Project Team: Warren Farr
“Using the proven Strategy Dynamic process focused our limited resources on organizing strategic issues, identifying the critical resources, and developing the insight to more rapidly create intuitive and actionable business strategy.”―Warren Farr, CEO, Refrigeration Sales Corporation
System Dynamic’s benefits do not have to be driven by outside consulting efforts. Sometimes, internal education and training lead to insights which cause a company to thrive and even evolve during the most tumultuous times. Take Refrigeration Sales Corp (RSC), during the downturn of the late 2000s. CEO-turned-modeler, Warren Farr, was able to see RSC’s revenue go up 25% in a market that saw sales go through a 25% decline, thanks to forecasts driven by SD modeling and insights gained during the process of developing those models.
RSC is a third generation privately owned family business located in Ohio, USA specializing in wholesale sales and support of Heating, Ventilation, AC and Refrigeration (HVACR) units. Over the years the company had expanded its business beyond just equipment sales to credit lending, technical support and training for many thousands of statewide contractors. At the time the HVACR industry itself had enjoyed three decades of continuous growth, and RSC had seen annual AC sales increases averaging 10%. But for the first time in years, these annual sales rates were decreasing. Competitors, and the company’s own customers, believed this to be a “temporary lull” caused by the economy as a whole and, along with more cautious suppliers, most parties were optimistic for the future. However, Warren, who had just taken the first of several courses in System Dynamics, believed that recent trends in 2001 signaled a new phase in the market’s development and that this historic growth would inevitably slow down more permanently.
Applying System Dynamics
Inspired by the notion of feedback stemming from inside the system rather than external causes, Warren applied the SD methodology to create long-term market forecasts; ones which were crucial to a company whose sales and profits were tied directly to the fortunes of market volume. Working off of the three views of the future in the figure below, Warren sought to find the underlying truth of the industry. While suppliers and customers were sticking with the “hope” prediction, and the company’s management was preparing for the “best guess”, information on the market’s installed base showed clearly that the “fear” scenario was in fact the appropriate response to prepare for.
A key systemic insight is this. The nature of durable goods, is that they are, well, durable. Sales levels for a durable product have the shape of a bell curve; sales of new units increase until the market becomes saturated, at which point new sales are mostly replacement or upgrades. Warren’s competitors and customers were blind to this basic truth, and fully expected demand to start rising again with the next recovery in the economy and in construction.
This saturation effect is simple enough to state – but how significant would it be, and over what time-scale would it play out? The dynamic model quantified the accumulation of the installed-base and internalized the feedback of declining first-time sales, allowing managers to observe reality playing out and refine their intuitions about the ebbs and flows of the industry’s growth cycle.
Not only was the approaching saturation point reducing the installation base for new AC units, but in a northern U.S. region like Ohio where an AC unit could last for 15 to 20 years, replacement sales were unlikely to fuel growth either. The model was predicting a 20% to 30% contraction in sustainable annual unit sales, which would be considered catastrophic in most markets.
U.S. Regulations Delay Market Contraction
The reaction to these insights was swift and not without controversy. In the years 2004 and 2005, sales spiked even higher than the hopeful prediction thanks to new US Dept. of Energy regulations which were causing contractors to stock up on older AC units before new minimum efficiency requirements took place and raised prices. But RSC stuck to its forecasting model. In an industry which had known nothing but growth for decades, RSC was downsizing its workforce, tightening customer credit limits and consolidating its inventory and warehouses. Both suppliers and customers thought these actions were short-sighted according to common industry logic and current record sales. But RSC’s uncommon logic would swiftly prove itself as the predicted market contraction came true. After 2005, sales entered an unprecedented crash, setting the market average back down to 1994 levels.
“The modeling provided a longer-term perspective, allowing RSC management to make these dramatic changes in company bandwidth over a period of about 3 years, avoiding the excess cost associated with rapid ‘cutting’.” ―Warren Farr, CEO, Refrigeration Sales Corporation
Forecasting Leads to Market Share Growth
Thanks to their SD forecasting efforts, RSC changes were proactive and planned, and not reactive with unintended consequences. Though RSC would have survived the industry downturn without modeling, the company was now in a prime position to diversify and take on new market share as competitors sought to shed expenses in a bid to stay afloat. New talents, territories, and market shares were taken from unprepared firms scrambling to adapt to the crash. In territories where the AC market saw a 30% contraction from 2000 to 2010, RSC saw a 25% increase in annual revenues and 30% increase in its sales locations.
