“Experimentation Works — The Surprising Power of Business Experiments” by Stefan H. Thomke (Harvard Business Review Press, $32).
Get Instant Access to This Article
Subscribe to Hartford Business Journal and get immediate access to all of our subscriber-only content and much more.
- Critical Hartford and Connecticut business news updated daily.
- Immediate access to all subscriber-only content on our website.
- Bi-weekly print or digital editions of our award-winning publication.
- Special bonus issues like the Hartford Book of Lists.
- Exclusive ticket prize draws for our in-person events.
Click here to purchase a paywall bypass link for this article
“Experimentation Works — The Surprising Power of Business Experiments” by Stefan H. Thomke (Harvard Business Review Press, $32).
While business lauds disruptive big ideas, most advancements come from the cumulative effect of incremental improvements in products and services. Example: While Apple has been a disrupter, it kept its “disrupters” fresh by introducing apps and product upgrades, which has sustained profit growth.
Organizations of all sizes thrive when they focus on continuous improvement, which involves constantly doing things differently and doing different things. “Doing” means experimenting. Will every experiment succeed? No. Amazon’s Jeff Bezos stated, “Failure and invention are inseparable twins … organizations embrace the idea of invention but are not willing to suffer the string of failed experiments necessary to get there. Why? They equate failure with “money and time down the drain.”
Thomke believes otherwise; he sees failure as an opportunity to learn more about how a company improves its business. That said, he believes there are productive and unproductive ways to experiment. The productive way, which finds its roots in scientific method, starts with measuring a cause and effect (what happens when … ) hypothesis. The mandate — narrow the focus to one causal variable.
Example: Kohl’s, when it was looking to reduce operating costs, conducted a controlled experiment. It could potentially save money by opening an hour later M-S. It could also lose sales if it did so. A test involving 100 stores showed that opening an hour later had a negligible effect on sales.
Expanding the reduced hours to the rest of the stores had the same result. Had the expansion produced significantly different results, Kohl’s would have had to figure out why it worked in the control group but not in others. More experiments would be needed.
What about unproductive experiments? A poor-quality problem statement results in a shotgun approach that generally drifts toward “more is better.” Adding things only makes a difference if they add value.
