```

Does Working From Home Affect Performance? What HR Data Can Actually Tell Us

September 25, 2026

---

INTRODUCTION

Working from home has become a normal part of work for many employees. But there is still one question that HR keeps hearing: "Are employees actually working when they are at home?" Like, if a manager sees an employee sitting at their desk in the office for eight hours, it can feel like they are being productive. But if the same employee is working from home, the manager cannot physically see them. So it becomes very easy to confuse being visible with being productive.

This is where HR data becomes useful. Instead of asking whether employees are online, sitting at their desks, or attending meetings, HR can look at actual work outcomes. Things like completed projects, sales, customer response time, quality scores, performance reviews and other role-specific measures can tell us much more about performance. So the real question is not simply, "Does working from home reduce performance?" It is, "What does the performance data actually show?"

PROBLEM

The biggest problem with measuring remote work is that managers may use visibility as a shortcut for performance. For example, imagine two employees are given the same project. Employee A works from the office from 9 to 6 every day. Employee B works from home three days a week and finishes the project early. If the manager gives Employee A a better performance assessment simply because they were physically present, the company is not really measuring performance. It is measuring visibility.

This can create proximity bias. Proximity bias happens when employees who are physically closer to managers receive an advantage simply because they are seen more often or have more face-to-face interaction. This can affect performance ratings, promotions, important projects and even career opportunities.

There is another problem too. Companies sometimes measure remote employees using activity instead of output. For example, they may look at hours online, number of meetings attended, emails sent or how often someone appears active on a communication platform. But being online for ten hours does not automatically mean someone produced more value than an employee who completed the same work in six hours. So HR needs better measures.

ANALYSIS OF THE PROBLEM

The first concept here is output-based performance measurement. Instead of focusing on where or when someone works, HR looks at what they actually achieve. For a salesperson, this could be revenue, conversion rate or customer retention. For a recruiter, it could include positions closed, time taken to fill roles and candidate quality. For a software engineer, it could include completed development work, product releases or code quality. This does not mean using one metric for everyone. The metric needs to match the job.

The second concept is productivity metrics. Productivity is basically about the relationship between what an employee produces and the resources used to produce it. HR can use different measures depending on the role, such as output per employee, turnaround time, error rate, customer satisfaction or goal completion.

The third concept is presenteeism. This is when an employee is physically present at work but is not necessarily performing at their full capacity. So, simply seeing someone at their desk does not prove that they are productive. This is why office attendance should not automatically be treated as a performance measure.

The fourth concept is proximity bias. This is particularly important in hybrid work because managers may naturally interact more with employees who are physically around them. For example, a manager might remember an employee who regularly speaks up in office meetings but forget the employee working remotely who contributed just as much through written work. So HR needs to check whether employees working remotely are receiving similar opportunities, feedback, performance ratings and promotions.

The fifth concept is causal analysis. This is basically asking whether working from home actually caused a change in performance, instead of simply observing that two things happened at the same time. For example, if remote employees have lower sales, we cannot immediately say, "Working from home caused lower sales." Maybe they were given more difficult customers. Maybe they had less experienced managers. Maybe they were in a different business unit. This is why good HR analytics needs to separate correlation from causation.

REAL-LIFE CASE: TRIP.COM

One of the strongest real-life examples comes from Trip.com, a large travel technology company in China. Researchers from Stanford, Peking University and Trip.com conducted a randomized controlled experiment involving 1,612 employees from the company's engineering, marketing and finance functions. Employees were randomly assigned either to a hybrid schedule where they could work from home two days a week or to a five-day office schedule. The experiment ran for six months, and researchers continued tracking outcomes for two years.

The interesting part is that the researchers did not just ask employees, "Do you feel productive at home?" They used actual performance data. They looked at performance reviews, promotion outcomes, detailed performance measures and lines of code written by software engineers. And what did they find?

There was no significant difference in performance reviews between employees working under the hybrid arrangement and employees working entirely from the office. There was also no overall difference in promotions over the following two years. For the software engineers included in the analysis, there was also no evidence that hybrid work reduced lines of code written.

So, basically, the employees were not sitting in the office five days a week, but the performance data did not show that hybrid work damaged their performance.

There was another interesting finding. Employees working under the hybrid arrangement had higher job satisfaction and their quit rate fell by about one-third compared with the control group. This makes the case more interesting for HR. The data did not simply say, "Working from home is better." It showed something more specific: For this group of Trip.com employees, working from home two days a week did not reduce measured performance, while it was associated with higher satisfaction and lower quitting.

That is a much more useful HR conclusion because it is based on actual evidence.

