“Using Systems Thinking to Shift Mindsets - A Fortune 500
Company’s Multi-year Journey”
Kathleen Clark, Assistant Director, Continuous Learning & Improvement
Northwestern Mutual
Michael Goodman, Consultant, Applied Systems Thinking
Chris Soderquist, Consultant, Pontifex Consulting
Context
Organizations can be defined as any group of individuals and/or groups that have
“organized” for a common purpose. That purpose could be to sell products,
provide services, or even achieve a common good. This definition can apply to
any organization — from Fortune 500s to community-based collective impact
groups. Some are paid to participate; some volunteer. In all cases, the individuals
and groups that participate do so because the collective effort is designed to
achieve an objective.
This is the simplest way to think about what the “collective” is up to. One of our
colleagues suggests it’s important for the organization to answer this question:
“What are we up to?” That’s another way of saying “Why do we exist?” But
sooner or later — sooner if well organized and aware — the organization will notice
that it is up against something(s). |t could be up against an unknown
environment, entering uncharted waters. It will most certainly be up against
forces, problems, issues, challenges, etc... that will make it difficult to achieve its
objectives.
The purpose of strategy is to achieve what the organization is up to in the face of
what it’s up against. Most individuals will relate to what usually happens next.
They hear the dreaded words: strategic planning. (And often cringe.) The
planning often follows a predictable pattern of interviews, off sites, goal setting,
responsibility relegating, and metrics monitoring. The word execution becomes
bandied about.
Most strategies aren’t executed — remaining impotently printed and organized in
binders relegated to the shelf — and those that are executed, rarely achieve
intended results. Why?
Roger Martin claims that most strategy creation processes aren't intended to dive
into the tough questions.' They tend to rely on the same thinking that got the
organization to where it was. Improvement is looked at as taking what we have
and making a few changes here and there. It becomes a process of just barely
changing a few deck chairs on the Titanic...when it might be better to reconsider
the safety of the ship, or perhaps even questioning the wisdom of the voyage in
the first place! Leaders prefer slightly tweaking what they’re doing as opposed to
throwing out faulty assumptions and starting anew. They like comfortable, don’t
rock the boat processes — they like familiarity.
Martin claims organizations need a better theory of the business. In other words
they need to have a shared picture of what they are up to and what they are up
against. But the theory they are currently using tends to be outdated, sometimes
useless, and often counterproductive. And the familiar, comfortable — don’t rock
the boat — planning processes they use are designed to avoid confronting what
we Cling to that may be contributing to the problem.
Why?
Because we tend to assume that all issues — what we’re up against — are routine
problems. Ron Heifetz refers to these routine problems (he calls them technical
problems) as something with a clearly defined answer or protocol.? They are
routine in that there’s a routine to address them, not that they happen routinely.
There’s an expert that can come to our aid...the problem is often just finding the
expert.
In reality, Heifetz asserts the biggest issues organizations are up against are
usually more adaptive in nature. An adaptive challenge isn’t well defined. ..there’s
little agreement on what the challenge is. There’s no expert. No off the shelf
solution.
In the case of a routine problem, our colleague Craig Weber asserts there’s little
reason to sit there, to think about it. The exhortation should simply be:
Implement. Go do it. But in the case of an adaptive challenge, the exhortation
must be: Learn. What’s needed is an orchestrated process of learning. And to
address adaptive challenges, that process must meet a few important needs.
' Martin, Roger. The Big Lie of Strategic Planning, Harvard Business Review, January 2014,
blog at https://hbr.org/2014/01/the-big-lie-of-strategic-planning
2 Heifetz, R. Leadership Without Easy Answers. Harvard University Press; | edition (July 22,
1998)
ak
. Those participating must develop a shared understanding of the challenge.
What’s the issue? How can we pool multiple perspectives to better define
and describe the challenge? To do this requires an openness to seeing
different perspectives, a willingness to get up on the balcony and expand
the boundaries of discourse beyond the typical bounds.
2. The group needs an ability to develop the theory of the issue. What’s the
cause? How does it work? And this theory needs to be operational — a “how
it works” theory.
