Three people can each make perfectly consistent choices, yet produce a group decision with no consistent winner.

In ordinary reasoning, if you prefer A to B and B to C, you should prefer A to C. Mathematicians call this transitivity. It feels natural to expect a group’s preferences to work the same way. They do not always.

In his 1785 work on collective decisions, the Marquis de Condorcet examined how majority voting could produce cycles. The problem is now called Condorcet’s paradox. It belongs to social choice theory, which studies how individual preferences become collective decisions.

1. The mechanism

Suppose three people must rank three alternatives:

  • Person 1: A > B > C
  • Person 2: B > C > A
  • Person 3: C > A > B

Each person has a consistent ranking. Compare alternatives two at a time.

A versus B: People 1 and 3 prefer A. So A wins, 2–1.

B versus C: People 1 and 2 prefer B. So B wins, 2–1.

C versus A: People 2 and 3 prefer C. So C wins, 2–1.

The group says A beats B; B beats C; C beats A. There is no option that defeats every other option in a head-to-head vote. Such an option, if it exists, is called a Condorcet winner.

The surprise isn't that anyone changed their mind. Different pairs assemble different majorities. A consistent individual ranking does not guarantee a consistent majority ranking.

2. Outside CPG: A city's public square

Imagine a city committee choosing among three hypothetical plans: A, a landscaped park; B, a market plaza; and C, a performance space. Three equally sized neighborhood groups rank the options exactly as above.

A two-stage vote can make the result depend on the order of decisions. Put A against B first: A wins, but C then beats A. C becomes the final winner. Put B against C first: B wins, but A then beats B. A becomes the winner.

Nothing about the voters changed. The agenda changed.

That is why deciding how to choose can influence what gets chosen.

3. What this means for beverage innovation

Imagine evaluating three hypothetical can designs. A is strongest from across the aisle. B has the clearest flavor navigation. C best preserves familiar brand cues.

Different groups may rank those priorities differently. A vote could produce a clean-looking winner under one procedure and a different winner under another.

This does not establish that packaging panels always exhibit voting cycles. It also doesn't show which design will sell best. A ranking is a preference in a defined test, not a sales forecast.

Before collecting votes, identify the actual decision: recognition, comprehension, trial intent, or something else. Measure it in conditions that approximate the intended retail setting. When preferences conflict, make the tradeoff visible rather than treating the winning vote as objective proof.

4. An assumption to challenge

“If most people are rational, the group's preference must be rational too.”

Condorcet's paradox shows that conclusion doesn't follow from individual consistency alone.

A meaningful counterargument: majority cycles are possible, not inevitable. Real voting groups may have a clear head-to-head winner. Nor is pairwise majority voting the only method. Other decision rules can select outcomes, but they answer somewhat different questions and carry tradeoffs of their own.

The lesson isn't “ignore democracy” or “ignore consumer preferences.” It's to understand what the decision procedure can and cannot establish.

5. Mental operator

When three or more options compete, test the head-to-head comparisons and ask whether changing the order could change the winner.

Then separate preference from performance evidence.

6. Reasoning exercise: Work it out first

Twelve respondents rank three concepts. Four rank A > B > C; three rank B > C > A; five rank C > A > B.

Who wins A versus B? B versus C? C versus A? Is there a Condorcet winner?

Worked answer

A beats B, 9–3: the first and third groups prefer A.

B beats C, 7–5: the first and second groups prefer B.

C beats A, 8–4: the second and third groups prefer C.

There is no Condorcet winner. All twelve people can hold consistent preferences, but majority comparisons still cycle.

Keep this idea: A vote can decide what happens without proving there was one uniquely best option.

Research sources and notes

Educational editorial content. Hypothetical examples are labeled in the original edition.