Game Theory in Management

When I saw ProjectManagement.com’s theme for June, Problem Solving/Wicked Problems, I asked myself what, exactly, makes a given problem “wicked,” by which I assumed to mean “excessively difficult,” as opposed to a problem that goes away if you pour water on it, as in The Wizard of Oz. I think the answer lies in the specific problem’s novelty: problems that are easy to address and correct may not have started out that way, but, once a workable solution was developed and demonstrated, subsequent appearances of said problem were readily addressed with a known remedy. Problems without precedent are a different matter. I remember seeing a definition of intelligence that asserted that true intelligence lay in the ability to encounter a novel situation or problem, recognize how it is similar to an analogous, previously-encountered difficulty, and providing a response that is either consistent with the previously-observed remedy, or the opposite of a previously-attempted but failed response. If you need to drive to work, and your Mazda automobile won’t start, and you recall that a Honda Accord of similar vintage had occasional battery problems, a suitable remedy might be to get a jump from a neighbor. If, however, your Mazda has somehow transformed itself into a large pterodactyl, seeking a pair of jumper cables would have to be considered a sub-optimal response.

Meanwhile, Back In The PM Robot World…

This aspect of truly difficult problems is one of the things that gives me hope that we PMs won’t be replaced by robots, at least not in the near future. Robots are essentially mobile computers, and computers (in 2021, anyway) aren’t that good at nuance, or recognizing analogous situations. Yes, I’m aware of the concept of Fuzzy Logic, where certain parameters do not have to be precisely stated or met in order to illicit a specific response, but this characteristic is a long way away from an ability to address unusual or unique problems. Interestingly, one of the definitions of the word “robot” refers to humans who behave in an automatic, reflexive fashion, which brings me back to the PM Robot world. I have worked with my share of PM robots (going by the latter definition), people utterly convinced that the solution to the problems they are facing is fully documented in procedures, the PMBOK Guide®, or even ProjectManagement.com blogs, and only awaits discovery. For these people, deviating from an approved template approach to PM problems represents some form of heresy, an unacceptable bending of the rules, automatically exposing its perpetrators as ignorant or incapable.

My take is more aligned with the Pareto Principle, this derivative being that, yeah, around 80% of the problems most PMs face has been dealt with before (“What has been will be again, what has been done will be done again; there is nothing new under the sun.”[i]). As for the other 20%, most of that set has also been encountered previously, it’s just that the most appropriate way to approach said problem hasn’t been agreed to, or documented, rendering it effectively unique. As for the residual, now we’re talking about the truly singular problem, in need of some sort of problem-solving technique.

Unfortunately, many problem-solving techniques don’t necessarily perform well in the PM world. Mazda is said to have engaged in an extensive trial-and-error approach to finding a suitable material to make its rotary engines’ rotor apex seals, resulting in hundreds of ruined engine casings before a suitable apex seal was discovered. But that was for an engineering process project, meaning that, once the apex seal problem was resolved, its “answer” could be replicated over and over. Most PMs, by definition, face work that is itself unique, aimed at delivering a product or service over a set period of time with a known quantity of resources. The trial-and-error approach, depending on the number of parameters being evaluated in combination, can take a very long time, and burn through many resources. Indeed, the team of engineers working the rotor seal problem at Mazda gained the reputation for wasting money that could have been better spent on other projects.[ii] Another problem-solving technique, using the process of elimination to discover a suitable technical approach from among a known set of alternatives, has the same time-and-resource-consuming issues.

So, what do you do if your Mazda has turned into a pterodactyl? Catch it, and sell it (it’s yours, after all, assuming you owned the car from whence it came). A living pterodactyl is, no doubt, worth millions, meaning you probably won’t have to drive to work for some time. Catching it shouldn’t be that big of a deal, as average adult pterodactyls were about the size of a large seagull. Now, if your car has turned into a Pteranodon, which could have a wingspan of over seven meters, that’s an entirely different problem.

But fetching a set of jumper cables is still not the solution.

[i] Ecclesiates 1:9

[ii] Retrieved from https://www.mazda.com/en/innovation/stories/rotary/newfrontier/ on June 14, 2021.

