Research Roundup: CEO Start Dates, Online Reviews, AI’s Impact on Morale, and More

To make the most informed decisions, your company needs to stay up to date on the latest findings from analysts, academia, and beyond. We’ve adapted some of the most timely research covered in the September-October issue of HBR to probe some pressing challenges, including: Why a new CEO might want to start their job in January How good supervisors boost performance, even after they’re gone When responding to customer reviews makes sense—and when it doesn’t How to minimize gen AI’s dampening impact on worker motivation Why, in workplace disuptes, managers tend to take hearsay seriously For several of these write-ups, we’ve also included links to related HBR.org articles to help you dig deeper. The article wraps up with three short “Talking Points” that cover research on CEO worry over AI, the link between grocery carts and credit risk, and middle-manager layoffs. You can use them in your next meeting or share with a team to start a conversation. Why a New CEO Should Start in Early January There are many ways to set new chief executives up for success, but one simple tactic may help: Have them begin during the first week of the year. That was the finding of a study examining 690 successions that took place from 2005 to 2019 at S&P 500 firms. The researchers set out to discover whether aligning start dates with calendar landmarks (in their study, the first 10 days of the calendar year or a firm’s fiscal year) had any impact on the firm’s post-succession performance. They analyzed firm performance over the first three years of each CEO’s tenure, which previous research had found was needed to capture the effects of a new leader. To eliminate potential confounding events that might impact their findings, their sample excluded turnovers caused by unpredictable circumstances like CEO illness or death, corporate restructuring, and mergers and acquisitions. And their analysis showed that performance was stronger for CEOs who took the helm in the first 10 days of the calendar year (or of a firm’s fiscal year) than for CEOs who began at other times. The researchers say that a start date at the beginning of a new year helps align a CEO’s vision with a firm’s natural planning and goal-setting cycles. That can help minimize disruption and decrease the time pressure faced by CEOs and employees working to achieve measurable gains before end-of-year evaluations. Taken together, these factors can help improve firm performance. The positive effect of starting at the beginning of the year was especially strong when new CEOs were under 46 years of age, women, members of racial or ethnic minority groups, or firm outsiders, the researchers found. They say this is because atypical CEOs are more likely to initiate strategic change, which can take time for employees to understand and coalesce around. Go Deeper: “Research: Why CEOs Should Start with the New Year,” by Diego Villalpando et al. Good Supervisors Can Boost Performance Even After They’re Gone Talented managers understand people’s unique skills and match their employees to tasks and positions where their strengths will be put to use best. That ability has a big impact on the success of employees and on the productivity of organizations—and not just in the near term. A new study shows that the positive effects continue long after capable managers move on. To gauge the impact of a good manager, a researcher analyzed data from a large multinational firm and tracked 200,000 of its employees and 30,000 of its managers in more than 100 countries from 2011 to 2021. She identified managers as good if they themselves were promoted quickly within the firm. Her analysis showed that employees who had good managers moved into different jobs inside the company at a much higher rate than other employees did. (Lateral transfers were 30% higher.) These new roles improved both their career trajectories and earnings. Even seven years after good managers had left or transitioned to new teams, the employees who had worked under them earned about 15% more, on average, than colleagues who hadn’t worked for good managers. And even when employees later had less-high-performing leaders, the beneficial effects of a good manager persisted. At the firm level, good managers helped generate higher profits by boosting the individual productivity of workers, the researcher found by examining data on 2,000 sales workers at the company. Moreover, the sales-performance increases inspired by good managers continued to grow for up to four years after they had moved on. “The impacts of a worker’s exposure to a good manager extend far beyond [their time together],” she writes. For firms, gains are the result of “a more productive allocation of workers and occur potentially at zero cost, as they do not require any firing, hiring, or training of workers.” Go deeper: “New Research on How the Best Managers Shape Employees’ Careers,” by Virginia Minni. Responding to Positive Customers—but Not Negative Ones—Can Boost Revenue It’s natural to want to defend your company when customers leave bad reviews, but a new study shows that it might