A World Without Email: Reimagining Work in an Age of Communication Overload

by Cal Newport

If you’re working on one of these projects, the rhythm of your work is unambiguous. You check in on the task cards assigned to you in Flow, then put your head down to work on these tasks, updating the cards when done.

These examples of effective production processes share the following properties: It’s easy to review who is working on what and how it’s going. Work can unfold without significant amounts of unscheduled communication. There’s a known procedure for updating work assignments as the process progresses.

A good production process, in other words, should minimize both ambiguity about what’s going on and the amount of unscheduled communication required to accomplish this work.

A big part of how he pulls off this feat is immediately visible when you walk into his office. Dominating one of the walls is a three-by-eight-foot chalkboard. It’s divided into five columns: plan, ready, blocked, work, and done. The work column is further divided into two sub-columns: in development and testing.

This is the third time we’ve encountered a similar pattern: information about knowledge work arranged into columns of cards on a board. Alex’s team uses both physical chalkboards and virtual boards implemented by Asana. Optimize Enterprises relies on Flow. Devesh, from the last chapter, uses Trello.

The key caveat in this belief, however, is that we’re able to effectively apply our planning instinct only if we have a good grasp of all the relevant information—what tasks are already being worked on, what needs to be done, where there are bottlenecks, and so on. Cards stacked on boards turn out to be an amazingly effective method for quickly communicating this information.

context. When in doubt, start with the default setup from the Kanban methodology, which includes just three columns: to do, doing, and done.

assigned. But how should these decisions be made? A foundational idea in agile methodology is that short meetings held on a regular schedule are by far the best way to review and update task boards. Agile rejects the idea that you should let these decisions unfold informally in asynchronous conversations on email or instant messenger. When using task boards for your own knowledge work production processes, you should abide by this same rule.

Task Board Practice #4: Use Card Conversations to Replace Hive Mind Chatter

In the knowledge work organizations I observed that used digital task boards, these card conversations proved a critical part of coordinating work on specific tasks.

Devesh, for example, described his shift from email to card conversations as “flipping the script” on communication. When you have a general email inbox through which all discussion flows, you’re forced to continually check this inbox, which then confronts you with discussions about many different projects. When you rely on card conversations, on the other hand, the only way to encounter the discussion surrounding a given project is to navigate to that project’s board.

There’s a slower pace and peacefulness that seems to accompany this shift of discussion toward card conversations. To avoid the need to wrangle an always-filling inbox is a benefit that shouldn’t be underestimated.

The key to this column—and a big part of the secret sauce of Kanban systems in general—is that you should maintain a strict limit on how many tasks you’re allowed to be doing at any given time. In Kanban-speak, this is called the works in progress (WIP) limit.

If you want to unlock the power of personal task boards to minimize hive mind–style back-and-forth messaging, this hack is probably the most important one you’ll encounter in this chapter. A regular rhythm of efficient meetings can replace 90 percent of hive mind messaging, if you have a way to keep track of what needs to be discussed in these meetings. The task board makes this simple.

For example, back when I used to write student advice books, I suggested that students create an individual automatic process for each type of regular assignment: problem sets, reading assignments, lab reports—anything they knew in advance they were going to have to do again and again throughout the semester. At the core of these processes was timing.

Maybe Tuesday from four to six is when you write up your lab reports for BIO 101, and you tackle your statistics problem sets in the free blocks between classes you have on Mondays and Wednesdays from ten thirty to eleven thirty, and so on. I then recommended detailing how this work gets done during these set times, including where on campus you go to work, as well as any methods or materials you regularly use. The key was to reduce cognitive energy wasted on planning or decision making, allowing the student to focus simply on execution. This advice often proved revelatory for the students. Whereas they used to wander through their week, always guilty that they were behind, driven by impending deadlines into all-nighters, they could now confidently execute their automated schedule, secure in their knowledge that what needed to get done would get done, week after week. The reduced overhead and cognitive toll made the same amount of work suddenly seem to require a lot less energy.

Similarly, because the consultant now knows when she’ll work on the report each week, she can have a standing agreement with her boss about when the report will be ready for review. For example: “I’ll always have the report ready for your review in our shared Google Docs directory by 11:00 a.m. on Tuesday; if you have any comments, add them to the document during the day; I’ll check for any notes at 4:00 p.m. before I send off the final version to the client at the end of the day.” A weekly task that might have once generated multiple back-and-forth urgent emails now adds no extra messages to our consultant’s inbox. It also takes up much less cognitive

stick with our consulting firm scenario, the hyperactive hive mind protocol for evaluating new client requests probably generates several dozen back-and-forth emails, with each message corrupting a different five-minute bucket, creating a large overall cognitive cycle cost. The meeting protocol, by contrast, requires only one meeting per week.