Referring to GTIM’s derivative of the Pareto Principle, that the 80^th percentile best managers with access to 20% of the information needed to obviate a given decision will be consistently out-performed by the 20^th percentile worst managers with access to 80% of the information so needed, I would like to point out that that does not necessarily mean that the latter category will always get it right. Even with 80% of the information needed to obviate a given decision, for the poor management set, a 20% missing or unavailable information window is more than enough to mess up any decision, particularly big ones. So, if you happen to be lucky enough to belong to an organization that takes its PM information systems (PMISs) seriously, and basic scope, cost, and schedule performance is available, timely, and accurate, what more can be done to ensure that the best available options are taken, and overall success virtually guaranteed? The answer is as brutal as it is inescapable: you’ve got to get rid of the bad PMs.

But, given that so much of management in general and PM in particular is at least somewhat subjective, how are the “bad” PMs to be recognized apart from “good” ones that just happen to be on an unlucky streak? I think I can help here. What follows is a few simple tests that can reveal the truly proficient and talented managers and leaders from the poor ones, those that, due to miseducation, poor attitude, or just plain lack of talent, should be removed from the PM pool.

Test #1: Are they trustworthy? Plenty of PMs will encounter an overrun or a late project finish, and it’s usually rather difficult to discern if that condition came about due to poor management, or a myriad of other factors. But here’s the Litmus Test: did the PM provide early warning of the overrun/late finish? Both Critical Path and Earned Value Methodologies (CPM/EVM) have the (uncanny, really) ability to accurately forecast overruns or late finishes, months in advance, if they are allowed. CPM networks can have this ability thwarted by imposing no-later-than constraints on key milestones, returning plenty of the so-called negative float when the forward pass/backward pass calculations are performed, but leaving the all-important finish milestone unmoved. Similarly, EVM systems that allow a “manager’s Estimate at Completion” (EAC) to be used instead of the calculated version will obscure the overrun until it’s too late for anyone to do anything about it. So, if the project in question did overrun or come in late, AND the PM provided no early warning, they’re untrustworthy, and should probably be avoided.

Test #2: Given the choice between successful project completion and strict adherence to procedure, which would they choose? I want to make clear right off the bat that I’m not talking about safety or security procedures or rules here. Those often come about due to something really unfortunate having already happened, and are in place to prevent a recurrence. Rather, what I’m talking about are things like a decree issued by a retail store where I worked for about nine months, that all employees had to put a company-touting bumper sticker on their personal cars. The top two salesmen refused. They were fired. The person doing the firing was a very poor manager. That instance was rather blatant, but the take-away test I wish to assert is this: if the manager cares more about being able to deflect or defend weak performance by pointing to strict adherence to non-safety or security-related policy, then they are probably not in the aforementioned 80th percentile best manager set.

Test #3: Is the manager extremely popular? This test is a bit more nuanced, but I believe it still applies. Good managers as well as talented Project Team members will be attracted to organizations that behave more like meritocracies. Conversely, poor managers and underachieving Project Team members will avoid such environments, preferring instead to “work” in an environment where things other than merit represent the vehicle for advancement, such as demonstrated personal loyalty to the upper levels of management. In the Maccoby archetypes[i], it’s the difference between arranging to have Craftsmen on your team, and allowing Company Men. It’s not that the Company Man archetype – the kind of team member who tends to assume the persona of the company’s culture – are necessarily inadequate. Rather, it’s the poor managers who will value personal loyalty above actual performance. A team of truly talented individuals is bound to have some differences. A group of sycophants usually won’t.

According to The Free Dictionary, one of the definitions of the word “tell” is:

Games An unintentional or unconsciously exhibited behavior that reveals or betrays one's state of mind, as when playing poker.[ii]

In the big game of Good PM/Bad PM poker, these three tells may well inform Project Team members (and executives) of the hand their PMs are holding, and how they’re likely going to play them.

Bet accordingly.

[i] Maccoby, Michael. The Gamesman: The New Corporate Leaders. New York: Simon and Schuster,1976

[ii] Retrieved from https://www.thefreedictionary.com/tell on June 7, 2021, 19:05 MDT.

