Why corporate penny-shaving backfires. (Also, how to do a layoff right.)

One of the clearest signs of corporate decline (2010s Corporate America is like 1980s Soviet Russia, in terms of its low morale and lethal overextension) is the number of “innovations” that are just mean-spirited, and seem like prudent cost-cutting but actually do minimal good (and, often, much harm) to the business.

One of these is the practice of pooling vacation and sick leave in a single bucket, “PTO”. Ideally, companies shouldn’t limit vacation or sick time at all– but my experience has shown “unlimited vacation” to correlate with a negative culture. (If I ran a company, it would institute a mandatory vacation policy: four weeks minimum, at least two of those contiguous.) Vacation guidelines need to be set for the same reason that speed limits (even if intentionally under-posted, with moderate violation in mind) need to be there; without them, speed variance would be higher on both ends. So, I’ve accepted the need for vacation “limits”, at least as soft policies; but employers expect their people to either use a vacation day for sick leave, or come into the office while sick, are just being fucking assholes.

These PTO policies are, in my view, reckless and irresponsible. They represent a gamble with employee health that I (as a person with a manageable but irritating disability) find morally repugnant. It’s bad enough to deny rest to someone just because a useless bean-counter wants to save the few hundred dollars paid out for unused vacation when someone leaves the company. But by encouraging the entire workforce to show up while sick and contagious, they subject the otherwise healthy to an unnecessary germ load. Companies with these pooled leave, “PTO”, policies end up with an incredibly sickly workforce. One cold just rolls right into another, and the entire month of February is a haze of snot, coughing, and bad code being committed because half the people at any given time are hopped up on cold meds and really ought to be in bed. It’s not supposed to be this way. This will shock those who suffer in open-plan offices, but an average adult is only supposed to get 2-3 colds per year, not the 4-5 that are normal in an open-plan office (another mean-spirited tech-company “innovation”) or the 7-10 per year that is typical in pooled-leave companies.

The math shows that PTO policies are a raw deal even for the employer. In a decently-run company with an honor-system sick leave policy, an average healthy adult might have to take 5 days off due to illness per year. (I miss, despite my health problems, fewer than that.) Under PTO, people push themselves to come in and only stay home if they’re really sick. Let’s say that they’re now getting 8 colds per year instead of the average 2. (That’s not an unreasonable assumption, for a PTO shop.) Only 2 or 3 days are called-off, but there are a good 24-32 days in which the employee is functioning below 50 percent efficiency. Then there are the morale issues, and the general perception that employees will form of the company as a sickly, lethargic place; and the (mostly unintentional) collective discovery of how low a level of performance will be tolerated. January’s no longer about skiing on the weekends and making big plans and enjoying the long golden hour… while working hard, because one is refreshed. It’s the new August; fucking nothing gets done because even though everyone’s in the office, they’re all fucking sick with that one-rolls-into-another months-long cold. That’s what PTO policies bring: a polar vortex of sick.

Why, if they’re so awful, do companies use them? Because HR departments often justify their existence by externalizing costs elsewhere in the company, and claiming they saved money. So-called “performance improvement plans” (PIPs) are a prime example of this. The purpose of the PIP is not to improve the employee. Saving the employee would require humiliating the manager, and very few people have the courage to break rank like that. Once the PIP is written, the employee’s reputation is ruined, making mobility or promotion impossible. The employee is stuck in a war with his manager (and, possibly, team) that he will almost certainly lose, but he can make others lose along the way. To the company, a four-month severance package is far cheaper than the risk that comes along with having a “walking dead” employee, pissing all over morale and possibly sabotaging the business, in the office for a month. So why do PIPs, which don’t even work for their designed intention (legal risk mitigation) unless designed and implemented by extremely astute legal counsel, remain common? Well, PIPs a loss to the company, even compared to “gold-plated” severance plans. We’ve established that. But they allow the HR department to claim that it “saved money” on severance payments (a relatively small operational cost, except when top executives are involved) while the costs are externalized to the manager and team that must deal with a now-toxic (and if already toxic before the PIP, now overtly destructive) employee. PTO policies work the same way. The office becomes lethargic, miserable, and sickly, but HR can point to the few hundred dollars saved on vacation payouts and call it a win.

