly workers particularly salient and
contentious, these later moves made Proactive Manufacturing vulnerable to appearing as
a management endeavor (and CI Forum as a “management meeting”) instead of
something more integrative.

In addition to these events, several other actions on the part of SmallCo’s senior
executives (above the plant management level) served to reward some people while
alienating others. All of these occurred after Proactive Manufacturing had begun, and
shaped the way management and management initiatives were perceived. These include:

e Personnel cuts: Due to shrinking margins and cost pressures prior to the
company’s IPO, a 25% reduction in salaried workforce occurred in 2002, known
internally as “Black Monday.” Hourly workers were protected by their unions but
supervisors and staff were not.

e Changes in gainsharing plans from a constant 8% up and down the organization to
a stratified plan where hourly workers stayed at 8% maximum while supervisors
could get 25%, superintendents 50%, and the plant manager 125%.

e The CEO retiring with a $93 Million pay out, at roughly the same time that this
stratified policy occurred.

e Exclusion of hourly workers and supervisors from participating in the company’s
IPO, after the expectation had been set that participation would be possible

e Changes in benefits programs that went around standard union negotiation
processes, an action perceived as using loopholes in union contracts to impose
managers’ agenda

Although the personnel cuts directly affected the Proactive Manufacturing activities by
removing advocates and organizers, the other events do not directly affect the continuous
improvement process. All of them, however, were cited by multiple personnel as
relevant to the credibility of worker involvement efforts, and as grounds for non-
participation on the part of union committee members and their hourly constituents.

Page 33 of 48
Much of the research on high performance work practices, high-involvement
employment practices, lean manufacturing, and similar “post-bureaucratic” organization
styles (Macduffie, 1995) emphasizes the importance of coherent logic with consistent
practice “bundles.” In Japanese or Japanese-influenced manufacturing plants, this logic
involves the kind of employee involvement in improvement efforts that we see at
Redberg Refinery, along with human resource management practices that ensure job
security: long term employment; training for new and continuing workers; job rotation;
and pay based on skills, training, and company performance. Vallas (2003), drawing on
Barley and Kunda (2001), emphasizes that these practices constitute normative control,
providing a moral grounding for the exercise of managerial power. These moves, while
not carried out by the immediate implementors of Proactive Manufacturing, nonetheless
affected the way it unfolded. The easiest way to see a decline in confidence, particularly
among hourly employees, is by looking at the Continuous Improvement Forum, or CI
Forum.

CI Forum - From Continuous Improvement to Centralized Information

Although the original activities required the participation of everyone in the plant, the
ongoing CI Forum is a voluntary activity, offering a free lunch for those attending. Its
purpose is to sustain an ongoing focus on continuous improvement activities, to share
successes and progress with plant personnel, and to handle issues and challenges arising
in the course of defect elimination efforts. While some hourly workers continue to attend
CI Forum, the number appears to hover around 1-2 out of the 800 in the plant. Of the 40-
50 who do attend regularly, the vast majority are supervisors and staff. These results
parallel those by Vallas (2003), who found similarly narrow participation in continuous
improvement activities at a paper mill. There are a number of causes that people cite,
which we outline below.

The simplest cause for worker non-participation, and one of the most frequently cited, is
that it can be difficult for hourly workers to leave their regular posts. The CI Forum
happens the third Wednesday of every month, at lunch time; for people on a day-job
rhythm, this can work. For hourly workers on 4*12 hour shifts each week, it means
staying significantly late or arriving significantly early at the plant. The choice of
meeting time is, of course, a political question - when asked whether the time might
make it difficult for hourly workers to attend, the organizer responded “maybe if we
played with the time we might get more hourly folks,” but there is little active
engagement with this question that we could see.

The more provocative explanations focused on the policies described above as creating
divisions between managers, supervisors, and hourly workers. The fact that CI Forums
are now attended by far more supervisors than hourly workers may be a symptom of this
division. One trainer, who had an earlier career as a well-respected supervisor, said it
succinctly: “It’s a management meeting.” One of the union leaders, whose opinions and
actions are no doubt influential for other workers, told the following story.

