A panel on distributed work, pt. 2

The following is a continuation of last week’s post, wherein I extract some teachings from my experience in a distributed company such as Automattic.

9-to-5 sponges

Our brains didn’t evolve to operate like machines. Machines are predictable and togglable systems. They are especially suitable for 9–5 operation, whereas our brains are not. That’s not to say they can’t be so, but that’s a function of each person’s propensity and conditioning.

By understanding the rhythms of one’s brain, one can shape one’s routines around these. As a personal anecdote, if I have to work on a deep, focused problem that I am to explore on my own, I may allocate time for it in the evening and work a late night (ca. until 3 or 4 in the morning), clearing time to unwind and rest in the afternoon instead. I typically accomplish more in that intensive session than in a couple of “regular” days. It would be naïf to attribute this merely to the absence of noise and distractions that is expected at night. Indeed, this is a pattern of mental operation that I—and surely many other night owls—have experienced since my childhood.

Note that allocating such an expensive time slot for anything less than a tough problem would have been a waste of energy and motivation. Tasks considered more menial—perhaps pruning a team’s task board—are to be left for e.g. the post-lunch slump.

Attention spans

The previous section alludes to distractions, hinting at the following: managing attention is important. Returning to our mantra, communication is oxygen, note the caveat that there is such a thing as oxygen poisoning. At Automattic, the expression drinking from the firehose illustrates vain attempts to keep up with too many channels of information (P2s, Slack channels, etc.) at once. Automatticians are advised to regularly tweak the streams of information they consume in a way that makes sense for the job they need to perform. Furthermore, there is a fundamental understanding that everyone’s time and attention are scarce resources to respect, which translates to:

  • A stated preference for non-targeted communication (i.e. a question asked in a channel that reaches the right audience without directly mentioning anyone) over targeted communication (mentioning one or more people, or sending private messages).
  • A preference for asynchronous communication (e.g. P2, or a non-urgent nudge on Slack to check out some other async resource) over synchronous (Slack, or denser real-time means such as voice and video).
  • It’s fine to close or mute Slack and disable other sources of notifications for certain periods. Absolutely urgent matters have a way of reaching their audience, anyway.


There is an expectation that being a distributed worker is necessarily very lonely. That might apply to a remote worker, by the definition of that word, but hardly describes my colleagues’ or my days.

It starts with team cohesion. Even across multiple timezones, teammates gravitate towards one another. Simple greetings in the morning, quick one-line summaries of their day, or miscellaneous jokes make up the baseline for daily chatter, on which more substantial discussions happen spontaneously. There’s ample room for nonsense, in my experience essential for rapport. Teams usually have at least one weekly realtime, preferably video-borne meeting; in these, although work is the primary object, we celebrate face time for the sake of face time.

In the broader human network of the company, nooks and crannies abound where Automatticians find forum for all sorts of topics: we call these watercooler P2s or watercooler Slack channels, a nod to their meatspace counterpart. Topics include general interests—fitness, cats, Linux—but notably include more difficult ones, e.g. mental illness and depression, where supportiveness is the word. For a little extra exposure, when joining the company, Automatticians are invited to present themselves in an intro video; Pedro from GitLab shared that their intros happen in realtime at their weekly company hangouts. Our company-wide happenings are monthly and known as town halls.

Outside of the Automattic ecosystem, most will still find their peers, either on the social Web—blogs, Twitter, et al.—or IRL, getting together for coffee or co-working when one is visiting another’s town, even if their paths don’t usually cross in work.

To be taken from this is the striking contrast between the autonomous, hands-off operation of each distributed worker and their readiness to interact for the sake of socializing. Hence my answer to the question, “Don’t you get lonely?”: no, not at all.

“There’s an async way for that”

It’s natural to find apparent roadblocks and think: “this requires a meeting”. However, synchronous time is expensive, harder to schedule, sometimes exhausting, and more opaque to those not present. So we do our best to remain asynchronous.

One way is to honor our tradition of the P2. P2 or it didn’t happen reminds us that P2s are the ultimate place for accessible, traceable, archivable discussions and decisions. Just say no to the urge to grab someone for a realtime chat and instead articulate your thoughts or questions on a P2 thread. If possible, anticipate a couple of answers and scenarios and reply to those too—as in chess, anticipation is advantage. For me, I balance this against my tendency to be over-verbose—as can be gathered from my blog posts—since excessive verbosity has an impact on mine and others’ efficiency.

If P2 doesn’t satisfy the requirements, one should consider richer but still asynchronous methods: screencasts or other forms of video, collaborative platforms such as whiteboards, pads (e.g., Hackpad), documents (Google Docs), Post-It boards (Stickies) and make sure to reference these in the appropriate P2(s). Incidentally, some of the above tools are also very fitting for those times when synchronicity is indeed chosen.


Perhaps antithetically, travel is a cornerstone of the company culture. Somehow, despite our trust in distributed processes, we regard meetups and other gatherings as essential. Historically, team meetups have held intensive weeklong meetup projects; nowadays, there is much more flexibility, and more time is allocated for focused discussions, evaluations, planning, projects. The only constant, really, is that teammates spend some fourteen hours a day, for a week, immersed in a very fertile setting for discussions, epiphanies, and the tightening of bonds that surpass work and constitute lasting friendships. The result is a job done better, but also an indescribable deep breath that instills renewed motivation, urgency and a sense of possibility. The annual company meetup—the Grand Meetup—is the apex of meetups, doubly taxing and doubly rewarding.

Final words

At an organizational scale, distributed work is novel and challenges a lot of established conventions of what work is, how teams should (self-)manage, and how space and time can be compressed. At the individual level, it demands much more self-awareness if one is to adapt to it successfully.

As with the actual work to be done, going distributed is a challenge that, hopefully, is never truly “done”.

Author: Miguel Fonseca

Engineer at Automattic. Linguist, cyclist, Lindy Hopper, tree climber, and headbanger.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s