On that, it’s worth noting that these pooled-leave policies aren’t actually about sick employees. People between the ages of 25 and 50 don’t get sick that often, and companies don’t care about that small loss. However, their children, and their parents, are more likely to get sick. PTO policies aren’t put in place to punish young people for getting colds. They’re there to deter people with kids, people with chronic health problems, and people with sick parents from taking the job. Like open-plan offices and the anxiety-inducing micromanagement often given the name of “Agile”, it’s back-door age and disability discrimination. The company that institutes a PTO policy doesn’t care about a stray cold; but it doesn’t want to hire someone with a special-needs child. Even if the latter is an absolute rock star, the HR department can justify itself by saying it helped the company dodge a bullet.

Let’s talk about cost cutting more generally, because I’m smarter than 99.99% of the fuckers who run companies in this world and I have something important to say.

Companies don’t fail because they spend too much money. “It ran out of money” is the proximate cause, not the ultimate one. Some fail when they cease to excel and inspire (but others continue beyond that point). Some fail, when they are small, because of bad luck. Mostly, though, they fail because of complexity: rules that don’t make sense and block useful work from being done, power relationships that turn toxic and, yes, recurring commitments and expenses that can’t be afforded (and must be cut). Cutting complexity rather than cost should be the end goal, however. I like to live with few possessions not because I can’t afford to spend the money (I can) but because I don’t want to deal with the complexity that they will inject into my life. It’s the same with business. Uncontrolled complexity will cause uncontrolled costs and ultimately bring about a company’s demise. What does this mean about cutting costs, which MBAs love to do? Sometimes it’s great to cut costs. Who doesn’t like cutting “waste”? The problem there is that there actually isn’t much obvious waste to be cut, so after that, one has to focus and decide on which elements of complexity are unneeded, with the understanding that, yes, some people will be hurt and upset. Do we need to compete in 25 businesses, when we’re only viable in two? This will also cut costs (and, sadly, often jobs).

The problem, see, is that most of the corporate penny-shaving increases complexity. A few dollars are saved, but at the cost of irritation and lethargy and confusion. People waste time working around new rules intended to save trivial amounts of money. The worst is when a company cuts staff but refuses to reduce its internal complexity. This requires a smaller team to do more work– often, unfamiliar work that they’re not especially good at or keen on doing; people were well-matched to tasks before the shuffle, but that balance has gone away. The career incoherencies and personality conflicts that emerge are… one form of complexity.

The problem is that most corporate executives are “seagull bosses” (swoop, poop, and fly away) who see their companies and jobs in a simple way: cut costs. (Increasing revenue is also a strategy, but that’s really hard in comparison.) A year later, the company is still failing not because it failed to cut enough costs or people, but because it never did anything about the junk complexity that was destroying it in the first place.

Let’s talk about layoffs. The growth of complexity is often exponential, and firms inevitably get to a place where they are too complex (and, a symptom of this is that operations are too expensive) to survive. The result is that it needs to lay people off. Now, layoffs suck. They really fucking do. But there’s a right way and a wrong way to execute one. To do a layoff right, the company needs to cut complexity and cut people. (Otherwise, it will have more complexity per capita, the best people will get fed up and leave, and the death spiral begins.) It also needs to cut the right complexity; all the stuff that isn’t useful.

Ideally, the cutting of people and cutting of complexity would be tied together. Unnecessary business units being cut usually means that people staffed on them are the ones let go. The problem is that that’s not very fair, because it means that good people, who just happened to be in the wrong place, will lose their jobs. (I’d argue that one should solve this by offering generous severance, but we already know why that isn’t a popular option, though it should be.) The result is that when people see their business area coming into question, they get political. Of course this software company needs a basket-weaving division! In-fighting begins. Tempers flare. From the top, the water gets very muddy and it’s impossible to see what the company really looks like, because everyone’s feeding biased information to the executives. (I’m assuming that the executive who must implement the cuts is acting in good faith, which is not always true.) What this means is that the crucial decision– what business complexity are we going to do without?– can’t be subject to a discussion. Debate won’t work. It will just get word out that job cuts are coming, and political behavior will result. The horrible, iron fact is that this calls for temporary autocracy. The leader must make that call in one fell swoop. No second guessing, no looking back. This is the change we need to make in order to survive. Good people will be let go, and it really sucks. However, seeing as it’s impossible to execute a large-scale layoff without getting rid of some good people, I think the adult thing to do is write generous severance packages.