In last week’s blog I asked whatever happened to “life cycle estimating,” and essentially made the point that it was a management fad, similar in performance and (ironically) life cycle to other management fads. Conventional wisdom holds that whether or not a given hypothesis in the management science realm is valid or a craze is the proverbial test of time – if a business model theory is valid, it will probably make its way into a significant percentage of companies – but even here some invalid but long-lived hypotheses can become regarded as theories, taught by colleges and adapted by many (if not most) organizations. This state of affairs afflicts the management world more acutely than others due to the near-impossibility to conduct a controlled-environment experiment – there are simply too many parameters to identify, much less quantify or control, in the free, open, and competitive marketplace. So, what’s the reading PM population, exposed to hundreds of guidance documents, journal articles, and, yes, blogs, to do when evaluating all of these ideas? I believe I have a usable tool for performing such a discernment: the null hypothesis.

According to Ann Marie Helmenstein, PhD,

In a scientific experiment, the null hypothesis is the proposition that there is no effect or no relationship between phenomena or populations. If the null hypothesis is true, any observed difference in phenomena or populations would be due to sampling error (random chance) or experimental error. The null hypothesis is useful because it can be tested and found to be false, which then implies that there is a relationship between the observed data.[i]

As an example, let’s look at one of the aforementioned long-lived and widely-accepted theories that’s being taught at business schools and has been broadly used throughout many industries, but is, in my opinion, invalid, that the point of all management is to “maximize shareholder wealth.” Can this statement’s null hypothesis be disproven? If “maximize shareholder wealth” had a null hypothesis, I would posit that it would be “there exists several valid management techniques that are highly beneficial to the organization that have a negative impact on that organization’s shareholder wealth.”

Consider the following payoff grid:

Let’s dispense with Scenario C right off the bat. Any management action that harms the organization and decreases shareholder wealth is patently invalid. Similarly, I think we can all agree that Scenario B represents good management.

But to disprove the null hypothesis as stated, one would have to prove that both Scenarios A and D are empty sets, i.e., there are no valid management techniques that benefit the organization while decreasing shareholder wealth, or that the original assertion should have been re-phrased as “maximize shareholder wealth, even if it harms the organization.” Is it true that the A bin is empty, or the original theory cannot stand as phrased?

As for Scenario A, consider a new business. Its owner(s), technically speaking, deplete shareholder wealth beginning the accounting period after they have purchased inventory, due to depreciation. But this effect carries on beyond mere accounting nuance: typically, new businesses’ owners will work themselves excessively, with little or no recompence, in the effort to attract customers away from the competition. Sometimes they will mark down their inventories to cost, or even below, in order to attain market share, a widely-used tactic that, by definition, decreases shareholder wealth. Then there’s my oft-referenced example of the hostile takeover, where the acquiring business will often take a significant hit to their reserves and share price, while the target organizations will almost always see an increase in shareholder wealth. And yet, the organizations performing the takeover do so fully expecting this to happen, and the target organizations will typically resist. If Scenario A is empty, how does one explain these observable phenomena?

Strategies that fall in to Scenario D are common, from “going out of business” sales to actions that can lead to a “piercing the corporate veil.” According the The Balance Small Business website,

Corporations are separate entities from their shareholders, and in normal circumstances, if a corporation is sued, the individual shareholders and officers cannot be brought into the lawsuit. But there are cases in which the corporation's officers and shareholders could be sued for negligence or for debts; the action of bringing in these shareholders to be sued is called "piercing the corporate veil" or "lifting the corporate veil."

In the same way as corporate shareholders, the owners of a limited liability company (LLC), called "members," may also be sued personally for business debts and actions.[ii]

Given these observable management phenomena, Scenarios A and D are not empty sets; therefore, the null hypothesis is not disproven, indicating that a basic tenet of business school teaching is likely invalid.

So, that’s my management fad litmus test: just ask yourself, for any given hypothesis, has its null hypothesis been disproven, or even articulated? If not, you may want to be wary of participating in a fad.

[i] Retrieved from https://www.thoughtco.com/definition-of-null-hypothesis-and-examples-605436 on May 23, 2021, 17:50 MDT.

[ii] Retrieved from https://www.thebalancesmb.com/piercing-the-corporate-veil-definition-398410 , May 23, 2021, 18:23 MDT.