3. The group needs an ability to build confidence in the theory...and to refine
and revise along the way.
The system dynamics methodology is ideally suited for adaptive challenges,
precisely because it can contribute to each of the above needs. Here are three
important components of the methodology. Note: For the purposes of this paper
when we refer to “system dynamics” was are also including “systems thinking”
tools and methods such as causal loop diagrams, system archetypes and mental
models.
1. The system dynamics mind set helps participants with different perspectives
to move onto the balcony, to expand boundaries, to look over longer periods
of time and across a broader horizon.
2. The system dynamics language of stocks and flows is ideal for developing
the operational theory. Its forte is building a “how it works” map of the issue.
3. The system dynamics simulation (sometimes mental, and often computer)
capacity helps test the theory and identify places where it can be improved.
As a package, the above three components can turbocharge a process of
adaptive learning. This paper will describe how a large company consciously
developed the capacity to think systemically. This was the necessary foundation
needed for the company to eventually use the system dynamics methodology to
address difficult, seemingly intractable adaptive challenges. We will also describe
how other organizations (from private sector companies to collective impact
groups) can build and apply this learning capacity based on learning generated
during this sustained multi-year endeavor.
The System Dynamics Journey
The company’s journey to building this system dynamics-based learning capacity
began in 2008 and continues today. The journey began with an introduction to
systems thinking principles and causal diagramming. This eventually led to
prototyping and then applying system dynamics modeling. Here is a timeline:
Years 0-2 Systems Thinking capacity development (causal diagramming,
dialogue, mental models)
Years 2-5 Maturing Systems Thinking capacity (archetypes, more
sophisticated causal diagraming)
Year 5-current System dynamics projects
This progression was consciously chosen to improve the acceptance and
application of the full system dynamics methodology. Building (first) the capacity
to think systemically is key to leveraging more sophisticated tools such as
simulation modeling (later). It demands building skills in the practice of dialogue
versus the more common types of communication — discussion and advocacy.
Principle to this is the skill of framing better questions in order to more deeply
understand complex, chronic problems and ultimately discovering uncommon,
enduring solutions. Typical problem solving methods derive from questions of
“What” (is the root cause, implying a single linear connection from cause to
effect) or “How” (do we modify the process or execute on low hanging fruit
activities that provide immediate perceived gains). System dynamics requires a
commitment to “admiring the problem” which springs from questions that begin
with “Why” or “How Come”. Framing the question becomes an important genesis
for the learning journey. It sets the stage for building the following skills
necessary to practicing dialogue: suspending judgment, listening to understand,
reflection, and assumption identification (in oneself and in seeking to understand
the assumptions that others hold).
This sets a foundation for the level of inquiry and commitment necessary to build
and use causal loop diagrams effectively to “see” the system — enabling highly
successful executives, middle managers and subject matter experts (SMEs) to
stand outside of the system that they work in. This is critical to deep learning —
understanding cause and effect through the lens of multiple feedback loops,
building an appreciation for compounding effects and behavior over time.
Especially the appreciation of delays in the system (which often mask unintended
outcomes or positive results that take longer than expected to emerge.)
In our experience, deeper conversation leads to deeper learning. All of this
enabled the company to move beyond a rich systems thinking practice steeped
in the use of causal loop diagrams and archetypes to the rapid learning and
decision-making that occurs through the use of system dynamics.
System dynamics can be a key ingredient to effective strategy formulation and
execution. It does this by providing the antithesis to Roger Martin’s assertion that
most strategy processes ask easy questions and search for easy answers.
System dynamics is a rigorous, intensive, assumption questioning process that
rarely asks easy questions or finds common solutions.
Lessons Learned Along the Way
Along the way, through a process of continual learning, the company has
discovered several important lessons about how (and how not) to build system
dynamics capacity.
