One Company A/B Tested Hybrid Work. Here’s What They Found.

Authors: Nicholas Bloom, James Liang, and Ruobing Han

Amazon’s recent call for employees to return to the office (RTO) five days a week is just one example of high-profile companies pulling back from their remote-work policies. RTO advocates often cite the importance of in-person connections, with former Google CEO Eric Schmidt even claiming in a talk at Stanford University, “Google decided that work-life balance and going home early and working from home was more important than winning.”

Our real-world study, which randomly assigned employees to a three-day or five-day in-office schedule, provided hard evidence on the benefits of hybrid work when it comes to reducing turnover and increasing profits. Indeed, the results were so strong that the company’s middle managers reversed their previously skeptical views on working from home.

Experiment Results

The company we worked with was Trip.com, one of the world’s largest online travel companies, with 40,000 employees. One of us, James, is a cofounder and chairman of the company.

Approximately 1,600 China-based employees in marketing, finance, accounting, and engineering volunteered for the study and were randomized into two groups based on whether their birthdays fell on even or odd dates. One group, the control group, went into the office five days a week for six months. The other group, the treatment group, went to the office only on Mondays, Tuesdays, and Thursdays within the same time frame. The company designed the hybrid work schedule in this way to encourage collaboration.

We analyzed data from the six-month experiment and subsequent performance reviews for the next two years and found the two groups showed no differences in productivity, performance-review grade, or promotion.

Before the experiment, managers estimated that hybrid would reduce productivity by 2.6%. After the six-month experiment, they estimated it increased productivity by 1%. Those working under the hybrid model had a higher satisfaction rate and 35% lower attrition. Quit-rate reductions were largest for female employees. Nonmanagers and those with commutes longer than 1.5 hours also had significantly reduced quit rates under hybrid.

According to the Society for Human Resource Management, each quit costs companies at least 50% of an employee’s annual salary, which for Trip.com would mean $30,000 for each quit. In Trip.com’s experiment, employees liked hybrid so much that their quit rates fell by more than a third—and saved the company millions of dollars a year.

Idea in Brief

• The Problem

Since the pandemic, executives have had an ongoing debate about remote work. Many companies are reconsidering their work-from-home policies, but they have no real data on whether hybrid work reduces turnover or increases profits.

• The Solution

A six-month experiment conducted by Trip.com showed that employees who worked from home three days a week experienced higher satisfaction and lower attrition rates compared with their colleagues who worked from the office. This reduction in turnover saved millions of dollars in recruiting and training costs, thereby increasing profits for the company.

• The Benefits

Business leaders can learn valuable lessons from this study to implement a successful hybrid work model: establishing rigorous performance management systems, coordinating team- or company-level hybrid schedules, and securing support from firm leadership. Additionally, executives should A/B test their own management practices to find what works best for them.

Managerial Lessons

After our study came out in Nature, executives were very interested in finding out more. We think there are three critical ingredients that contributed to the success of hybrid working at Trip.com.

First, Trip.com has a rigorous performance management system that’s on par with best practices around the world. Managers don’t hover over employees at their desks to check their progress or give ad hoc feedback once a year. Instead, the company has an extensive performance-review process every six months to help employees correct course in real time. Employee performance data, as well as feedback from coworkers, clients, direct reports, and managers, are synthesized into a detailed, multidimension performance review on a five-point scale. Through a bell curve appraisal system that ensures a range of grades, Trip.com can effectively recognize and reward the top performers while identifying the bottom ones, who are then placed on performance improvement plans. Moreover, pay and promotions are directly linked to the review. So, the company’s managers can effectively motivate and reward high-performing employees whether they work from home or in the office.

Second, in Trip.com’s approach to hybrid work, employees have a clear, coordinated schedule of when their team will be in the office together. This prevents the frustration of coming into an empty office only to participate in Zoom calls they could have easily done from home. Trip.com enables all employees to work from home on Wednesdays and Fridays. Other businesses might choose to set their in-office days at the team or company level, but clear schedules are critical for successful hybrid policies.

Finally, Trip.com’s CEO and the full executive suite support a hybrid policy. As is common in many modern management practices, from lean manufacturing to organizational agility, having leadership buy-in is critical to support a successful strategy.

“At Trip.com Group, we are dedicated to unlocking the full potential of our employees,” said Jane Sun, CEO of Trip.com Group, the parent company listed on NASDAQ. “Our hybrid work model, refined over a decade of innovation and experimentation, is designed to support both personal and professional excellence. We remain committed to fostering an environment where everyone can thrive.”

The Value of A/B Testing

One other lesson that managers can take from our study is the value of organizational A/B testing. We’ve all been subject to A/B experiments probably dozens of times when using online services, but doing A/B tests with management practices is far rarer.

Trip.com has a tradition of A/B experimentation on management practices going back more than a decade, using these to continuously improve its productivity. For example, back in 2010 it ran experiments on remote working for call-center employees. The current study involved a more diverse group—computer engineers, accounting, marketing, and finance. And with positive results, the leadership took hybrid to the whole company.

Trip.com has been highly data-driven in its decision-making to avoid jumping to incorrect conclusions about the productivity of the hybrid work model. The data showed hybrid employees were working about 1.5 hours less per home-day, superficially suggesting these employees were working less. But a closer examination of the data indicated that hybrid employees put in longer hours on their office days and weekends to make up. Employees shared that they found home-days useful for important activities like a doctor’s appointment, taking their children to school or on trips, or leisure activities like golf. Because these workers were well motivated by rigorous performance evaluations, they made up for this with longer hours on office days and weekends.

Trip.com is also conducting more experiments to collect data in other hybrid-related aspects, such as fine-tuning the number of days in the office to find the optimal balance. The benefits of happier workers and better retention make it all very attractive.

Our results showed that under a hybrid work policy, Trip.com was able to generate millions of dollars of profits by reducing expensive attrition without any impact on performance, innovation, or productivity. Firms should expand their A/B testing from consumer experience to daily practices for continuous managerial improvement. Only companies that continually innovate and improve will survive—and to do that, they need to experiment and refine their own management practices.