Source: Nature, "Hybrid working from home improves retention without damaging performance." The study reports a six-month randomized controlled trial involving 1,612 Trip.com employees and two years of follow-up data.

WHAT HR CONCEPTS CAN WE SEE IN THIS CASE?

The first concept is output-based measurement. Trip.com did not use office attendance as its main measure of performance. It looked at performance reviews, promotions and work outputs.

The second is role-specific productivity measurement. The company did not try to measure every employee using the same metric. For software engineers, lines of code were one additional measure. For other employees, performance reviews and other work outcomes were available.

The third is proximity bias. The experiment is useful because employees were not simply compared based on what managers thought about remote work. Their outcomes were compared using a controlled research design.

The fourth is causal analysis. Because employees were randomly assigned to the hybrid or office group, researchers could make a stronger assessment of whether the work arrangement itself affected performance. This is much stronger than simply comparing employees who voluntarily work from home with employees who prefer the office.

The fifth is hybrid-work analytics. This means combining HR information about work arrangements with performance, retention and employee-experience data to understand what is actually happening.

SOLUTION USING HR CONCEPTS

So, how should HR measure performance in a hybrid workplace?

First, companies should create role-specific performance metrics. There should not be one universal productivity formula for every employee. For example, an HR recruiter could be measured using hiring outcomes, time-to-fill and quality-related measures. A customer service employee could be measured through response time, resolution rate and customer satisfaction. A software developer could be assessed through delivery, quality and project outcomes.

Second, HR should reduce its dependence on activity-based measures. Number of emails, hours online or meeting attendance can provide context, but they should not automatically become the definition of productivity.

Third, HR should regularly check for proximity bias. Companies can compare performance ratings, promotion rates, important project allocation and development opportunities between remote, hybrid and office-based employees. If one group consistently receives fewer opportunities, HR should investigate why.

Fourth, HR should use multiple performance measures instead of relying on one number. For example, an employee could have strong output but poor quality, or high customer satisfaction but slower delivery. Looking at several measures gives HR a more complete picture.

Fifth, HR can use before-and-after analysis when a company changes its work policy. For example, if a company moves from five office days to three, HR can compare performance data before and after the change. Even better, companies can run controlled pilots where possible. A smaller group can try the new arrangement while HR measures performance, employee experience, absenteeism and retention.

Finally, HR should separate employee activity data from employee performance data. If an employee is online from 9 a.m. to 7 p.m., that is activity data. If the employee completes their targets, delivers high-quality work and meets customer requirements, that is performance data. The second one is much closer to what HR actually needs to understand.

ACTUAL SOLUTION IMPLEMENTED IN REAL LIFE

Trip.com used an approach that was much closer to this idea of evidence-based workforce management. Instead of deciding the hybrid policy based only on opinions, the company ran a controlled experiment and examined employee performance, promotions, work output, satisfaction and retention.

After the experiment, Trip.com's executive committee decided to extend the hybrid work policy across the company. The decision was influenced by the finding that hybrid work did not reduce measured performance and was associated with lower quitting.

The case also shows why managers' assumptions can be different from actual employee data. Before the experiment, managers in the study expected hybrid work to have a negative effect on productivity. After experiencing the experiment and seeing the results, their views became more positive.

Microsoft research has also highlighted the importance of looking beyond simple activity measures when studying hybrid work. Its research on information workers found that remote work changed collaboration patterns, including a reduction in time spent collaborating across organizational groups. This shows that productivity analysis needs to consider not only individual output but also how work happens across teams.

So, HR should not look at hybrid work through only one number. Performance, collaboration, employee experience and business outcomes can all tell different parts of the story.

Sources: Nature, Trip.com randomized controlled trial; Microsoft Research, "The effects of remote work on collaboration among information workers."

CONCLUSION

Does working from home affect performance? The honest HR answer is: it depends on the work, the employee, the manager, the way performance is measured and the type of remote or hybrid arrangement being used.

The Trip.com experiment gives us a useful example. For the employees studied, working from home two days a week did not produce a measurable drop in performance, while satisfaction increased and quitting decreased. But the bigger lesson is not that every company should copy Trip.com. The bigger lesson is that HR should measure instead of assume.

If a manager says, "My remote employees are less productive," HR should ask, "What data shows that?" If another manager says, "Everyone needs to come to the office because people are more productive there," HR should ask the same thing.

Like, sitting at a desk where your manager can see you may make you more visible, but visibility is not the same thing as performance. So the future of hybrid-work HR is not about choosing home versus office. It is about building better ways to understand work, output, performance and employee experience, wherever the work happens.

```