The Decrease in Employee Productivity

According to the Bureau of Labor Statistics in the US, productivity per worker has declined for three quarters in a row.  Similar studies from other institutions have made similar claims.  First, we should apply some skepticism - how we define productivity and how we measure it is deeply nuanced, job sector specific, and flawed.  Even in "physic thing manufacturing" where there is some direct output to measure "how many things per hour are produced" there is fuzziness - if we produce fewer things per hour, but they are more sophisticated and complicated to build, has employee productivity dropped?  And when you get into less tangible roles - the "knowledge worker" - it becomes even murkier.  Also, there is an entire debate as to whether the fixation on three month increments is a healthy granularity (I am skeptical).

Regardless, there is the hand wringing about "what to do about it" ranging from "Maybe AI is the answer" to "We should cut spending so that contribution margin remains stable (or goes up)".  Unfortunately, "make sound decisions based on the available scientific evidence" isn't one of the things being recommended (uh, until now.  hey, try making sound decisions based on available evidence).

Focusing on knowledge workers, lets first look at what has changed in the last year and a half (since almost all productivity measures are trailing indicators):

  • Companies began ending work from home / flexible work arrangements either because of some unvalidated assumptions about in person productivity being higher (even though the pandemic was the largest increase in productivity in a generation), or because of financial ties to the commercial real estate market.
  • Companies adopted austerity philosophies in response to "macro economic conditions", despite DECADES of research showing that it is both short term AND long term harmful to companies.

The reasoning for why both of these would have a negative impact on productivity are pretty easy to sus out but before getting into those, it's important to look at what generates productivity in knowledge workers, both psychologically and organizationally so we can then talk about how those impacted things.  And spoiler "efficiency improvements" are not actually one of those things - it's a second order effect that comes from the primary impactors.

The first element is making sure employees are able to bring their full focus and energy.  It's one of those "it seems obvious, but it's useful to test even obvious things to make sure they also are true" sort of statements, but fortunately there are ample studies showing exhausted and emotionally or cognitively distracted people get less done.  Of particular relevance for this blog post:

  • There are two types of stress response (well, kind of three, but we will leave out episodic acute stress for now) and they are both, in the original conditions we evolved in, adaptive (i.e. we have them for originally beneficial reasons)(tangent - behaviors and psychology are absolutely evolved traits, but "evolutionary psychology" is a field of charlatans and quacks).  The first is for acute stimuli - what happens to your body and mind when "Holy shit, that's a man sized Snake/Spider hybrid and its holding 8 AR-15s while ranting about a Facebook post!".  We basically experience an incredibly rapid and energetic response centered around taking immediate action.  It is BAD for us if there is no immediate action to take though - going all ninja turtle on the Snake/Spider hybrid Oath Keeper is an outlet for said response.  But getting the exact same activation to an email, or in a meeting, and NOT having an outlet for that mobilization of energy is quite problematic - we neither expend the activated energy NOR get the neuro signals that a resolution to the stressor would bring.   The second form of stress is chronic stress - how our body responds to a persisting stressor - and in many ways the response is exactly opposite of the acute stressors.  Biologically, this response is optimized for dealing with prolonged exposure to a stressor that would likely create a significant period where energetic expenditure exceeded caloric intake.  Disaster (natural or artificial), famine, temperature (prolonged too cold or too hot), etc.  Varying degrees of depression is the response to chronic stressors - depression is adaptive in our original socio-environmental conditions.  It's the slow down of our metabolism, less impulse to expend energy, a decrease in cognitive and emotional activation (brain is the most calorie intensive system in the body - reducing its horsepower makes fuel last longer), a decrease in appetite (both because the digestive system is the second most calorie intensive system, and because most things that would naturally cause chronic stress involve food scarcity in some form), etc.  Triggering either of these stress responses when the conditions they are evolved to address are not present leads to all sorts of mental health problems, and for the purpose of this post, mental health problems make us less productive at work.
  • Humans vary by the time of day and the conditions they are most cognitively active, when task focus is easiest, etc.  Some people need a distraction free environment while others need to be around other people. Some people are morning or evening people, and some people are "bimodal" where they have an initial stretch of productivity first thing after waking up, and then a second hours later in the day, but are less able to task focus in the hours in between.  And so on - anecdotally, we all know this just from knowing ourselves and knowing how other people work, and this is a case where our anecdotal direct experiences are supported by large sample based studies.
  • Cognitive work burns calories, and burning calories tires people out.  See previous comment about the brain being the most energetically costly system in the body.  We know that asking people to do physically demanding tasks for long stretches of time wears them out.  If you push someone too much physically on day 1, you are going to get a lot less out of them on day 2 and 3.  We also know that about cognitive tasks, but for some reason ignore it.  Repeatedly study after study shows that knowledge productivity peaks somewhere in the 30 to 32 hours of work per 7 day stretch, and does not reach comparable levels of output until working more than 60 hours for most people, when comprising reoccurring weeks (a person can push themselves hard for one week here and there and do ok, but they suck at it when its regular).  The typical person will get less done working 50 hours than working 30.