One of the things they always did, whether it was contractors or company
personnel, they always provided you the resources in the area you were doing the
shutdown. Because what happens is your manpower is multiplied in that area...
So the facilities here such as water and sewer and all that is not adequate to take

Page 34 of 48
care of it. So ever since I’ve been here... they moved, you know, J ohn-on-the-
spots in, wash stations in, and more water, all that kind of stuff. Well, after
SmallRefCo got started with the CI Forum, I was working on a shutdown over on
the next unit... and they didn’t have water. So, you know, one of the questions I
had was I talked to the unit supervisor and said, “We need some drinkin water
stations over there.” And they said, “They can come over here and drink water.”
And I said, “ Well that’s counterproductive, because they’ re not working over
here, they're working over there.” He said he wasn’t gonna do it. So I called HR
department and said hey, we got some problems here. If we're going to be this...
get into this... CI Forum, with all this change and all this innovation... we can’t
even do what we’ve maintained in the past! I need some Gatorade or something
over here. I had to go pick up a case of Gatorade to bring over there. | tell you
what it was the biggest rigamarole you ever did see just to get water over to the
people who were working. After that is when I dropped out of the CI Forum. I
said, if y’all can’t make any more sense about providing for your working people
than you are, then there is no sense in me going to the CI Forum because y’ all
aren't really concerned about what happens to the working people. So, that was
the last time I went to the CI Forum.

Ina similar story, one hourly operator, Carl, who had been heavily involved in CI Forum,
was concerned about an issue in his part of the plant that affected operators’ quality of
life in working there. Carl tried to bring up the issue repeatedly, which generated friction
with his supervisor. The supervisor asked him to “shut up about it’ and then Carl
abruptly stopped receiving emails related to the Forum. It is difficult to uncover exactly
what happened, but it appeared that the supervisor told the organizer that Carl wanted to
be removed from the list. Following this incident, Carl never returned to CI Forum or
participated in any of the activities.

Carl’s exclusion is an interesting case for three reasons. First, because he was highly
active in the Proactive Manufacturing efforts prior to this incident, he is a good case for
analyzing strong reasons for non-participation. Second, his complaints were perceived
irrelevant by supervisors, highlighting the fact that hourly workers and supervisors have
different issues they consider important. Third, the supervisor's stance towards Carl and
deliberate exclusion may have been a result of the early supervisor-hourly conflicts, a
way of re-asserting supervisory authority (although this is speculation). The net effect is
to both exclude an hourly worker and exclude a voice willing to raise negative issues.
The accumulation of such events would tend to make CI Forum both a management
meeting and meeting more focused on success than challenges.

Elaborating on this idea of CI Forum as a pat-on-the-back management meeting, one of
the hourly workers put it more colorfully:

I was asked to go maybe 5-6 years ago and | went and it was like a big giant suck
ass session. Some people just couldn't control theirselves because the plant
manager was there. I’m like... you can’t tell the man the truth when he’s sitting
there, then what's the point? ... You know what I mean when I say kissing up to
the boss. These are supervisors or other salaried type people and their whole
agenda was to get in with the plant manager and get higher up. Kind of destroyed
whatever they were trying to do for maintenance. [The maintenance director]

Page 35 of 48
asked me why I didn’t go to more CI Forums and I said I couldn’t take all the
sucking ass and he said I can understand that, it’s pretty bad. So even upper
management knows it.

Colorful language aside, there are several features of the CI Forum as practiced that may
bias it towards an exercise in ingratiation. First, attendance is recorded and monitored,
and at one point there was a KPI for supervisors about their own attendance and the
attendance of their hourly people. This was a controversial move - some advocates of
the forum felt it critical to keep it voluntary. Monitoring and rewarding CI Forum
attendance may select for those supervisors most interested in ingratiating themselves
with the plant management.

Second, while the intention of CI Forum may have originally been to discuss issues
arising in the course of doing continuous improvement, the emphasis has been on
describing success stories. While this function of “celebrating” achievements may
provide an incentive, it also sets the tone of not raising issues perceived as negative, and
even suppressing them.

Third, over the course of the CI Forum’s evolution, it seems to have become increasingly
focused on information provided by the plant’s management team. One of the original
purposes of the Continuous Improvement Forum was to make improvement visible to the
supervisors and workforce by presenting data about plant performance. This tradition
was set by plant manager Eli Larson, who was the first to make some of the financial
numbers transparent, along with measures of equipment failure rates and reliability. As
the CI Forum has evolved, the tradition has been to have each of the management team
members present the data they know best: the financial manager presents revenues, costs,
and margins; the HSE manager reports incident rates; the operations services manager
presents rate and reliability; and so on. And, in fact, the people who do attend CI Forum,
including hourly workers, describe this as the most valuable part of attending, because it
gives them an indication of what is going on in the rest of the plant.

