By Cal Newport
In the early 1980s, IBM decided to deploy an internal email system. In typical careful IBM fashion, they began by measuring employee communication, so they could estimate how many messages would be sent on the new system. Based on this research, IBM provisioned a $10 million mainframe to run their email server — an amount of processing power that should have easily handled the typical volume of intra-office interaction.
Within a week, the machine was overwhelmed.
As an engineer who worked on this project recently recalled, the email team had gravely underestimated the load. Instead of employees simply transferring their normal offline communication to the more convenient online system, they began to communicate vastly more than they ever had before. “Thus — in a mere week or so — was gained and blown the potential productivity gain of email,” he lamented.
This story highlights a common misunderstanding about our current tempestuous relationship with email. Most knowledge workers believe that email is a passive tool they choose to use to make their real work easier. But as the Big Blue engineers discovered three decades ago, this technology is not passive; it instead actively changes what we mean by “real work.”
Accompanying the rise of this technology was a new, unstructured workflow in which all tasks — be it a small request from HR or collaboration on a key strategy — are now handled in the same manner: you dive in and start sending quick messages which arrive in a single undifferentiated inbox at their recipients. These tasks unfold in an ad hoc manner with informal messages sent back and forth on demand as needed to push things forward.
This unstructured workflow arose from the core properties of email technology — namely, the standard practice of associating addresses with individuals (and not, say, teams, or request type, or project), and the low marginal cost of sending a message. It spread for the simple reason that it’s easier in the moment. It takes significantly less effort to shoot off quick messages, for example, than it is to more carefully plan your work day, figuring out in advance what you need, from whom, and by when.
But just because this unstructured approach is standard and easy doesn’t mean it’s smart. It’s important to remember that no blue ribbon committee or brilliant executive ever sat down and decided that this workflow would make businesses more productive or employees more satisfied. It instead just emerged as an instinctual reaction to a disruptive new technology. Like the employees at 1980s IBM, one day we looked up and noticed that what we meant by “real work” had shifted radically under our feet.
The high cost of cheap messages
Given that no one planned the rise of the unstructured approach, it shouldn’t offend anyone when I claim that it’s been a disastrous development for the knowledge work sector. A consequence of this workflow is that an organization’s tasks become entangled in a complicated network of dependencies with inbox-enslaved individuals sited at each node. The only way to keep productive energy flowing through this network is for everyone to continually check, send, and reply to the multitude of messages flowing past—all in an attempt to drive tasks, in an ad hoc manner, toward completion. If you step away from your human network router duties, the whole apparatus can grind into deadlock. This reality forces modern knowledge workers to constantly check their inbox and feel great guilt or unease about the possibility of unanswered communication awaiting attention. This compulsion is not irrational, as these unrelenting messages are not supplemental to real work — they’re instead at the core of what we now mean by this term.
The negative impacts of this lifestyle are so widely felt that they hardly need elaboration. But for the sake of this argument, I’ll briefly note what I believe to be the two biggest harms.
First, this incessant communication fragments attention, leaving only small stretches left in which to attempt to think deeply, apply your skills at a high level, or otherwise perform well the core activity of knowledge work: extracting value from information. To make matters worse, cognitive performance during these stretches is further reduced by the “attention residue” left from the frequent context switching required to “just check” if something important arrived.
These behaviors are not just annoying; they have a substantial impact on productivity. I recently wrote a book called Deep Work, which details the immense professional benefits experienced when you allow people to spend long periods, without distraction, focused on cognitively demanding tasks. To eliminate the ability for knowledge workers to perform deep work is like putting assembly line workers in thick gloves that hamper their ability to manipulate their tools — it’s an absurd self-imposed handicap.
The second harm is more personal. As more knowledge workers now acknowledge, the inbox-bound lifestyle created by an unstructured workflow is exhausting and anxiety-provoking. Humans are not wired to exist in a constant state of divided attention, and we need the ability to gain distance from work to reflect and recharge. Put simply, this workflow, which can transform even the highest skilled knowledge workers into message-passing automatons, is making an entire sector of our economy miserable.
The syllogism here is inescapable, leading us to the conclusion that there’s great advantage for those organizations willing to end the reign of the unstructured workflow and replace it with something designed from scratch with the specific goal of maximizing value production and employee satisfaction.
Given the tangled relationship between email and our current approach to work, however, it’s also clear that this transformation is almost certainly going to require a radical first step: to eliminate email.
Tame efforts to curb the worst impacts of this technology — be it email-free Fridays or smarter inbox applications — are doomed to failure. Once you assign each employee a universally accessible address of the form email@example.com, an unstructured workflow will follow, and this workflow, by its very nature, demands the excesses that plague the knowledge economy. These problems cannot be tamed with better etiquette. The email weed, in other words, must be pulled out by the root.
