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

small an impact as possible on the specialists’ main work
obligations. If taken seriously, this metric might mean a given
support unit needs to make its own work less efficient to better serve
the organization.
The contact point where this idea becomes relevant is the various
interfaces through which specialists interact with support staff. As
underscored by my above mailing list example, if every unit is left to
design these interfaces in a manner that makes its own operation
easiest, soon everyone ends up deluged with more communication
than they can reasonably handle. A better interface in this case might
be a shared weekly newsletter that includes summaries of all relevant
announcements with links to more details for those who are
interested. This rule makes the operations of the support units
slightly harder, as they can’t simply blast out announcements
whenever they want, but the information still gets spread, and this
time it’s done in a way that reduces interruption.
For a more extreme example, imagine an organization in which
requests for a specialist’s time and attention—such as the parking
office asking them to fill out parking renewal forms, or the travel
office demanding preregistration of all travel—go to some sort of
attention capital ombudsman who can weed out unnecessary
requests, consolidate others, and perhaps even negotiate with
support units to make their requests easier to complete. This might

---

sound absurd, but is it? Google, for example, already invests heavily
in free food and subsidized dry cleaning to help its high-paid
specialist developers produce more value. Against that backdrop, the
cost of an ombudsman of this type might be minor relative to the
additional value it would unlock.
Going the other direction, one might imagine also optimizing the
interfaces specialists use to contact support units, with the goal of
minimizing the impact on the specialists’ time and attention. In the
world of consumer interaction, there has been a push over the past
decade toward what’s known as invisible UI—interfaces that are so
simple and flexible that the consumer doesn’t even think of them as
interfaces at all. Perhaps the most common current examples of
invisible UIs are digital assistant appliances like Alexa and Google
Home. 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
The examples above are meant only to prompt more concrete
thinking. The details of how you actually optimize these interfaces
depend on your specific type of work. A more abstract way to think
about this optimization is to imagine that each support unit
maintains a counter that, through some magic, is able to track the
total number of minutes of attention the unit has commandeered
from other employees so far that week. The goal could then be simply
to minimize that number as much as possible while still executing
core functions. No such counters exist, of course, but this neatly
captures the shift in thinking this approach to support induces.
Finally, I should admit to feeling some trepidation about these
concepts. An ethical pitfall in moving toward more specialization is
the fear of creating a sharp divide between specialists who enjoy
their work and an underclass of support professionals confined to
overload. My suggestion here that support units should be willing to
make their own work harder to make specialists’ work easier, in

---

addition to being self-serving, seems to nudge our discussion toward
this pitfall. With this in mind, I want to propose two defenses.
First, reorienting support staff to optimize the production of
specialists doesn’t necessarily need to make the work life of the
former more miserable. My first idea in this section dealt with ways
to introduce more structure into support processes to sidestep
hyperactive hive mind overload. That idea still applies: changing
your objective from making your own unit as efficient as possible to
helping your organization produce as much value as possible doesn’t
need to reduce the quality or sustainability of the work involved.
My second defense is that whether or not we like this suggestion,
it’s an economic reality. If a knowledge work organization is
producing valuable cognitive output in a competitive marketplace,
then it’s self-evident that having support units prioritize this output
will make an organization more successful than if it instead allowed
every unit to focus myopically on its own internal objectives. To be
clear, no unit should be disrespected or treated as less important,
and nobody should ever have to tolerate a misery-inducing work
environment. But beyond these fundamental principles, it’s also true
that companies aren’t democracies, and employees are not all
necessarily guaranteed the same types of liberties surrounding their
efforts. Put more bluntly: no knowledge work organization ever
conquered a market because of the internal efficiency of its HR
department.
Supercharging Idea #3: As a Last Resort, Simulate Your Own
Support Staff
The two preceding ideas concern the role of support staff in large
knowledge work organizations. Putting these ideas into action
requires that you’re in a position of power—perhaps CEO or head of
a large division. If you’re instead an employee without this control,
but still suffering because you don’t have enough support, you’re not
entirely out of luck. In this situation, as a last-resort measure, I
suggest simulating your own support staff.
One way to accomplish this goal is to partition your time into two
separate categories: specialist and support. For example, perhaps
12:00 to 1:00 p.m. and 3:00 to 5:00 p.m. are support hours. During

---

all other hours, you act as if you work in a specialized organization:
focus only on skilled work that directly produces value. Don’t answer
administrative emails or attend administrative meetings—just work
on what you do best, as if you’re an XP developer. During the support
hours, by contrast, act as if you’re a full-time support staffer whose
objective is to make your specialist alter ego as effective as possible.
Don’t simply get lost in emails during these times, but actually follow
the advice given above and put in place processes for minimizing the
sense of overload you experience juggling these logistical matters.
(The process principle chapter provides some specific strategies
individuals can deploy toward this purpose.) You can even optimize
the interfaces between these two sides of your work life by putting in
place simple collection bins where your specialist self can store
administrative work to be later tackled by your support self. Maybe
you keep a text file for this purpose, or an actual plastic collection bin
on your desk where you can drop forms, or reminders you jot down
on paper (an idea originally proposed by David Allen).
If you want to get more advanced, consider using two separate
email addresses. I do this to some degree in my role as a professor. I
have an email address for the georgetown.edu domain that was
assigned to me by the university. This is where I receive all official
university correspondence, and I use it, as much as possible, for
administrative issues. I also have an address hosted by a server in
our department for the cs.georgetown.edu domain. I use this to
interact with other professors, the students and postdocs I supervise,
and my research collaborators. The former address belongs to my
support self; the latter to my specialist self.
Another advanced tactic is to assign entire days to these roles.
Perhaps Tuesday and Thursday are support days, and Monday,
Wednesday, and Friday are specialist days. Not every job allows such
a dramatic split in your behavior, but if yours does, there’s great
clarity in such a clean division. I’ve even met practitioners of this rule
who use different locations—coming into the office for support days,
for example, and working from home when in their specialist role.
This idea of pretending to be two different types of workers
might seem heavy-handed, but there’s a surprising amount of
efficiency to be gained by isolating these distinct categories of effort.
As discussed in part 1, rapidly switching back and forth between
support and specialist work reduces your cognitive capacity, leading

---

to less quality work produced at a slower rate. An hour dedicated
exclusively to a hard project followed by an hour dedicated
exclusively to administrative work will produce more total output
than if you instead mix these efforts into two hours of fragmented
attention.
—
Technology helped push us down the road toward diminished
specialization and increased overload. Once personal computers
made it feasible for specialists to handle more support work, tackling
an overwhelming number of obligations became the new norm,
helping to cement the hyperactive hive mind workflow as the best
option for wrangling our hectic professional lives.
Reimagining work, therefore, first requires more specialization.
Let the knowledge workers with value-producing skills focus on
applying those skills, and put in place robust and smartly configured
support staff to handle everything else. This move toward less (but
better), built on a balance between specialization and support, is
fundamental for the evolution of knowledge work from its current
inefficient chaos toward something much more organized.

---

Conclusion