The trouble is that, as a result, CI Forum has therefore become a predominately one-way
communication device. Most of the time is taken up by these presentations. As one
supervisor put it:

I'll be honest with you, I have mixed feelings [about CI Forum]. Sometimes it’s
like a show. It’s entertainment...

Sometimes I just don’t want to go. I don’t want to go there because I don’t think
I’m getting a lot out of it. There are some things, like we might go over refinery
metrics or refinery profitability, but we spend way the heck too much time talking
about it. We need to show the data and get on with the show. We drag it out...
and with all due respect to the directors [names some], a lot of them like to the
talk. They get up there and they ramble, and I’m thinking “get to the point.” So
something that in my book should take 5 minutes takes 30. But on the other hand,
it’s communication, and a lot of people out here will say they don’t know what's
going on because they don’t get communicated to. The good thing is, you're
going to get communicated to over there! And you'll have an opportunity to
communicate some of your thoughts.

Page 36 of 48
This aspect of CI Forum was brought into relief most powerfully in another conversation
that the first author had with two operators - Frank, who is a relatively new hire, and
George, who has been at the plant for thirty years. We were sitting in a control room,
discussing the history of the plant. I brought up CI Forum. George was vaguely aware of
its existence, and asked what “C1” stands for. Frank, who has eagerly gone to CI Forum
several times because he is interested in hearing how the plant is doing, said “I’m pretty
sure it stands for ‘Centralized Information Forum.’” I was about to correct him and let
him know that it stands for “Continuous Improvement Forum” until I realized that his
interpretation of CI Forum was reasonably in line with how the forum had evolved
within the culture and context of the plant.

When we examine the graph below of attendance records at CI Forum, we can see a
fairly steady pattem of attendance, with a gradual increase in 2004 that peaks in 2005.
This period coincides with the IPO of the parent company of the refinery, and the record
peak coincides with the decision to sell the refinery and change plant managers. This
behavior is consistent with the idea that CI Forum is a venue for “centralized
information” about goings on in the company.

Cl Forum Attendance

160

140 f
120 4 |
100

vs FT SF F YF F£ Y EF SF /M
KC ae Ca & ae we FP
Voge MO eX ew

Figure 10 - CI Forum attendance, 12/2002 through 10/2006
Note: CI Forum not held in September-O ctober, 2005

CI Forum is not the only continuous improvement activity, and in fact is designed to
support the cross-functional teamwork happening out in the plant every day. Because it
is designed to be a cross-level and cross-functional governance body, however, it serves
as a microcosm for the plant as a whole. While Proactive Manufacturing was largely
successful in fostering continuous improvement, the evolution of Continuous

Page 37 of 48
Improvement into Centralized Information shows that some of its momentum has
decayed over time. In the discussion below we will more explicitly theorize about this
process.

Discussion

The purpose of this study was to examine the implementation of Proactive Manufacturing
at the Redberg oil refinery, and evaluate its effects on reliability and overall plant
performance. We hope to use the results to inform three things: the model underlying
The Manufacturing Game®, which was developed using system dynamics; its method of
implementation as part of a program of change; and attempts at workforce involvement in
continuous improvement more generally. Rather than using system dynamics as a
method, we have instead stepped back and treat the whole systems thinking intervention
described above as a phenomenon, and examined it from a sociological perspective using
the tools of grounded theory.

Our findings are that this program of change has had an important positive influence on
the refinery, including measurable improvements in reliability and the development of a
culture of proactive maintenance and defect elimination. Proactive Manufacturing has
helped contribute to the refinery’ s industry-leading performance gains, particularly in
tandem with a competitively priced heavy sour crude oil supply and significant
investments in the plant’s physical infrastructure. Nevertheless, this simultaneous
investment in organizational and technical improvements creates ambiguity about the
causes for the plant's success. This ambiguity demands an ongoing sensemaking process
among the plant’s personnel, from which multiple interpretations have emerged. Givena
history of some conflict among managers, supervisors, and hourly unionized workers,
these interpretations fall out along political lines - many hourly workers question the
value and sincerity of management's attempts at workforce engagement. These divergent
perspectives have led to a pattern of non-participation by hourly workers in the
Continuous Improvement Forum, which has contributed to its evolution into something
more like a “Centralized Information Forum.” While this does not completely undermine
continuous improvement efforts - they are alive and well in parts of the plant,
particularly in the newer south side complex - it does indicate a challenge faced by
continuous improvement as an organizational strategy.

