Module 4.

REVEALING SYSTEMIC STRUCTURES - DRAWING BEHAVIOR OVER TIME

GRAPHS

Introduction

In the previous chapters, we discussed the concept that systemic structures drive behavior patterns and
are consequently the source of many of our problems. In this part, we will discuss numerous techniques
for revealing these structures.

Objectives

In this chapter, you will be able to:

❖ Formalize the issue
❖ Determine the situation's essential variables, or the system's principal actors.
❖ Graph the temporal behavior of these variables

FORMULATE THE PROBLEM

Suppose you have recently completed a course on systems thinking and have identified a problem you
wish to solve. Could you sort it out using tools from systems thinking? Of course! All problems have
systemic causes; the key is to select one that is relevant and important to you. Here are some tips:

Criteria for Identifying Systemic Problems

1. The issue is persistent and recurrent.

2. The issue has existed long enough for it to have a history.
3. You or someone else may have attempted to solve this issue, but your efforts either failed entirely
or became ineffective over time.
4. You have not been able to identify a clear cause for the behavior pattern over time.
5. In Figure 4.1, "Patterns of Problem Behavior," one of the classic shapes represents the behavior
pattern of the problem over time.

Figure 4.1 Patterns of Problem Behavior

Yet another reason for conducting a systems thinking analysis is that the problem is significant to you or
your organization and warrants the investment of time and effort. Perhaps the issue is presently being
discussed, or you have a hunch that a past problem is about to resurface.

Listed below are examples of problems exhibiting typical systemic behavior:

● We are having difficulty assembling our refrigerators quickly enough to meet customer demand.
Therefore, we reconfigured the material flow on the manufacturing floor in an effort to enhance the
 assembly process. However, after implementing this modification, the assembly process took
longer than before. It appears that the change has somehow made everything more detrimental.
● Three years ago, we introduced a line of premium investment portfolio products, but our agents
still promote the older products. We've tried modifying the incentive programs and distributing vast
quantities of marketing and informational materials, but nothing seems to motivate agents to
prioritize the newer line.
● Every quarter, we implement a new round of cost-cutting initiatives, extended maintenance
intervals, and marketing simplification. Costs decrease for a time but then begin to rise again.

Guidelines for Problem Formulation

After identifying a problem for a systemic approach, develop a clear, concise statement of the issue. This
is frequently the most difficult aspect of systems thinking, but it is extremely valuable. The greater the
clarity and specificity of your problem statement, the more focused your systemic analysis will be. Ensure
that you engage in group brainstorming with individuals who can contribute their perspectives on the
issue as well as their knowledge of its history. If necessary, proceed with two or three problem
formulations and learn from the various perspectives. It is normal and even advantageous for this phase
of the process to require some time. You and your group will gain the most insight into the problem if you
take the time to ask a large number of penetrating questions, share your perspectives on the issue, and
revise your problem statement multiple times, if not numerous times.

The following are instances of problem statements:

● Customer service issues have increased by thirty percent in the past year.
● Prior to the last four training conferences, we were late in returning one-third of the registration
confirmations to customers.

Problem statements frequently (but not always) include the following components:
● the behaviors (example: customer-service problems)
● a time-based account of the behavior (example: the problems have increased)
● a measurement of the change in behavior over time (example: the problems have increased 25
percent)
● the duration of the conduct (example: the problems have increased 25 percent in the last year)

IDENTIFYING VARIABLES

Once the problem has been formulated, it is time to identify its main variables. (Variables are the
components of the problem whose value can vary over time; that is, go up or down.) To commence this
process, briefly describe the problem. Storytelling entails elaborating on your problem statement by
fleshing out some of the details so that you have a more complete understanding of the issue and the
variables at play.