Cutting complexity is hard. It requires a lot of thought. Given that the information must be gathered by the chief executive without tipping anyone off, and that complex organisms are (by definition) hard to factor, it’s really hard to get the cuts right. Since the decision must be made on imperfect information, it’s a given that it usually won’t be the optimal cut. It just has to be good enough (that is, removing enough complexity with minimal harm to revenue or operations) that the company is in better health.

Cutting people, on the other hand, is much easier. You just tell them that they don’t have jobs anymore. Some don’t deserve it, some cry, some sue, and some blog about it but, on the whole, it’s not actually the hard part of the job. This provides, as an appealing but destructive option, the lazy layoff. In a lazy layoff, the business cuts people but doesn’t cut complexity. It just expects more work from everyone. All departments lose a few people! All “survivors” now have to do the work of their fallen brethren! The too-much-complexity problem, the issue that got us to the layoff in the first place… will figure itself out. (It never does.)

Stack ranking is a magical, horrible solution to the problem. What if one could do a lazy layoff but always cull the “worst” people? After all, some people are of negative value, especially considering the complexity load (in personality conflicts, shoddy work) they induce. The miracle of stack ranking is that it turns a layoff– otherwise, a hard decision guaranteed to put some good people out of work– into an SQL query. SELECT name FROM Employee WHERE perf <= 3.2. Since the soothsaying of stack ranking has already declared the people let-go as bottom-X-percent performers, there’s no remorse in culling them. They were dead weight”. Over time, stack ranking evolves into a rolling, continuous lazy layoff that happens periodically (“rank-and-yank”).

It’s also dishonest. There are an ungodly number of large technology companies (over 1,000) that claim to have “never had a layoff”. That just isn’t fucking true. Even if the CEO was Jesus Christ himself, he’d have to lay people off because that’s just how business works. Tech-company sleazes just refuse to use the word “layoff”, for fear of losing their “always expanding, always looking for the best talent!” image. So they call it a “low performer initiative” (stack ranking, PIPs, eventual firings). What a “low-performer initiative” (or stack ranking, which is a chronic LPI) inevitably devolves into is a witch hunt that turns the organization into pure House of Cards politics. Yes, most companies have about 10 percent who are incompetent or toxic or terminally mediocre and should be sent out the door. Figuring which 10 percent those people are, is not easy. People who are truly toxic generally have several years’ worth of experience drawing a salary without doing anything, and that’s a skill that improves with time. They’re really good at sucking (and not getting caught). They’re adept political players. They’ve had to be; the alternative would have been to have grown a work ethic. Most of what we as humans define as social acceptability is our ethical immune system, which can catch and punish the small-fry offenders but can’t do a thing about the cancer cells (psychopaths, parasites) that have evolved to the point of being able to evade or even redirect that rejection impulse. The question of how to get that toxic 10 percent out is an unsolved one, and I don’t have space to tackle it now, but the answer is definitely not stack ranking, which will always clobber several unlucky good-faith employees for every genuine problem employee it roots out.

Moreover, stack ranking has negative permanent effects. Even when not tied to a hard firing percentage, its major business purpose is still to identify the bottom X percent, should a lazy layoff be needed. It’s a reasonable bet that unless things really go to shit, X will be 5 or 10 or maybe 20– but not 50. So stack ranking is really about the bottom. The difference between the 25th percentile and 95th percentile, in stack ranking, really shouldn’t matter. Don’t get me wrong: a 95th-percentile worker is often highly valuable and should be rewarded. I just don’t have any faith in the ability of stack ranking to detect her, just as I know some incredibly smart people who got mediocre SAT scores. Stack ranking is all about putting people at the bottom, not the top. (Top performers don’t need it and don’t get anything from it.)

