Sears battled the challenges of big data…in 1932

Sears battled the challenges of big 1932
Courtesy Library of Congress

In his 1962 book Strategy and Structure, former Harvard business professor Alfred Chandler Jr. chronicled the growth of four iconic companies: DuPont, General Motors, Standard Oil of New Jersey, and Sears, Roebuck. Specifically, he looked at how they redesigned their organizations in the early 20th century to meet emerging strategic challenges.

I have already written several posts about Strategy and Structure. (You can find a list of links to these posts below.) What amazes me is how many of our “modern” business problems aren’t modern at all. So many of today’s problems are variations on yesterday’s problems. The challenge around the clever use of data falls in this category.

Below is an excerpt from Strategy and Structure. Chandler describes the findings in a spring 1932 report from James M. Barker, Sears’ Vice President of Retail Administration:

[In his report,] Barker was stressing the vital importance of statistical data for efficient administration at every level of operations. With Pollock, the Comptroller, Barker continued to improve these statistical tools and informational controls and to instruct the auditors and other officials to show store managers how to analyze them properly. In this work, both men received [Sears President Robert E.] Wood’s strong support, for although [Wood] was primarily concerned with obtaining competent managers, he realized that the supervision of their work required some sort of thorough statistical controls. He later pointed out that:

As a business becomes larger and has branches all over the country, it becomes impossible for the executive to follow it and to control it except by figures. These figures have to be carefully and rapidly compiled, available immediately, and the executive must be able to interpret them properly — read behind the figures.

Those last four words from the president of Sears are key: “read behind the figures”. That was the core data challenge that Sears faced.

Mr. Barker specifically called this challenge out in his report (Strategy and Structure, page 263):

A couple of hours in the office on the reports enable the diagnosis of the Store’s case, usually with clarity and definiteness. The time spent in spot-checking actual conditions in the Store after that was efficiently directed through this information and it often seemed almost uncanny to the Manager that the inspection of actual conditions in the Store confirms the diagnosis from figures so accurately.

Clearly, part of what qualified Mr. Barker for his position as vice president was his competence with data. It is also clear that many of Sears’ managers didn’t share that competence.

The actual data sets that Sears used in the 1930s wouldn’t qualify as “big data” today. Their data sets were much too small. But the challenges were the same…namely, how do you harvest the most business value from your existing data?

Big data is unique because it requires elaborate computational methods to harvest the trapped value. From the perspective of a business leader, though, the computational machinery isn’t the interesting part. The interesting part is knowing what questions the data can help you answer.

Further, not all questions are created equal. Answering some questions will only satisfy your curiosity. That isn’t good enough for a business. To make the data harvesting worth it, you need to be confident that you will make different (read: better) business decisions with the data than you would without the data.

In the case of Sears, the data was crucial. As Mr. Wood wrote, given Sears’ scale, the only way to diagnose all the relevant problems was through studying data. And the executives had real concern that too few managers could deftly analyze the data.

This brings me back to how so many seemingly “new” problems are actually old. Take big data. Yes, the computational angle is relatively new. What’s old, though, is the most important part…the harvesting of value.

The hardest, and most interesting, part is knowing what you might learn in the first place. You have to know the most impactful questions you can ask. Once you have precisely posed questions, you can build up the computational framework to divine the answers.

So much of the big data conversation lingers in the depths of the computational issues. And I understand why. Those issues are interesting from a technical perspective. But they’re only relevant to the extent that the analysis will help you drive meaningful change in your business. That’s the hard part. And that’s the part that Sears executives were exploring deeply…in 1932.

P.S. Below are my other posts inspired by Chandler’s Strategy and Structure. It’s a thoroughly thought-provoking book.


Leave a Reply

Your email address will not be published. Required fields are marked *