The Variability of Human Behavior — Why Isn’t the Average the Reality?

VZ Lens

Through a VZ lens, this analysis is not content volume - it is operating intelligence for leaders. The article explains why averages are insufficient for understanding reality, and why standard deviation hides the most critical decision points in market research and behavior. Its advantage appears only when converted into concrete operating choices.

There is no such thing as the average consumer. No one fits that description.

---

TL;DR

Traditional market research relies on averages: what is the target group’s average preference, average anxiety level, or average reaction? This is useful, but it paints a distorted picture. The average hides the variance—and it is precisely in that variance that the interesting, real moments of decision-making lie. This article is about why the average isn’t enough, and what we should model instead.

---

The Silence of the Inner Courtyard

I sit on the stone bench, feeling the cold through the damp stone. Raindrops patter quietly on the grass, and the wet stones take on a darker hue. The college walls rise high, as if guarding the centuries-old silence. I can smell the damp earth and decaying foliage in the air. My eyes follow the path of raindrops on a windowpane as they converge and then diverge. Each drop takes a different direction, sliding downward at a different speed. A chaotic movement emerges from the static image. And at this moment, a thought occurs: what if we overlook this very dispersion when we’re looking for the average?

1. The Myth of the Average Person

In the mid-19th century, a Belgian statistician, Adolphe Quetelet, coined the concept of the “average person.” The idea was that the natural state of society is the average—and individual deviations are merely noise that causes unpredictable fluctuations around the average.

This way of thinking has become ingrained in modern measurement culture. In statistics, medicine, economics, and—yes—market research as well.

But there is a fundamental problem: the average person does not exist.

This is not a metaphor. In 1945, the U.S. Air Force conducted a large-scale measurement study aimed at determining the dimensions of the “typical pilot” and designing the aircraft cockpit accordingly. More than four thousand pilots were measured across ten body dimensions. They then examined: how many fell within the 30% range of the average for all dimensions?

Zero. Not a single pilot was “average” in every dimension at the same time.

The average person doesn’t exist. Everyone is non-average in their own way.

---

2. What does research do with averages?

Most market research tools are forced to work with averages—because that’s the basis of statistics. Average opinion levels, average preference strength, average brand loyalty.

This isn’t a mistake. This reduction is necessary if we want to obtain comparable data from large groups.

The problem begins when the average becomes the basis for operational decisions—and the variance is ignored.

Here are a few examples where the average is misleading:

Price point: The average willingness to pay is 4,500 HUF. But if the standard deviation is large—say, half the responses fall between 2,000 and 7,000 HUF—then a price of 4,500 HUF loses both the lowest-spending customers (for whom this is too expensive) and the premium customers (for whom this isn’t premium enough).

Message Effect: The average target group reacts neutrally to one of the packages. But if 40% are strongly positive and 40% are strongly negative (and 20% are neutral), the average is neutral—even though it masks deep polarization.

Decision Speed: The average decision time is 3 days. But some decide immediately (impulsive buyers), and others wait up to 3 weeks (risk-averse). These two groups require completely different communication.

---

3. Standard Deviation as a Research Dimension

If the average isn’t enough, what else is needed?

At least two things:

1. The degree of dispersion. How wide is the distribution? A mean with a narrow dispersion (where everyone is roughly in the same place) means something different than a mean with a wide dispersion (where everyone is in different places, but the mean falls in the middle).

2. The logic behind the activation of outliers. When does someone move toward the high end, and when toward the low end? This is what we call trigger dynamics—and this is what is missing from the classic persona.

[!NOTE] Why is standard deviation important? The mean says: “On average, the target group reacts this way.” Standard deviation says: “How much is lost if communication is optimized for the mean?” In the case of wide standard deviation, mean optimization loses the segments at the extremes.

---

4. Intraindividual vs. Interindividual Variability

Psychology distinguishes between two levels of variability:

Interindividual variability: Differences between different people. Peter is different from Catherine. This is the classic segmentation level.

