---
title: "Working From Home Holds Your Learning Back"
description: "Output is not the same as potential. Full-time home working weakens passive learning, flattens the development curve, and does younger workers a disservice."
pubDate: 2026-08-01
canonical: "https://sjg.io/writing/working-from-home-holds-your-learning-back/"
---

# Working From Home Holds Your Learning Back

This article began as one section of a much broader piece called [The Rules I Try to Do Business By](/writing/the-rules-i-try-to-do-business-by/).

It was supposed to explain why I spend three days a week in the office with my team, despite appreciating the flexibility of working from home myself. It kept growing. The more I wrote, the more obvious it became that it did not quite fit with the rest of the article. More importantly, I realised I felt far too strongly about the subject to hide it halfway down a much longer page.

So I broke it out and gave it the space it deserves.

My position is straightforward: working from home full time holds your learning back.

There, I said it, and I stand by it.

I have tried working fully remotely. People working for me have tried it. Over the years I have permitted myself and others all sorts of working patterns, from almost entirely remote through to almost entirely office-based.

Nothing I have experienced has beaten the feedback, pace of learning, and strength of relationships you get from a group of people regularly working together in person.

That does not mean people cannot be productive at home. In fact, I absolutely believe that immediate individual output is often higher when somebody works remotely and has a quiet, distraction-free environment.

There is no commute. Fewer people interrupt you. You can settle into a difficult piece of work and stay there for hours. For certain kinds of focused individual activity, home can be an excellent place to work.

I appreciate those benefits myself. I work from home two days a week.

The mistake is treating immediate individual output as the whole measure of whether a working arrangement is effective.

I think that is a profoundly short-term view.

## Output is not the same as potential

It is easy to measure what somebody produced this week. It is much harder to measure the knowledge they did not acquire, the relationships they did not build, and the opportunities they never discovered.

A person can produce a great deal of work while learning very little. They can complete every assigned task, attend every scheduled meeting, and appear perfectly productive while becoming only marginally more capable than they were six months earlier.

Output tells you what somebody can do now. It does not necessarily tell you what they are becoming capable of doing.

Think about how talent scouts assess players in a team sport. They are not only interested in the score from the last match or the player's output at that exact moment. They are looking at movement, awareness, decision-making, temperament, coachability, and potential. They are asking what this person might become in the right environment.

We should think about people at work in much the same way.

Of course current performance matters. Businesses have work to do and customers to serve. But a good organisation should also be increasing the capability of the people inside it.

If somebody joins your company, gives you several years of their working life, and leaves with roughly the same breadth of capability they arrived with, I think you have failed them.

One of my favourite things about running a business and leading teams is developing people. I enjoy discovering what somebody is interested in, giving them room to explore it, and watching that interest gradually become capability, confidence, and sometimes a completely new direction in their career.

I think we owe that to one another.

Work should not be a purely transactional exchange in which somebody completes the tasks attached to their job description and receives a salary in return. We spend too much of our lives working for that to be the full extent of the relationship.

The people around us should help us become more capable, and we should be looking for opportunities to do the same for them.

I do not believe we can do that as effectively when the seed of every conversation first depends on somebody booking a meeting.

## Most workplace learning is not scheduled

A huge amount of learning at work is passive.

A junior employee sitting near a finance team will absorb fragments of how finance works without anybody arranging a training session. They hear people discussing budgets, margins, forecasting, purchase orders, and the difference between revenue and cash.

Much of it may have nothing to do with their immediate role. Over time, those fragments accumulate into a broader understanding of how a business actually operates.

The same thing happens across engineering, sales, product, operations, and leadership.

You overhear a problem being discussed and learn how somebody more experienced approaches it. You notice how a difficult conversation is handled. You see which details senior people care about and which they dismiss. You hear a customer objection, a commercial trade-off, or an operational problem that would never have appeared in your development plan.

None of this is formal training. Much of it is never deliberately taught. It is still a significant part of how people become good.

