Want to Understand the Power of Decision Analysis? Here's All You Need to Know…
In 2010, Dr. Carl Spetzler, CEO of Strategic Decisions Group, delivered a talk at the Lessons in Decision-Making
The story took place back in the 1970s, when Standard Oil of Indiana, later known as Amoco and eventually BP, was grappling with a significant strategic decision. The company had the technical capability to switch to unleaded gasoline—a costly transition that required massive investment in refineries. But the risks were daunting. Would consumers buy it? Would competitors jump on the bandwagon? Would the government force their hand by banning leaded gasoline? This wasn’t just about adopting a new product; it was about reshaping their entire business for the next decade.
Standard Oil had studied the issue before—four times to be exact. Each time, the conclusion was the same: too risky, too uncertain. But now, a new CEO was in charge, and he was determined to finally get an answer. So, they called in Carl and his decision analysis expertise
The Complex Web of Factors
One of the first things Carl’s team did was to help Standard Oil map out the interconnected web of factors that would influence the decision. They visualized all the elements at play through an influence diagram—everything from government action and market share to manufacturing costs and competitive response.
This diagram was a game-changer. It wasn’t just a visual; it was a roadmap to understanding how each decision and uncertainty would impact the company’s overall profitability. The diagram helped clarify the decision space and made it easier for everyone involved to see the interactions between key variables.
The Critical Drivers: Market Share and Government Action
From there, Carl’s team conducted a sensitivity analysis
This finding was pivotal because Standard Oil had already invested significant time and resources in refining their manufacturing processes. The realization that market share was the real make-or-break factor shifted the company’s focus. They now understood that the key to their success wasn’t about cutting costs—it was about ensuring customer adoption.
Divergent Expert Opinions
With market share identified as the critical uncertainty, Carl interviewed a panel of experts to get their views on how unleaded gasoline might perform in the market. The results were, to put it mildly, wildly different.
Some experts were deeply pessimistic, convinced that consumers wouldn’t buy in. Others were far more optimistic, predicting widespread adoption. These divergent opinions made the uncertainty even more challenging. If you believed the optimists, the switch to unleaded gasoline looked like a golden opportunity. But if you sided with the pessimists, it seemed like a disaster waiting to happen.
The President’s Judgment Call
Faced with these conflicting views, the president of Standard Oil made a bold decision. He synthesized what he had heard from the experts and drew his own compromise distribution. It was a middle ground—an acknowledgment of both the upside potential and the risks.
This was a key moment in the decision-making process. Despite the lack of consensus, the president understood that waiting for perfect information wasn’t an option. He had to make a judgment call based on the best available data, and that’s exactly what he did. His distribution became the foundation for further analysis, a reflection of the uncertainty but also a clear step forward.
Exploring the Scenarios
Armed with the president’s distribution, Carl and his team created a probability tree
The tree was an essential tool for navigating the complexity of the decision. It helped the team see how each uncertainty fed into the others, and it revealed the range of possible outcomes—both the good and the bad. By modeling the probabilities and outcomes of each scenario, they could better understand where the greatest risks and opportunities lay.
With the advantage of our knowledge from the future, we can see that while decision trees offer solid insights, they quickly become unwieldy for complex problems with many uncertainties. The "combinatorial explosion" of possible outcomes makes it nearly impossible to evaluate every scenario—especially when multiple layers of uncertainty are involved.
领英推荐
That's where influence diagrams and Monte Carlo simulations
The Product-Line Switch: A Risky Bet
One of the options on the table was a product-line switch to unleaded gasoline. The cumulative probability distribution for this decision was sobering. The analysis showed that while there was a possibility of significant profits, the expected value was negative. In the worst-case scenario, the company could lose over $600 million, and even the best-case scenario didn’t offer enough upside to justify the risk.
This realization was critical. The product-line switch wasn’t just a tough decision; it was a bad decision. The numbers made that clear. And because of this analysis, Standard Oil was able to avoid making a costly mistake.
Manufacturing Costs vs. Market Share
Another important takeaway from the analysis was the relative impact of manufacturing costs versus market share. The sensitivity analysis showed that even if the company managed to lower manufacturing costs, the upside would be minimal. On the other hand, a very high market share would transform the entire decision, turning what looked like a risky bet into a highly profitable venture.
This insight shifted the team’s focus. Instead of pouring more resources into optimizing costs, they realized they needed to double down on understanding and influencing market share. The upside potential was far too large to ignore.
Misallocated Resources: A Hard Truth
One of the most surprising findings from the decision analysis was that the task force had misallocated its resources. They had spent $750,000 simulating various manufacturing cost scenarios, even though the value of perfect information
This misallocation of resources was a wake-up call. It reinforced the importance of focusing on the right uncertainties. If the team had spent more time understanding market share, they could have made better decisions sooner.
Testing the Market: A Smart Gamble
In the final stages of the analysis, the team considered conducting a nationwide test market to gather more data on consumer demand for unleaded gasoline. The decision tree showed that while there was an 80% chance of losing $10 million to $20 million if the market share fell short, there was also a 20% chance of hitting a very large market share. If that happened, the company would move forward with full-scale marketing, with an expected value of $360 million.
This analysis made the test market a worthwhile gamble. The downside risk was relatively modest compared to the upside potential. And with more reliable data, the company would be in a much stronger position to make a final decision. But not everyone in the room was on board. The Executive VP of Marketing was vehemently against it, insisting that the company “never takes less than a 50/50 bet in marketing.” For him, the potential downside of market rejection felt too risky, and he wasn’t about to wade through the backlash of a failed test.
Carl, however, wasn’t backing down. The CEO, ever the pragmatist, jumped into the debate. "That’s when you can make $360 million and only risk $15 million?" he asked, clearly seeing the value of the test. What unfolded was a heated 15-minute exchange, where the tension became clear: the VP was seeing the situation through a personal lens—he would be the one dealing with the fallout if the test failed. Meanwhile, the CEO viewed the decision from a corporate strategy perspective, weighing the potential upside for the company as a whole.
In the end, Carl’s data-backed argument carried the day. The decision analysis highlighted the opportunity, and despite the emotional resistance, the CEO saw the strategic value of testing the market. It was a moment where decision analysis revealed not just the numbers, but the human factors—the individual risks and rewards that play into decision-making at the highest levels.
Decision Analysis: A Power Tool for Complex Problems
As Carl Spetzler shared this story, it became clear that decision analysis is like bringing power tools to an organization accustomed to using hand tools. It allows you to peel back the layers of complexity, identify the critical uncertainties, and focus on what truly matters. In the case of Standard Oil, decision analysis didn’t eliminate the uncertainty—it made it manageable. It gave the team the structure and insight they needed to make an informed choice, even in the face of massive unknowns.
If you want to hear Carl tell the story himself and learn more about how decision analysis was applied in this high-stakes scenario, you can watch the full recording of his talk here. It’s an eye-opening lesson in how powerful decision analysis can be, especially when the stakes are high and the future is uncertain.