The danger of garbage data (and, #YesAllData generated by stack ranking is garbage) is that people tend to use it as if it were truth. The 25th-percentile employee isn’t bad enough to get fired… but no one will take him for a transfer, because the “objective” record says he’s a slacker. The result of this– in conjunction with closed allocation, which is already a bad starting point– is permanent internal immobility. People with mediocre reviews can’t transfer because the manager of the target team would prefer a new hire (with no political strings attached) over a sub-50th-percentile internal. People with great reviews don’t transfer for fear of upsetting the gravy train of bonuses, promotions, and managerial favoritism. Team assignments become permanent, and people divide into warring tribes instead of collaborating. This total immobility also makes it impossible to do a layoff the right way (cutting complexity) because people develop extreme attachments to projects and policies that, if they were mobile and therefore disinterested, they’d realize ought to be cut. It becomes politically intractable to do the right thing, or even for the CEO to figure out what the right thing is. I’d argue, in fact, that performance reviews shouldn’t be part of a transfer packet at all. The added use of questionable, politically-laced “information” is just not worth the toxicity of putting that into policy.

A company with a warring-departments dynamic might seem like a streamlined, efficient, and (most importantly) less complex company. It doesn’t have the promiscuous social graph you might expect to see in an open allocation company. People know where they are, who they report to, and who their friends and enemies are. The problem, with this insight, is that there’s hot complexity and cold complexity. Cold complexity is passive and occasionally annoying, like a law from 1890 that doesn’t make sense and is effectively never enforced. When people collaborate “too much” and the social graph of the company seems to have “too many” edges, there’s some cold complexity there. It’s generally not harmful. Open allocation tends to generate some cold complexity. Rather than metastasize into an existential threat to the company, it will fade out of existence over time. Hot complexity, which usually occurs in an adversarial context, is a kind that generates more complexity. Its high temperature means there will be more entropy in the system. Example: a conflict (heat) emerges. That, alone, makes the social graph more complex because there are more edges of negativity. Systems and rules are put in place to try to resolve it, but those tend to have two effects. First, they bring more people (those who had no role in the initial conflict, but are affected by the rules) into the fights. Second, the conflicting needs or desires of the adversarial parties are rarely addressed, so both sides just game the new system, which creates more complexity (and more rules). Negativity and internal competition create the hot complexity that can ruin a company more quickly than an executive (even if acting with the best intentions) can address it.

Finally, one thing worth noting is the Welch Effect (named for Jack Welch, the inventor of stack-ranking). It’s one of my favorite topics because it has actually affected me. The Welch Effect pertains to the fact that when a broad-based layoff occurs, the people most likely to be let go aren’t the worst (or best) performers, but newest members of macroscopically underperforming teams. Layoffs (and stack ranking) generally propagate down the hierarchy. Upper management disburses bonuses, raises, and layoff quotas based on the macroscopic performance of the departments under it, and at each level, the node operators (managers) slice the numbers based on how well they think each suborganization did (plus or minus various political modifiers). At the middle-management layer, one level separated from the non-managerial “leaves”, it’s the worst-performing teams that have to vote the most people off the island. It tends to be those most recently hired who get the axe. This isn’t especially unfair or wrong, for that middle manager; there’s often no better way to do it than to strike the least-embedded, least-invested junior hire.

The end result of the Welch Effect, however, is that the people let go are often those who had the least to do with their team’s underperformance. (It may be a weak team, it may be a good team with a bad manager, or it may be an unlucky team.) They weren’t even there for very long! It doesn’t cause the firm to lay off good people, but it doesn’t help it lay off bad people either. It has roughly the same effect as a purely seniority-based layoff, for the company as a whole. Random new joiners are the ones who are shown out the door. It’s bad to lose them, but it rarely costs the company critical personnel. Its effect on that team is more visibly negative: teams that lose a lot of people during layoffs get a public stink about them, and people lose the interest in joining or even helping them– who wants to work for, or even assist, a manager who can’t protect his people?– so the underperforming team becomes even more underperforming. There are also morale issues with the Welch Effect. When people who recently joined lose their jobs (especially if they’re fired “for performance” without a severance) it makes the company seem unfair, random, and capricious. The ones let go were the ones who never had the chance to prove themselves. In a one-off layoff, this isn’t so destructive. The Welch Effected usually move on to better jobs anyway. However, when a company lays off in many small cuts, or disguises a layoff as a “low-performer initiative”, the Welch Effect firings demolish belief in meritocracy.