Some of the most valuable development starts with a conversation nobody would ever have thought to schedule. Somebody mentions a problem, asks a small question, or shows an interest in something happening nearby. That creates an opening for a story, an explanation, a piece of advice, or an invitation to become involved.

Those moments are seeds. Some go nowhere. Others alter careers.

A calendar invitation creates a threshold. The conversation needs a purpose, a title, and enough perceived importance to claim time from two diaries.

Many of the conversations that help people develop never cross that threshold. They are too small, too speculative, or too loosely formed.

In an office, they can begin with somebody leaning across a desk and saying, "Can I ask you something?"

That question might take thirty seconds. It might become a useful five-minute conversation. Somebody else nearby may hear it and add the missing piece. A misunderstanding may be corrected before it becomes a week of wasted work.

Remotely, the same question has to become an event.

You check calendars. You send a message or arrange a call. You wait. By the time the conversation happens, the moment may have passed, or the person may have decided the question was not important enough to ask.

Remote working does not make coaching or career development impossible. A good manager can schedule regular one-to-ones, create mentoring schemes, and make deliberate space for development.

The problem is that deliberate development mostly addresses needs and interests we already know exist.

It is much worse at discovering the curiosity, misunderstandings, and unexpected interests that become visible simply because people spend time together.

## We are doing younger workers a disservice

I think putting younger people into full-time home working does them a profound disservice. We take somebody out of education, place them at a desk in their bedroom or spare room, and then expect them somehow to become competitively capable in the ways of work and business.

We give them a set of scheduled meetings, a stream of messages, and a narrowly defined list of responsibilities. Then we wonder why they do not develop the broader instincts that previous generations acquired through years of being surrounded by working life.

It puts the narrowest possible set of blinkers on them.

They see the work assigned to them. They attend the meetings somebody else has decided they need to attend. They speak mostly to the people directly connected to their role.

What they miss is almost everything happening at the edges.

They do not overhear the sales team discussing why a deal was lost. They do not hear finance challenge the assumptions behind a forecast. They do not watch an experienced manager calm a difficult situation or notice how a senior engineer approaches a problem before touching the technology. They do not build the dozens of weak relationships across a business that gradually become an understanding of how the whole thing really works.

People who were established before COVID can easily underestimate this cost.

We entered remote working with years of accumulated context, confidence, and relationships already behind us. We knew how organisations behaved, who to contact, how to read a room, and which unwritten rules mattered. We could make remote working function because we had built much of the foundation in person.

Somebody entering working life from home does not have that foundation. Asking them to recreate it through Teams calls and Slack messages is not an equivalent experience. We benefited from an environment and then removed that environment for the generation following us.

I do not think that is fair.

## Remote work can flatten development

One of my concerns with fully remote employment is that the development curve can become remarkably flat.

A person joins a company and learns a great deal at first. They have to understand the systems, people, products, and immediate responsibilities of the role. There is an initial burst of development because nearly everything is new. Then the curve begins to level out.

They continue to become more efficient at the work already assigned to them, but they receive much less exposure to adjacent areas. They may become better at that particular job inside that particular company without becoming significantly broader in their trade or more commercially capable. Their world becomes the work directed towards them.

In that situation, moving to another company can become one of their largest development events. Suddenly there is a new set of systems, people, expectations, and problems. Their learning accelerates again because the environment has changed.

That is not necessarily bad. Moving between companies has always been an important source of development.

The problem is when changing employer becomes one of the only reliable ways to achieve meaningful breadth.

When you work in an office full of people, you are passively developing all the time. You encounter ideas, professions, problems, and perspectives beyond your immediate responsibility. You learn your own trade, but you also learn enough about the work around it to become more effective. You begin to understand finance, sales, people management, operations, and customers through exposure rather than a formal course.

Remote work can make a job narrower without anybody consciously choosing to narrow it.

