Is This a Risk Worth Taking?

By Azhar Syed

Categories

Your team wants to change a supplier. The new price would cut costs, but a disruption during the transition could stop production. Some colleagues see an opportunity. Others see a risk the business should not take.

Arguments like this tend to stall because each side is defending a conclusion. I would ask the team four questions instead. What could happen? How likely is it? What would it cost if it did? What can we do to make it less likely or less costly? Then I would ask the same questions about staying with the current supplier.

That leads to the question that matters: after the actions we can realistically take, is the remaining risk worth the opportunity?

Define the decision and the alternative

“Should we change suppliers?” is usually too broad. We could qualify a second supplier, move a small share of orders, or switch completely. Each choice exposes the business differently.

Be specific about the objective and the period. Are we assessing disruption in the first year, or reliability over five years? A probability without a period means very little.

Then describe what happens if we stay as we are. Current arrangements carry their own risks: rising costs, slipping performance, dependence on one provider. Because they are familiar, they are easy to overlook. Give them the same scrutiny.

Separate likelihood from consequences

“The transition is risky” gives us nothing to work with. “The new supplier may miss an early delivery, and we would have to stop a line or buy emergency supplies” does. Now there is an event to assess. How likely is a missed delivery during the transition? If it happens, what does it cost, and who else is affected?

For a financial consequence, a simple calculation helps:

Expected loss = probability of the event × estimated loss if it occurs [1]

The second number should be the average loss if the event happens. If the possible losses vary widely, work through a few scenarios instead of one figure. Write down what each estimate rests on: comparable transitions, supplier records, specialist advice, or assumptions nobody has tested yet.

When the evidence is thin, use ranges. Writing 23.7% does not make a guess more reliable. A low, central and high estimate shows whether a reasonable change in assumptions would change the decision.

Check what the controls actually change

Some actions make the event less likely. Qualification testing and verified production capacity reduce the chance of a missed delivery. Others make it less costly. Temporary stock or a second source limits the interruption if a delivery fails.

Risk managers call the exposure before controls gross or inherent risk, and what remains after controls net or residual risk. [2] If your first estimate already assumes the controls you have today, call it current residual risk and assess any new controls separately. Otherwise the same protection gets counted twice.

Here is a deliberately simple, hypothetical assessment of one transition risk:

Assessment over the first yearProbability of disruptionEstimated loss if disruption occursExpected loss
Gross risk before controls25%$400,000$100,000
Net risk after effective controls10%$150,000$15,000

These figures are illustrative, not benchmarks. The lower probability assumes the qualification work supports it. The lower loss assumes the contingency stock and second source would work when needed. Suppose those measures cost $20,000 in the first year. That cost belongs in the business case, and the $85,000 fall in expected loss is an estimate, not cash in the bank.

Then separate controls that are already working from controls that are only planned. A backup supplier on a list is not the same as one with verified capacity and agreed terms. Name who will put each control in place, by when, and what evidence will show it works. Until then, the lower figure is a target. [3]

Question the number you want to believe

Look again at that 10%. A manager keen on the savings could accept it too readily.

We sometimes favor an answer before we have assessed it and then find reasonable arguments for it. Research on motivated reasoning describes how the conclusion we want shapes the evidence we look for and accept, usually without our noticing. The same research adds a point that matters here: we reach the conclusion we want only when we can build a case that would look reasonable to someone else. [4] A neat table can be that case. Structure helps, but it does not remove the bias. It can give the bias somewhere to hide.

So ask what you are feeling and where it comes from. In two experiments, anxiety from an unrelated source made people less willing to take risks. The effect disappeared among people who were better at understanding emotions, and among those who were told where their anxiety came from. [5] If you feel uneasy, ask whether the unease is about this supplier or about something else entirely. If it is about the supplier, investigate it.

