Explore a better way to work – one that promises more calm, clarity, and creativity.

On the Law of Diminishing Specialization

On Productive Technology and its Discontents

Recently, I’ve been dipping in and out of Edward Tenner’s provocative 1996 book, When Things Bites Back. In following one of Tenner’s footnotes I came across a fascinating 1992 academic study from the National Review of Productivity, authored by the Georgia Tech economist Peter G. Sassone

The paper has an innocuous title, “Survey Finds Low Office Productivity Linked to Staffing Imbalances,” but its findings are profoundly relevant to our recent discussion of attention capital theory, and the value of deep work more generally.

Beginning in 1985, Sassone began a series of twenty office productivity case studies spread over different departments in five major U.S. corporations. His initial goal was to measure the bottomline benefits of the front office computer systems that were new at the time, but as he notes, this soon changed:

“[I]t became apparent that [my] data collection and analysis techniques were yielding important productivity insights beyond the cost justification of office computer systems.”

Deploying a technique called work value analysis, Sassone measured not only the amount of work conducted by his subjects, but also the skill level required for the work. He found that managers and other skilled professionals were spending surprisingly large percentages of their time working on tasks that could be completed by comparably lower-level employees.

He identified several factors that explain this observation, but a major culprit was the rise of “productivity-enhancing” computer systems. This new technology made it possible for managers and professionals to tackle administrative tasks that used to require dedicated support staff.

The positive impact of this change was that companies needed less support staff. The negative impact was that it reduced the ability of managers and professionals to spend concentrated time working on the things they did best.

Among other examples uncovered in his case studies, Sassone highlighted:

  • A corporate marketing department where senior marketing professional were spending more than a day per week of their time preparing charts and graphs for presentations.
  • A large commercial bank where corporate bankers were devoting more than a quarter of their time to handling routine interactions with clients.

This reduction in the typical deep-to-shallow work ratio (see Rule #1 in Deep Work) became so pronounced as computer technology invaded the front office that Sassone gave it a downright Newportian name: The Law of Diminishing Specialization.

What makes Sassone’s study particularly fascinating is that he used rigorous data collection and analysis methods to answer the question of whether or not this diminishing specialization was a good trade-off from a financial perspective.

His conclusion: no.

Reducing administrative positions saves some money. But the losses due to the corresponding reduction in high-level employees’ ability to perform deep work — a diminishment of “intellectual specialization” — outweighs these savings.

As Sassone explains:

“The results of a comparison of a ‘typical’ department, with a department with a reasonable high level of intellectual specialization were startling. The typical office could save over 15 percent of its payroll costs by restructuring its staff and increasing the intellectual specialization of its workers.”

To make this more concrete, he calculated:

“[T]he typical office can save about $7,400 [around $13,200 in 2018 dollars] per employee per year by restructuring its office staffs and improving its levels of intellectual specialization.”

In other words, Sassone found that the corporate divisions he studied could produce the same amount of valuable output by reducing the number of managers and professionals while increasing the number of administrative staff.

This rebalancing works because more administrative support means the higher level employees can spend more time working deeply on the activities that produce the most value. Because the former are cheaper to hire than the latter, the result is the same work for less total staffing costs.

An important lesson lurks in these results that’s just as relevant now as it was then, back in the early days of the front office IT revolution: optimizing people’s ability to create value using their brains is complicated. Just because a given technology makes things easier doesn’t mean that it makes an organization more effective, you have to keep returning to the foundational question of what best supports the challenge of thinking hard about valuable things.

25 thoughts on “On the Law of Diminishing Specialization”

  1. Hi Cal – thanks for another great post. This idea can connect to your idea of the API idea you mentioned recently. I remember either in one of your books or in a post, or both, you talked about that office at MIT that was sort of a hub and spoke system, with offices on the outside and hallways connecting to a center lobby. I think a physical space like this, where you had rooms for doing deep work on the outside, combined with administrative ‘bouncers’ that are there to handle the shallower work, and also to server as gate keepers to protect the deep workers from being interrupted, might be one way to have an API could be created to protect and grow an organizations attention capital. In this system you would literally have the attention capital ‘bouncers’ positioned in the hallways going to the study rooms, so that people would have to go through admin person to interrupt. If you have to pay someone $30 – 50 K to sit and protect the time of say two engineers being paid $100K, if they each get twice as much work done, it would be well worth it. Thanks for all your great posts! Now to figuring out how to implement them in a world of open offices … that’s the challenge!

  2. Great post!

    I think the meeting clients problem extends greatly to CEOs of big companies, here is an excerpt from “Peter Thiel: Zero to One, notes on startups.”

    > If your average sale is seven figures or more, every detail of every deal requires close personal attention. …
    Elon Musk persuaded NASA to sign billion-dollar contracts to replace the decommissioned space shuttle
    with a newly designed vessel from SpaceX …
    But since complex sales requires making just a few deals each year, a sales grandmaster like Elon Musk can use that time to focus on the most crucial people—and even to overcome political inertia.

    In this example it seems like Elon could better spend his time doing more valueable stuff, like making sure production for Tesla is ramped up. As sales happen only a few times a year for him it might take just a small portion of his time, but for him it comes down to thousands in costs/revenue.

