The way to avoid turning staff into lab rats in a hybrid work experiment

Thomas Edison’s 14-acre industrial laboratory in New Jersey was an “invention factory”, built on a culture of rigorous experimentation. It is appropriate, then, that the great innovator was honorary chair of the US National Research Council’s unglamorous-sounding Committee on Industrial Lighting in 1924 when it collaborated in one of the most famous, and controversial, series of experiments in industrial history.

The mundane question researchers set out to answer was whether better lighting enhanced productivity at Western Electric’s Hawthorne Works in Chicago. They found that each time they changed the lighting, telephone equipment workers’ performance improved, even after the lights were dimmed back to the original setting. Similar results were noted for other changes, including variations in working hours and breaks.

The idea that merely consulting workers and adjusting their conditions might make them more efficient and productive became known as the “Hawthorne effect” and entered the lore of leadership, human relations and workplace design studies.

Unfortunately, the experiments lacked Edisonian rigour. When Steven Levitt and John List in 2009 uncovered and reanalysed the results of those first Hawthorne experiments, they found that “existing descriptions of supposedly remarkable data patterns [proved] to be entirely fictional”.

Workplace testers are again out in force today. This time, we are all guinea pigs. I am guilty of having described the past 12 months as a great homeworking experiment. In fact it meets none of the criteria for a disciplined lab test. Companies have, though, laid the groundwork for some serious experimentation after blanket-bombing remote-working staff with questionnaires and engagement surveys asking how they feel about returning to the workplace.

Professional services firm PwC’s UK arm last week became the latest employer to lay out plans for a “blended working” model based on a more flexible “empowered day”. Workers will have more freedom to decide start and finish times, for instance.

It makes sense for employers themselves to stay flexible during this period. Even if the pandemic itself is held at bay, nobody really knows what will happen next. The latest Morgan Stanley survey suggests European office workers still want to work from home about two days a week. When McKinsey reopened its London office during a lull between lockdowns last year, more staff returned than the consulting firm had expected. In some Australian cities, staff stayed away from reopened workplaces, even though Covid-19 case numbers were lower than in Europe.

At this point, investing colossal amounts in rebuilding offices, or committing to leaving them altogether, would be rash. As Stefan Thomke points out in his recent book Experimentation Works, “when management aims for big results . . . they cannot rely on lucky guesses, experience, or intuition alone”. Mounting a small series of experiments and assessing the impact they have on staff is the best way to fumble forward through uncertainty.

Thomke says it is important to be narrow and precise about what you aim to measure. Companies may not be able to apply the gold standard of randomised, controlled, double-blind experiments used in, say, vaccine trials. But they can follow a scientific method. He provides a template used by Booking.com, the online travel site: first formulate a theory or belief, then a method of validation (“we will know this when we see these effects happen to these metrics”), and a clear objective (how this will change the business).

The Hawthorne experiments underline some of the pitfalls of testing. In the modern workplace, there are others.

Consent is vital. If you use your staff as lab rats, you should not be surprised if they bite back. The temptation to conduct covert experimentation, monitoring workstations, keyboard strokes and even individuals’ movements, is great. Companies must resist it.

Not all business dilemmas can be resolved by experiments. Employers need to buy into the testing process, though. Simply ignoring the outcome of surveys into staff’s preferences for hybrid working would be hugely counter-productive.

Finally, be ready for unintended consequences. Some may be positive. My (untested) hunch is that a version of the Hawthorne effect partly explains the improved productivity reported by remote-working staff. Some of these workers have basked for the first time in the warm glow of attention from their bosses. If managers can work out how to carry those relationships forward into the murky future of blended working, it really will be a lightbulb moment.

andrew.hill@ft.com

Twitter: @andrewtghill