How do we do remote work across time zones?

How do we work together when our people are spread across multiple time zones? It seems like this is one of the core questions that companies ask when they're getting into remote work.

I was curious about the "best practices" that are out there right now for handling the time zone issue, and how some well-known companies are thinking about it.

It's all about meetings

I found it interesting that almost every discussion on this topic is about how to schedule meetings effectively.

  • Do we have "core hours" in which everyone must be available for meetings?
  • Do we limit which time zones we hire in so that no two time zones are too many hours apart?
  • How do we inconvenience the fewest people when scheduling meetings?

On the GitHub blog, they talk about the importance of "overlap" for scheduling meetings:

We both have team members based all over the world, so typically we try to find times that overlap for various projects. A good rule of thumb is to create teams with four hours of time zone overlap, with the bare minimum being two hours.

The Dropbox team talks about their "non-linear" workdays:

We’re embracing what we call “non-linear workdays.” We’re setting core collaboration hours with overlap between time zones, and encouraging employees to design their own schedules beyond that.

The GitLab folks want to alternate who is most inconvenienced when scheduling meetings:

Synchronous meetings should be inclusive of those who want to attend and are in different time zones. For example, a team's recurring weekly meetings, alternate between a time which is ideal for EMEA and Eastern AMER (8:00AM Pacific) and a time ideal for APAC and Western AMER (3:00PM Pacific).

Why so many meetings?

In the post I wrote on this blog in 2014 about how I work remotely, the very first section was about the importance I placed on effective asynchronous communication.

Remote work requires a paradigm shift. Instead of thinking how can we go on working the same way we always worked without these pesky distances and time zones in between us, the best companies are starting at first principles. A true time zone-agnostic organization is one that, on a basic level, does not rely on meetings for collaboration.

GitLab deserves a lot of credit here, because they have written extensively on the topic of asynchronous communication, and it's a competency at the core of their organization. As the quote I included from them above indicates, they recognize some meetings as prudent, but they operate on an asynchronous-first model, with each meeting requiring a rigorous justification as to why it necessitates an exception to their default asynchronous mode of collaboration.

The easiest way to enter into an asynchronous mindset is to ask this question: "How would I deliver this message, present this work, or move this project forward right now if no one else on my team (or in my company) were awake?"

Instead of thinking about how to schedule meetings effectively when people are in different time zones, it's way more interesting to think about how do we…kinda…not need meetings?

Is Colocation Still Relevant to Agile?

As 2020 has undoubtedly been the watershed year for remote work, I've been thinking a lot lately about colocation and its relevance to Agile.

When the principles of the Agile Manifesto were written in 2001, the authors believed in the importance of colocation:

The most efficient and effective method of conveying information to and within a development team is face-to-face conversation.

I was curious to know what Agile thought-leaders were making of the remote work phenomenon these days, and if the idea of colocation was still imperative to them.

Martin Fowler, one of the original signatories of the Agile Manifesto, makes the point that although face-to-face communication is more effective, a more important concern is to hire the best people you can and figure out how to make their collaboration more efficient later:

The first value of the agile manifesto is “Individuals and Interactions over Process and Tools”, which we should read as encouraging us to prioritize getting the best people we can on the team and helping them work well together. While we recognize face-to-face communication is more effective, that recognition cannot act to override the importance of individuals and interactions.

The fact that you can get a better team by supporting a remote working pattern has become increasingly important during my time in the software business and I expect its importance to keep growing.

It's interesting to think about this topic in the "necessity is the mother of invention" vein. When we all presupposed that a daily commute to the same physical building with your coworkers was the normal thing to do, we had best practices that presupposed that requirement.

In the current climate, we're all presupposing that work will carry on as normal with no one commuting anywhere and all communication being digital. What best practices will emerge in this world? Will the primacy of hiring good people quickly obliterate the former primacy of physical proximity in these conditions, with the world quickly moving on?

¯\_(ツ)_/¯

The Golden Rule of Interviewing

We've all heard the Golden Rule, which goes something like:

Treat others the way you want to be treated.

Well, my Golden Rule of Interviewing goes something like this:

Treat candidates the way you would want to be treated if you were the candidate.