RSC’s story highlights a number of points. First, managers need to regularly ‘step outside’ the day-to-day crush of business activity, and understand the fundamentals of what is happening in the market around them. Second, just being aware of principles – like market saturation – may not be enough. Often what is needed is modeling work to figure out the quantified, time-based consequences of these dynamics. Third, the company’s work highlights the advantage of making decisions with forethought and examination, rather than waiting for the impact of new market conditions to throw off business plans. The story reflects the simple but powerful insights that System Dynamics can bring to such questions. Clearly, RSC’s strategic change decision was very significant, enabling them to create new advantages over the competition.
Lastly, we can see that System Dynamics is not some highly technical tool only accessible to math whizzes, but a practical, accessible method that smart managers like CEO Warren Farr can learn to use to great effect.
Air Conditioning Industry Case Study by Warren Farr
Project Team: James Rogers, Advance Management Group; Edward Gallaher, Advance Management Group; Craig Hocum, Mayo Clinic; David Steensma, Dana-Farber Cancer Institute; T. Ronald Chrisope, Advance Management Group; David Dingli, Mayo Clinic; James McCarthy, Mayo Clinic
Systemic, long-view approach to End Stage Renal Disease treatment increases patients’ well-being while decreasing drug costs by $2 million per year at Mayo Clinic in Rochester, MN.
Craig Hocum, a Physician’s Assistant focused on Nephrology and Hypertension at Mayo Clinic, along with Jim Rogers and Ed Gallaher of the Advance Management Group, were blown away by the insights their team came to as a result of the integration of systems thinking and clinical practice in End State Renal Disease (ESRD). Their team developed a tool that uses System Dynamics to provide improved, individualized Erythropoietic Stimulating Agent (ESA) dosing regimens for dialysis patients, creating positive effects on patient health and well being, while reducing Mayo’s yearly drug and hospitalization costs by $2 million. Though not precisely quantified, additional savings of about the same scale have been realized through reduced hospitalizations.
Rogers and Gallaher were originally commissioned to develop an ESRD reporting system, but their long history of Systems Dynamics work helped them find potential leverage in the longitudinal per-patient data that was not normally considered in other dialysis treatment plans. Combining this big-picture view with Hocum’s direct experience managing ESA treatment led the team to a strong understanding of the underlying biophysical dynamics and a highly effective individualized treatment solution.
Assessing the Data through System Dynamics
The administration of ESA is intended to raise red blood cell levels for patients undergoing dialysis treatment. When Hocum, Rogers, and Gallaher looked at the historical dialysis patient data, they saw oscillations both above and below the optimal red blood cell count as a result of both over- and under-administration of ESA. Both too high and too low red blood cell levels have negative effects: low blood cell levels are associated with fatigue and poor quality of life, while high red blood cell levels are associated with increased risk of stroke and cardiovascular disease. (1, p3)
By mapping the situation using Systems Dynamics tools, the team identified that the wide oscillations occurred due, in part, to a delay of 20 days or more in the effect of the ESA. This delay in impact resulted in patients being administered a second dose before the first had taken effect, which meant over-administration of the ESA and higher than intended red blood cell levels. Once the caregiver noted these high red blood cell counts, they would then delay additional administration of ESA in order to allow the red blood cell levels to lower on their own. Again, due to the significant delay in the effect of the ESA, it would then be “too late” at the next administration to prevent very low levels in the red blood cells. In Systems Thinking, this system of impact is referred to as a delayed feedback loop, where each factor impacts the next with a delay.
Misunderstandings Mapped and Corrected
Once they identified the delayed feedback loops in the system of ESA administration, the team then created a stock and flow diagram of the key components (see figure below). When talking with David Steensma, a hematologist, this clear representation of the system and its elements provided a clear, accurate way to communicate the idea of systemic delays, which allowed Steensma to easily identify and correct a misunderstanding present in the team’s model.
Figure 1: This stock and flow diagram allowed the researchers to effectively communicate with specialists and resolve key misunderstandings in their work.
Systems Thinking provided the tools to clearly communicate across discipline-boundaries and to gain insight into the true nature of what they were trying to understand. In Rogers’ words: “As Ed astutely pointed out…we [didn’t] have a nephrology problem, we [had] a hematology problem.” And as Ed Gallaher himself later noted, the team was “collaborating from different disciplines…each discussion was a teachable moment for everyone involved.”
With their misunderstanding corrected and a cross-disciplinary arsenal of knowledge, the team was able to develop a System Dynamics model that used an individual’s historical data to generate an optimal dosing regimen, customized for each patient.
New Protocol Leads to Improved Health and Savings
The use of the dosing protocol defined by the System Dynamics model provides a high level of accuracy and customization that has reduced overall hospitalization time by 25% while also reducing the amount of ESA administered to patients. If this dosing method were to be applied across the United States, it could cut the use of ESA in half, saving approximately $2 billion dollars a year.
