Three Options, Not Two
Section I · THE FORD CONTRADICTIONS · Henry Ford · Volume I
The Mechanism
Sorensen did not begin with a theory about optimal assembly-line speed. He tested three speeds, measured the output, and let the data choose. The middle speed won, a discovery that contradicted every prior assumption about faster-is-better. The critical insight: two options create a binary that confirms existing conviction. The third option introduces a curve, and the shape of the curve reveals the system's actual behavior rather than the operator's expectations.
The Story
April 1, 1913. Twenty-nine men. A rope dragging flywheels across a factory floor. Sorensen's experiment looked laughable: a rope-and-pulley rig that would embarrass a high school science fair. It produced the most productive factory in human history. Most organizations skip the test because testing implies uncertainty, and uncertainty feels like weakness to people whose careers depend on projecting confidence. Sorensen's genius was treating uncertainty as data.
Application Scenarios
Pricing decisions.
Test three price points, not the two your finance team proposed. The relationship between price and demand is rarely linear, and the inflection point, where the curve bends, is invisible without the third data point. Specifically: your finance team probably proposed "current price" and "10% increase." Add a third option that changes the pricing structure entirely (usage-based, tiered, or bundled). The third option exists not because you expect to adopt it but because it reveals which axis of the pricing curve your customers actually respond to. If the structural change outperforms both incremental options, you have been optimizing within the wrong frame.
Hiring strategy.
When filling a critical role, interview candidates from three distinct talent pools rather than the one pool your VP already favors. If the VP wants ex-consultants, also source from operators and from an adjacent industry entirely. The value is not that the third pool will necessarily produce the hire. The value is that comparing candidates across three pools reveals which qualities actually predict performance versus which qualities your organization has mistaken for predictors because they happen to correlate with familiarity.
Product launches.
Three feature sets, not the one engineering fell in love with six months ago. Build the full version, a stripped version that ships in half the time, and a version that changes the core value proposition entirely. The third version forces a conversation about whether you are solving the right problem or merely solving the familiar one faster.
Critical Warning
When was the last time your organization tested three versions of a critical decision before committing? If you are sitting in a company that spends millions on decisions validated by a single pilot program, you are running on assumption where Sorensen ran on evidence.