Dynamics of implementation at Redberg

To illustrate the dynamics we observed, we can retum to the original system dynamics
model underlying The Manufacturing Game® and the Proactive Manufacturing Program,
depicted above in Figure 1. In the model, one of the highest leverage points for plant
performance is workforce participation in defect elimination efforts. Proactive work on
defect elimination avoids the expensive and chaotic trap of reactive maintenance and
unplanned repairs. By reducing such breakdowns, the plant can free up resources for
further investment and stimulate a reinforcing dynamic of Capability Growth. At
Redberg, implementation of this strategy occurred in parallel with capital investments
and new supplier relationships designed to create a “world class facility.”

At the risk of over-simplifying, we can therefore compress Figure 1, and depict the
intended rationality of the managers of Redberg Refinery in the following causal loop

Page 38 of 48
(Figure 11). Investment in workforce engagement was intended to bring about Capability
Growth through continuous improvement at the front lines of the organization.
Simultaneous investments in Plant Expansion would further augment performance and
perhaps realize a synergistic effect. For the most part, this strategy seems to have been
successful - the refinery has enjoyed significant improvements in equipment reliability
and plant performance, making it the flagship refinery for both SmallCo and its acquirer,
NewCo.

Actual

Performance An) * available

A pike Resources
Expansion
+
‘Synergistic Investment in

Effect_+ Physical Plant
4 =

®)

Capability

Growh
Workforce 9
Participating in Investmentin
Continuous Workforce
Improvement, Engagement

Figure 11 - Intended Rationality of Proactive Manufacturing

It is critical to recognize, however, that the policies depicted here were carried out in two
distinct management epochs at Redberg in very different way. Effects accumulated from
the first epoch - under SmallCo’s first plant manager, Sam Taylor - set the context for
those of Eli Larson and the Proactive Manufacturing Program. Some of these effects
drove the Capability Growth reinforcing loop, as indicated by improvements in plant
reliability beginning in 1996. Some effects instead created limits to growth - balancing
dynamics that would later compromise workforce participation in continuous
improvement activities.

For example, as discussed above, Taylor's approach to workforce engagement was
through an open-door policy with hourly workers. This had the unfortunate consequence
of undermining supervisor authority and spurring worker-supervisor conflict. We believe
this shaped the way workers perceived the next wave of engagement efforts by
supervisors as part of Proactive Manufacturing, a dynamic depicted as the Conflict
balancing loop in Figure 12.

Page 39 of 48
Actual =
Performance 1) * available
& Plant Resources
Expansion
"
Synergistic
Effect + Investmentin

Yo fee
ais
t)

Capability
Growth
Workforce 4
Participating in Investmentin
Continuous Workforce
Improvement. Engagement

a Mss

Workers Assessment
of Management 7
p mgroms B Supervisor
Conflict Authority

Bias Towards|
Negative Worker-S upervisor

Interpretation a Conflict

Figure 12 - Side effects of first-wave workforce engagement efforts

As time progressed, and the plant began to reap the benefits of these investments in
capability, events transpired that further undermined workers’ trust in management-led
initiatives. These included changes in compensation practices that favored supervisors
and management, as well as actions perceived as unilateral on the part of management,
such as changes in benefits programs. Some of these decisions emerged from corporate
headquarters, and the plant management team saw them as beyond their control.
Nevertheless, in combination with an espoused philosophy of valuing and engaging the
workforce, these decisions made management vulnerable to attributions of hypocrisy.
This dynamic is depicted as the Distrust balancing loop in Figure 13.

Page 40 of 48
+
Avwilable
A Resources

Plant ——
+

Actual ey hg,
Performance 1)

Synergistic Investment in
Effect + Physical Plant
eee oe
ts)
Capability
Growth
a
Workforce
Participating in Investment in
Continuous Workforce
Improvement Engagement
f* WS eee

Workers Assessment of

Management P rograms
B

\ Conflict /-

Bias Towards

Negative Worker-S upenisor
+

Interpretation | — Conflict
:
Ae)