1. Make use of the full range of system dynamics and systems thinking tools
2. The problem (issue) chosen matters
3. Surfacing and challenging mental models (human “software”) is very high
leverage
4. Less is more generates better learning
We will now use a system dynamics modeling project as a case study to illustrate
these lessons. The project concerns an issue that relates to nearly all
organizations (from private to public): workforce development. The project was
designed to address several areas of workforce development that were not only
applicable to a specific area within the company, but were relevant to all areas.
And all organizations should see the relevance of the modeling work to an
important adaptive challenge faced by all.
Case Study — Claims Administration
Last year, a team of executives, managers, and SMEs assembled to deeply
explore the following question:
Why — despite our best efforts — are we challenged to attract, develop
and retain a workforce capable of meeting varied and dynamic
business needs, now and in the future?
The intent of this effort was twofold:
+ Understand the underlying systemic forces that were impeding the company’s
ability to attract, develop and retain benefits staff and
+ Identify high leverage approaches to address the issue.
The team identified the most important dynamics the model should generate.
Staff learn on the job as well as through formal training. The staff's skills
determine their productivity. However, work overload can build stress over time.
Stress generates several negative consequences including:
+ Reduced workforce learning
+ Reduced productivity
+ Increased attrition and
+ Increased errors (potentially a huge risk)
A key dynamic in the model is that these consequences of stress further raise the
level of stress in the staff, exacerbating the problem and generating a vicious
cycle (feedback loop).
In order to generate these dynamics, the team developed a “general theory” (how
the system works) of the Claims area that specifically links the following four core
elements
+ Workflow (Claims)
* Staffing (people)
+ Skills (capabilities and productivity) and
+ Stress (sources and impacts)
The model generates an array of output variables that unfold over time including
decision times for claims, errors (growth opportunities) and stress levels among
the staff.
Both the modeling process as well as the model itself generated some very
powerful insights and sometimes “counterintuitive” results for the modeling team.
We will now describe the general theory and then highlight the most interesting
dynamics it generates.
General Theory of Claims Workforce Dynamics
WORKFLOW
seavarg
STRESS
STAFFING
PIPELINE
Overview
The above map represents the major assumptions built into the model. Below is
a step-by-step build of this general theory to help develop understanding of these
assumptions and their implications. The general theory necessarily simplifies the
detail complexity of the Claims process to maintain the simplicity required for
ease of understanding. The actual model contains significantly more detail and
data in order to run the required simulations.
The model includes four major elements: workflow, the staffing pipeline, skills,
and stress. Based on their experience — and external research — the team
proposed a set of relationships between these major elements that could
generate the historical behavior experienced by the claims area.
The workflow sector focuses on Claims: Pending (including Pending & Active),
and Continuing Claims. The map shown here includes two stocks representing
these claims. A stock is an accumulation that represents the current level of a
variable “at a point in time”. Think of a stock as a bathtub containing “stuff”. For
example, right now there is a backlog of Pending Claims that could be counted.
And just like bathtubs have wommrtow 7 =
inflows and outflows, the tea J rr
stock of Pending Claims has @:O-/"22") Tes oe :
an inflow of arriving pcs j si
(pending claims). These
claims are then approved or
rejected (approving pcs and
rejecting pcs). Along the
way, approved claims will be
Active (not shown) and then
become Continuing Claims.
Continuing Claims generate
additional work that needs to sTAFne
be done beyond the
approval process. The workflow element of the simulation model also included
an appeals process, complaints, and growth opportunities (errors).
Staff and their Skills determine the ability to process claims and complete other
work (e.g. completing cc work). For simplicity, the general theory shows two
classifications of staff (the simulation model included multiple levels of staff, each
with different types of tasks and time allocations). The two classifications are:
Junior Staff and Senior Staff.
STRESS
WORKFLOW staal a
STAFFING
PruUNe
The general theory mapped here tracks people and their associated skills/
capacities separately. The Junior Staff stock is the count of the people at that
level; the Skills of JS represents the “total” skills of junior staff. Think of each
person as having a briefcase of all their skills. If everyone classified as Junior
Staff emptied their briefcases of skills into bathtub, this is the sum of all their
skills.