That, right there, explains why I get so much flak over how I left Google. Technically, I wasn’t fired. But I had a disliked, underdelivering manager who couldn’t get calibration points for his people (a macroscopic issue that I had nothing to do with) and I was the newest on the team, so I got a bad score (despite being promised a reasonable one– a respectable 3.4, if it matters– by that manager). Classic Welch Effect. I left. After I was gone I “leaked” the existence of stack ranking within Google. I wasn’t the first to mention that it existed there, but I publicized it enough to become the (unintentional) slayer of Google Exceptionalism and, to a number of people I’ve never met and to whom I’ve never done any wrong, Public Enemy #1. I was a prominent (and, after things went bad, fairly obnoxious) Welch Effectee, and my willingness to share my experience changed Google’s image forever. It’s not a disliked company (nor should it be) but its exceptionalism is gone. Should I have done all that? Probably not. Is Google a horrible company? No. It’s above average for the software industry (which is not an endorsement, but not damnation either.) Also, my experiences are three years old at this point, so don’t take them too seriously. As of November 2011, Google had stack ranking and closed allocation. It may have abolished those practices and, if it has, then I’d strongly recommend it as a place to work. It has some brilliant people and I respect them immensely.

In an ideal world, there would be no layoffs or contractions. In the real world, layoffs have to happen, and it’s best to do them honestly (i.e. don’t shit on departing employees’ reputations by calling it a “low performer initiative”). As with more minor forms of cost-cutting (e.g. new policies encouraging frugality) it can only be done if complexity (that being the cause of the organization’s underperformance) is reduced as well. That is the only kind of corporate change that can reverse underperformance: complexity reduction.

If complexity reduction is the only way out, then why is it so rare? Why do companies so willingly create personnel and regulatory complexity just to shave pennies off their expenses? I’m going to draw from my (very novice) Buddhist understanding to answer this one. When the clutter is cleared away… what is left? Phrases used to define it (“sky-like nature of the mind”) only explain it well to people who’ve experienced it. Just trust me that there is a state of consciousness that can be attained when gross thoughts are swept away, leaving something more pure and primal. Its clarity can be terrifying, especially the first time it is experienced. I really exist. I’m not just a cloud of emotions and thoughts and meat. (I won’t get into death and reincarnation and nirvana here. That goes farther than I need, for now. Qualia, or existence itself, as opposed my body hosting some sort of philosophical zombie, is both miraculous and the only miracle I actually believe in.) Clarity. Essence. Those are the things you risk encountering with simplicity. That’s a good thing, but it’s scary. There is a weird, paradoxical thing called “relaxation-induced anxiety” that can pop up here. I’ve fought it (and had some nasty motherfuckers of panic attacks) and won and I’m better for my battles, but none of this is easy.

So much of what keeps people mired in their obsessions and addictions and petty contests is an aversion to confronting what they really are, a journey that might harrow them into excellence. I am actually going to age and die. Death can happen at any time, and almost certainly it will feel “too soon”. I have to do something, now, that really fucking matters. This minute counts, because I may not get another in this life. People are actually addicted to their petty anxieties that distract them from the deeper but simpler questions. If you remove all the clutter on the worktable, you have to actually look at the table itself, and you have to confront the ambitions that impelled you to buy it, the projects you imagined yourself using it for (but that you never got around to). This, for many people, is really fucking hard. It’s emotionally difficult to look at the table and confront what one didn’t achieve, and it’s so much easier to just leave the clutter around (and to blame it).

