Hybrid Work’s Sweet Spot

Illustration showing that a blend of working from home and in the office leads to better-performing and happier employees

Illustration by Nathan Daniels

Does hybrid work harm the bottom line? As organizations across the United States wrestle with post-COVID office policies, researchers at Harvard Business School (HBS) present evidence for an optimally productive blended office-and-home schedule. Their findings may help managers cope with the shift to remote work that is transforming the professional landscape.

Within corporate America, firms including Spotify and Airbnb have allowed employees to go permanently remote, while Netflix and Goldman Sachs have made it clear they prefer their workers in the office more often than not.

But a hybrid model may be the best solution for everyone. Workers who spent between 23 and 40 percent of their week in the office and the remainder working from home were happier, more highly ranked by their managers, and shared more novel information with their colleagues, according to the HBS scholars. “Hybrid is the sweet spot where you get that flexibility without being isolated,” said Lumry Family associate professor of business administration Prithwiraj Choudhury, who led the nine-week study with Lemann professor Tarun Khanna, doctoral student Kyle Schirmann, and Christos A. Makridis, a research professor at Arizona State University’s business school.

The team ran their experiment in the summer of 2020 at BRAC (formerly, Building Resources Across Communities), the world’s largest nongovernmental development organization, which is based in Dhaka, Bangladesh. They studied 108 employees in the company’s human resources department, comparing their work habits during a national lockdown with a nine-week period immediately following the lifting of restrictions when a daily lottery determined whether people worked from home or went to the office. The authors reviewed the data and assigned employees to three categories based on how often they had been selected to work remotely or on site. Workers in the low group went to the office 0-8 days during the period of the experiment, while those in the intermediate group went in 9-14 days, and those in the high group worked on-site more than 15 days.

Analyzing the text of 32,745 interoffice emails sent prior to and during the trial period, as well as the content of more than 30,323 attachments, the researchers found that the workers in the intermediate group sent 0.8 more emails per day. The number of their distinct email recipients also increased 58 percent, countering an oft-heard corporate concern that isolation associated with working from home hampers communication.

Companies often fear that remote workers are “communicating with a smaller subset of people, or the communication network is becoming more siloed,” said Choudhury. “In our study, we looked at how many people each person emailed, or the sparsity of the network, and found that the intermediate workers sent more emails to more numbers of people than the low and high work-from-home groups.”

Many employers also fear remote workers will miss out on key, in-person connections that can spark innovation. But according to the study, workers in the intermediate group excelled. Using machine learning and text-based analysis, researchers examined the contents of both emails and attachments for rare combinations of words and unique information. They found the intermediate group scored the highest in both categories. The novel text generated by the workers was also independently ranked more useful by a senior manager in the organization.

“I think that is the most interesting finding of the paper,” said Choudhury, “that the intermediate group does better in generating novel and useful information.”

Responses from managers and employees in the study further highlight the benefits of the hybrid model. Supervisors at BRAC rated the performance from the intermediate work-from-home category as being of higher quality based on a seven-point scale ranging from “unsatisfactory” to “excellent.” Three months after the experiment, workers in the intermediate category received higher wage increases compared to their peers. Workers in the intermediate category also expressed greater job satisfaction, he said, “reporting both an increase in work-life balance and the least sense of professional isolation, compared to the two other groups.”

But will the findings from the summer of 2020 based on one non-governmental organization in Bangladesh hold up elsewhere? Choudhury, citing earlier studies, thinks the basic principle that working from home doesn’t reduce, and might even boost, productivity still applies. He points to a 2017 report by Stanford researchers that found call center workers in a Chinese travel agency were more productive, and experienced greater job satisfaction, when working remotely. Similarly, Choudhury’s pre-pandemic study of U.S. patent workers found a 4.4 percent increase in productivity when the workers were allowed to work from anywhere. His latest study is the first of its kind, he believes, to measure hybrid schedules and worker performance based on jobs that routinely require collaboration. “If you’re working in a team,” he said, “then you have to balance flexibility and isolation. And I think what we are showing is that some kind of mix is the ideal.”

Above all, Choudhury stresses that this study is a starting point, and he encourages other organizations to use the model to “run their own experiments.” A limitation of this research, he acknowledges, is the absence of an objective measure of productivity. And he notes hybrid work schedules will likely differ depending on the team, the nature of the task, and even the country where the work is being done. He hopes other researchers will also examine the effect of hybrid work in various contexts and its impact on mentoring relationships.

“With our study, we are starting a conversation,” Choudhury said. “We’ve outlined how such an experiment can work, and how to measure outcomes,” he added. “So, if I was Aetna health or Prudential, or any company in the world, I could run my own 10-week experiment now and find what the sweet spot is.”

Read more articles by: Colleen Walsh

You might also like

Reparations as Public Health

A Harvard forum on the racial health gap

Unionizing Harvard Academic Workers

Pay, child care, workplace protections at issue 

Should AI Be Scaled Down?

The case for maximizing AI models’ efficiency—not size

Most popular

AWOL from Academics

Behind students' increasing pull toward extracurriculars

Why Americans Love to Hate Harvard

The president emeritus on elite universities’ academic accomplishments—and a rising tide of antagonism

The Broken Social Contract

Danielle Allen on America’s broken social contract

More to explore

Darker Days

The current disquiets compared to Harvard’s Vietnam-era traumas

Making Space

The natural history of Junko Yamamoto’s art and architecture

Spellbound on Stage

Actor and young adult novelist Aislinn Brophy