Example ➤ The Case of Shamzung

At Shamzung, an electronic semiconductor company, we are perplexed by a series of events that

transpired during our most recent quarter. The majority of our sales force met or exceeded their sales
quotas for the quarter, resulting in quarterly sales records. All scheduled products have been released,
with additional products available for early release in the following quarter. Our administrative costs as a
percentage of sales reached an all-time high, which resulted in a decline in profits for the first time in our
company's history. What are the significant variables in the narrative of Shamzung? The following
summary was compiled by the managers of Shamzung after discussion:

● Sales
● Recent product launches
● Profits
● Sales force

Guidelines for Determining Variables

As essential as clearly defining the problem is deciding which variables to use, because your selection of
variables influences the remainder of your analysis.
As is the case with problem formulation, identifying variables is typically an iterative process. You may
begin by listing several, and then, after much discussion and consideration, decide to delete some, add
new ones, or revise your explanation of the problem.

Here are some guidelines for getting started:

● Quantitative and qualitative variables that could be reasonably included should be listed. For
instance, your list could include measurable variables such as "Sales" and "Sales Force Size" as
well as difficult-to-measure variables such as "Morale" and "Commitment to Company Goals".
Again, if possible, conduct this activity as a group in order to obtain a wide range of perspectives.
The plan is to begin with a lengthy list of variables.
● Reduce your inventory to the most essential variables. You can accomplish this by combining
variables that essentially represent the same type of information, such as "Morale" and "Job
satisfaction." You may also eliminate some variables from your list if you and your team determine
that they are not as strongly connected to the central problem as other variables. Determine which
variables appear to play the most prominent roles in the central issue you described in your
problem statement in order to determine the most relevant variables. These variables are likely to
have a relationship that you can either describe or investigate. For instance, in the case of
Shamzung's declining profits, the variables ``Profits", "Sales", and "Recent product launches" are
related to what the company identified as the core issue, and they have a significant dynamic
relationship.

Guidelines for Identifying Variables

After selecting your variables, it is time to give them precise names. The following guidelines will be
particularly essential when creating causal loop diagrams in Chapter 5.

1. Use nouns and noun phrases rather than verbs and verb phrases

Example ➤ New products in the pipeline

Revenue
Experience level of engineers

Not:

Developing new products

Being profitable
Sell
Produce

2. A well-named variable fits into phrases such as “the level of,” “the amount of,” “the number of,”
“the size of”.

Example ➤ The number of new products in the market

The amount of revenues

The experience level of the engineers
The size of the profit margin

3. Use a neutral or positive term whenever possible to name a variable.

Example ➤ “Job satisfaction” rather than “Job dissatisfaction”

“Morale” rather than “Negative feelings”

Thus, you will be able to explain how the variable changes ("increases," "decreases," "improves,"
"worsens") without confounding double-negatives. For instance, "Job satisfaction declined" is
conceptually much simpler to understand than "Job dissatisfaction declined."
 4. Bear in mind that variables can be tangible entities such as memory chips, structures, and
production workers, as well as intangible entities such as morale, job satisfaction, and company
value alignment.

After formulating your problem, selecting and

naming its variables, the next step is to plot the
behavior of the variables over time. The graphs
can then be used to generate hypotheses about
the interrelationships between variables and to
generate additional graphs that lead to greater
comprehension of the problem. To create behavior
over time (BOT) graphs as shown in Figure 4.2,
we suggest the following three steps:

1. Identify a time horizon.

2. Sketch the graph.
3. Construct hypotheses concerning the
interrelationships of the graph's variables.

Figure 4.2 The Behavior Over Time Graphs

The Time Horizon

Deciding on a time horizon for your graph is crucial, as the time horizon affects the quantity and type of
information your graph will ultimately display. Here are some guidelines:

1. Select the variable with the longest time cycle, such as new product development, production, or
sales cycle, and if feasible, extend the time horizon to encompass three or more cycles.

2. Attempt to work for at least two years, and explore employment for five or more years. If you need
to work with a significantly shorter or lengthier time horizon, discuss your justifications with your
colleagues.

3. Create a timeline similar to the example below, which includes:

Now is the present moment in which you are analyzing the problem.