That may not appear in this quarter's productivity figures. It becomes visible later, in the gap between somebody's current output and the potential they might have developed.

## Relationships make work faster

The value of working together is not limited to learning.

Relationships built through regular proximity create shortcuts.

You learn how somebody thinks, what they know, and how they prefer to communicate. You understand their intent more quickly because each conversation rests on months or years of shared context.

A brief exchange can carry a surprising amount of information when the relationship underneath it is strong.

That matters when work becomes difficult or urgent. Teams that know one another well can move quickly without every interaction becoming formal. They can challenge one another more directly because trust already exists. They can resolve ambiguity without writing a miniature contract around every request.

Before COVID, I experienced tightly connected teams that could move at remarkable speed.

People shared context almost continuously. Problems were surfaced early. Knowledge travelled quickly. The distinction between working and learning was often invisible because both were happening at the same time. I have not been able to recreate those dynamics fully since COVID.

Remote tools have improved enormously. Disciplined distributed teams can do excellent work. I work for a global company and rely on those tools every day. Geography gives us access to people and expertise that would otherwise be unavailable.

Even so, every person who cannot regularly work alongside the rest of a team represents a compromise of some form.

The compromise may be entirely justified. The individual may be exceptional. The skill may be scarce. The company may operate across several countries, with no realistic prospect of putting everybody in one place.

It is still a compromise.

We should be honest about that rather than pretending every working arrangement produces the same result.

## This is not what I mean by a hybrid team

The language around hybrid work has become unhelpfully vague.

A company can describe almost any arrangement as hybrid. People may work all over the country, visit an office occasionally, and meet once a week on a video call. A globally distributed team may gather in person every six months. Another team may attend a building on different days and rarely see one another. That is not the model I am describing. I am talking about an in-office team that is allowed to work from home twice a week.

The centre of gravity is the office. The team spends most of the working week together. Home working provides flexibility around that shared rhythm rather than replacing it.

I currently work from the office with my team from Tuesday to Thursday. Monday and Friday are my home-working days.

Those days are coordinated with everyone else. It's not "Pick 3", we are all in on the same days. Some more.

Five people who each attend the office on a different day have not formed an office-based team. They have simply changed the location from which they work alone.

For a while, I encouraged people to spend time in the office. I no longer merely encourage it. For people working locally on my team, I require three days out of five together. I make that requirement because I believe it creates better conditions for learning, development, relationships, and long-term team performance.

It is not a rejection of flexibility. I value flexibility and use it myself.

It is an attempt to keep the benefits of home working without pretending there is no cost when people stop spending meaningful time together.

## Office days do not need to justify themselves

I also reject the argument that every day in the office needs to be carefully programmed to justify the commute.

If a globally distributed team meets in person once every six months, or worse, once a year, then of course you should make the most of it.

Plan the time carefully. Bring the right people together. Have the difficult conversations, build the relationships, and squeeze as much value as possible from a scarce opportunity. That is not the same thing as regularly working together in an office.

When people are together three days a week, every week, the greatest gift is simply that they are there.

The value does not need to be scheduled in advance. It emerges from the ordinary rhythm of people working alongside one another.

Before COVID, we did not create elaborate agendas to prove that an office day had delivered sufficient value. We went to work. The value happened around us.

Somebody overheard a conversation and contributed the missing piece. A junior employee asked a question because the right person happened to be nearby. Two people realised they were solving related problems. A misunderstanding was corrected before it became a week of wasted effort.

Somebody joined a discussion they would never have been invited to because nobody knew in advance that their experience would be useful.

These are not distractions from the value of an office. They are much of the value.

Serendipity requires repeated proximity.

It works because people share space often enough for conversation to be cheap, informal, and unremarkable. Nobody has to decide whether a thought deserves a meeting. Nobody has to identify every relevant participant before the conversation begins.

The connections form themselves.

When office attendance is rare, it becomes an event. Events need agendas, objectives, and visible returns.

