RTO vs WFH: my default recommendations for remote vs colocated teams

I’ve been disappointed with the tenor of the conversation from many of my peers about so-called “Return to Office” (RTO) policies. So with full understanding that I’m inviting flames, I want to try to articulate a more nuanced way of looking at working from an office versus remote or working from home (WFH).

My core position is:

Neither remote nor office-only are globally “better”: the “correct” model is contextual.
That context is: the individuals’ roles, and the collective work of the team.
Much of the tension comes up when the needs of different roles within the same organization conflict. In particular: many management roles are better suited to colocated teams, and many individual roles are better suited to WFH/remote. Much of the manager/IC conflict around WFH comes down to this difference.
Because this difference is contextual, any large organization that tries to impose a one-size-fits-all policy is going to fail. The only broad policy that can possibly succeed is one that is flexible based on the needs of the teams and the individuals on that team.

In this post, I’ll try to articulate those contextual differences, and provide “default” guidance for which kinds of roles are better suited to shared-office work, and which are more amenable to remote-first or frequent WFH arrangements. I’m using the word “default” here deliberately: I’m not arguing for hard and fast rules, but instead trying to provide a starting point for deciding how to locate teams.

These “defaults” are for intended for mixed organizations – those that have some offices, and some people working remotely, and are trying to decide how to most effectively use both tools. I’m not suggesting that organizations that have chosen to be fully-remote have made a mistake: they’ve chosen different (perfectly valid) priorities. I’m also not speaking to organizations that require everyone work from an office with no exceptions. That’s either a company with incredibly specific work that requires everyone be in person, or an organization that’s pathologically inflexible. The advice that follows is for organizations that have offices and are trying to decide how to use them most effectively.

If you just want those recommendations, feel free to skip to the summary.

Background

Where I’m coming from

Before I get into it, I should establish my bonafides and biases. I’ve been on remote-first teams for over 15 years, the majority of my career. I’ve long been a proponent of remote teams and/or very permissive WFH policies, dating back well before the COVID-19 pandemic accelerated industry moves in that direction. I’ve built and run remote and globally-distributed teams for several employers, and written policy and procedure for remote work for many of them. So I’m far from neutral here: I’ve benefited immensely from the ability to decouple my physical location from my employer’s location, and as you’ll see below I’m an advocate for broad application of remote work/WFH.

Hyberbolic arguments about RTO and WFH

With that background: I’m disappointed by the arguments that some of remote-first peers are making about RTO policies. These arguments, made by people I agree with about remote work in general, make sweeping, hyperbolic generalizations about RTO policies that can’t possibly be true globally. Some arguments I’ve heard include:

“RTO policies are just layoffs in disguise; they’re a way for companies to force people to resign without having to pay severance.”
“It’s a power play: managers want workers back in the office because those managers are petty tyrants and are better able to micromanage or otherwise wield that power in person.”
“It’s a power play: RTO policies are all about Management trying to seize back power from Labor.”
“Managers want workers back in the office because they paid for these offices and are too stupid to understand the sunk cost fallacy”
“There’s literally no good reason for anyone to work from an office any more, anything other than fully-remote is just idiotic.”

These are all arguments I’ve actually seen — not strawmen. I’m not linking to anyone making them because my intent here isn’t to shame; it’s to point to a trend.

I’m sure that all of these statements are true about some situations – for example, I know for a fact of organizations that have used RTO policies as explicit “soft layoff” tactics. But as broad explanations for why all managers might want colocated staff, they’re uncharitable and incorrect readings of that desire. They’re just as a inaccurate as some of the bad-faith arguments that some pro-office proponents make, such as:

“WFH is just code for slacking off. People arguing for WFH want to get paid to sit on the couch and watch Netflix.”
“These entitled millennials are too lazy to get dressed every day and make an appearance, they want us to cater to that laziness.”
“Everyone’s more productive working at an office so WFH is just a waste of money.”
“Nobody wants to work any more.”

Once again: I’m sure that there are specific instances of these statements being true about individual workers – I’ve certainly known of people who used WFH as an excuse to watch Netflix instead of working! – but they’re not accurate explanations of the trend of a growing desire to work from home. Indeed most people who want to work remote want to do so because they know they’ll be more productive, not less.

So all of these arguments, on both sides, are equally ludicrous. The existence of individuals who want to WFH because they want to slack off doesn’t invalidate the myriad good-faith reasons for people to want to work remotely. And the existinance of bad managers who use bad-faith RTO arguments to lampshade layoffs and discrimination likewise doesn’t invalidate the good-faith reasons why a good manager might want colocated teams.