Earlier is the point in time, two to five years in the past, from which you will begin to trace the
behavior of the variables.

Earliest is a point in time even earlier than the earliest, at which an event may have initiated the
problem.

Check again at your list of variables. Visualize the behavior over time of each variable between the
“Earlier” point and “Now.” Following this, reconstruct each variable’s “Earliest” behavior, if possible. Lastly,
picture how the variable will function in the future if nothing around it changes. These “thought
experiments” can help you visualize the behavior of your variables over a broad span of time—to get the
“big picture” of how they changed.

Bear in mind that BOTs may be more free-form than other graphs you are used to seeing, particularly if
you work primarily with quantitative data. The lines are intended to indicate qualitative patterns over time
rather than precise values. Certainly, these patterns may eventually need to be verified by quantitative
data at some point throughout the process.

Guidelines in Sketching the Graph

Here are the guidelines for drawing a BOT graph once you’ve decided on a time horizon:
1. Graph your important variables together on the same graph. That way, you can see the variables’
interrelationships—parallel variations, opposite variations, and delayed effects. Even though the
variables are measured in different units, BOTs show how variations in behavior of variables could
be dynamically interrelated.
2. Label the lines concisely. If possible, use distinct colors to draw each variable.
3. On the horizontal axis, indicate the time horizon, either the number of years encompassed or the
dates.

NOTE: If you notice that a series of significant events associated with variables in the graph have
occurred, you may have found another variable to include in your graph. For instance, assume your
graph showed that about every four years, a significant number of individuals left the company voluntarily.
In that case, you may add “Resignations” as a new variable and graph it in combination with your other
variables.

Utilizing Your BOT Graph to Create Testable Hypotheses

When you’ve outlined your initial BOT graph, the next stage is to hypothesize about how the variables’
behavior may be interrelated. This step often leads to concepts for new variables, additional BOT graphs,
yet there are more theories about how all the variables are interconnected. As with earlier stages of this
process, at times you may feel as if you’re “going backwards,” but in fact it’s this iterative quality of
systems thinking that makes it so valuable for generating insights.

In the Shamzung semiconductor story, the managers chose a time horizon, constructed their initial graph,
and drew lines representing the behavior of each variable (Figure 4.3, "Shamzung's Performance Over
Time"). Then, as they continued to investigate the problem, they posed queries, proposed hypotheses
regarding the relationships between the variables, and drew additional graphs.

Figure 4.3 Shamzung’s Performance Over Time

The initial graph created by Shamzung illustrated the behavior of sales, sales force, profits, and new
product releases. The managers of Shamzung hypothesized that the issue may lay in the relationship
between the total number of new products and the unit cost of carrying new products after noting that
while sales were increasing, profits were declining. According to their second hypothesis, the issue may
originate from the correlation between the number of low-revenue new products and the average selling
price. These new hypotheses prompted the creation of their second graph, which is depicted in Figure
4.4 as "Pressure on New Product Development."
 Figure 4.4 Pressure on New Product Development
This second graph depicts variables that are subsets or refinements of the original variables. The
managers identified these new variables as they collected data, discussed the issue, and made their first
graph. As you endeavor to draw BOTs in the subsequent learning activities, be mindful that identifying the
problem and variables and creating a graph may require multiple attempts. However, the effort will pay
off.

After completing these steps, you will be able to draw the causal loop diagram discussed in Chapter 5 to
represent the system structure that has been generating the patterns you observe on your BOTs.

Over the past decade, sales revenue at Shamzung had increased every quarter, but profit growth had
slowed in recent quarters. Meanwhile, both the number of new product launches and the size of the sales
personnel increased annually.

When Shamzung's managers analyzed their product sales over time, they discovered that while the
number of products had increased steadily, the cost of transporting them had increased even more
rapidly. Drawing this BOT helped identify a self-perpetuating cycle in which rising sales pressures caused
new product development to churn out less innovative and potentially more profitable products that were
simpler to develop and launch.