Figure 2. Behavior over time graphs generated at the conclusion of the study show that the oscillation patterns of the new treatment plans (indicated with a grey background) now fall within the optimal red blood cell levels (indicated by the two horizontal lines).
In addition to saving money, the new protocol also leads to dramatically improved health outcomes for patients. “It’s important to go back to individualized medicine – [the model is] based on how [patients] respond, which gives insight into their overall health,” says Hocum. When using the new model, fluctuations outside of the expected range of red blood cell levels generally indicate other bodily system failures or complications. In this way, the model allows doctors to monitor and address these complications earlier on, adding to an overall decrease in costly hospitalizations.
In Hocum’s words, systems thinking allows users to predict a patient’s trajectory: “If we have a good model and a good fit, not only do we get stabilization…[but,] if a person has the time to look at it, there is predictability.”
With billions of dollars to be saved, and opportunities to increase a patient’s’ well-being, “we’re surprised that diffusion of this thinking has taken as long as it has” says Rogers. After they were nominated for the System Dynamics Society bi-annual Application Award, the team is moving forward with the Biomedical Systems Dynamics Special Interest Group and looking forward to sharing these tools and insights in related fields where Systems Dynamics can yield similar results. Fields like methadone treatments, glucose and insulin treatments, thyroid hormone care, and long-term pain management all seem likely to yield substantial benefits from a similar systemic approach to dosing protocols.
Thanks to Jim Rogers, Ed Gallaher, and Craig Hocum for taking the time to share their story.
- Rogers, Jim, Ed Gallaher, Craig Hocum, et al. “Individualized Medicine and Biophysical System Dynamics: An Example from Clinical Practice in End Stage Renal Disease.”
“ …We could not get past…, ‘We have been doing this for 20 years, why shouldn’t it work now?’”―David Starr, Senior Director Information Technology, Company A
“This work affected a 7-point market share increase.”―David Starr, Senior Director Information Technology, Company A
Company A, a major credit card company, once held a dominant market position with only one primary competitor. They woke up one day and realized that, after six years of steady decline in revenues and market share, they had lost their leadership position. The company was stymied. Everyone was blaming each other. The erosion continued despite every effort to turn things around. Conventional wisdom, based on past experience, was not working. It was pretty desperate, nobody knew what to do, and it was feared that they would be out of business in the next five years if they couldn’t turn the “death spiral” around. It became clear, given the changes in the marketplace, that a new way of understanding the forces at play was necessary.
The company called in the System Dynamics group of PA consulting as part of a larger project to change the company’s fortunes. System Dynamics (SD) and PA consulting were chosen because of their ability to look at the problem in a whole new way: people within in the company were thwarting each other’s efforts, and a more holistic approach for the organization was necessary. SD provided the necessary perspective by looking at the system as a whole, without losing crucial details.
Among the early modeling insights, was a simple recognition of the perilous situation in which Company A found themselves. The fact was that the company’s clients, credit card issuers, tended to focus all of their marketing efforts on the market leader, while ignoring the runner-up. In this case, when Company B was king, clients only promoted Company B’s product to the detriment of Company A. This was a case of a common systems trap called “Success to the Successful” and, if left unchecked, would surely have led to the demise of the less successful entity, Company A.
Company A also learned that there was a limit on spending due to minimal payments. Customers would not use the cards to the full advantage if they were not being paid off. Counterintuitively, the best observed solution for this problem was to create a higher minimum payment, to allow for increased monthly spending.
PA consulting worked to collect real world experience, expert interviews, quantitative and qualitative data, and cultural factors. The information was cross checked, and at first, client focused. But a unique benefit of SD is its ability to scale upwards to encompass the entire market and capture the inherent complexities within the system. This model was scalable and could be moved from department to department to test a wide variety of initiatives. Hundreds of factors could be tested within the model’s nonlinear and time delayed system, to find the true leverage points which would allow Company A to regain and improve their market share. Consequently it was discovered that efforts such as increased value added assurance and increased issuer preference held the most value while conventional solutions like increased advertising held little leverage. This led Company A to be the first to market with the technique called co-branding, a partnership between a credit card brand and company, which allowed them to regain 6 points of market share. To the Company A member banks the increase was worth billions in relative market share. Since then, co-branded cards have taken off in the United States and have redefined profits within the industry.
For Company A, System Dynamics changed not only the way the company saw itself, its competitors and its products, but it also helped drive the creation of an entirely new product line. Company A was able to efficiently understand their strategic resources and move on to a new field of growth by recognizing the key leverage points within their own system, identified thanks to the work done by PA consulting.
“In the end, everyone took credit for the work.”―Sharon Els