(If you’re sure that you have one of those bad bosses, go unionize.)

The truth is that there’s no single “right” answer. People who claim that WFH is always better are as full of shit as those who claim colocated teams are always better.

How roles map to remote work

Some roles are better suited to remote work than others. This is fairly obvious in some cases: my wife, for example, is a veterinarian; remote work isn’t a possibility for her. It’s pretty hard to sew up a laceration over Zoom. While it’s much more subtle when it comes to tech industry work, it’s still true that different roles have differing levels of effectiveness remotely versus in a shared office.

I’m going to share two ways of looking at these differences: how remote work maps to different kinds of technical work, and how remote work maps to management responsibility. Those differences lead to my “default” recommendations for those roles.

I’m using the word “default” deliberately: these recommendations are meant to be starting points, a place to start, all else being equal. As I’ve said all along: there’s no single one-size-fits-all “answer”. There are all sorts of exceptions for different teams with good reasons to prefer in-person or remote. And there individuals for whom one style of work or another will be a better choice. Forcing those individuals into the wrong kind of work for them is a bad idea –and potentially illegal if their requirement to work from home or an office is a formal disability accommodation!

How remote work maps to different types of technical work

First, how remote work works out for different types of technical roles. For this, I’m going to build off of Graham’s scientist/engineer/mechanic model for technical positions. Briefly, Graham’s model breaks down technical roles into three archetypes:

“Scientists” research novel capabilities.
“Engineers” turn theory into functional infrastructure.
“Mechanics” focus on ensuring that infrastructure produces the desired business outcome.

(Read Graham’s article for a more complete treatment.)

When it comes to remote versus office work:

“Scientists” often benefit from being in the same physical space as their peers. Research often requires high levels of collaboration, and being able to communicate with the highest possible bandwidth – sharing the same physical space – streamlines research. There are certainly notable cases of solo inventors making novel discoveries in their basement, but far more often research breakthroughs are the result of intense collaboration, which is facilitated by having everyone in the same space.
That said, there are also parts of research that benefits from substantial heads-down time with minimal interruptions: reading, writing, iterating on prototypes, etc.
Thus: scientist roles should default to colocated, but have generous no-questions-asked WFH allowances for heads-down reading and thinking and writing.
“Engineers” are similar to “scientists”, but with more of an emphasis on solo work. Building out new systems takes long stretches of uninterrupted work, which is best done wherever the person doing the work is most comfortable. There often is some need for high-bandwidth collaboration, especially at the beginning of a project when the whole team – which often includes a mix of roles – needs to agree on important design decisions.
Thus: engineer roles could default to colocated or remote, but if colocated should again have generous no-questions-asked WFH policies. I have a slight preference for people in these roles being within easy travel distance of a central office – say, no more than 3-5 hours travel time – so that meeting in person occasionally is relatively easy to pull off logistically.
“Mechanics” typically have the least need for high-bandwidth collaboration. Most of their work requires long periods of focus — writing code, debugging, working through a queue of customer requests/bug fixes/feature additions/etc. Most of the synchronous meetings needed in these roles are tactical – “who does what by when”-style – and thus are easily done over voice/video chat.
Thus: mechanic roles should default to remote-first (with obvious exceptions for roles that require manipulating physical devices, like office network administrators, etc.).

I’ve very carefully used the word “default” here in my recommendations here: these are starting points. not strict rules. Once again: there’s no one-size-fits-all policy; what you do has to be tailored to the team. But that said, all else being equal, the above is how I’d structure these kinds of teams and roles.

How management responsibility maps to remote work

Remote-friendliness also breaks down along management lines. Generally speaking, as management responsibility and scope increases, remote effectiveness in that work decreases. This is because the bulk of management work involves communication – meetings, feedback, coaching sessions, strategy, … – and, once again, communication is better with higher bandwidth, and face-to-face communication has very high bandwidth.

This is probably the root for much of the current problems: managers, the ones with the most power to set policy, genuinely benefit from being in the office. If they lack empathy or insight into other kinds of work, they’ll fail to recognize that what works best for them won’t be best for everyone.

