Introduction

Cal Newport has written a book on ‘Deep Work’. I’m usually archly sceptical of ‘productivity gurus’ or ‘productivity guides’, but this book was recommended and I found it was very good – so much so I made notes on the way through.

I put them here in case others find them valuable: I’ve done a bit of rearranging of the line of argument, as well as some commentary on the areas I take to be less well-supported. Nonetheless, the notes follow the general structure in the book of section 1: Why deep work is great; section 2: How to do it. Those after the armamentarium of suggestions Newport makes on how to do it can skip down to the third section. Avoiding anything in italics spares you most of my editorialising.

What is deep work?

Newport describes deep work thus:

Professional activities performed in a state of distraction-free concentration that push your cognitive capabilities to their limit. These efforts create new value, improve your skill, and are hard to replicate.

Newport also offers in contradistinction an idea of ‘shallow work’:

Noncognitively demanding, logistical-style tasks, often performed while distracted. These efforts tend not to create much new value in the world and are easy to replicate

Note that both of these definitions carry both a descriptive claim of what they are, but also an evaluative claim (‘easy to replicate’, ‘create new value’ etc.).

Newport goes on to suggest getting better at the former is an extremely important skill.

Deep work hypothesis: The ability to perform deep work is becoming increasingly rare at exactly the same time it is becoming increasingly valuable in our economy. As a consequence, the few who cultivate this skill, and then make it the core of their working life, will thrive.

The argument for deep work

To motivate the evaluative claims in his definitions (i.e. you should aim to do more deep work, and less shallow work), Newport offers a mix of arguments and examples.

I’m principally going to ignore the examples (e.g. Rowling, Jung, Stephenson, etc.) Although these anecdotal cases are commonplace in popular non-fiction, I generally find them of limited value because of selection: maybe for every Rowling or Stephenson who shunned social media I could find 10 similarly successful authors who thrive whilst buzzing avidly around the internet. Happily Newport offers a more principled argument, and also (to his credit) attempts to account for and explain adverse examples.

Deep work makes you rich

Newport suggests the modern economy increasingly favours three things.

1. High-skilled work. Lots of routine tasks are increasingly automated, so having exotic and hard to substitute skill is more valuable.
2. Superstars: Thanks to modern technology top performers can scale-up their work and outcompete less-super peers. The canonical case is musicians after good recording and playback was commonplace. If you were a good but not exceptional pianist, one could still make a good living playing at less prestigious concert halls and so on. Now consumers of music can listen to an exceptional pianist on CD (or now, online), and so the demand for not-as-good pianists has fallen.
3. Owning capital: Cf. the Piketty thesis.

But note the automation literature doesn’t equate ‘high skill’ or ‘automation resistant’ work to be (as Newport later implies) ‘working well with intelligent machines’. Many ‘lower level’ programming jobs, although technically hard, are nonetheless at high risk of automation. Frey and Osborne (2013) suggests alongside skilled ‘intellectual tasks’ (e.g. CEO) other careers with hard social or manual skills (consider psychotherapists and electricians) are also resistant to automation. Many of my former colleagues in medicine remain pretty tech illiterate, but they probably have less to fear from the rising tide of automation than a typical data scientist.

If one wants to enter one of the first two winning categories (Newport doesn’t claim to have tricks to amassing vast amounts of capital), he suggests the core abilities required are:

1. Quickly mastering new skills
2. Producing at an elite level

So becoming Nate Silver required not only mastering some pretty tricky statistical knowledge and know-how about how to use databases, SQL, and STATA (1), but also putting these together into a unique product which was very valuable (2).

Newport argues deep work is a key for both of these abilities:

For the first, he points to the literature on deliberate practice, following Ericsson suggesting that long periods of focus are necessary (alongside short feedback loops) for developing expert performance.

