3 tips to get the most from your experiments in the workplace
By James Elfer and Dr. Laura M. Giurge
What’s the best way to design hybrid working? How can we build a psychologically safe culture? What should we do to reach our 2030 diversity and inclusion targets? These are the types of big questions leaders often ask about the future of work. But how should we answer them?
As the founder of a behavioral science practice and an organizational scholar focused on the future of work, we are excited to see leaders being more open to the idea of running experiments as the means to that end. But, at the same time, we are worried that leaders are still seeking solutions that can quickly solve everything in the next 12 months or less; a challenge so unrealistic that it leads to overly ambitious campaigns more likely to inspire cynicism than change behavior. Indeed, research tells us that change takes time and careful testing and re-testing. So, what can leaders do to avoid the speed trap and ensure that they create sustainable positive change in their organization?
In this piece, we outline 3 things leaders should do better when designing experiments in the workplace:
1 — Turn your big vision into many small actions.
One of the most common reasons leaders fail to learn from experiments is because they expect them to address big cultural questions all at once. This emerges from the idea that when you change people’s ‘hearts and minds’ (i.e., so they care more about inclusion or they believe in a business case), you will also change their behavior – i.e., they will behave more inclusively across different contexts and over the long-term. However, the last decade has seen billions of dollars spent on this approach with little evidence for its return.
Attitudes and engagement are an important yet often insufficient driver of behavior change. As behavioral scientists, we encourage our partners to break down big cultural themes into a series of everyday behaviors that can be targeted directly. For example, instead of tackling employee motivation (which is vague and can lead to inaction), think about some of the specific actions or processes that matter for employee motivation. It might be that you focus on virtual team meetings, or on one-on-one interactions, or change how a performance evaluation process is carried out.
Whatever challenge your organization faces, breaking it down into smaller pieces will allow you to follow through on the rule of thumb for effective behavior change: the closer your intervention (e.g., workplace initiative) is to the specific behavior you care about, the more likely you are to positively shape that behavior.
2 — Clarify what success means.
When talking to organizations it’s been common to hear their passion for improving outcomes like well-being, diversity, or sustainability. But when we ask them to explain what they mean by “improving”, the room goes silent. Why?
As we discovered, organizations often fail to think about their key KPI or success measure before they run an experiment. When we push our partners on this question, a loose topic like well-being suddenly becomes about decreasing the proportion of employees regularly working above eight hours a day, or that are regularly communicating by email at the weekend, or that score under a certain threshold on a validated belonging scale in their engagement survey.
By thinking smaller, you’ll focus on realistic interventions that help you predict the type of measurable change you’d like to see. That’s the sentence that leads you directly to experimental research where you can ask yourself: does our intervention have an impact on our measure of success?
3 — Include a ‘business as usual’ (control) group
A randomized control trial (RCT) is the gold standard in determining whether and why a particular intervention (e.g., workplace initiative) is effective – meaning it had a causal impact on the outcomes of interest. There are different types of RCTs but the most basic involves randomly assigning employees to either the treatment group (e.g., those who would engage with the new workplace initiative) or the control group (e.g., those who would not yet engage with the new workplace initiative –i.e., the “business as usual” group).
Although you might think this methodology is more suitable for vaccine trials than workplace initiatives, platforms like Netflix, Facebook and Amazon relentlessly experiment with your viewing, socializing, and shopping behavior because it’s the most effective way to learn, iterate, and maximize the outcomes they care about.
Despite this, the requirement for a control group (and the need to randomize people to the control vs. the treatment group) is by far the biggest pushback we receive from our partners. But here’s the thing: without a control group and randomization, you have little way of knowing how successful (or not) your intervention was. Take, for example, a recent experiment where the researchers provided some employees with sales bonuses before knowing if their sales target had been reached (they would have to give this back if not). Decades of research in behavioral economics would suggest that this type of framing (i.e., loss aversion) should result in a positive outcome. Yet, in a nationwide field experiment the researchers found that people randomly assigned to the loss-framed incentives had significantly lower sales than those assigned to the control group. Without knowing how employees who trialed the new work initiative compared with employees who did not, this organization could have lost billions of dollars across years.
As the workplace becomes increasingly digitized and therefore measurable (e.g., through the use of Microsoft Viva Insights), we have the potential to use these three principles to design happier, healthier lives and more productive, inclusive organizations. Like the tech giants, organizations that learn to run robust experiments are more likely to evolve their culture in today’s hyperconnected and diverse world and thrive in the long-run.
This is not a vision for the future but a reality happening today – our experimental trials cover everything from effective remote induction to mental health provision and virtual 1-to-1 meetings, at everywhere from big banks to big pharma to big tech. But we need more organizations to engage with the research community and expedite learning at this critical juncture of redesigning the workplace.
Of course, no experiment in isolation can reveal the answer to ongoing challenges in the future of work such as how to effectively design hybrid work. However, as knowledge via experimentation is accumulating, experiments can illustrate an iterative, adaptive approach that will help us design better organizational cultures. Organizations who aspire to be frontrunners in the future of work will invest in their capability to acquire and apply knowledge. To those pioneers, we say welcome to the world of experimental research!