Regular attendance is different. It creates an environment, and that environment produces value continuously, including value you could never have planned because you did not know in advance that it was available.

There will be days when I spend much of my time on calls with people elsewhere in the world. That does not make being in the office pointless.

I am still available to the people around me before the call, after it, and in the gaps between. I still overhear, notice, answer, ask, and contribute. So do they.

The test of an office day should not be whether every hour contained an activity that could only have happened in person. That sets an absurd standard we never applied before COVID.

The value is in making shared presence normal again. Once it is normal, the valuable moments stop needing to be engineered.

## Convenience is not the only responsibility

I know this is not the most popular position in every organisation.

The conversation about office working has often become moralised, as though flexibility is inherently enlightened and asking people to work together in person is inherently regressive.

I do not accept that framing.

There are genuine trade-offs.

Home working gives people valuable autonomy. It can improve the balance of their lives and make focused work easier. It can remove a long commute, allow somebody to walk the dog, or make it possible to run an errand during the day.

Those benefits are real.

Working together in person accelerates learning, creates broader context, strengthens relationships, and produces a kind of team cohesion I have not seen technology reproduce.

Those benefits are real too.

My job as a leader is not simply to choose the arrangement people find most convenient at this exact moment. It is to create the conditions in which people, teams, and the organisation can become better over time. That requires thinking beyond immediate output. It requires asking whether people are developing towards their potential, whether knowledge is travelling across the team, and whether younger employees are being given a fair opportunity to learn how work actually works.

Most importantly, we have a responsibility not to offer younger people convenience at the expense of their development. They deserve the chance to build the breadth of knowledge, confidence, and relationships that many of us acquired by being surrounded by working life.

After trying the alternatives, I believe that requires spending a meaningful part of the week together.

Not gathering periodically. Not meeting once a week. Not working remotely and occasionally attending an office event.

Working together, in an office, as the normal way the team operates, with the option to work from home twice a week.

So much of what I understand about how a business actually works, I picked up in rooms I never had to book. A conversation I overheard. A question I asked because the right person happened to be sitting nearby. An argument I watched somebody handle better than I would have.

Nobody planned any of it for me. It was simply there, because I was.

I would like the people starting out now to have the same chance.

## Addendum: What the research says

After re-reading this a few times, I thought to myself, "Yep, I stand by all of this strongly. But… it's all rooted in my own observation." I got actually curious if this is just me or this is what the research shows too, so used my robot friend to help with some research. Imagine starting a research task with Google search in 2026.

Turns out, I am far from alone in this line of thinking. Here are some up-to-date research outcomes supporting the claims I'm making here.

**Immediate individual output really can be higher at home:** [Does Working from Home Work? Evidence from a Chinese Experiment](https://www.nber.org/papers/w18871)

In this nine-month randomised experiment, call-centre employees working from home delivered 13% more output. Nine percentage points came from working more minutes per shift, with fewer breaks and sick days, while four came from handling more calls per minute in a quieter environment. Once people were allowed to choose the environment that suited them best, the gain rose to 22%, although home workers were still less likely to be promoted after accounting for their performance.

This matches my observation. Give somebody clearly measurable individual work and a quiet, distraction-free home environment and their immediate output may rise substantially. That does not mean they are learning, they are helping, or they are becoming capable of doing much more.

**Working together trades some immediate output for development and future quality:** [The Power of Proximity to Coworkers](https://academic.oup.com/qje/article/141/3/1825/8676722)

Before offices closed, engineers sitting alongside their teammates received 23.9% more code-review comments than those whose teams were split between buildings. The advantage largely disappeared when everybody became remote, and the feedback lost was disproportionately useful, actionable, and well reasoned. Experienced engineers produced less code when sitting with colleagues, but younger, and less-tenured engineers received more help and subsequently produced better-quality work, achieving roughly twice the average improvement in code quality. Following a three-day office mandate, co-located engineers were also 1.4 percentage points less likely to introduce bugs and 2.2 points less likely to add files that were subsequently deleted.