Here are some corollaries to the Golden Rule of Interviewing:

  • Don't ask a question you couldn't answer yourself if you didn't already know it was coming.
  • Don't ask a question you wouldn't want to answer yourself.
  • If you like to ask trivia questions about your favorite subject, think how you'd fare if the candidate asked you trivia questions about their favorite subject.
  • Don't give a candidate homework you wouldn't do yourself.
  • Don't ghost a candidate unless you like being ghosted.
  • Don't require more preparation from the candidate than you're doing yourself.


You're a busy, high-status, well-paid professional. And so is the candidate you seek to hire. Do unto candidates as you would have them do unto you.

Start with a Walking Skeleton

Agile pioneer Alistair Cockburn introduced the concept of the Walking Skeleton:

...a tiny implementation of the system that performs a small end-to-end function. It need not use the final architecture, but it should link together the main architectural components. The architecture and the functionality can then evolve in parallel.

In the book Growing Object-Oriented Software, Guided By Tests the authors elaborate on this concept:

A “walking skeleton” is an implementation of the thinnest possible slice of real functionality that we can automatically build, deploy, and test end-to-end [Cockburn04]. It should include just enough of the automation, the major components, and communication mechanisms to allow us to start working on the first feature.

The Walking Skeleton is usually described as the best way to start a new project, by confirming first that you can get a small amount of functionality from end-to-end—in other words, that your delivery pipeline is working. Since we know that the value of software is realized in production, we establish from the very beginning of the project that we have the means to release something.

I believe that not only is a Walking Skeleton the best way to start a project, but it's also a great way to start a user story. When adding a vertical slice of functionality to the software, I begin by establishing that I can simply communicate anything through all the layers. Once I have a Walking Skeleton built for my user story, I'll then start adding meat to the bones, knowing that I have a mini-pipeline in place.

It's a delivery-focused mindset that can be applied at the micro and macro levels of a software project.

2020: The Watershed Year for Remote Work

As someone who's been advocating for remote work for almost a decade and working fully remote himself for half that time, I never could have guessed that the tide would turn so dramatically within the span of a few months. Due to a once-in-a-century pandemic, we're suddenly inside the greatest experiment in the history of...well...work.

When I was writing my to-this-day most popular blog post ever--But Where Do People Work in This Office?--back in 2015, Facebook was in the midst of building the world's largest open-plan office. Mark Zuckerberg had a dream of "the perfect engineering space: one giant room that fits thousands of people, all close enough to collaborate together."

And in May of this year, Facebook in an about-face announced that it will let many employees work from home permanently. Zuckerberg is quoted as saying, "We’re going to be the most forward-leaning company on remote work at our scale."

Twitter, another open-plan bandwagon company I highlighted in my aforementioned post, also announced in May their employees are now encouraged to work remotely permanently if they please.

The CEO of Shopify said about his company's COVID-initiated transition to permanent remote work, "We choose to jump in the driver’s seat, instead of being passengers to the changes ahead. We cannot go back to the way things were. This isn’t a choice; this is the future."

With all the horror of 2020, I'm hoping we can at least look back on this year as a landmark of the Information Age when the world finally woke up to the liberating force of remote work.

Autonomy: Not Just the Good Parts

It's a common lament from management that development teams don't take ownership of the product they're developing.

This may mean:
  • Caring about the quality of the product
  • Proactively implementing improvements to the product
  • Not waiting to be told what to do
The Agile Manifesto talks about self-organizing teams:
The best architectures, requirements, and designs emerge from self-organizing teams.
It's a beautiful concept: teams that rally around a product, nurturing it like it was their own child. Why is it so hard to find people like that?

Actually, those people are everywhere. They've probably just had their hand slapped or their heart broken too many times.

Let's think: What is it that makes a development team want to take ownership of a product? It's probably because they like creating something in their own image, in their own way.


Who wants to own an ugly product that has been foisted on them by the decisions of people whose command always takes precedence over their own?

This isn't the team's baby anymore, it's yours. No one is going to raise your child.

Perhaps this goes without saying, but micromanagement and autonomy are incompatible. If you want the team to own the product, then they own the details, too.

If you want the team to take ownership, then the product has to look like them, warts and all. You don't just get the good parts.