Replacing the chaos of email with a structured workflow
The natural follow-up question, of course, is what qualifies as a “better” workflow. Even the most strident email opponents recognize that we need some way to coordinate and communicate with colleagues. To validate the idea that organizationscan thrive without this tool, let me offer a concrete alternative inspired by my own experience in academia: office hours.
The concept is simple. Employees no longer have personalized email addresses. Instead, each individual posts a schedule of two or three stretches of time during the day when he or she will be available for communication. During these office hours, the individual guarantees to be reachable in person, by phone, and by instant messenger technologies like Slack. Outside of someone’s stated office hours, however, you cannot command their attention. If you need them, you have to keep track of what you need until they’re next available.
On the flipside, when you’re between your own scheduled office hours, you have no inboxes to check or messages demanding response. You’re left, in other words, to simply work. And of course, when you’re home in the evening or on vacation, the fact that there’s no inbox slowly filling up with urgent obligations allows a degree of rest and recharge that’s all but lost from the lives of most knowledge workers today.
Notice that the workflow induced by an office-hours scheme replaces on-demand messaging with structured communication. People now know exactly when someone might need their attention and exactly when they can command the attention of others. This freedom from a constant background hum of interaction will increase the intensity of concentration achievable when people need to work deeply, and the efficiency with which shallower tasks can be batched together and dispatched.
This workflow also replaces asynchronous interactions with synchronous conversation. This change is crucial. Synchronous conversation is efficient and nuanced: not only does it allow you to handle in three minutes decisions that might have otherwise taken three days of attention-snagging messages, but it tends to also produce more thoughtful conclusions. Imagine, for example, that Alice and Bob need to work together to write a report. If they use email, the process would likely unfold in an inefficient manner, as both Alice and Bob, under the Sisyphean pressure of an ever-filling inbox, keep dashing off quick responses to each other so as to temporarily clear the issue out of their psychic space. In the office-hours scenario, however, Alice and Bob would be forced to talk in real time about the report project. This interaction, though taking more time than sending a quick message at first, is more likely to lead to a complete and coherent plan for how the work should best unfold in the days that follow.
Answers to common objections
There are, of course, issues with replacing email with office hours. Consider, for example, client communication. I accept that this is an area that an organization might need to leave untouched. It’s perfectly reasonable, in other words, to keep this office-hours strategy confined to internal communication, allowing your interactions with clients to still meet their expectation for your availability. (Though it should be noted that when it comes to external communication, many issues related to the unstructured workflow are already solved: there exist many popular client management systems that provide significant structure to such interactions.)
Another issue is team communication. An advantage of email is that it allows you to communicate with multiple people at once. It would be a burden to have to attend multiple office hours to spread the same message to all members of a team. A solution to this issue is to synchronize office hour slots within teams—creating periods every day where you know you can talk to a whole team at once, using a Slack chatroom or conference call.
There’s also the issue of transferring files, which many now accomplish using email messages. Fortunately, there are no shortage of shared-folder technologies, such as Dropbox or Google Drive, that make it simple to pass files between different users.
Perhaps the biggest concern generated by this proposal is the fear that there are some situations that really do seem to require the asynchrony provided by email. I want to emphasize, however, that office hours do not eliminate asynchrony — they just shift the responsibility such interactions generate. In an email-driven organization, for example, if I have some feedback to give you on a report draft, I would simply send you these notes when I was done compiling them. This action places the responsibility for keeping track of the information falls to the receiver. In an office-hours organization, by contrast, I would instead hold onto these comments until your next convenient office hours, at which point I could bring them to your attention in real time. The responsibility for keeping track of this information now falls to the sender — but the asynchronous nature of the interaction remains.
There will, of course, be some circumstances where the urgency of an issue dictates that you cannot wait until office hours to interact with someone. In such cases, however, the best solution is an old one: call. In other words, I would suspect that an organization using this strategy would have a policy that you can and should call someone’s office or cell phone if there’s a truly urgent matter. I would conjecture that such emergencies would be much rarer than most might predict.
More generally speaking, when I’ve floated this idea in business circles, many of the complaints that are presented as reasons why this will not work for me turn out to bereasons why this would make certain situations harder for me. There’s a key difference here. The goal for most organizations is not to make work as easy as possible; it is, instead, to organize work in a way that allows it to be effective, productive, and satisfying. The unstructured workflow that currently dominates satisfies the former, while solutions such as office hours satisfy the latter.
Office hours might not work for every organization — although, as I’ve argued, they would probably apply in more settings than you might at first assume. The broader goal for this discussion is to illuminate the true depth of the problems generated by email, and to underscore the feasibility of radical solutions.
Email, as a technology, is not intrinsically bad. But the unstructured workflow it engenders is disastrous. We need to fix it—and I’m doubtful this can be accomplished while email still plays a core role in our business culture. It’s this reality that brings me back to the modest proposal that titles this essay, which, if workplace trends continue as they are, might one day soon seem less like an interesting thought experiment and more like a necessary call to action.
This article originally appeared in hbr.org