Successful simplicity leads to, “What now?” The workbench is clear; what are we going to do with it? For an organization, such simplicity risks forcing it to contend with the matter of its purpose, and the question of whether it is excelling (and, relatedly, whether it should). That’s a hard thing to do for one person. It’s astronomically more difficult for a group of people with opposing interests, and among whom excellence is sure to be a dirty word (there are always powerful people who prefer rent-seeking complacency). It’s not surprising, then, that most corporate executives say “fuck it” on the excellence question and, instead, decide it suffices to earn their keep to squeeze employees with mindless cost-cutting policies: pooled sick leave and vacation, “employee contributions” on health plans, and other hot messes that just ruin everything. It feels like something is getting done, though. Useless complexity is, in that way, existentially anxiolytic and addictive. That’s why it’s so hard to kill. But it, if allowed to live, will kill. It can enervate a person into decoherence and inaction, and it will reduce a company to a pile of legacy complexity generated by self-serving agents (mostly, executives). Then it falls under the MacLeod-Gervais-Rao-Church theory of the nihilistic corporation; the political whirlpool that remains once an organization has lost its purpose for existing.

At 4528 words, I’ve said enough.

About these ads

16 thoughts on “Why corporate penny-shaving backfires. (Also, how to do a layoff right.)

  1. I love your rants. I love the definitiveness of much of this — and I agree with your conclusions.

    I just wish I could figure out a way to translate them sufficiently into corporate-ese to express them usefully to mainstream corporate workers :-)

    • I wish I could figure out how to deal with the complexity without feeling like I’m going insane. I am caught in a territorial war right now. I am being asked to do things that make not a bit of sense except as a way of protecting territory, and I am not sure I can manage to motivate myself to do them.

      • The problem is that “lean” and “agile” often increase micromanagement and, by penalizing long-term-oriented work, decrease morale. The result is that you end up getting more of the worst kind of complexity: personnel/relational complexity.

  2. We are really on the same page. I hate pooled time off. Jesus, if I need to go to the doctor or just rest for a day or two, don’t make me eat up my vacation, especially when I’m a new hire or coming in because of an acquisition. Don’t make it hard to take vacation and then cap the number of hours I can bank. Don’t put in high-deductible health plans and then dribble out the HSA employer contribution over 26 pay periods when the law allows you to put the money in at the beginning of the year. Don’t shrink cubicle sizes because “we did a study and our cubicle sizes are a little bigger than average.”

    Don’t fix a problem by generating two extra hours of work for every engineer, every week. Don’t make me give you my status in three different formats every week.

    • >our cubicle sizes are a little bigger than average
      Remember, if a company is at the median, than 1/2 of other companies are better. This applies to salary, and anything else.

      Another money saving technique that often backfires, is starving engineering of equipment, in a setting where people need more than their a computer and a monitor or three to do their work. Maybe not enough tests are run, or maybe people just sit around waiting for their turn.

      If California the sick time issue is not quite so intense, as the winters are milder, but you still people in the office who should not be. Or people who work from home when they clearly should be resting for the whole day.

  3. Doing a bunch of complicated crap that doesn’t really need to be done is the sin qua non of modern American society at large, not just the corporate micro level. It’s textbook Joseph Tainter, “The Collapse of Complex Societies.” I’d guess that nearly half of “economic activity” is essentially waste of resources at this point. It’s all hack fixes layered on hack fixes.

    Businesses that have transitioned to mining diminishing returns will inevitably grow in complexity and pain. No human society anywhere ever resorts to the simplicity of cutting back and rationalizing expectations. No, we’re always one or two more policy tweaks away from restoring the good old days of high growth.

  4. For the ‘I have an annoying cold’ level of illness most people just work from home and we tell them to work from home. When you can do that, it’s pretty rare to need to take real sick days as an employee. I think I maybe get 1 cold a year as a result? And the place I work at has a bunch of other bad-ish policies such as open offices, ‘unlimited’ pto, some retarded executive management moves, etc.

    • I typically work at home if I don’t feel well, and I will even sign off and take a nap if I feel I need to (without putting in FTO hours–considering the number of late nights I work otherwise). The problem is other people sitting near me come in and you can hear them hacking and coughing. Then I will get sick 2-3 days later.

  5. You sort of go off on a tangent there at the end.

    I think what you call the “Welch Effect” existed before Welch. Welch just created a tool to facilitate it. Other tools to facilitate it include tenure and union contracts.

  6. This is a fantastic site to read. I would love to hoist beers with you some day, you have a fascinating intelligence and perspective.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s