**Younger workers have more to lose from remote work:** [How Do Workers Learn? Theory and Evidence on the Roots of Lifecycle Human Capital Accumulation](https://www.nber.org/papers/w35199)

This 2026 working paper uses German and US data to examine how learning changes over a career. It found that learning directly from coworkers matters most earlier in working life, and concluded that remote work disrupts this internal learning, and early-career wage growth. Formal training can offset some of the loss, but not all of it.

**Learning from coworkers is a huge part of learning at work:** [Production and Learning in Teams](https://doi.org/10.3982/ECTA16748)

Published in *Econometrica* in 2024, this research estimated that learning from coworkers accounts for around two-thirds of the human capital people accumulate while working. Less knowledgeable employees tend to catch up when they work with more knowledgeable colleagues, while the more knowledgeable person is not dragged backwards. In other words, putting capable people around developing people really does matter.

**Remote work makes organisations more siloed:** [The Effects of Remote Work on Collaboration Among Information Workers](https://www.nature.com/articles/s41562-021-01196-4)

Researchers analysed emails, messages, meetings, and calls involving 61,182 Microsoft employees. Full-time remote work made collaboration networks more static and siloed, with fewer bridges between different parts of the organisation. Communication also became less synchronous, making it harder for new information to travel across the business.

**Serendipity and weak relationships have measurable value:** [The Effect of Co-location on Human Communication Networks](https://www.nature.com/articles/s43588-022-00296-z)

During the loss of physical co-location at MIT, researchers measured the disappearance of more than 4,800 weak ties over 18 months. These are the loose connections between people in different parts of an organisation through which unfamiliar information and opportunities often travel. Partial co-location regenerated some of them, providing pretty compelling evidence that simply being around one another creates connections that remote tools do not reliably reproduce.

**Video calls are not an equivalent environment for creating ideas:** [Virtual Communication Curbs Creative Idea Generation](https://www.nature.com/articles/s41586-022-04643-y)

Across a laboratory experiment involving 602 people and a field experiment involving 1,490 engineers, pairs working over video generated fewer ideas and fewer creative ideas than pairs working face to face. Interestingly, virtual participants were not worse at selecting which idea to pursue once the ideas existed. The research suggests that different working modes may suit different kinds of work, but that video particularly inhibits the messy, expansive process of generating something new.

**A loosely coordinated hybrid team is not the same as an office-based team:** [Employee Innovation During Office Work, Work From Home and Hybrid Work](https://www.nature.com/articles/s41598-024-67122-6)

This study followed innovation activity across more than 48,000 employees. During full-time home working, people submitted a similar number of ideas but the average quality fell; during hybrid working, the number of ideas also fell. The effect was particularly bad in teams whose members attended the office on different days, supporting the point that five people independently visiting the same building do not constitute an in-office team.

**A genuine three-and-two arrangement can preserve the useful flexibility:** [Hybrid Working From Home Improves Retention Without Damaging Performance](https://www.nature.com/articles/s41586-024-07500-2)

This randomised trial involved 1,612 graduate employees working in engineering, marketing, accounting, and finance. Allowing people to work from home on two coordinated days each week reduced resignations by one-third and improved satisfaction, with no detected damage to performance ratings, promotion rates, or code output over the following two years. That is strikingly close to the model I advocate: an office-based team that works together for three days, with the option to work from home for two.

None of this proves that every person, role, or organisation should operate in exactly the same way. There are jobs that suit remote work brilliantly, exceptional people worth hiring wherever they live, and enormous quality-of-life benefits from flexibility. However… Immediate individual output is not the same thing as long-term capability. Communication is not the same thing as connection. Scheduled training is not the same thing as learning continuously from the people around you. A distributed team that periodically meets is not the same thing as a group of people who normally work together.