Skills determine how productive these employees are [A]. For Senior Staff, their
skills determine their productivity at helping with (mostly) Pending Claims. For
Junior Staff, their skills also determine how productive they are at completing cc
work.
As they do the work, staff build “on the job” skills (shown in the building js skills
ojt and building ss skills ojt). The amount (and quality) of mentoring provided by
Senior Staff can greatly influence the ability of Junior Staff to learn [B]. This is
represented by the variable time formally & informally mentoring.
However, since time is limited, the more time mentoring, the less time Senior
Staff can allocate to other work. And the reverse is true. The more work “that
must get done” accumulates, the more time Senior Staff would allocate to that
work — and there would be less time for mentoring. As workloads increase, the
WORKFLOW euisaatal ~
7, kvane ** ” *
result is Junior Staff are likely to learn less because of reduced mentoring. When
they leave, Senior Staff will take their skills with them. And the same is true of
Junior Staff [C]. One scenario the team wanted to understand was what would
happen if they lost a few Uber mentors, who really could help the staff learn.
Avery important type of skill that is also in the Ski/is SS bucket concerns
emotional intelligence (EQ). EQ directly influences mentoring ability. Low EQ —
even with high levels of skill at “doing the work” — makes for an ineffective
mentor. The Uber mentors can be considered to have a high level of EQ, so their
loss could be acutely felt.
EQ is also a skill that anyone can possess, and the greater that skill, the more
capable they are at handling Stress, an important — and often overlooked —
workplace element to be discussed.
As workloads increase, employees at all levels / positions can build up stress.
Stress JS and Stress SS are stocks that impact system performance in multiple
ways. [A more thorough explanation of the following points will be provided in the
next section.] First, stress can reduce a mentor’s ability to help others learn,
as well as diminish the ability to retain new information. Second, Stress JS
and Stress SS can reduce productivity [D]. A stressed employee will be less
WORKFLOW seprnea racter +
STRESS
SKILLS 4
STAFFING > 2] an ° (
PIPELINE 4 a f
productive — for an equivalent number of hours, a person who is stressed will
process less work than if that same person wasn’t stressed. Third, Stress can
cause employees to leave, shown in the js leaving and ss leaving flows. It will
reduce average tenure time [E]. Fourth, Stress can cause employees to make
errors, measured as Growth Opportunities in Claims [Growth Opptys].
The team explored various ways to reduce Stress on the workforce. The model
assumes that the greater the mentoring capacities (EQ) of Senior Staff, the more
work Staff overall could handle without building up stress. Also, the greater the
EQ of Junior Staff, the less stress they would build up.
Summary
In summary, the high level “general theory” of Claims includes:
+ Workflow
+ Staffing Pipeline
+ Skills
* Stress
Staff learn on the job, and their skills determine their productivity.
Overload can build stress; stress causes several negative consequences:
+ Reduced workforce learning
+ Reduced productivity
+ Increased attrition
+ Increased errors (potentially a huge risk)
EQ (through mentoring and personal self-management) can reduce stress.
An Important Structural Dynamic in the General Theory
Making Connections between Skills and Stress — and their impacts on
“doing the work’
This map captures the main mathematical relationships in the model connecting
work, staff, skills, and stress. For simplicity, it focuses on one type of work (e.g.
Pending Claims) and one staff position within Claims.
The diagrams in this section will explain these relationships to better understand
how the model represents stress and skills.
Productivity and Cumulative Learning
Work is arrayed to represent complexity. The stock of Work has a 3-D effect to
represent that some types of claims are more complex than others.
As work is completed, it accumulates in a stock of Cume Learned Tasks. There’s
an s-shaped relationship between Cume Learned Tasks and maximum
productivity. The complex work contributes more to staff learning.
Initially, employees learn a lot with each task completed. Eventually, their
learning “maxes out” [s-shaped growth “learning curve”] and they can’t really
become much more productive at a given task.