Unsurprisingly, Newport sides with the stronger version of Ericsson’s view where the main thing is deliberate practice, and relatively little emphasis should be paid to ‘natural talent’. I’m not entirely persuaded given the pervasive influence of g (it seems pretty hard to suppose becoming a computer science professor like Newport wouldn’t require one to be much smarter than average). I also don’t really buy the sketched neurological explanation around more oligodendrocytes etc. That said, these are only niche points. Perhaps given whatever level of natural talent one drew from the birth lottery, more deliberate practice makes the most of it.

For the second, Newport notes Abam Grant’s high performance, and suggests ‘batching’ tasks to focus on each in turn (i.e. doing only teaching in the first term, doing only research the next) helps improve performance. Newport offers the following toy model:

(High quality) work produced = Time spent * Intensity of focus

With the further rider that intensity of focus is also a function of time – it has an initial ramp up period, so lots of interrupted periods add up to less than a single uninterrupted stretch, and suggests Leroy’s work on ‘attention residue’ may explain why: switching from task A to task B usually has ones brain ‘thinking a bit about’ task A, so impeding performance on B. So switching a lot is bad for focus.

Counter-examples

Newport recognises that some ‘winners’ in the economy seem to follow the opposite of his suggestions. Top CEOs are often ultra-shallow, flitting between meetings, always connected, and so on. Newport suggests that these people’s elite skill is rapid and accurate judgement, and so although for them deep work is a bad idea, they are the exception rather than the rule. Newport also adds the caution that many jobs have some parts where deep work may help, even if it is not mission-critical.

There’s a nice quote by Knuth – who Newport later cites as an example of practicing deep work – that might help distinguish these cases, “Although some people need to stay on top of things, my job is to get to the bottom of things“. Some roles – think CEOs, but also emergency physicians, international conference managers, perhaps ‘day to day’ operations roles – are much more ‘stay on top of things’, and so for them being in easy contact and rapid action are the crucial value they add. Yet for other roles – research most obviously – are much more ‘getting to the bottom of things’, where success is measured more in a few large achievements rather than lots of smaller tasks that need to all be done well.

Newport also offers a preemptive undercutting defeater to arguments along the lines of, “If deep work is so great, how come so many companies love using Slack? How come open-plan offices are in vogue?” He diagnoses the mistake as a metric black hole – improvements to shallow work metrics are easy to measure and seem good, whilst any costs they impose on deeper work much harder to measure (even if they’re really bad):

1. Principle of least resistance: if everyone runs their day out of their inbox, then people don’t need to task manage and always feel productive, even if this stops them planning out and accomplishing important larger tasks (e.g. the case of the Boston Consulting Group, where unplugging for a day led to better results. (Newport also directs some ire to non-informative email forwards and frequent ‘status meetings’)
2. Busyness as proxy for productivity: Unlike the industrial age (or some areas of academia) there’s not a clear metric for performance like how many widgets you made or your h-index or whatever. Yet some shallow work indices are more clearly measurable (how rapidly you reply to emails, how many meetings you attend, etc.) so workers either do this so they satisfy themselves they are busy, or to satisfy management metrics which focus on the same (cf. Yahoo).
3. Technopoly (or the cult of the internet): Technology (like the internet) looks shiny, new and hip, and so being ultra-connected and using lots of communication tools projects the right sort of image, even if the important values for the work (craftmanship, mastery, etc.) are in truth a bit old fashioned.

Of course, maybe (2) is rational management with not-that-motivated employees: cajoling them into regular meetings and making sure they come into work might push them more into doing something, whilst if you don’t your employees might slack off all the time!

The upside of this is that this opens a gap in the market for people who can resist these urges and do more deep work.

Deep work is meaningful

Craftsmen enjoy work where success is easy to define (e.g. can you forge a good sword?) but difficult to perform (e.g. the hours spent carefully hammering the metal). By contrast, in many ‘knowledge work’ jobs success is much harder to define (what distinguishes a great communications strategy from mediocre one?), even if the performance of subsidiary steps (e.g. opening Word, typing in the right headings) is straightforward. Newport suggests that deep work can lead to increased satisfaction by moving to work more analogous to a master-craftsman than the stereotypical office desk-jockey.