  3. Hi Cal! It is a interesting reflection you make towards the IT revolution. It is something that I see very often here in Peru: Managers are obssessed with IT revolution, believeing that would make their organizations more effective. I don’t believe that embrassing IT is bad, but I was wondering how can we encourage people to take these new technologies while at the same time improving the value they give to deep work. I will be glad to have some ideas on these.

  4. This is a great post! But I wonder if it endorses the IT revolution in the sense of using AI for low-level specialization tasks instead of increasing human staff. Ultimately, does this means we should all become high-level specialists in the future? What is the role of people who don’t have the opportunity for having a high-level specialization? I think this idea complements well with the importance of becoming learners instead of specialists. Everyone can improve his/her learnability if properly develop his/her learning mind.

  5. Hi Cal, Great post! I think this idea can be combined with the human API idea, and the MIT office you wrote about a while back that was the anti-open air office. I remember you saying that the office was sort of a hub and spoke system, where the offices were on the outside of the building, and the halls ran into something like a main lobby. An office that had closed door offices at the end of hallways, and each hallway could be guarded by a receptionist or admin staffer who serves as an ‘attention capital’ bouncer or body guard. Want to interrupt the engineer? You’re gonna have to get through her attention bouncer, big Mike. In all seriousness though I think that if you brought back admin staff to replace things like slack and email, and in the process double the productivity, or increase the quality of thinking of an organization, it would be well worth paying those guards to serve as the ‘ports’ for the api.

    • sorry – the first post I put up didn’t appear when I refreshed after I submitted it, I thought I had lost it, so I posted again. Thanks again for a great blog.

  6. Wow – this is a very timely post. We have been dancing around this conclusion as we are revising our accountability chart at my engineering consulting company (kind of an org chart, but it focuses on what needs to get done and whose accountable). This is the support I need to push my leadership team to try to break our work into more rational pieces vs. having engineers do projects start to finish.

    Thanks for the great work Cal! (as an aside – we are also debating our open office floor plan. You previously posted about less interaction happening as folks where headphones and keep their heads down in an open office floor plan. Wouldn’t that be a positive indicator of being in Deep Work and therefore argue for an open office floor plan?)

  7. Great post, and unfortunately very true.
    From my experience, the more a person gains knowledge and becomes good at his/her work, the more he/she is offered shallow work in order to advance in his/her career.
    What puzzles me is that there is a financial case for this not to happen, so the question is whether this knowledge is unknown to decision takers or whether it is ignored.

  8. This is very definitely at work in the medical field. Electronic medical records (EMRs) have turned physicians into data and order entry clerks. They now spend significant portions of their time serving the EMR master. Some have gone so far as to hire a “scribe” to do this work for them, but usually at their own expense (note that more and more physicians work for hospitals or other health care corporations). Often these tasks are pushed into after work hours, resulting in less personal and family time. Failure to complete these tasks leads to official sanctions, including being suspended from medical duties, and certainly is a major driver in physician burnout. This has similar impacts for nurses, PAs, and nurse practitioners.

    • I was going to write the same thing. It’s endemic in the medical field. For a while my husband paid an assistant out of his own pocket and finally the hospital saw the value and hired her, then my husband changed hospitals. He’s back where he began – entering his own data and doing far more administrative tasks than he should, time taken away from the medical tasks he’s educated, and paid, to do.

    • You beat me to it, Erik. I completely agree. We used to have transcriptionists to write fairly brief notes. Now we have scribes (or more often, our own fingers) to write notes that are often pages long, with resulting substantial losses in productivity and only marginal (at best) gains in quality of care.

  9. Interesting article.

    However, I think that the reality is more complex for organizations and teams. Efficiency and focus of individual contributors is important, but it’s not as important as team throughput. The risk of optimizing for efficiency is to increase the “inventory” of a team. An interesting book about this subject is The Goal by Eliyahu Goldratt. What do you think of this?

  10. Hello Cal! I just have a question regarding Deep Work. I do use the 50/10min ratio for studying but is that a good ratio for when working or trying to learn other skills? For example, I am planning on learning to start coding and would following this type of deep work/break ratio produce just as well? I find people differ on this topic regarding coding and working on projects where they say it’s best to work until you feel you need to get up.

  11. Hi Cal,

    This could be an interesting counter-argument to the administrative bloat that many claim encumbers Higher Ed and astronomically raises prices. As you know, the cost of 4-year universities has dramatically outpaced the inflation rate of 2-3%/yr and has reached troubling levels; for example, NYU topped 70K this year. Many speculate that this is because of excessive administrative expenditures. However, one could argue that at least some of these administrative outlays may in fact protect the deep workers therein (professors, researchers, students, etc) from shallow work, thereby counter-intuitively optimizing the productivity of the university per dollar spent.

    • It depends what you mean by “administrator,” which can really mean two very different types of work in academia. When abbreviated to “admin” it typically means someone who is there to assist faculty and to take secretarial work off of their plates; but the numbers of those admins in academia are, if anything, declining, with a greater shift to part-time work and single admins serving more and more people. On the other hand, you have “administrators” who are really *managers*, and a lot of them are hired to run new initiatives that have nothing to do with research. In a sense maybe you could argue that they’re taking stuff off faculty’s plates because they do things like bring the university into compliance with federal regulations — but they’re mostly doing things that were never the faculty’s responsibility in the first place.


Leave a Comment