This is a powerful “productivity” dynamic.
saved (Easy, Complex),
Productivity
/
V4
U “
i ped
i H ancunt of
frencs 3 completed
1 A Sie Snise
i! oan |
ve +
F saregte sree tom acres
OP? core wregaed © reave D Presnett , hoy Sbem
1 ; : ao
\ | ‘
\ +
Ly . ee = ‘
v } ot] oe wt
\ * prwran: aueare
x
ee
me gay gone
~. " Stress dampers
eg warnny
a
==> n
more learning
Work Overload, Mentoring and Learning
The more workload per staff, the greater the need for mentoring to learn
(required coaching & mentoring...esp. from complex work]. The amount of
mentoring relative to the level required (coaching time relative to required)
determines how much /earning occurs.
Work s arraved (Easy, Complex).
What could accelerate
desipation? 4
Stress, Learning and Growth Opportunities (Errors)
Additionally, the more workload per staff, the greater the need for mentoring to
keep Stress from building. And Stress reduces learning and negatively impacts
productivity. This can lead to a vicious cycle (red) where workloads continue
to build with more Stress resulting. The reduced learning also contributes to
more work accumulating and more stress (loop of grey arrows + red arrow to
generating).
Additionally, errors (Growth Opptys) can result...creating more work.
Productivity
ume Learned Tasks
exp Pom rompiox work
| [Cx perenne Sf 4
wren \ al
i proactety
iid
rettve Wrequred = retstvelh Presto
in the absence
\ of sufficrent ete a
e \ siapiin %
i ee, What could accelerate
‘ Qaaity Pore cespation?
Stress dampens
~ earning
Resulting Dynamics
The relationships previously shown can generate behavior similar to the chart
here.
Bm Stress
B Growth Opportunities (errors)
B Decision Times
TIME
There’s a confounding effect where rising stress can lead to greater decision
times (due to decreasing productivity). Simultaneously, stress increases growth
opportunities (errors). These two dynamics create a “perfect storm” — where they
continue to worsen — as both greater workload (due to increased decision times)
and errors generate stress. Stress then accelerates both dynamics.
These dynamics have been recently been experienced not only in the claims
area, but across several other areas of the company. The team understood that
a successful workforce strategy would need to take into account these dynamics.
Model Insights
Among the key findings from the scenarios and interventions that were tested:
o The modeling team established the business scenario. This business
scenario was based on four “future world” assumptions — about the internal
workings of the company and the external business environment — under
which they would like to see potential dynamics of claims:
1. The business environment will become more complex, requiring
more time and resources.
2. The business will continue to grow.
3. It will be challenging to find the right people at right time.
4. Demographics make it more likely to lose Senior Staff with high EQ
The business scenario generated by the above would see decision times,
errors and stress all continue to rise over the next five years
o Through experimentation with the simulation model, the modeling team
determined that a highly effective response to the business scenario
involved a combination of interventions:
Y Provide EQ training for Senior Staff to improve mentoring
Y Provide stress management EQ training to all
¥ Implement process improvement
This combination of interventions directionally kept errors (growth opportunities)
from rising beyond current levels over the next five years and actually reduced
both stress and decision time relative to current conditions. The simulation
graphs from the model are shown below.
© ree tema
Growth Opportunities (Errors)
etm |
Reasonable
Business
Séenario Combined
ae ae Recommendations
Se eran eee |
5 years
te a = = =a
ser 2 tossisitnta tocar
rere rears
Decision Time i] Avg. Stress ———t
Reasonable
Reasonable i
Business ‘Scenario
‘Scenario
Combined Combined
Recommendations Recommendations
te io =a = = te T= = == =a
Ae a sa 7 cea
Project Insights
Team members noted that having a simulation model where scenarios and a
variety of interventions could be readily tested was a very powerful learning
process. The modeling process provided important insights into not only why it
was difficult to attract, develop and retain a capable Claims workforce but ways
to most effectively respond to these issues going forward. The team members
made the following observations
+ This method of problem solving combines data, experience, intuition and
judgment. It results in a richer understanding of the problem and uncovers
very different solutions.
+ This process brought a different frame of mind and was eye opening.
Specifically, the modeling team determined that
* The current status quo is not acceptable — Doing more of what we've been
doing is not enough to solve the problem.