Tenner notes that economics textbooks used to introduce the idea of efficient labor markets by telling the story of the best lawyer in town who also happens to be the best typist. The obvious conclusion of the textbook story is that the lawyer would be foolish to not hire a typist. If the lawyer bills $500 an hour and a typist costs $50 an hour, then the lawyer will clearly end up better off outsourcing the typing so she can spend more time on legal work. The arrival of computers in the workplace, it seems, obscured this once obvious reality. We’ve all become the lawyer spending hours at the typewriter.

One of the key ideas from our extreme programming case study was the importance of working on one objective at a time, without interruption, until it’s complete.

Knapp told me that he suspects a more extreme approach, in which communication outside the team is banned altogether throughout the week, could “yield even deeper focus and better results.” But he thought it might be a “hard sell” to persuade a group of modern knowledge workers to sign up for five days of complete disconnection. After a beat, however, he noted that once they “experienced the benefits” of such disconnection, the idea might start seeming less extreme.

Jake Knapp’s design sprint process works great for making high-stakes decisions about future directions for your business, but there are many other areas in knowledge work where sprints could prove effective.

Indeed, in Deep Work, I discuss how Wharton professor Adam Grant used exactly this strategy to become one of the youngest professors to earn tenure in Wharton’s history.

The problem, of course, is that these requests accumulate. If two dozen other units and committees all make these same reasonable requests, suddenly we’re desperately overwhelmed by work that has little to do with our main objectives of research and teaching—a recipe not just for inefficiency, but outright frustration.

“One solution is to directly confront the zero sum trade-off generated by service obligations,” I wrote. “Professors have a fixed amount of time. . . . Instead of ignoring this reality, we should clearly articulate these trade-offs by specifying the exact amount of time a faculty member is expected to devote to service each year.” As I then explained, in this plan, professors would not be allowed to exceed whatever time budget they had agreed on with their department chair for the semester.

The task boards discussed in the process principle chapter also provide a powerful tool for implementing workload budgets. Using a task board to organize work offers two benefits in this context: it makes it easy to determine how much work each person is currently doing, and it has a structured system for how these work assignments are updated, usually in the form of a status meeting attended by everyone. Imagine you’re working on a team that uses task boards. If you’re already tackling a heavy workload, this will be immediately clear on the board—making it much harder for your team leader to overload you, especially if other people have lighter loads. In the situation where overloading you is necessary, the magnitude of what you’re being asked to do is unambiguous, meaning you’ll receive the credit due for your efforts. In a hyperactive hive mind workplace, on the other hand, where these tasks are distributed in an ad hoc manner through emails, you could easily find yourself not only overloaded, but unrecognized for this sacrifice.

The issue at my university, which is common in most large knowledge work organizations, is that each of the support units operates more or less as a standalone entity, focused on trying to accomplish its own internal objectives as efficiently as possible. For the more than twenty-seven units that regularly send me emails, it makes perfect sense to send those messages. They have information they need to spread, and putting it in a bulk email is clearly an efficient way for them to accomplish their goal. These same issues occur when interactions go in the opposite direction. Anyone who works for a large organization is familiar with the pain of struggling with a complex and ambiguous form that a support unit requires you to fill out to request some service. Once again, when we treat each such unit as a mini standalone entity trying to accomplish its individual objectives as effectively as possible, these complex forms make sense—if support staff can get everyone to enter information in a format maximally useful to them, they’ll be able to process it more easily.

Given this reality, a better objective for support units would be the following: to effectively fulfill their administrative duties with as small an impact as possible on the specialists’ main work obligations.

Instead of requiring you to navigate through menus on a computer screen to find some information, or send a message, or play some music, you can ask out loud for what you want, and the appliance will figure out what you need. In the context of a large organization, imagine if instead of wrangling a complex web interface to request a vacation or submit a grant proposal, you could just type into a chat window what you’re trying to do, and someone will stop by your office or call you to get the additional information they need.16

In this situation, as a last-resort measure, I suggest simulating your own support staff.

The office of 2021 is not the office of 1991 plus some extra capabilities; it’s instead a different office altogether—one in which work unfolds as a never-ending, ad hoc, unstructured flow of messages, a workflow I named the hyperactive hive mind.