Loveman’s Control Groups
We often find ourselves revisiting material we were exposed to years ago and this week, when thinking about how to quantify the impact of change at our portfolio companies, we recalled learning about Gary Loveman. The Harvard professor and MIT Economics PhD turned business executive was quoted in a 2011 MIT Technology Review when he was CEO of Caesars Entertainment saying “there are three ways to get fired from the hotel and casino company: theft, sexual harassment, and running an experiment without a control group.” While we added the emphasis, the third criteria for being fired may catch many readers off-guard.
For those unfamiliar, whenever scientists run experimental designs, there are at least two groups of people: the treatment and control groups. For example, if a scientist believes wearing the color blue makes supervisors appear more friendly, the scientist will recruit 60 people for a study. The group of 60 will be randomly assigned to the treatment or control group. If assigned to the treatment group, the supervisors will be forced to wear blue t-shirts. If assigned to the control group, the supervisors would continue to wear the shirts they have always worn. After the experiment concludes, analysis allows the scientist to determine whether blue shirts had a statistically significant impact. Without the control group, where nothing changes, there would be no way of knowing whether differences would be because of the blue shirt or something else entirely.
This is why Loveman, who believed strongly in promoting a quantitative approach to making decisions, expected employees to quickly scale small tests into company-wide initiatives, and likewise saw change initiated without the benchmark of a control group as grounds for termination. In the MIT article, Loveman described the most important thing about the culture he was cultivating:
“We need to overcome hunch and intuition with empirical evidence. We’ve set up a process and a discipline for evaluating our intuitions and improving our understanding of what our customers prefer. We can start with a hunch or strong belief, but we act on it through experiment. We want evidence. We’ve gone from the introduction of experimentation as a technique to a culture of experimentation as a business discipline. We hire people predisposed to do this by temperament and by background. Organizationally, we’re committed—and I’m committed—to making sure we have the discipline to have the decisions we make informed by this evidence.”
Since 2011, this mentality of experimentation has become much more mainstream in certain industries. Often referred to as A/B testing in a business context, the technology industry has relied heavily on this method to gauge the efficacy of its product offerings, which makes sense given the ease with which products can be tweaked and the resulting impact accurately measured. While it seems straight-forward to run two versions of a website and determine which color scheme results in the highest user engagement, it can be much harder (but not impossible) to run two versions of snow shoveling techniques during a blizzard. Chenmark tends to invest in industries where establishing a control group can be difficult, but we still see the value of implementing such practices when feasible: we can, for instance, easily test different methods for the sending and collection of invoices, and can even limit such testing to one portfolio company prior to suggesting a new practice across the board.
In 2015, an LBO-induced debt load of $18 billion combined with weak Vegas tourism led to Loveman’s stepping down as CEO of Caesars amidst bankruptcy proceedings (although he stayed on as the Chairman of the Board). Interestingly, during the workout process, the most valuable asset was the company’s proprietary data, which was estimated to be worth around a billion dollars. To us at Chenmark, the experience at Caesar’s demonstrates that while a culture of experimentation, ideally featuring prominent use of control groups, can create significant economic value, it remains only one piece of the puzzle. Perhaps Loveman would have been better off quipping that “there are four ways to get fired from the hotel and casino company: theft, sexual harassment, running an experiment without a control group, and taking out $18 billion of debt in a cyclical industry.”