The second element is that it is import for employees and teams to have enough buffer in their work commitments so that they have slack.  The necessity of this won't make sense to people who manage by Excel - "isn't uncommitted work time wasted time" - but will make sense to anyone who understands just how much work can't be anticipated or accounted for in advance.  Slack gives people the freedom to investigate and adopt efficiency improvements when they make themselves apparent - efficiency improvements need to be made to happen, they don't just happen magically, and employees need enough slack in their committed work that they can pursue them.  It gives people the ability to say "yes" rather than "I'm sorry, I am overcommitted" when another team asks for collaboration or assistance.  It gives them time to look at the bigger picture, ask if they should be doing things differently, to address things that have been a constant tax, and so forth.  It also gives them time to take vacations, or be sick, or just to reset.  Without slack, either these things don't happen, or they happen at the expense of the employee, which is ultimately at the expense of the employer (see first element).

My usage of "slack" isn't a perfect map to this book, which is a pretty compelling read with a more refined thesis than I put forth.

A work arrangement that maximized productivity would look like 30ish hours a week most weeks, with the employee having flexibility to decide how they spread those 30 hours over 7 days - what days they worked, what hours of the day they worked, and in what environment they worked those hours.  The company would be very conservative about creating either acute or chronic stressors, and intentionally create a culture and managing decisions intended to minimize them (it is true that the "activation -> action -> resolution" flow for acute stress can have beneficial elements in moderation, but the corporate world is note remotely wise enough to engineer that intentionally; "thrive on stress" is mostly not a thing, and even when it is, is probably not a good thing).  It would additionally be ensuring employees had sufficient slack when thinking about the project planning (how many of the OKR sessions have a "I intentionally undershot every commitment so I have slack for my people to jump on other opportunities"? - we all get pushed to be more ambitious, right? And that's even presuming OKRs are good in the first place, which, well, citation needed as Wikipedia says ).

Instead companies have done the EXACT opposite of all of that.  They took people's flexible work arrangements and tossed them (not all of them have - MS is still mostly team "flexible" but it's also team "lets sublet all of our leased space and spend net less on real estate"), created tremendous acute AND chronic stress with fabricated job insecurity and benefits/compensation constrictions, instituted sweeping austerity that didn't just completely annihilate slack (no solving problems if it requires expending money - actually if there are problems currently solved with spending money spend less and use employee time instead, no staffing up when scope of work changes, etc.) but set project deliverables that exceed what can be delivered in 40 hour work weeks.   So yeah, productivity has declined - it declined because businesses made specific decisions that lead to it declining and are now looking for any other explanation than "we completely ignored all of the scientific data that would lead to better productivity because it didn't confirm our accounting biases". Maybe AI and other efficiency improvements will reverse the decline in some areas, if we have time to implement and adopt them, but even if they do it is also true that we would get more productivity if we adopted them AND created a work environment that humans flourish in.  

Some of this stuff individual managers have little control over - I can't control the overall policies of my company (but I will absolutely email executives with suggestions). But I can wield the new Discretionary Time Off policy for both the good of my employees and of my employer, set team arrangements so that we support how each member works best, and be both aware of the acute and chronic stressors in my employees' lives (work related or otherwise) and both give them room to deal with them so that I am not adding to said stressors, and where possible obviate those stressors.  However, it should not just be individual managers trying to create the conditions for improved productivity - it should be something companies intentionally engineer in their policies, processes, and culture, and very few do.