GTIM Nation knows of my theory that there are three kinds of management, each with its own goals and Management Information System (MIS) needs:

• Asset Management seeks to “maximize shareholder wealth,” and uses the General Ledger as its main information source,
• Project Management (yaay!) is all about accomplishing the customers’ goals of scope, within their constraints of cost and schedule (the famed “triple constraint,”), using Earned Value and Critical Path information, and
• Strategic Management targets greater market share, which is a company’s income divided by its industry’s total revenue, and use all of the tools from the marketing industry to advance towards that goal.

Meanwhile, Back In The Movie Industry…

I want to talk about movies that have as a key component an aspect of these types of management. First up, let’s take a look at The Accountant (2016), a film that has, yes, a forensic accountant as its central character and protagonist. But this is no ordinary accountant, no siree. This accountant is not only brilliant, but is advanced in the martial arts, is an art connoisseur, and is physically attractive. The character is, in fact, played by none other than Ben Affleck – think of Batman as your genius accountant, and you have a good grasp on this character. The Accountant is an action-adventure film, as Affleck’s character uncovers massive amounts of evil-doing and evil-doers, who, naturally, want to kill him and his loved ones. I’m going to go out on a limb here and assert that the vast majority of Certified Public Accountants do not have, as a significant component to their everyday professional lives, an aspect of immediate physical danger, or at least not one that requires advanced martial arts capabilities in order to survive day-to-day. Indeed, when it comes to likely protagonists for action-adventure movies, the typical lists include secret agents, military personnel, police officers or private detectives, mutants, aliens (Superman), and even children seeking to become famous guitar players (Coco, 2017) – but accountants, in general, do not immediately pop to mind when one considers jobs that entail ubiquitous personal danger.

The next movie sort-of deals with Strategic Management. Pretty Woman (1990) is a romantic comedy, starring Richard Gere and Julia Roberts as a “high-powered corporate raider”[i] and a prostitute, respectively. As described (rather disingenuously) in the movie, corporate raiders seek to obtain a majority share in struggling companies, then force them in to bankruptcy in order to sell off their assets, presumably for more value than was paid to obtain the majority share of stocks. But this is a rather naïve view of the purpose of “corporate raiding,” also known as a hostile takeover. If a profit could be made from simply buying out companies that appeared to be “undervalued” based on their balance sheet, then it wouldn’t happen nearly as often as it does. The key component receiving short-shrift here is market share. It’s obviously valuable, but can’t be quantified in the general ledger as an asset – otherwise, it could be taxed. By buying out a competitor and either assuming or liquidating its assets, the organization performing the takeover frees up market share. It’s why performing a hostile takeover remains attractive even in those instances where the price per share jumps up (as it almost always does) to the point that the acquiring organization takes a loss in their own books. And yet, here’s a highly prominent movie, with A-list Hollywood stars, based on a misinterpretation of basic Strategic Management goals and techniques.

Meanwhile, Back In The Project Management World…

Which brings us to Project Management. Where’s our movie? We wouldn’t even have to stretch its predicate to anything as improbable as accountants being the first line of detection for massive funding of terrorists, or “corporate raiders” performing a business model analogy to being a prostitute.  I readily concede that first-rate movies have been made about famous projects (e.g., Fat Man and Little Boy, 1989), but project management itself has not garnered a treatment analogous to the other two types of management.

Besides, PM is way more dangerous to implement than Asset Management. Don’t believe me? Watch what happens when a PMO Director mandates that all PMs are required to have filled-out and signed Work Packages prior to starting work! PM is also way more virtuous than Strategic Management. Just watch a few episodes of Mad Men, and compare those goings-on to your own PMO. I think PMI® Publishing should stop with the academic analysis stuff for a little while, and send out a call for movie scripts that highlight PM. No, I don’t want any renumeration from the Institute for this brilliant, potentially millions-making idea.

I would, however, like for my character to be played by Ben Affleck.

[i] Wikipedia contributors. (2021, May 15). Pretty Woman. In Wikipedia, The Free Encyclopedia. Retrieved 01:42, May 17, 2021, from https://en.wikipedia.org/w/index.php?title=Pretty_Woman&oldid=1023299360

GTIM Nation is well aware that, in my self-appointed role of defense-against-management-fads guard dog, I have snarled at, barked at, and thrown up on several popular initiatives, mostly contra risk management (no initial caps), misapplication of Generally Accepted Accounting Principles, and those who would eschew Earned Value’s capability to produce an Estimate at Completion (EAC) in favor of using burn rates, or re-estimating remaining budget and adding that figure to the cumulative actual costs. Despite my reasonable opposition, unfortunately, these practices have yet to be roundly denounced or, better still, jettisoned from and ignored by the community of management scientists writ large. So, when a management fad does actually receive the left-behind treatment, I think it’s illustrative to go back and see what all the original fuss was all about, how the fad gained traction, and what ultimately led to its demise, particularly and especially if the fad was largely germane to the Project Management world. Such an example is the “Life-Cycle Cost Estimating” craze.