Three lines of argument:

1. Attention and happiness: If (as Newport suggests) a lot of what makes you happy is what you pay attention to, spending more time doing deep work is a pleasant residence for your attention (cf. flow), but also avoids it being harried by lots of trivial niggling frustrations – extant or imagined – in day to day life (consider the emails loitering in your inbox, conversations you are putting off, etc.)
2. Flow: The experience sampling literature suggests that people are at their happiest when being stretched to their limits doing something challenging. For this reason work can often be fun, and ‘relaxing’ surprisingly unsatisfying. So as deep work seems likely to engender these flow states, deep work can lead to personal happiness.
3. Craft and sacredness: Newport suggests there is a source of satisfaction (‘glimpses of the sacred’) that can emerge even from pretty mundane roles if they involve making something well (whether it wheels, swords or code). So deep work, which focuses on doing just this, can provide another source of satisfaction.

How to work deeply

Having motivated the theoretical claim of the value of Deep Work, Newport now goes on to offer advice to how to achieve it. Newport counsels that a ‘I’ll just do deep work now’ is unlikely to work out. Experiential sampling of people shows they not only have competing desires (e.g. food, sleep, sex), but also a desire not to be working. He also counsels that ‘willpower’ strategies are unlikely to work due to the literature on willpower being a finite resource. So he now provides an armamentarium of habits, tips, and other stuff that enables more deep work

We now lnow ‘ego-depletion’ hasn’t weathered the replication crisis well – Deep Work was written pre-replication crisis, and Newport is hardly the first to be burned by trusting decades of accumulated psychological research. It doesn’t seem fatal – commonsensically, most people find working in a focused and dedicated way hard, so might like some tips on how to do this better.

Segregate deep work into the smallest number of biggest chunks you can

Newport uses case studies to highlight a variety of ‘philosophies’ of deep work (‘monastic’, ‘bimodal’, ‘journalistic’), but I think they can be condensed to basically a simple optimisation problem under a scheduling constraint.

For deep work, the ideal is the biggest chunks of uninterrupted time possible: better a dedicated hour than a dedicated five minutes, better a dedicated month than a dedicated week, and so on.

So why not just make one’s entire working time comprise deep work? If it were the case professional success was solely a function how much deep work one does, then this is a call to monasticism: cut email, cut meetings, etc. (Newport uses Stephenson and Knuth – people who’s all-but-only priority is to write books – as examples). Most jobs either require, or workers benefit from, some shallower stuff (even as a researcher, it might be worth keeping an eye out for grants). So they’re either constraints or complements. So how to interleave the two?

The key consideration seems to be latency. How long (for example) could you ‘ignore everything else’ before there was a risk of bad things happening? For some lucky academics, this might be weeks (longer if on sabbatical), recommended long stretches of time devoted to depth. For busy managers (or parents of young children), this might be no more than a morning, or an hour. But whatever it is, the best strategy (not accounting for taste) is to divide up whatever quantity of deep work one needs to do as few times as possible into the biggest chunks ones other commitments (and their respective latencies) allow.

Rituals

Contra stereotypes of free-willing geniuses gamboling around from one burst of inspiration to the the next, Newport reports many instances of famed thinkers having very regimented habits. Although Newport suggests each individual may be somewhat idiosyncratic, he suggests making sure yours ‘sets in stone’ the following.

• Where you’ll work.
• How long you will work.
• What you’ll do (i.e. ‘no internet’, ‘no email’, etc.)
• Preparation: start with coffee, walking, etc.

One bit of ‘preparation’ that might be helpful sometimes is a grand gesture, which could be thought of as a particularly emphatic way of initiating a period of deep work to ensure you actually do it. Newport gives examples of Gates’s ‘think weeks’, or Rowling booking herself into a hotel to complete Deathly Hallows.