be more profitable to focus on replying to good ones. Researchers analyzed how 4,782 hotels in Texas responded to 804,911 Tripadvisor reviews that were posted from August 2001 to December 2016, looking at the impact those responses had on their revenues. The researchers found that when managers replied to positive reviews, it led to a significant boost in their ratings on Tripadvisor. That in turn allowed them to charge more per room. However, replying to negative reviews resulted in a decline in ratings and revenue per room. Each time the ratio of positive-review replies to negative-review replies rose by 10%, the researchers found, monthly revenue per room increased by $164. However, each time the ratio of negative-review replies to positive-review replies rose 10%, monthly revenue per room fell $36. In subsequent experimental studies with 597 participants in China, the researchers found that a reply to a review amplified awareness of its content, increasing its influence. When there were more replies to positive reviews, their influence grew, and the same was true for negative reviews. Additionally, participants felt more encouraged to write their own reviews when their experiences resembled those that management responded to most. (For example, people who’d had a bad stay were more likely to say they’d leave a review if they saw that management had replied to a larger share of negative reviews.) All of that helps explain why companies saw their ratings drop the more they replied to negative comments. “Our research provides important guidance on how to allocate…resources,” the authors write. “We recommend prioritizing responses to positive reviews due to their significant positive impact on subsequent ratings and financial performance.” Using Gen AI Can Make Workers Feel Disengaged Generative AI can help employees produce higher-quality work with less effort. But new research shows they can later feel bored and unmotivated when they have to do tasks without it. In studies with 3,562 participants from the United Kingdom, researchers tested what happened when humans and gen AI collaborated on various assignments, including writing performance reviews, drafting marketing posts for social media, and crafting a welcome email to a new employee. Participants were sorted into two groups: One did a task with the help of ChatGPT and then worked on a different task unaided. Participants in the second group tackled both tasks alone. After each task, all participants responded to a survey that measured their levels of intrinsic motivation, sense of control, and boredom. The researchers found that the people who used ChatGPT on their first task performed better on it than people who worked solo. For example, performance reviews written with the assistance of gen AI were significantly longer and more analytical and had a more helpful tone than reviews written without it. However, when the gen AI users shifted to the second, unaided task, they experienced a decline in motivation and an increase in boredom. Across the studies their intrinsic motivation dropped by 11%, on average, and their boredom rose by 20%. In contrast, those who performed both tasks without gen AI maintained relatively steady levels of motivation and interest. The studies’ authors think that the first group became more bored and less motivated because collaborating with ChatGPT reduced the effort required of workers while producing strong results. Returning to solo tasks may then have felt not only more laborious but also less stimulating—dampening their motivation and engagement by comparison. To harness AI’s benefits—without causing worker motivation to suffer—the researchers suggest that companies emphasize to employees the unique skills and perspectives they bring to a task and why those insights are crucial to its successful completion. “Clear communication about how AI is assisting—not replacing—their contributions can help workers maintain a sense of ownership and fulfillment,” they write. Go deeper: “Research: Gen AI Makes People More Productive—and Less Motivated,” by Yukun Liu et al. Leaders Find Hearsay More Credible Than Firsthand Reports In a courtroom, when witnesses try to present information they heard secondhand, lawyers will often cry out, “Objection, your honor—hearsay!” But new research shows that in workplace disputes, managers tend to take hearsay seriously. Researchers began by examining how much managers think they value firsthand versus secondhand information. Across six studies they explored what made some reports of misconduct more credible to managers than others. In a survey of 300 U.S. managers and compliance officers, nearly all participants said that firsthand reports were more convincing than secondhand reports, and most managers said they were more likely to act on complaints lodged in firsthand accounts. To test whether this was true in practice, the researchers analyzed some 2 million employee reports of misconduct that had been made from 2004 to 2017, which they obtained from a major provider of workplace reporting systems. The data included whether a report was firsthand or secondhand, the type of incident (for