Just so everyone’s clear on what’s being evaluated here, according to acqnotes.com,

A Life-Cycle Cost (LCC) is the total cost of a program from cradle to grave. (also referred to as Total Ownership Cost (TOC)) LCC consists of Research and Development (R&D) Costs, Investment Costs, Operating and Support Costs, and Disposal Costs over the entire life cycle. These costs include not only the direct costs of the acquisition program but also include indirect costs that would be logically attributed to the program. In this way, all costs that are logically attributed to the program are included, regardless of funding source or management control.[i]

My recollection is that Life-Cycle Estimating or Life-Cycle Cost first became a thing in the mid-1990s, and internet searches for papers on the topic show a sudden increase in titles published late in that decade. While perhaps noble in purpose, its stated goal of returning the “total cost of a program from cradle to grave”[ii] is impossible, as a couple of thought experiments will demonstrate.

Consider two (American) football stadiums: Soldier Field in Chicago, home of the Chicago Bears, and the Seattle Kingdome, home (for a time) to the Seattle Seahawks. I’ll contrast them so[iii]:

To further quantify the differences between these two projects facilities, consider the variance between them if each had had just one game, which sold out, per year, and that the ticket price was an inflation-adjusted $10 (USD), and that this was the sole source of revenue. What would have been the result?

Soldier Field:    $64,935,680

Seattle Kingdome: $15,180,000

…or, a delta of $49,755,680, which represents a whopping 62% advantage to Soldier Field, and even that figure gets larger each year Soldier Field stays open. I wonder what the respective Life-Cycle Estimators would have had to say at the project’s kickoff meeting had they been asked about each facility’s Return on Investment.

The second thought experiment that I would like to propose is based on the parameters needed to generate the “Life Cycle Costs.” From the definition in the first paragraph, these include:

• Research and Development
• Investment
• Operating and Support
• Disposal

Prior to project initiation, none of these improbably reduced number of parameters can possibly be known, or even estimated to any reasonable degree of precision (“…regardless of management control”? Puh-leeze). And yet, here’s the Life-Cycle Estimating crowd not only laying claim to an ability to perform such a capture, but actually asserting that such an “analysis” ought to be part of PM strategies going forward.

So, how did Life-Cycle Estimating become popular? I think it’s because of its implied (?) claims to be able to reasonably quantify far into the future, an ability that would automatically enrich any of its practitioners. With such a potentially beneficial capability, sprinkled with officious-sounding jargon, how could it not gain immediate and widespread recognition, if not complete acceptance, within the management science realm?

Did Life-Cycle Estimating end up going away, like other management fads? Sort of. There are still recent papers being published (not to be confused with a recent offering from my esteemed colleague Elizabeth Harrin, who blogged about the life-cycle of the estimate itself), but it seems to me that it doesn’t receive nearly the attention it did a decade or two ago. It may well be that it’s finally dawning on people outside of GTIM Nation that the future cannot be quantified. Not by risk managers, and not by estimators.

It’s why the term “precise estimate” is an oxymoron.

[i] Retrieved from https://acqnotes.com/acqnote/careerfields/life-cycle-cost-estimate on May 9, 2021, 18:38 MDT. This source cites the source of the quote as the Defense Acquisition Guidebook (DAG).

[ii] Ibid.

[iii] Wikipedia contributors. (2021, May 7). Kingdome. In Wikipedia, The Free Encyclopedia. Retrieved 01:57, May 10, 2021, from https://en.wikipedia.org/w/index.php?title=Kingdome&oldid=1021985574 and Wikipedia contributors. (2021, May 9). Soldier Field. In Wikipedia, The Free Encyclopedia. Retrieved 01:58, May 10, 2021, from https://en.wikipedia.org/w/index.php?title=Soldier_Field&oldid=1022224914