Execution

Newport suggests Christensen’s approach for big business (The 4 Disciplines of Execution – 4 DX) can also provide useful insight into how to do more deep work.

Although the precise link-up between ‘deep work’ and these remains a little unclear…

1. Focus on the wildly important. ‘The more you try to do, the less you actually accomplish’. Ergo try and limit oneself to a few key achievements, and target deep work to these rather than spreading oneself thinly across of laudable tasks (cf. 80:20 rule, Buffet’s maxims about ‘do not do’ lists, and ‘Hell yeah or no’ decision making).
2. Act on the lead measures. ‘Lag’ measures are often outcomes (e.g. I will write 3 maths papers this year), ‘Lead’ measures tend to be processes (e.g. I will spend 4 hours each day writing maths papers). Although the former are ‘gold standard’ for assessing overall success or failure, they aren’t necessarily very useful during: if at the end of the year you’ve written only one paper, you can’t retroactively change what you did that year to have written two more. In contrast, if you can find helpful lead measures to target, these have a much shorter feedback loop to ensure good performance. (Newport naturally enough suggests ‘hours spent in deep work’ is a good lead measure).
3. Keep a compelling scoreboard. Monitor these (generally) lead measures and review them frequently.
4. Cadence of accountability. Also have other people on you team or similar review each others progress. (cf. accountability buddies etc)

Downtime

Newport makes a broader call for work hygiene where one sets clear boundaries about when one is and is not going to try and do work work (e.g. “After 6pm I’m going to relax until 9am the following morning”). He argues as follows:

1. Unconscious thought theory: Maybe ‘sleeping on it’ lets one’s unconscious brain filter through the various heuristical judgements necessary, and so insights or understanding can steadily emerge whilst one does something else. Alas, seems not to have weathered replication well on a quick skim.
2. Attention restoration theory: One has a limited stock of directed attention, which can be depleted (by work, or by stressful environments, or other things). So having downtime, to ‘recharge’ might be a better strategy than trying to run on near empty. Walking through nature comes highly recommended. Doesn’t seem to have fared as badly in the replication crisis, but plenty to worry about, and it sounds a bit ‘ego depletion’ esque.
3. Evening work tends to suck so you might as well relax anyway: Newport returns to Eriksson suggesting there is a limited store of ‘deep work’ one can do in a given day (1-4 hours). So if you’ve hit this during the day, you’re probably going to spend the evening doing nothing very deep anyway.

Independent of the motivation (“relax sometimes” has pretty strong face validity even if the rationales above don’t go through), Newport makes some suggestions:

• Have a strict (and regular) endpoint: I.e. the rule should be a clear time, and that nothing work related after it (e.g. ‘quickly checking emails’).
• Shutdown rituals: Newport suggests a GTD-style ‘task cleanup’ (e.g. check email for anything urgent, check next few days in calendar, look at task lists, then make sure it is either done or there’s a note made about it), then some habit or ritual (“Shutdown complete”) to close.  

Training deeper focus

Newport suggests that although deep work is hard (cf. Erikson on a novice can only do it for 1 hour a day), one can develop the skill of doing so for longer, and that this mental training might bring other benefits. One skill might be to better resist distraction – easier said than done in the age of the smartphone. How?

Scorched earth internet/distraction avoidance

Rather than segregate time one won’t be distracted (e.g. ‘Internet sabbaths’), Newport goes further and suggests the default should be to not have distractions, which are only indulged in for ones breaks. Newport simplifies distraction as ‘internet access’, and so suggests:

• Schedule when you will use the internet
  • But what if my job needs lots of internet use?
    • Fine, then you need more internet blocks to do work in. But you should still have some non-internet time to do non-internet work, and you should keep this pure of internet contamination. (Even if its 5 minutes every 15) – having some time helps build the habit.
• Without exception do not use it except for those times. (e.g. every other pomodoro is a non-internet one, the other ones I can use it for work, in the breaks I can do whatever).
  • I was doing non-internet work, but I’ve come to a point where I could really use internet access for X.
    • Tough it out: the benefits of resisting the temptation are worth it, even if you promise to yourself you’ll only go online for X and leave afterwards. Switch to something else, work around, or relax instead.
    • If you absolutely have to use the internet (X is time-critical), at least impose some finite delay.
• Do the same thing at home, and for leisure.
  • Maybe allow very limited pragmatic exceptions (e.g. texting friends to meet in the right place, google maps to navigate to evening drinks).
  • Again, if you’re leisure involves a lot of internet, schedule more internet time, but again maintain the hygiene of non-internet time.