To explain this a bit further:

Line managers mostly manage “down”: the bulk of their time is spent working with the people on their team, doing things like project management, feedback, performance management, etc. As with all management work this is typically easier to do in person but it’s not that much more difficult to do remotely. For example, I’ve found giving tough feedback over video chat particularly difficult relative to having those conversations face-to-face – but I’ve been able to learn techniques to have those conversations effectively over vide or the phone.
Much of the communication work that line managers need to do is purely tactical, easily done over text chat or a quick call. A manager with some training or practice in the skills specific to managing remotely can easily be almost as effective as one in the office.
Thus: line management roles should default to whatever’s best for the people they manage, as laid out above. If based out of an office, like many roles, line managers benefit from generous WFH allowances – there’s not a ton of heads-down work, but come performance season they can certainly use the peace and quiet.
Directors/middle managers primarily manage “sideways”: the bulk of their work is intertwined with the other groups that are their peers. That is, a great deal of the work a director does is dependent on other teams: they’re working on long-term plans and strategy (which are intertwined with the goals and strategy of their peer teams), manage budgets (which trade off against money earmarked for other teams), define the SLAs/SLOs for fielding requests from other teams, and so forth.
So the most critical work a middle manager does is the work of communication with their peer middle managers, and people at similar levels throughout the organization. Much of this work resembles negotiation. At healthy companies it’s friendly negotiation, but even friendly negotiation can be tense at times. And at less-healthy companies, the office politics between middle managers can be intense.
Thus: middle managers/Director roles should default to be physically co-located, and should be expected to work out of the office the majority of the time. This can be true even if their organizations are mostly remote: the “manage-down” responsibilities of a middle manager are similar to that of a line manager, so like for those roles they can be performed remotely.
It is workable to have remote Directors — when I ran the security team at Heroku, I was remote. But it’s not easy: for me it required frequent flights to San Francisco, some at short notice, and I leaned pretty heavily on a member of my team who was in the office daily to sort of be my “eyes and ears” for the kinds of context I’d miss not being there in person. Overall, it made my job more difficult than that of other people at my level.
Finally, Executives: their work is hard to define; it doesn’t fit neatly into the “up/down/sideways” metaphor — they do all of those things. To an outside observer, the work of an executive is hard to spot: they’re in a lot of meetings, but they don’t seem to be doing much there, listening more than speaking. As Andy Grove writes in High Output Management:
It’s obvious that your decision-making depends finally on how well you comprehend the facts and issues facing your business. This is why information-gathering is so important in a manager’s life. Other activities—conveying information, making decisions, and being a role model for your subordinates—are all governed by the base of information that you, the manager, have about the tasks, the issues, the needs, and the problems facing your organization. In short, information-gathering is the basis of all other managerial work[.]
You often do things at the office designed to influence events slightly, maybe making a phone call to an associate suggesting that a decision be made in a certain way, or sending a note or a memo that shows how you see a particular situation, or making a comment during an oral presentation. In such instances you may be advocating a preferred course of action, but you are not issuing an instruction or a command. Yet you’re doing something stronger than merely conveying information. Let’s call it “nudging” because through it you nudge an individual or a meeting in the direction you would like. This is an immensely important managerial activity in which we engage all the time, and it should be carefully distinguished from decision-making that results in firm, clear directives. In reality, for every unambiguous decision we make, we probably nudge things a dozen times.
[Emphasis added. Also: Grove writes “manager”, but the majority of his book, including the section I’m quoting, is most applicable to executives. His perspective as CEO sort of warps his thinking about how lower-level managers operate. He’s not wrong, exactly, it’s just that he’s most insightful when speaking to and about executives.]
Executive work is the least remote friendly of all. It’s wall-to-wall meetings, some of them very quick, and it’s often subtle, nudges rather than big decisions. All of this is very, very difficult to do over Zoom.
Thus: executives should default to be physically co-located, and should rarely (if ever) work from home. Unless a company has chosen to be fully-remote, as I mentioned in the intro, making exceptions for individual executives is likely a bad decision organizationally.

Summary

To wrap up, these are my default/starting point recommendations for deciding if teams should be remote or colocated (see above for full explanations and definitions of terms).

For teams of individual contributes, decide based on the type of technical work:

“Scientist” roles (those that research novel capabilities) should default to colocated.
“Engineer” roles (those that turn theory into functional infrastructure) can be either. I have a slight preference towards colocated, but either is a good choice for the role.
“Mechanic” roles (roles that operate existing infrastructure) should default to remote-first (with obvious exception for roles that require hands on physical devices).

All of these roles should come with generous, no-questions-asked work-from-home policies.

For management roles, decide based on the organizational level and scope:

Line management roles (managers of ICs) should default to whatever arrangement is best for the team they manage, as above. If colocated, line managers should, like their teams, have generous WFH allowances.
Directors (managers of managers) should default to physically colocated with their fellow directors, and should be expected to work from the office the majority of the time.
Executives should default to being physically colocated with fellow executives, and should rarely (if ever) work from home.

Questions/comments/hate mail? Toot at me, or send me an email: jacob at this domain.