* Hiring more people is not the answer. We need to stop “dabbling” in
growing EQ.
* The model moved us from a linear, reactive approach to be more proactive.
It enabled us to “see” the future.
In addition the team recognized that the insights as well as process could be
applied beyond Claims
+ This reinforces a broader focus on continued training and succession
planning.
* The model raises the visibility of so called “soft” issues (e.g. stress) and the
quantifiable outcome of addressing it
* The dynamics in this model are common across many organizations
o Dependence on the few best performers
o Observed high stress, lower productivity and attrition over time
o Lower capacity and fewer opportunities to grow staff with potential.
Lessons to Guide Your Organization’s System Dynamics Journey
As mentioned previously, the company has discovered several useful and
transferable learnings about the journey. Here they are again.
1. Make use of the full range of system dynamics (and systems thinking) tools
2. The problem (issue) chosen matters
3. Surfacing and challenging mental models (human “software”) is very high
leverage
4. Less is more generates better learning
We will now describe these lessons in more detail and use the case study to
illustrate.
1. Make use of the full range of system dynamics (and systems thinking) tools
Our multi-year experience at the company has demonstrated the value of using
the whole range of system dynamics / systems thinking tools: the iceberg, causal
loops, mental models, archetypes as well as stock and flow diagrams, simulation
modeling and standalone flight simulators These were all applied throughout the
journey; this approach to capacity building continues as we write this paper.
The tools used have evolved over time, as the organization has developed a
learning culture — with both an appetite and capacity to use the more
sophisticated tools of system dynamics.
When systems thinking was initially launched, causal loop diagramming was
introduced and used extensively. Introducing mental models and building
capacity for dialogue was an integral part of this work. A few years later, the
organization matured the use of systems archetypes (e.g. shifting the burden,
limits to growth and accidental adversaries) in order to identify leverage and
interventions more efficiently. Stock and flow diagramming was introduced a few
years later which made possible another level of learning and insight. This was
quickly followed by simulation modeling (using primarily iThink). Overall
participants in this work were able to see how the systems thinking tools (e.g.
causal looping and archetypes) are largely qualitative and descriptive tools and
even when combined with data analysis, lack the ability to simulate results over
time. Through the use of computer simulation, company leaders began to
actually see how causal loop structure could generate behavior over time to allow
them to test out “what if’ questions and potential leverage points in a more
rigorous way than was possible with causal loop diagrams. What was learned
through this was that the core systems thinking tools - causal loop diagrams,
archetypes, stock & flow diagrams and simulation - each have their role and in
combination are very effective when in service of learning and capacity building
to find uncommon solutions to complex, intractable problems.
The project described here made use of the full range of tools, from causal
diagramming to simulation modeling.
2. The problem (issue) chosen matters
We have found that the systems framework and tools work best on highly
“adaptive” (human, subjective) problems. At the company, leaders appreciated
the issues they faced had intangible, subjective human and often qualitative
dimensions that systems dynamics could easily address. Further, SD brought a
reasonable level of rigor to the process not possible otherwise. We strongly
believe that all the prior causal looping/mental modeling work established a
culture of openness among company leaders to the criticality of qualitative
dimensions when trying to model (understand) tough business issues.
With each modeling effort we made sure participants were truly committed to the
focusing question. The focusing question is meant to be a clear declaration of
what the modeling team is trying to understand - why were we investing our time
and effort. This is typically stated in the form of a “why” or “how come” question.
Hence the question mentioned earlier, “Why despite our best efforts are we
challenged to attract, develop and retain a workforce capable of meeting varied
20
and dynamic business needs? We made sure that the participants understood
that the focus of the systems thinking work was to generate insights into why we
were having the problem BEFORE putting our attention on “how’” to fix it. We
refer to this as “admiring” the problem. The intent here is to more fully “admire”
the problem and not short-circuit the process by jumping to solutions.
J don't have
any solution,
but J
certainly
admire
the problem.