One can imagine some circumstances one might have to relax the bright line of no internet into a very narrow whitelist (e.g. I understand writing code without the internet is very counter-productive)

Practice bursts of effort

On the other side of practising not getting distracted is practising to work intensely. Newport suggests training bursts of effort with artificial deadlines as one way of ‘building the muscle’ of intense effort.

‘Productive meditation’

Newport suggests light physical activity which isn’t mentally taxing to spend time focusing on a particular deep problem without other distractions.

This and some more stuff on memory palace techniques I’ll skip as it feels a bit dubious yield over and above ‘if you’re stuck on something, take a walk/ active travel is pretty good to do if you can’.

Against Social Media

Newport is famed for his anti-social media practice and preaching. (And unlike the psychological results reported pre-replication crisis, Newport’s objections to social media have aged extremely well). In sketch, he argues that we can explore a middle ground between Thoreau style disconnecting from everything online, and gluing ourselves to our smartphones and social media accounts. He suggests we weigh each ‘bit’ of the internet on its merits, and the balance in the case of social media (he also dubs them ‘network tools’) is often unfavourable.

Limited benefits

Newport firstly suggests the ‘upsides’ of social media like facebook are minor, e.g.:

• Mediocre entertainment (e.g. looking at a meme group).
• Lightweight freindships (e.g. joining a group and chatting about a shared interest online)

We could probably add more, especially around network effects. Many social events are advertised exclusively on FB, but that these aren’t life changing is generally fair.

Significant downsides

Newport sketches the downsides of social media as these:

• Distracting (cf. above about deep work, avoiding distraction, etc.)
• Potentially addictive, so stealing time away from more valuable pursuits (cf. attention engineers)
• Maybe bad for one’s mental health in getting burned out and distracted

Definitely more plausible in light of some further interesting (albeit correlational) studies.

Contextual weighing-up

Newport goes on to argue the right thing to do is to look at ones particular circumstances and weigh up the costs and benefits of ‘network tools’ for that, and make a decision accordingly. He points to a few heuristics which suggest in many cases the judgement should be against social media (I’ll skip the somewhat belaboured walk-through):

You could also imagine subdivide given bits of a network tool (e.g. ‘FB events/newsfeed/messenger)

• A few big things matter the most: In our lives there are a few big goals which matter far more than other minor things: E.g.:
  • Professional success: as (see above) this often relies on deep work, social media isn’t helpful.
  • Personal relationship satisfaction: For most, we get the most out of deep connections with our family, partner, and close friends, so it is better to prioritise them for our time versus a cloud of light touch e-acquaintances.
  • (Maybe exceptions (e.g. you’re a college freshman who wants to get in contact with a lot of people to make friends), but for many people this will apply.)
• Yet even if social media isn’t contributing to the most important thing, why not keep it in the mix for its minor but diffuse benefits? Newport argues there are usually costs with limited resources of time and attention, and so social media will tend to substitute rather than complement more important stuff with less important stuff.

Trial FOMO

Newport suggests an easy way to see whether you’d miss out without social media is to quit all of them for thirty days (but don’t say publically you are quitting them) and see what happens. Did you miss anything important? Did anyone miss you? If no to both, quit permanently.

Newport’s snarky, “Trust me, people will survive without your querulous social media commentary – most of the time you and your correspondents are playing a negative sum game of ‘I’ll pay attention to you if you pay attention to me’” is amusing but (slightly) over-reaches.

