4 traps of categorical thinking – and how to avoid them

Article based on research by Bart de Langhe

Humans are categorisation machines. We constantly take in vast amounts of data from a chaotic and unpredictable world then simplify and structure it. To make sense of our senses, we divide the information they provide us into categories.

Our brains have evolved to apply this kind of heuristic because it speeds up decision making – and it can also be the difference between life and death; for example, the ability to instantly differentiate a snake from a stick in a dark forest would have been useful to our ancestors.

Unfortunately, we’re too good at categorisations. We often see groupings where none exist, or we make other categorisation errors that lead to poor decision making.

Leaders must be aware of their own categorisation instincts

As the data revolution transforms the business world, leaders must be aware of their own categorisation instincts. Esade Associate Professor Bart de Langhe has explored ways to assess – and avoid – the dangers of categorical thinking.

What defines a valuable categorisation?

According to de Langhe and his research partner Philip Fernbach of the Leeds School of Business at the University of Colorado, a valuable categorisation fulfils two criteria: validity and usefulness.

Validity means asking: Is this a real category? One cannot arbitrarily divide a homogenous group – the divisions must be made in a meaningful way. Usefulness means asking: Does creating these categories serve a purpose?

Invalid categories lead to errors. A high-profile example of this is the Myers-Briggs Type Indicator (MBTI), used to inform HR decisions at a reported 80% of Fortune 500 companies. It is easy to see why recruiters are tempted by what the MBTI claims to provide: a way to distil the infinitely variable scope of human personality into four distinct categories and 16 unique types. It seemingly removes the need for intuition, offering a measurable way to match people to the roles that suit them best.

Thinking in terms of categories can cause you to forget about variety

Unfortunately, it is pseudoscience. It compresses complex, variable traits and responses to circumstances into categories that are not just invalid but also useless. “Personality type does not predict outcomes such as job success and satisfaction,” wrote de Langhe in Harvard Business Review.

4 dangers of categorical thinking

De Langhe outlines four common elements of categorical thinking that can lead to bad decision making.

1. Compression

Thinking in terms of categories can cause you to forget about variety and treat members of a category as if they were more alike than they really are. A classic example of this is when marketers carry out segmentation studies to define target customers.

Based on responses to questions about wants, behaviours and demographic details, marketers use algorithms to divide customers into categories. They then typically move on to creating and labelling customer profiles (e.g. "minivan mom") without stopping to ask if the category clusters are valid. This can blind them to the variation that exists; customers within a single segment may behave very differently. The segments that most businesses work with, says de Langhe, are not as clear-cut as they seem.

Stereotyping isn’t just a problem in political and social groups, it can also have consequences for companies

To resist the effects of compression, de Langhe urges companies to remember to calculate the probability of two customers from the same cluster being more similar than customers from two different clusters. The likelihood is frequently lower than intuition would suggest.

2. Amplification

Exaggerating differences across boundaries can also lead to mistakes. Stereotyping isn’t just a problem in political and social groups, it can also have consequences for companies. “Success often hinges on creating interdepartmental synergies,” says de Langhe. “But categorical thinking may cause you to seriously underestimate how well your teams can do cross-silo work together.”

He cites as an example a leader who assumes that data scientists have lots of technical expertise but little understanding of how business works, and who believes that marketing managers have domain knowledge but can’t handle data. As a result, the leader might fail to consider teaming them up – which is one of the reasons why many analytics initiatives fail, according to de Langhe.

3. Discrimination

Favouring one category and failing to pay sufficient attention to the others introduces the risk of overtargeting. If a segmentation study identifies a target customer, advertising spend will typically be partially or fully directed toward them, even if this option is more expensive (in terms of cost per click for Facebook ads, for example).

Advertising Facebook
Favouring one category and failing to pay sufficient attention to the others introduces the risk of overtargeting (Photo: Eddie Espinal/Twenty20)

The evidence, however, does not support this. “Targeting broadly often yields a higher ROI than targeting narrowly. Researchers have found that online ads tend to increase purchase probability by only a small fraction of a percent,” argues de Langhe. 

Tight targeting with limited reach can also exclude potential customers that the segmentation study failed to identify. This leads neatly onto the final element of categorical-thinking failure: 

4. Fossilisation

When we allow categories to frame our worldview, we limit our horizons. When beer brands fail to spot the growing number of women customers, or when bicycle manufacturers perish because they continue blindly marketing their products to children even as millions of adults take up cycling, we see examples of categorical thinking killing innovation.

Thinking only within existing categories can slow down the creation of knowledge

Companies use categories to become more efficient by assigning tasks within disciplinary boundaries. But future business problems don’t fall precisely within the boundaries that were created to help solve past problems. “Thinking only within existing categories can slow down the creation of knowledge,” says de Langhe, “because it interferes with people’s ability to combine elements in new ways.”

4 ways to avoid the traps 

De Langhe suggests four steps to avoid the dangers of categorical thinking:

1. Increase awareness

You can’t avoid thinking categorically – and nor should you want to. To avoid it distorting your decision-making and encouraging false certainty, however, it’s worth remembering the value of complexity, doubt and nuance. Before acting, pause to consider if a categorisation is both valid and useful. 

2. Analyse data continuously

Don’t outsource analytics then misinterpret the information. Metrics exist to evaluate the validity of defined segments. Learn to use them and develop the in-house expertise to do so continuously. 

3. Check your decision-making criteria

If your company uses predetermined criteria as action triggers, it could lead to problems. For example, if consumer evaluations above a certain threshold in a market research exercise are used to trigger a product launch, a tiny and insignificant random variation could result in a significant – and possibly wrong – course of action. Perform company-wide audits of decision-making criteria and look for go/no-go triggers, seeking alternatives such as staged triggers wherever possible.

4. "Defosillise" often

Schedule regular meetings to examine your assumptions. Question your beliefs about your industry, your customers and your company. Are they still relevant? Is anything changing?

De Langhe recommends reflecting on the individual components that comprise existing categories and imagining new functions for them. The innovation that defines the future of your company could be outside an existing category, but right under your nose. 

All written content is licensed under a Creative Commons Attribution 4.0 International license.