The Perils of Polling Part II: Neglecting Non-Customers in Market Studies

Add bookmark

Editor’s Note

The Perils of Polling Part I defined the term statistical universe and warned of the dangers of statistical ignorance in polling results.

The statistical universe consists of all things about which conclusions are to be drawn. If you take a sample from an improperly defined statistical universe, your conclusions will be meaningless – downright wrong!

Part II continues the discussion of the colossal blunders that can be made by improperly defining the statistical universe. We suggest reading Part I (unless readers are familiar with the concept of a statistical universe) before beginning Part II.

RELATED: The Perils of Polling Part I: Can Polling Really Tell Us Who Will Win the 2020 American Presidential Election?


Peter F. Drucker taught us the importance of studying both customers and non-customers. A non-customer can be defined as somebody who should be a customer but is not.

Most companies are focused on existing customers. Endless surveys are conducted on what existing customers buy and how they buy.

But–and this is a very big "but"–non-customers are neglected. Yet non-customers always outnumber customers.

In accordance with Part I of this article, we can classify the neglect of non-customers as an error in defining the statistical universe (in many kinds of market research studies).

Big Data/Predictive Analytics Focuses On Existing Customers: Non-Customers Are Sometimes More Important

Today's new emphasis on big data and predictive analytics focuses on knowing as much as possible about one's existing customers.

Drucker said:

But the first sign of fundamental change rarely appears within one's own organization or among one's own customers. Almost always they show up first among one's non-customers…

In fact, the best example of the importance of the non-customer is U.S. department stores.

At their peak department stores served 30 percent of the U.S. non-food retail market. But they paid no attention to the 70 percent of the market who were not their customers…

They questioned their existing customers constantly, studied them, surveyed them.

But they failed to question non-customers.

They saw no reason why they should. Their theory of the business assumed that most people who could afford to shop in department stores did. Many years ago, that assumption fit reality.

When baby boomers came of age, the game changed.

For the dominant group among baby boomers–women in educated two-income families–it was not money that determined where to shop…

Time was the primary factor, and this generation's women could not afford to spend their time shopping in department stores…

Because department stores looked only at their own customers, they did not recognize this change until [it was almost too late]…

By then, business was already drying up… And it was difficult to get a significant share of this baby boomer market back.

Sony's Colossal E-Reader Market Blunder: A Textbook Case Of A Poorly Defined Statistical Universe

In a March 2015 Harvard business review article, W. Chan Kim and Renée Mauborgne provided this extremely insightful example of an incomplete definition of the statistical universe:

Consider Sony's launch of the Portable Reader System (PRS) in 2006. The company's aim was to unlock a new market space in books by opening the e-reader market to a wide customer base…

To figure out how to realize that goal, it looked to the experience of existing e-reader customers, who were dissatisfied with the size and poor display quality of current products…

Sony's response was a thin, lightweight device with an easy-to-read screen...

Despite the media's praise and happier customers, the PRS lost out to Amazon's Kindle because it failed to attract the mass of non-customers whose main reason for rejecting e-readers was the shortage of worthwhile books, not the display of the devices…

Without a rich choice of titles and an easy way to download them, the non-customers stuck to print books…

Amazon understood this when it launched the Kindle in 2007, offering more than four times the number of e-titles available from the PRS making them easily downloadable over Wi-Fi…

Within six hours of their release, Kindles sold out, as print book customers rapidly became e-reader customers as well…

Though Sony has since exited e-readers, the Kindle grew the industry from around a mere 2 percent of total book buyers in 2008 to 28 percent in 2014. It now offers more than 2.5 million e-titles.

Lessons Learned: Summary & Conclusions

The media praised Sony's PRS and customers were delighted beyond expectations.

But the PSR suffered the ultimate defeat on the corporate battlefield – they failed to convert indifferent buyers (non-customers) into solid paying customers.

Why? They sampled a poorly defined statistical universe. They neglected to drill down on the reasons non-customers initially rejected the e-reader.

Drucker taught us to always ask the question "What is the product/service being bought for?" Remember this question. It could save you from making a colossal blunder!

Had this question been asked, in all likelihood, the issue of non-customers (people who should be customers but were not) would have suggested itself.

Bottom line: Non-customers were not purchasing e-readers because there was a shortage of worthwhile books. Simple? Yes!

But make no mistake: This is a frequently occurring polling error committed by institutions of all kinds and sizes (i.e., government, public service, business, health care, and universities).

Conclusion: Lessons from well-taught statistics courses were forgotten or worse, never learned. This is called the law of slow learning and/or fast forgetting.

By sampling already existing customers of e-books, Sony misdirected their efforts by focusing primarily on improving the user device experience of e-readers.

Stated differently, non-customers wanted a rich choice of titles and a quick and easy way to download them. That was the unfulfilled need that had to be addressed. Sony failed to meet that need.

Amazon asked the right question: Who is the non-customer, the person who does not buy an e-reader, even though he/she is (or might be) in the market? And can we find out why he/she is a non-customer?