Distrust

Attribution of +
Hypocrisy

Ratio of Management to Unilateral Actions by
Worker Rewards Management

Figure 13 - Effects of policies differentially affecting supervisors and workers

The final dynamic that seems to have operated at Redberg is the interpretative flexibility
discussed throughout this paper. Although the Conflict and Distrust dynamics may have
fostered a feeling of alienation and a Bias Towards Negative Interpretation on the part of
workers, such a bias should not necessarily lead to negative assessments of management
programs. In the presence of clear and unambiguous data about the effectiveness of such
programs in sustaining plant success, such biases might be weakened. In the Redberg
context, however, the simultaneity of investments in the physical plant and workforce
engagement created exactly the opposite condition. Causal ambiguity about Redberg’s
success increased the importance of workers’ negative bias as they assessed management
attempts at engagement, particularly the CI Forum. This dynamic is depicted as the
Interpretative Flexibility balancing loop in Figure 14, which serves to amplify the effects
of the other balancing loops in limiting workforce engagement.

Page 41 of 48
Actual ee eg
Performance 1)

Available
i, Resources
Plant Expansion
Performance y
Measurement Synergistic Investment in

Effect + Physical Plant

Causal
Ambiguity Capability
Growth
i
Workforce
Participating in Investment in
Continuous Workforce
Interpetative Improvement =. Engagement
+ Flexibility Wee ee
Effect of Bias on +
Sensemaking

Workers Assessment of
Management Programs

: Supenisor
\ Conflict f-

Bias Towards

Negative Worker-Supenisor
+

Interpretation a Conflict
:
1)

Distrust

Attribution of +
Hypocrisy

Ratio of Management to Unilateral Actions by
Worker Rewards Management

Figure 14 - Interpretative flexibility as side-effect of synergy strategy

Although these diagrams emphasize the negative side of the story, it is important to
recognize that the dynamic here is not one of declining performance - Redberg has been
a high performing organization during the period of study, and Proactive Manufacturing
was, on the whole, successful. We simply posit that the growth in that performance may
have been partially limited by an underlying dynamic of worker-management conflict
and distrust. That dynamic has a stronger effect in a context of a synergistic business
strategy that produces causal ambiguity and interpretative flexibility. These dynamics
have eroded workforce participation in continuous improvement, and have helped
transform the “CI” Forum from Continuous Improvement to Centralized Information.

Page 42 of 48
Implications for The Manufacturing Game and defect elimination

While these findings speak to the particular implementation of The Manufacturing
Game® and the larger Proactive Manufacturing program at Redberg, how might they
inform the underlying model and approach?

In its simplest form, the original system dynamics model at Du Pont was based on a co-
flow with two stocks - the equipment in the plant, and the defects in the equipment (See
Figure 1). One critical insight from this framing is that defects start accumulating from
the very point of manufacture at the supplier level, so defect elimination must be a
proactive and continuous process. It is, however, very high leverage, because it frees up
resources and yields increasing returns. The Manufacturing Game® attempts to create an
embodied experience of these dynamics, with defects accumulating as chips that
overflow and produce a shutdown and/or safety/environmental incident. The cultural
credo created around TMG at Redherg reinforced this point - "Don't Just Fix It, Improve
It," the imagery of "Boris the Bug" and "Fix it right, Don't let the bug bite."

In its enactment in action teams, the notion of a "defect" was extended to include not just
physical defects (e.g. bad welding) and defects in the interconnections between physical
items (e.g. mismatch between metallurgy and process), but in more abstract processes. If
we look at the "Action team status" spreadsheet used at Redberg, we can see items like
"refinery KPI's," "Utilities maintenance planning," even "Improving communications."
The official definition of a defect among facilitators of The Manufacturing Game® is
“any deviation from perfection.” Although some defect elimination activities are more
abstract than others, they all provide a concrete point of focus for a group of people
(ideally cross-functional and cross-sectional) to work on.

The trouble, as our study illustrates, is that people come into these cross-boundary
contexts (the Game, the action teams, the CI Forum) with an already-accumulated stock
of experiences, narratives, and perceptual lenses through which they view the others in
that context and the very context itself. Like equipment coming in the door from
suppliers with defects in tow, the relationships - out of which collaborative efforts are
composed - come with their own historical baggage. Depending on the nature of this
history, it might help or impede collaboration, or affect it in complex ways.

In the causal loop diagrams above, the unintended consequences of managerial action
resulted in the accumulation of a stock we labeled as “Bias Towards Negative
Interpretation.” While in this case, the stock referred to workers’ negative interpretation
of management programs, it is an example of the more general phenomenon of relational
conflict shaping perception and action. We can stylize this idea through a notion of
“relational defects” in order to create a rhetorical parallel to the idea of defects in
equipment. We might depict it as an additional co-flow like the one below.