Intraindividual variability: The same person behaves differently at different times. Katalin on Monday morning is different from Katalin on Friday afternoon. Katalin during a good month is different from Katalin during a stressful quarter.

Traditional market research measures almost exclusively the interindividual level—it compares people with one another. The intraindividual level is almost entirely missing from the toolkit.

This is a problem because consumer decisions are typically made at the intraindividual level—the question isn’t whether Katalin is different from Péter, but how Katalin decides at this moment, in this state.

---

5. The Sigma Consumer

In statistics, sigma is the symbol for standard deviation (σ). The “sigma consumer” metaphor refers to a consumer who is not at the average but somewhere on the fringes.

Sigma consumers are particularly important for three reasons:

1. They make extreme decisions. Brand loyalty, word-of-mouth, boycotts, impulse buying—these thrive on extremes. The average consumer rarely does such things.

2. They reveal the breaking points. The weak points in a product’s communication usually become apparent when someone reacts in an unexpected way. This extreme reaction—not the average.

3. They form the basis of the early adopter and early rejecter segments. The logic of innovation diffusion (Rogers, 2003) relies on extremes. The first 2–3% of adopters and the first 2–3% of rejecters are not the average.

---

6. The Density Illusion

In market research, we almost always simplify: we focus on the large segment because it is more important in terms of market impact. This is a practical approach.

But there is a side effect: the density illusion.

If all research seeks the average and filters out the extremes (treating them as statistical outliers), a false picture emerges: consumers appear more homogeneous than they actually are. The decision-making landscape appears more uniform. Polarization and internal contradictions disappear from the picture.

This is dangerous. Because the real market is not homogeneous. Especially not during stressful periods.

---

7. What does this mean in terms of synthetic personas?

When building synthetic personas, we must steer clear of the “average persona” trap. Three specific consequences:

1. We need multiple personas. A segment cannot be covered by a single persona if the variance is wide. We need at least three: the “typical” (average), the “leaning toward high activation,” and the “leaning toward low activation” variants.

2. The variance must be explicit. The persona should not only store the average but also the activation conditions: what pushes it toward the extremes? What keeps it stable in the middle?

3. The simulation must be situation-dependent. The same persona behaves differently in different situations. In addition to the “average state,” it should be possible to model a stressed state, a state following a surprise, and a state under peer pressure.

State	Decision speed	Risk tolerance	Information need
Stable, relaxed	Normal	Normal	Normal
Stressed, pressed for time	Reduced	Reduced	Increased
Post-surprise	Blocked, delayed	Very low	Maximum
Under group pressure	Accelerated	Increased	Reduced

---

8. The Benefits and Limitations of the Average

The average isn’t bad—it has its limitations. It accurately indicates expected average behavior. It allows for good comparisons between segments and time periods.

But as soon as the question is not “how does the target group generally behave,” but rather “how does it behave under stress, in unexpected situations, or under decision-making pressure,” the average is insufficient.

Standard deviation—and the logic that activates outliers—is the area where market research needs to evolve. Synthetic persona systems aim to cover precisely this area.

---

9. Summary

The average person does not exist. Everyone lives at the extreme end of certain dimensions. Most market research tools are optimized for the average—introducing a bias that becomes most apparent in stressful, high-pressure decision-making situations.

A synthetic persona could be based on the average—and most systems do just that. But a well-functioning system also handles variance, activation logic, and situation-dependent state transitions. This is the difference between a portrait and a model.

---

This article is the sixth part of the Synthetic Personas series. Next part: Stress as a behavioral organizer — how prolonged stress rewrites decision-making patterns.

---

Zoltán Varga | vargazoltan.ai — Market research, artificial intelligence, synthetic thinking

Strategic Synthesis

• Convert the main claim into one concrete 30-day execution commitment.
• Set a lightweight review loop to detect drift early.
• Review results after one cycle and tighten the next decision sequence.