Generally avoid the internet for entertainment

Newport expands his case to ‘e-entertainment’ generally (think reddit, buzzfeed, ‘clickbait’ generally, and whatever one’s reflexively does on ones smartphone besides social media), through a parallel argument: these are low-quality means of leisure which often crowd out something better to do with your free time. Newport suggests structured and deliberate leisure time (e.g. reading a book, exercise, team sports, etc.) is strictly superior both for a happier life and being in a better state to do work later.

Limiting shallow work

Even if deep work is best, some shallow work is almost inevitable. How can one limit one’s shallow work as much as possible, leaving the most time for depth?

• Schedule everything/timeboxing: As people are terrible at estimating how much time they spend(/waste), so explicitly block out your day with what you plan to do, and then follow it. Easier said than done:
  • New priorities: That’s fine, just reschedule – the important thing is to be following a conscious plan for how to spend your time, not that this plan can’t change.
  • Planning fallacy: As compensating for it seldom works, put in overflow blocks which can either be used for the preceding task or a smaller one if you finish (vaguely) ‘on time’.
  • “You can’t schedule flashes of insight!”: You sort-of can (e.g. blocks of ‘thinking time’), and you can treat a sudden inspiration as a new priority prompt rescheduling (cf. rigid schedules often followed by famous creatives).

• Identify shallow work (and how much time it takes): If one wants to minimise shallow work, one needs to know what it is and how much of it one does. Newport offers a handy rule for identifying shallow work when one isn’t sure: “How long would it take to train a college graduate to do this?”
  1. Editing an academic paper: deep (even if the editing looks pretty shallow, knowing what to edit needs a lot of academic background)
  2. Turn the paper into a powerpoint: shallower (you can probably teach the hypothetical college grad what to extract and how to run powerpoint without imparting lots of extensive domain knowledge.
• Work out how much time you should spend on shallow work: One can ask oneself this, but one can also use it to try and negotiate with boss/colleagues (e.g. “You’ve hired me as a PhD computer scientist, yet I’m spending 40% of my time on emails, so we should figure out a way of getting that down”)

 

 

 

 

 

Cynics might find this idyllic, as a Boss might not go, “Ah, then don’t sweat the weekly update email”, but rather “You need to be more efficient, and these ‘shallow’ things are non-negotiable commitments”. Newport’s reply in that this tells your job isn’t good for deep work and you should look for something else isn’t that helpful.

• Work backward from a quota of time to what you will do in it: As work can always fill available space, set limits on ones hours and work backwards from them, prioritising ruthlessly.
  • E.g. set quota’s on shallow/less important tasks (for academics: I’ll only review X papers, give Y non-academic talks, etc.)
  • Learn to say no better:
    • Don’t give an excuse for refusal people can argue with.
    • Don’t offer time-consuming ‘consolation prizes’ (I can’t do X, but can do Y).

• Limit access: As lots of meetings/email threads sap time and are generally shallow, one can limit them by increasing the barrier to entry.
  • Don’t make it that easy to email you
    • Sender filters (e.g. ‘for inquiries re. X, go here, FAQs, etc.)
    • Make it clear unsolicited emails will be default be unanswered.
    • Triage ruthlessly (e.g. if it takes 30 seconds to figure out what they’re asking, delete it instead).

 

1. • Comprehensive replies can be a true economy: For project-relevant emails, work out the path which gives the smallest rally of back-and-forth emails between now and completion. Sometimes more time spent can save time on net.
     • E.g. “Grab coffee sometime” versus, “Let’s grab coffee at T, X, if you can’t make that, book a time I’m available in my calendly”.
     • “Where are we on X?” versus “Here’s a google doc with my understanding of X and next actions, I’ve arranged a time to discuss to check I understand, feel free to reschedule at my calendly if the time doesn’t suit”.
     • “I’ve taken a stab at the paper” versus, “I’ve revised the paper and commented the areas which I think should be changed but you’re better placed”.