Page 43 of 48
>) Equipment

Installation Decomissioning
{ Equipment
Defects
Defects from eed Defects removed
Design and T through
Suppliers decomissioning

+ Breakdowns

Defects Defect Elimination
| __———» through Use Through Maintenance
Cross-F unctional ~ and Poor iS
improvement Teams Maintenance + Capability
; )
Workforce Participation
Proactive in Continuous Capability
Manufacturing ——— Improvement _ + Growth 4
P
eM Available Performance
Resources ~¢—
\
|
Relational Defects,
Perceived Perceived injury
Successful
Collaboration 4
ot 500)
Healing through Pre-existing
Time Attitudes
le se
Attrition Workforce Hiring

Figure 15 - Speculative Model - Defects in Equipment and Relationships

This framing allows us to think about a dynamic view of relational conflict that
undermines collaboration and continuous improvement activities. In so doing, we
illuminate a critical fact - that the time scales upon which these processes are working
are quite different. Defect elimination in equipment occurs on the timescale of weeks to
months. A Proactive Manufacturing program takes on the order of 3-5 years to settle into
a stable rhythm and interpretation within the organization. But relational conflicts and the
generation of stable interpretive frames and narratives happens on the timescale of
decades. We can see this quite clearly in the data from Redberg Refinery, because the
rise and fall of entire programs is taken as data in the formation of narratives and attitudes
about trust.

These layers are not, of course, separate, and there important feedback loops. As we
indicate through stories about action teams and cross-functional collaboration on the
south side of the plant, collaborative efforts can actually serve to improve morale and
foster further collaboration. This dynamic is depicted as the “Healing Through
Collaboration” loop below. High performance of the organization can also be used as

Page 44 of 48
evidence for successful collaboration, as it seems to be for many of the supervisors
involved in Proactive Manufacturing. This is the “Success Stories” reinforcing loop.
The underlying idea here is that successful collaboration is an important driver of trust,

an idea supported by research in the social exchange theory tradition (Lawler & Y oon,
1996; Molm, Takahashi, & Peterson, 2000).

Equipment
Installation Decomissioning
Equipment
Defects
Defects from Defects removed
Design and through
Suppliers decomissioning
+ Breakdowns
=
{ Defects DefectEtininaton~
| ___——» through Use Through Maintenance
Cross-Functional > and Poor
improvement Teams Maintenance me Capability
B.
)
Workforce Participation
Proactive in Continuous Capability
Manufacturing ——— Improvement _+ Growth 4
Program
9 ie aoe

Performance
(R \
Gs

Resources
Healing Through \
Collaboration \
Narratives of \
Colaboration \

| na | |

Relational Detects |
| Perceived Perceived inury |
\ Successful
\ Colaboraton I ry |

o <> |
Healing through Pre-existing
ge Time Attitudes
a ~
‘tion Workforce Fram

4s)

Success Stories

Figure 16 - Collaborative, cross-functional activity can gradually heal relational defects
At the same time, Narratives of Fragmentation also exist - stories and attributions that
depict the organization as a site of conflict and performance as benefiting some people
over others. Such narratives originate or are fueled by certain kinds of events, for
example decisions that produce or highlight inequality among groups. These narratives
create perceptions of injury or injustice among groups, which we see as an inflow to
relational defects.

Page 45 of 48
Equipment

Installation Decomissioning

|
| Equipment a
ar bees -
debce fom Delis removed
Designand frown
© Supper decomislning
= breakdowns
es
Defies detect etnnatan
__ a tieathlse thee Hainmrarce
Cross Ftnctonal and oor :
improvement Teams waitenance 7% Capbtiy
ji

Workforce Participation
in Continuous
improvement

Proactive
Manufacturing —
Program

== z

—— Available

apability
Growth

Performance
+ ee
Resources —@—___—

B
Fragmentation

Healing Through
Collaboration

Uniateral
| feton \
Narratives of \
Narratives of
\
epbiboratcn \ Fragmentation CD Tnequal |
\ Rewards i
| Terttory and |
42 i ‘Authority,
Contes
Relational Defects
Perceived Perceived inuy |
Successful
Collaboration J ry
owt a)
Healing trough Pre-existing
reine ities
X
Oo ~t
iifon Workforce ts

Success Stories

Figure 17 - Performance can also perpetuate relational conflicts when rewards are shared unequally

The implication here is that any program or social technology that seeks to reorganize
work arrives in an already generated relational landscape. That landscape will influence
both how the new actors and activities are seen and how they play out. Defects in
relationships must at least be approached in parallel if not before any other defects in
equipment or processes. While collaboration can heal relationships, it can also perpetuate
fragmentation when rewards are shared unequally. Furthermore, the expectation must be

that such processes will take the same order of magnitude of time as it took to accumulate
the relational stocks.