Does this feel sustainable?

Agile processes promote sustainable development. The sponsors, developers, and users should be able to maintain a constant pace indefinitely.

- The Principles of the Agile Manifesto

I'm kind of obsessed with the idea of sustainability. In fact, you could say it's a major theme of my writing on this very blog.

Below are some unsustainable practices I've covered in the last few years, with links to my writing on them.


We've been collecting paychecks for several months without delivering a valuable feature to a real user.

We complain often about the need for improvement in our processes, but we don't make time for the work or hold anyone accountable.

We're paying highly-trained technical experts to do tedious, manual tasks that a computer could do.

We're pulling our hair out trying to forecast release dates, and working overtime to fit in all the work.

The QA people never have enough time to test all the code the developers are writing by the end of the sprint.

We're rushing to meet deadlines, and we insist on getting every detail right the first time.

We're slowing down the delivery of features to do code reviews of every commit, but the code quality is still low and we have tons of bugs.

Maximize the Amount of Work Not Done

My favorite principle of the Agile Manifesto addresses simplicity, or as they describe it...

the art of maximizing the amount of work not done

This principle is so core to the way I approach my work as a software developer, that I often take it for granted.

"Maximizing the amount of work not done" almost sounds like a definition for laziness, right? No, it's not about laziness--we just don't want to work on things that no one will care about.

Software is such a pain in the ass to describe, develop, test, budget for, maintain, operate in production, etc. I can stomach all these things because I think about the people who will use the software and how it will save them time and make their lives easier.

Cat

It's hard to get motivated to work on things I don't feel confident any real user will value.

In fact, I'd say a thread that is woven throughout the Agile Manifesto is this idea of never drifting too far away from what we know will be valuable to real people.

Let's examine a few more of the principles of the Agile Manifesto...

Principles:

Our highest priority is to satisfy the customer through early and continuous delivery of valuable software.

Deliver working software frequently, from a couple of weeks to a couple of months, with a preference to the shorter timescale.

Translation: At every stage, let's make sure that the customer sees exactly what we're building for them, so that we don't waste time building things they don't like.

Principle:

Welcome changing requirements, even late in development. Agile processes harness change for the customer's competitive advantage.

Translation: We don't want to keep building something the customer doesn't like just because the requirements say so.

Principle:

Business people and developers must work together daily throughout the project.

Translation: Let's make sure the people who will actually use the software we're building actually like what we're building for them and find it useful.

Principle:

Working software is the primary measure of progress.

Translation: The people we're building the software for only know that we've built something valuable for them when they have real software at their fingertips. Until then, they must take it on faith that we're building something valuable.

Maximize the amount of work not done. The customer gets the most bang for their buck, and the team stays motivated. Everyone wins.

Does this feel agile?

This is how Merriam-Webster defines agile:
  1. marked by ready ability to move with quick, easy grace
  2. having a quick, resourceful, and adaptable character

It's common for software development teams to get lost in the artifacts and ceremonies associated with "Agile" methodologies. You know--standups, backlogs, sprints, etc. They can lose sight of the very basic idea of what agility means.

I've always thought of agility as being fundamentally about change--how well you deal with change.

The Agile Manifesto talks about change explicitly:
Responding to change over following a plan
Welcome changing requirements, even late in development. Agile processes harness change for the customer's competitive advantage.

It doesn't matter how well you're following "the process" if you're not responding well to change. If you're not responding well to change, then you're not agile.

How well do we handle changing requirements, for example? Do we grumble and complain, or do we take it in stride?

If we're grumbling, why are we grumbling? Maybe it's one of these reasons:
  • We wrote very detailed user stories or acceptance criteria that we have to change now
  • The QA people have test cases they have to rewrite now
  • We're mid-sprint and we've already committed to a certain scope of work
In the spirit of responding well to change, we could examine those pain points:
  • Are the user stories or acceptance criteria written at the level of value an end user would recognize, or do they include low-level implementation details?
  • Are the QA people focused on testing business value delivered, or following a rote checklist focused on implementation details?
  • Why don't we deliver the business value that we committed to at the beginning of the sprint and capture the new requirements as an item/items for a future sprint (which is at most 2 weeks away)?