ACTIVITY 1. THE PROBLEM WITH PRICE PROMOTIONS

Purpose: To identify the central problem in a complex situation

Outcome:
● A statement of the central problem in the case study.
● A description of a deeper problem that may lie beneath the central problem described in the case.

Instructions: Read the case below and then answer the questions that follow.

Case Study ➤ The Problem with Price Promotions

Slower U.S. population growth in the early 1990s resulted in lesser annual increases in consumer
consumption, particularly of food products. Moreover, manufacturers' product innovation slowed, and
businesses struggled to differentiate their brands in ways other than price. Their response: Offer price
reductions to increase sales.

Manufacturers and retailers acknowledged that constant promotions can undermine brand image and
encourage consumers to shop based solely on price. In addition, companies could develop a
dependence on temporary sales promotions. Moreover, the promotions of food product manufacturers
gave supermarkets a disproportionate amount of power, as they controlled these promotions.
Supermarkets could request a variety of subsidies, such as fees for prominent shelf space and funds to
pay for promotional materials and newspaper advertisements.

The emphasis on price promotions produced practices such as "forward buying," in which a supermarket
purchased more of a discounted product than it intended to sell during a promotion and then increased its
profit margin by selling the remaining product at its regular price. Some supermarkets also "diverted"
some of their low-priced shipments to supermarkets outside the promotional area at a minor markup (but
still well below the wholesale list price). This resulted in a significant portion of manufacturer-initiated
discounts that were intended for consumers ending up in the wallets of retailers.

QUESTIONS

1. What are all of the issues outlined in the case? Include as many as possible.
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________

2. From the perspective of food manufacturers, what would you characterize as the overarching
issue that encompasses many of the specific issues you listed in Question 1?
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________

3. Is there an even deeper problem other than the one you named in Question 2? If so, what is it?
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________

ACTIVITY 2. THE CASE OF THE ENERGY DRAIN

Purpose: To practice drawing a behavior over time graph

Outcome:
● A statement of the overall problem in the story
● A list of key variables
● A graph of those variables’ behavior over time

Instructions: Read the case below and then answer the questions that follow.

Case Study ➤ The Energy Drain

Monday morning, at 6 a.m. The alarm sounds, jolting you awake. You make your way to the kitchen for a
cup of coffee. A few sips, and... ahh! The curtain begins to lift and your eyes begin to open. 10:30 a.m. is
the weekly personnel meeting time. "I feel so groggy," you ponder. "I need something to keep me awake
throughout this"You pour another cup of coffee before proceeding to the conference room. You are
speaking with your coworkers during a brief lunch break at noon. Someone refers to a newspaper article
about the fitness regimens of fashion models. "It was stated that those high-priced runway models must
rely heavily on caffeine. 'Beauty sleep,' I'll wager, is how they maintain their high levels of energy.

Who has time for daily exercise, getting paid to exercise, etc. is the subject of much discussion. You are
experiencing a mid-afternoon energy drop at 3:30. You approach the congested coffee cart to obtain a
second cup. You say to a friend in line, "I really ought to cut down on this stuff," meaning that you should
consume less of it. He agrees. "I drink five cups of coffee per day myself," he admits. "I just am unable to
give it up."

QUESTIONS

1. What is the issue with this story?

______________________________________________________________________________
______________________________________________________________________________
 ______________________________________________________________________________
______________________________________________________________________________
______________________________________

2. What are the top three or four variables in this case?

______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________

3. What is the temporal behavior of these variables? Create a graph in the space provided.

4. What observations do you make regarding the conduct of the variables? For instance, do any of
them consistently increase or diminish over time? Are there significant fluctuations (sudden
increases or decreases)? Does it appear that any of the variables cycle, as indicated by
up-and-down patterns? Are any of them stable?
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________

5. Do you observe any relationships between the variables' behavior patterns? Moreover, what? For
instance, does one variable rise or decline, followed by the rise or fall of another variable?
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________
______________________________________