Again, while the worker-supervisor relationships were salient in the Redberg context,
relationships as framed in this conceptual model exist across multiple boundaries in the
organization. In particular, there are functional lines (operations, maintenance,
engineering, HSE, finance), employment tiers (corporate, plant management, supervisors,
hourly workers, contractors), and interactions between the two (e.g. between IBEW vs.
USW unions). "Defects" can accumulate in any "tie" between people in these groups.

Page 46 of 48
Clinical choices

From a practical or clinical perspective, the question thus becomes one of "how to
eliminate defects in relationships between human actors so as to produce integrative
action?"

We don't have any answers or even hypotheses here. What we do have is an analogy
(John Carroll, personal communication) between the inter-group conflicts at Redberg
Refinery and individual therapy. In the individual therapy realm there exists a dichotomy
between Freudian psychoanalytic approaches and cognitive-behavioral therapies.
Psychoanalysis focuses on history of relationships back to childhood, and say that you
can't move forward without addressing and re-narrating them. Cognitive Behavioral
approaches say that it might even be counterproductive to unearth the past and mentally
reinforce trauma, so the important thing is to focus on moving from the present forward
through new behaviors.

Although this study does not resolve or necessarily even inform this question, it does
pose it quite nicely in the organizational realm.

References

Barley, S. R. & Kunda, G. 2001. Bringing Work Back In. Organization Science, 12(1):
76-95.

Carroll, J., Sterman, J. S., & Marcus, A. 1998. Playing the maintenance game: How
mental models drive organizational decisions. In J. Halpem & R. Stern (Eds.), Debating
rationality: Nonrational elements of organizational decision making: 99-121. Ithaca, NY:
Cornell University Press.

Czamiawska-J oerges, B. 1997. Narrating the organization : dramas of institutional
identity. Chicago: University of Chicago Press.

Eisenhardt, K. M. 1989. Building Theories from Case Study Research. The Academy of
Management Review, 14(4): 532-550.

Ewick, P. & Silbey, S. 1995. Subversive Stories and Hegemonic Tales: Toward a
Sociology of Narrative. Law and Society Review, 29(2): 197-226.

Glaser, B. G. & Strauss, A. L. 1967. The discovery of grounded theory; strategies for
qualitative research. Chicago,: Aldine Pub. Co.

Griffith, J. D., Kuenzil, D.J., & Monus, P. 1999. A New American TPM: Leadership
requirements for breakthrough change, National Petrochemical and Refiners Association.
Houshower, H.; A voyage beyond the horizon and back: The Heartland Refinery’s
continuous improvement story;
http://homepage.mac.com/paulmonus/.Public/LH%20Sail% 20updated% 207-1-99.PDF
Jay, J. In Press. Social technology: A lens for the theory and practice of organizational
change. Cambridge, MA: Massachusetts Institute of Technology.

Kim, D. 1989. Learning Laboratories: Designing a Reflective Learning Environment,
Computer-Based Management of Complex Systems: Collected Papers from the 1989
International System Dynamics Conference, Berlin et al: 327-334.

Page 47 of 48
Kim, D. H. 1995. Managerial Practice Fields: Infrastructures of a Learning Organization.
In S. Chawla & J. Renesch (Eds.), Learning Organizations: Developing cultures for
tomorrow's workplace: 350-363. Portland, OR: Productivity Press.

Lawler, E. J. & Y oon, J. 1996. Commitment in Exchange Relations: Test of a Theory of
Relational Cohesion. American Sociological Review, 61: 89-108.

Ledet, W. J. & Ledet, W. P. 2002. Dynamic Benchmarking: Experiencing the Best
Practices of Others in Y our Plant.

Ledet, W.J., Monus, P., Cardella, T., & Burgess, W. 2005. Modeling Sustainable
Organizational Change: Why Did Change at BP Lima Sustain While the Change at
DuPont Faded A way?