A gut feeling can also carry real experience. Ask whether you have seen comparable transitions and learned clearly how your earlier judgments turned out. Intuition is more trustworthy where the situation has regular patterns and you have had the chance to learn them. Feeling confident does not show that you are right. [6]

Then look for the strongest case against your preferred answer. In experiments on social judgment, asking people to consider the opposite reduced their bias more than asking them to be fair and unbiased. [7] Ask a colleague to challenge the qualification evidence and the backup supplier’s capacity. What would make 20% more plausible than 10%? At 20%, the expected loss is $30,000. Add the $20,000 the controls cost and the total is $50,000, still below the $100,000 expected loss without them. Under these assumptions the controls still earn their cost. Whether the change of supplier is worth making is a separate question, and it depends on the opportunity. Apply the same challenge to the case for staying with the current supplier.

Weigh the opportunity with the same care

The new supplier may offer savings, extra capacity, or capabilities we do not have. Ask what has to happen for those benefits to arrive. How likely are the savings once you count transition costs, quality requirements and management time? Would a smaller commitment keep most of the benefit?

The opportunity can be improved too. A trial lets both organizations learn to work together. Better specifications can raise quality.

Compare the alternatives over the same period and on the same basis, and do not count the same loss on both sides. Where cash flows run over several years, compare their present values using consistent assumptions. Some consequences, such as customer trust or management attention, still need explicit judgment even when you cannot price them.

Decide what you can afford to accept

The residual expected loss in the example is $15,000. If the disruption happens, the loss is $150,000. Expected loss is an average for comparing options. It does not limit what you can lose. [1] Worse outcomes than the one modeled are also possible.

Ask whether the business could absorb a credible severe outcome. Think about safety, legal obligations, customer commitments, and several problems arriving together. A favorable average does not make every outcome acceptable.

You may decide to proceed in stages, wait for one particular piece of evidence, or decline. More analysis is worth doing only if it could change the choice. It costs time, and delay has its own price.

Before you commit, write down the assumptions behind the decision, the controls that must be working, and the conditions that would make you pause or change course. Give someone the job of watching those conditions. Then ask whether you used the analysis to decide or to justify what you already preferred. If you cannot name any evidence that would change your mind, the analysis has not been tested yet.

In my book When Trusted with More, the chapter on Judgment looks at how leaders make a consequential decision before they are certain, and how they get others to carry it. A structured assessment gives colleagues something specific to challenge: the 10%, the $150,000, the backup supplier’s capacity. Their knowledge has to be able to change your answer while the decision is still open. And the people who must act on it should carry it as their own, not just do what was decided.

Learn more about When Trusted with More.

Sources

1. Alan Gump, “Using decision models in the real world,” PM Network 15, no. 1 (2001): 43–45, Project Management Institute. Defines expected monetary value as probability multiplied by impact, and notes that the cost actually incurred if the event occurs is the full impact. Read source

2. ProtectUK, “Stage 3: Risk Treatment,” March 21, 2024. Defines residual risk as the risk remaining after additional controls. Read source

3. HM Treasury, The Orange Book: Management of Risk, Principles and Concepts, updated July 29, 2026, sections 5.5 (risk treatment) and 5.6 (risk monitoring). Covers weighing treatment against its costs, implementation plans, accountability, and monitoring whether controls operate as intended. Read source

4. Ziva Kunda, “The case for motivated reasoning,” Psychological Bulletin 108, no. 3 (1990): 480–498. Read source

5. Jeremy A. Yip and Stéphane Côté, “The emotionally intelligent decision maker: Emotion-understanding ability reduces the effect of incidental anxiety on risk taking,” Psychological Science 24, no. 1 (2013). Read source

6. Daniel Kahneman and Gary Klein, “Conditions for intuitive expertise: A failure to disagree,” American Psychologist 64, no. 6 (2009): 515–526. Read source

7. Charles G. Lord, Mark R. Lepper and Elizabeth Preston, “Considering the opposite: A corrective strategy for social judgment,” Journal of Personality and Social Psychology 47, no. 6 (1984): 1231–1243. Read source

Further reading: HM Treasury, The Green Book (2026). Read source

SHARE WITH YOUR NETWORK