instance, financial misconduct, harassment, or theft), and whether a manager marked the report as substantiated or not. The researchers found that secondhand reports were nearly twice as likely to be substantiated as firsthand reports were—even after controlling for such factors as who was involved in the incident (customers, management itself, or a supervisor), the tenure of the reporter, and the type of incident. In follow-up experimental studies, the researchers uncovered what lay behind this tendency: Managers believed that secondhand reporters were motivated by helpfulness and commitment to the common good rather than by self-interest. They also assumed that secondhand reporters had vetted the information they’d shared, giving it an extra layer of credibility. This research can help correct a disconnect between what managers say (that firsthand reports are more believable) and how they actually act, the study’s authors write. They suggest that organizations remedy it by changing their policies to align with what managers find most valuable when investigating misconduct. This can include encouraging workers to come forward if they have useful information about an incident—even if they weren’t directly involved in or impacted by it. Employees Who Unplug Are Seen as More Productive—but Less Promotable One of the simplest ways to prevent burnout is to set boundaries between work and home and stop answering office emails at night and on weekends. But when it comes time to discuss a promotion, disconnecting from your job after hours may cost you. Across 16 studies with 7,860 U.S. participants, researchers probed how employees who unplugged during non-work hours were perceived by managers. In some experiments actual managers were shown profiles of employees who had identical performance-evaluation statistics but differed in availability. (For example, one employee left an out-of-office reply during a weekend trip while another did not.) Across the studies managers consistently reported that the employee who disconnected from work nights and weekends would be more recharged and productive after returning to work. However, they also penalized that same employee in evaluations, consistently rating that person as less committed and less promotable than the other employee. The penalty happened even when the unplugged employee was a manager’s direct report or was objectively better at the job. It also occurred when neither employee actually did any work after hours and when the reason for disconnecting was noble (for example, caring for a sick family member). The effect was just as strong among managers who said that they valued work-life balance as it was among those who didn’t. Even managers who reported explicitly encouraging employees to disconnect from work penalized employees who did so in these studies. While the researchers found the promotability penalty for those who unplug to be pervasive, they also discovered that it was smaller when managers were made to read a company policy stating that employees should not receive emails over weekends or breaks. This shows the importance of aligning company policies with managerial actions, they argue. “If your company claims to value employee well-being, your promotions and policies must reflect that,” the researchers write. “In the long run, your best employees are not the ones who burn out trying to prove their worth. They are the ones who know when to switch off so they can consistently show up at their best.” Go deeper: “Research: Are You Penalizing Your Best Employees for Unplugging?” by Eva Buechel and Elisa Solinas. CEOs Worry They’ll Be Ousted Over AI In a survey of more than 500 chief executives around the globe, 70% of respondents said they expected a fellow CEO to be fired this year because of a failed AI strategy or AI malfunction. Seventy-four percent believed their job could be at risk if they didn’t deliver gains through AI implementation in the next two years. (“Global AI Confessions Report: CEO Edition,” by Dataiku and the Harris Poll.) Items in People’s Grocery Carts Predict Their Credit Risk Do your customers buy a lot of canned sardines? If so, they may be more likely to default on credit card payments. In a study of how grocery store habits predict credit behavior, researchers found that people who regularly buy cigarettes, tinned fish and meat, and energy drinks are the most likely to default. The greatest predictor of creditworthiness? Buying significant amounts of vinegar-based salad dressing. (“Using Grocery Data for Credit Decisions,” by Jung Youn Lee et al.) Where Are the Middle Managers? Many organizations start cost-cutting initiatives by eliminating layers in the middle, a recent survey of more than 15,000 employees worldwide revealed. In it 41% of respondents said that their company had slashed the number of managers this year—and 37% reported that it had left them feeling directionless. (“Workforce 2025: Power Shifts,” by Korn Ferry.) These research write-ups are adapted from the September-October 2025 issue of our magazine.

Research Roundup: CEO Start Dates, Online Reviews, AI’s Impact on Morale, and More

Summary.

Partner Center