Software is soft. If we're working with a process and artifacts that we experience as hardened, ossified concretions, then we might stop and ask: what would it take to get back to our software being soft? What would it take to get to a place where it felt easy to respond to change. What would it take to be agile again?

Keep Code Reviews Focused on the Code

Hopefully we're all doing code reviews, right? Whether it's part of a pull request process, or something else, it's important to get another set of eyes on one's code.

There are a million checklists and blog posts out there on what to look for and how to give constructive feedback.



The thing I don't see discussed enough is how to appropriately scope a code review. When I'm reviewing commits, what should I leave out?

Here are some things I believe are out of scope for code reviews:

- Hypothetical business requirements

"What if the business wants X as well?" Then let them ask for it. That's what the product backlog is for.

- Broad architectural debates

"Should we change the way that we're doing X in general?" Maybe we should. Let's put an item in the backlog to find a new approach. If we adopt a new pattern for similar functionality, we can revisit this code and refactor.

- Fear of change

The perspective that the code under review must represent the last word on a particular topic. If the code under review is high quality, well-tested, and meets a known requirement, then it represents no danger to master. There's no shame in changing or even deleting the code in question with a future pull request if requirements shift.


As with so many aspects of software engineering, setting the appropriate scope is key. For code reviews, stay focused.

"Environmental Issue"

I really like a post I read recently from Paul Osman called Production Oriented Development.

He touches on a lot of topics near and dear to my heart, but the section on non-production environments stood out:

Environments like staging or pre-prod are a lie. When you’re starting, they make a little sense, but as you grow, changes happen more frequently and you experience drift. Also, by definition, your non-prod environments aren’t getting traffic, which makes them fundamentally different. The amount of effort required to maintain non-prod environments grows very quickly. You’ll never prioritize work on non-prod like you will on prod, because customers don’t directly touch non-prod. Eventually, you’ll be scrambling to keep this popsicle sticks and duct tape environment up and running so you can test changes in it, lying to yourself, pretending it bears any resemblance to production.

I've heard the heartbreaking phrase "environmental issue" too many times in my career. You know, like when a developer spends days investigating a bug reported by QA only for the team to decide it was due to a configuration drift between a lower environment and production.

With modern infrastructure-as-code tools like Puppet and Terraform, at least there's a chance of preventing the pitfalls with using non-production environments. Even then, only prod is prod. ¯\_(ツ)_/¯

Process Bikeshedding

Hopefully we’re all familiar with Parkinson’s law of triviality:

Parkinson observed that a committee whose job is to approve plans for a nuclear power plant may spend the majority of its time on relatively unimportant but easy-to-grasp issues, such as what materials to use for the staff bikeshed, while neglecting the design of the power plant itself, which is far more important but also far more difficult to criticize constructively.

One anti-pattern to watch out for in retrospectives is what I would call process bikeshedding.

Topics like these tend to get a disproportionate amount of coverage in retrospectives:

  • The length of sprints (2 weeks or 3 weeks?)
  • On which day to start the sprints (Monday or Wednesday?)
  • On which day to have the sprint planning meeting (first day of this sprint or last day of previous sprint?)

They come up over and over because they’re simple, easy to have an opinion on, and no one’s feelings will be hurt by discussing them. Unfortunately, this also means they’re trivial—i.e., not impactful to the delivery of working software to production (the purpose of a sprint).

So how does a team get beyond the trivial to discuss the real stuff? That’s the tough part. But I believe it comes down to trust and scope.

People need to trust each other in order to discuss deep concerns. There’s no magic to building trust. It just takes time.

In Parkinson’s example of the nuclear power plant, why are people so focused on the bikeshed while ignoring the design of the plant itself? I’d say it's an issue of scope. Luckily on a Scrum team we’re not approving designs for a nuclear power plant (hopefully); we’re just trying to improve the next two weeks on our software project.

In this video on Scrum retrospectives, Jeff Sutherland cuts the scope:

You only want to fix one thing at a time.

What is the one thing, that if we fixed it, would have the biggest impact on making this team go fast? And then commit to fix it.

Certainly the color of the bikeshed is not that one thing.