One Size Fails All: How Average-Based Decisions Deliver Below-Average Results

We've all been told to be data-driven. To let the numbers guide us. To trust in metrics over instincts.

But what if the numbers we trust most are deceiving us? What if our sophisticated dashboards and careful measurements lead us astray, not because we're measuring wrong but because we're averaging right?

Every business leader staring at their average metrics sees the same comforting illusion. Averages are mathematically sound, perfectly reasonable, and generally wholly misleading.

Designed for No One

In the late 1940s, the US Air Force faced a puzzling problem. Despite using the most advanced technology of the time, its planes were crashing at an alarming rate.

At its worst point, seventeen pilots crashed in a single day.

The brass initially blamed pilot error, training programs, and even the mechanics. None of the explanations made sense, so their attention shifted to the planes themselves. Officials wondered if there was a flaw in the design of the cockpit that would help explain the difficulty pilots had controlling their aircraft.

Initially designed in 1926, the cockpit in use across the fleet was built to conform to the average pilot's dimensions. Everything from the windshield height to the distance between the pedals and sticks was designed with the average pilot in mind.

Perhaps the average pilot had grown since 1926, and the solution was to update the design elements to reflect the new average. Thus, the most extensive study of pilots began, with more than 4,000 pilots evaluated on 140 dimensions. The researchers considered everything from thumb length, crotch height, and the distance from a pilot’s eye to his ear.

One of the researchers assigned to this work had his doubts. Lieutenant Gilbert S. Daniels wondered how many pilots were truly average and made it his mission to find out. When he compared the measurements of the individual pilots to those of the "average pilot,” not one of the 4,000 enlisted pilots fit within the average range across the 10 physical dimensions believed to be most relevant for design, including height, chest circumference, and sleeve length.

One pilot might have a longer-than-average arm length, but a shorter-than-average leg length. Another pilot might have a big chest but small hips. Even more astonishing, Daniels discovered that if you picked out just three of the ten dimensions of size — say, neck circumference, thigh circumference and wrist circumference — less than 3.5 per cent of pilots would be average sized on all three dimensions.1

The average pilot didn’t exist. The cockpit designed for everyone accommodated no one.

This revelation led to the birth of adjustable seats, pedals, and other innovations we take for granted today.

More importantly, it highlighted a fundamental truth: averages are informative but rarely actionable.

From Cockpits to Coffee Shops: The Flaw of Averages

The same fallacy that nearly doomed the Air Force continues to plague modern businesses.

Consider Starbucks, which recently learned this lesson the hard way when it experienced its first quarterly sales decline since 2020. The culprit? A staffing algorithm that fell into the same trap as those 1926 cockpit designers—optimizing for the average at the expense of reality.

While robust, accounting for order forecasts, product availability, and average preparation time, the staffing model failed to adequately account for the ballooning special customer requests. As a result, wait times skyrocketed from an average of less than 5 minutes to closer to 10 minutes, with 10% of customers waiting over 15 minutes!

In hindsight, it seems obvious that broader averages are misleading, but our attention is generally focused on what’s going wrong rather than what’s going on beneath the surface. Consider the cascading effect of managing the average at a call center amid a shift in customer strategy.

• The company’s "average order size" of $100 led them to optimize their website for SMB purchases – while completely missing that 80% of their profit came from the 5% of enterprise customers spending over $500. These valuable customers now struggled to find enterprise-specific features buried in a website designed for small businesses.
• Subsequently, the call center celebrated a reduction in "average handle time." However, it failed to see that this was a function of a shift in customer mix (e.g., more low-end customers with more straightforward problems to solve naturally have shorter call times). It also missed that the more profitable enterprise customers were now on hold longer, which drove up churn rates.
• Despite "average employee satisfaction" of 8/10, the smaller enterprise team dropped to a 3/10 after dealing with never-ending customer complaints from the long wait times.

Continue reading at CFO Frameworks ??

To learn more about why we keep falling for the average trap, when averages are ok to use, and how to leverage segmentation to turbocharge your results, check out the full post at CFOFrameworks.com.