The focusing question sets a reasonable and practical boundary on the modeling
effort. A critical function of the facilitators with active support from the participants
is to carefully manage the scope and detail of the model in order to produce a
rich learning model and avoid the classic trap in modeling of adding more “stuff”
that does not provide value to understanding the problem or identifying potential
solutions. This is about finding the “sweet spot” between too simple and too
much/too detail. We adhered to Barry Richmond's value to effort graph. There’s a
sweet spot on the curve where the amount of effort to apply system dynamics
achieves an asymptote around a mid-sized complexity model.
Barry Richmond's
Value per Effort graph
Ca
“ Complex model
lane Simple Model
Mother of ail models
j-—_——. Simple stock/fiow map
'+————._ Conversational use of skills
22;
3. Surfacing and challenging mental models (human “software”) is very high
leverage
The model and modeling process is the vehicle for challenging current thinking/
assumptions not a replacement for thinking. In fact at some level, the learning
process around modeling is more critical than the “product (model).” Our
approach at the company was based on our deep seated belief born out of years
of experience with clients from all sectors - both private and public that the
primary focus of the process as well all the systems tools (whether causal loops
or iThink models) is ultimately to engage leaders in having different
conversations than might be usual about important and often quite puzzling
issues and to create conditions that enable mindsets to shift. The bottom line for
this long-term effort was about generating uncommon solutions to complex
business problems by building collective (shared) intelligence (thinking) about
how the system works. This enabled company decision makers to make wise
choices about actions to take that will make a real difference.
Ultimately, the core of the modeling process we utilized facilitated a process of
“double-loop learning” where the participants were active participants in the
model building process. Developed by Chris Argyris, double loop learning is a
framework that makes the case that changing thinking (‘“Let’s think something
new”) is ultimately far more effective for bringing about change (or just better
results) than simply redoubling efforts, relying on conventional actions or reacting
(‘Let’s try something new’). To oversimplify, the idea is that we shift from “re-
acting” (single-loop) to “re-thinking” (double-loop).
23
Two Types of Learning*
Single-loop Learning vs. Double-loop Learning
Mental models
Problems Mental maps Behaviors
Results
Issues Assumptions Actions Ciunssinas
Challenges inferences Strategies Consanances
Opportunities (ladders) Plans peda 4
Beliefs Single-loop
“Let's try
something new.”
Double-loop
“Let's think something new.”
"Craig Weber, author of Conversational Capacity
Systems thinking and System Dynamics are highly suited to enabling people to
see the differences in the two learning modes and appreciate the power of
changing thinking in order to shift outcomes and performance.
One observation we have made in doing this work with clients over the years is
that embedding double-loop learning is not a one-time event. Making double-loop
learning stick requires repetition and sustained focused over time. The
momentum for most of us is to naturally revert to conventional or habitual
problem-solving (Old habits of thinking don’t die quickly).
4. Less is more generates better learning
When it comes to the model building process itself, we typically started off with a
simple causal map or simulation model versus building a bigger, more detailed
model. For example, we augmented the causal looping work that was done
during the first several years by using one or more of the classic system thinking
archetypes such as shifting the burden or limits to growth. Additional causal
looping work with the modeling teams generally then expanded on the core
archetype by building it out based on the stories they were sharing about the
issues.
24
In the more recent quantitative simulation modeling, we constructed a relatively
simple high level stock and flow model very early in the modeling process that
was intended to capture the essence of the “physics” of the issue we were
focused on. This often became the “general theory” that was our 30,000 foot
view of the issue. The follow-up modeling sessions then concentrated on
highlighting as well as exploring different aspects of the issue that seemed to be
dynamically important using the general theory map. We found that going slow
and focusing on a few intriguing, but easily digestible puzzles or koans (riddles)
was highly effective and engaging for the business leaders involved.
Conclusion
We were fortunate to be part of a long term process — still continuing — to build a
culture of double-loop learning. Our experience is that it is possible for all types
of organizations to undertake this process. The company has learned a few
lessons along the way. It is our hope that by conveying these we will encourage
and support other organizations as they undertake a similar journey.
25