. Paper presented at the 23rd International Conference of the System Dynamics Society
Cambridge, MA, USA.

Ledet, W. J.; The Manufacturing Game website; http://www.manufacturinggame.com/.
Macduffie, J. P. 1995. Human Resource Bundles and Manufacturing Performance:
Organizational Logic and Flexible Production Systems in the World Auto Industry.
Industrial and Labor Relations Review, 48(2): 197-221.

Molm, L. D., Takahashi, N., & Peterson, G. 2000. Risk and Trust in Social Exchange: An
Experimental Test of a Classical Proposition. 105(5): 1396-1427.

Nakajima, S. 1988. Introduction to TPM : total productive maintenance. Cambridge,
Mass.: Productivity Press.

Repenning, N. P. & Sterman, J. D. 2002. Capability Traps and Self-Confirming
Attribution Errors in the Dynamics of Process Improvement. Administrative Science
Quarterly, 47(2): 265-295.

Roth, G. L. & Senge, P. M. 1996. From theory to practice: Research territory, processes
and structure at an organizational learning centre. Journal of Organizational Change
Management, 9(1): 92-92.

Schein, E. H. 2004. Organizational culture and leadership, 3rd edition. San Francisco,
CA: Jossey-Bass.

Sterman, J. 2000. Business dynamics : systems thinking and modeling for a complex
world. Boston: Irwin/McGraw-Hill.

Sterman, J. D. 1989. Modeling Managerial Behavior: Misperceptions of Feedback in a
Dynamic Decision Making Experiment. Management Science, 35(3): 321-339.
USDOE-EIA; Financial Reporting System Survey;
http://www.eia.doe.gov/emeu/finance/.

USDOL-BLS. 2007. Industry Injury and Illness Data, 1994-2005: US Department of
Labor, Bureau of Labor Statistics

Vallas, S. P. 2003. Why Teamwork Fails: Obstacles to Workplace Change in Four
Manufacturing Plants. American Sociological Review, 68(2): 223-250.

Weick, K. E. 1995. Sensemaking in organizations. Thousand Oaks: Sage Publications.
Womack, J. P., Jones, D. T., & Roos, D. 1991. The machine that changed the world: How
Japan's secret weapon in the global auto wars will revolutionize Western industry (1st
HarperPerennial ed.). New Y ork, NY : HarperPerennial.

Page 48 of 48

Metadata

Resource Type:: Document
Description:: This mixed-method retrospective case study examines the implementation of a systems thinking intervention The Manufacturing Game as part of a comprehensive approach to Proactive Manufacturing and workforce involvement in continuous improvement at an American oil refinery. Notable gains in local reliability metrics such as Mean Time Between Failure were achieved as a result of these efforts, which contributed to financial, safety, and environmental performance. At the same time, the overall impact of the program on the refinerys performance is ambiguous, confounded by simultaneous investments in the physical capital and capacity of the plant that generated significant value. This ambiguity created interpretive flexibility for employees of the plant, who were able to both interpret and shape the continuous improvement activities based on prior political divisions, mental models, and interpretive biases. A structurational perspective is employed to understand these emergent and unintended consequences. From this perspective, a speculative model of relational conflict that undermines workforce participation in continuous improvement is developed.
Rights:: CC BY-NC-SA 4.0
Date Uploaded:: December 31, 2019

Using these materials

Access:: The archives are open to the public and anyone is welcome to visit and view the collections.

Collection restrictions:: Access to this collection is unrestricted unless otherwide denoted.
Collection terms of access:: https://creativecommons.org/licenses/by/4.0/

About the Archives

Access options

Ask an Archivist

Ask a question or schedule an individualized meeting to discuss archival materials and potential research needs.

Ask a Question Book a Meet

Schedule a Visit

Archival materials can be viewed in-person in our reading room. We recommend making an appointment to ensure materials are available when you arrive.

Schedule a Visit Hours

Collection

System Dynamic Society Records, 1978-2016

Collection ID	ua435
Total components	7460
Items available online	6307
Last indexed	2026-04-07

Schedule a Visit

Collections

Jay, Jason with George Roth, "From Continuous Improvement to Centralized Information: The Life and Times of a Systems Thinking Intervention", 2007 July 29-2007 August 2

Online content

Metadata

Using these materials

About the Archives

Access options

Ask an Archivist

Schedule a Visit