Game Theory in Management

A few weeks back I blogged about the appropriate way of calculating the impact of the COVID-19 shutdown on our organizations, and pointed out one method that will probably be pushed as valid, but really isn’t. This week I want to take on perhaps the mother of all wrong methods for calculating the impact of a Black Swan event[i], one rooted in several highly questionable but widely accepted premises of modern management science theories. To be clear, I’m only referring to these notions as “theories” due to their rate of acceptance, not because they have been suitably verified (or tested, really) with empirical data or repeatable experiments.

Let’s work backwards from the actual analytic approach under my microscope, shall we? My sense is that there are going to be hundreds, if not thousands, of executives who will request their cost performance experts to quantify the cost impact of the slow/shutdown by pouring through General Ledger data, line by line, and identify which costs are attributable to the slow/shutdown, and which are not. I hold this to be an invalid approach, for the following reasons:

• In order for such an analysis to be remotely reliable, the ones poring over the data must be fluent in the differences between proximate cause, and material cause. The former is a clear and direct causal factor, from one event to another, similar to domino #1 falling over and forcing domino #2 to also fall. Material cause, on the other hand, must pass the if-not-for test, as in domino #2 would not have fallen from domino #1’s toppling if not for someone standing it up in close proximity to domino #1 in the first place.
• If the case for material cause can be made, the next impossibility would be to quantify how much responsibility, expressed objectively, should be split between the proximate causes and material one(s). Commercial and civil liability cases involving a jury setting the percentages of the culpability of the various defendants usually turn on such an analysis, and even here the findings often defy logic. In the domino example, we know that domino #2 fell. Let’s pretend that this had a serious, negative impact on our organization. What percentage of the total damages should be assigned to the person who knocked over domino #1, and how much should be assigned to the person who placed them upright so close to each other?
• If the previous bullet appears next to impossible to quantify, consider how ridiculously simple that example is compared to the near-chaotic environment of the near-infinite working parts of the local economy, let alone the regional, national, or global variety.
• In the face of such dizzying analysis parameters, how many of the executives ordering just such an analysis do you suppose will do so equipped with a readily-understandable guide, or template, usable as a sort of litmus test for every single line item in the General Ledger?
• And, in those rare instances where such guidance is provided, I can guarantee that none of them can preclude multiple exceptions. Some costs believed to be attributable to the shutdown, but actually had nothing to do with it, will be included in the tally, and vice-versa. Metcalf’s Law, also known as The Butterfly Effect[ii], virtually guarantees it.

Pulling the thread a bit further, it’s easy to see why the poring-through-the-General-Ledger approach is often considered the obvious way of approaching the impact quantification problem: generations of business degree holders have been taught that the General Ledger is the source and residence of all management information that has a currency symbol in front of it. Recall, however, my oft-stated distinctions between the three types of management:

• Asset Management is centered on the organization’s assets, and its main tool is the General Ledger;
• Project Management focusses on customer-generated performance standards of scope, cost, and schedule, with its main information streams based on Earned Value and Critical Path Methodologies;
• Strategic Management is all about relative market share.

Now ask yourself, where in the balance sheet or profit-and-loss statement can we see information about your organization’s market share percentages? Of course, it’s not there. That’s not what those reports convey. Then why should anybody expect them to show the cost or schedule performance impact of a pandemic? Unless COVID-19 destroyed or harmed a company asset, the General Ledger simply cannot convey any meaningful quantification of the impact of the pandemic on the organization, save the singular data point, relative revenue changes.

And that, GTIM Nation, is NOT how one quantifies the impact of a shutdown.

[i] See Nassim Taleb’s outstanding book, The Black Swan; The Impact of the Highly Improbable, Random House, 2007.

[ii] Usually expressed as the question “If a butterfly flaps its wings in Brazil, does it cause a hurricane in Texas?,” the more formal definition is that relatively small perturbations in far-flung nodes of a large network can have a cascading effect, bringing massive changes to nodes on the other extreme of said network.

My oldest son was telling me a story about a time when he was watching a movie with his fiancée at her parents’ house, late at night. Without warning, a series of events unfolded, taking no more than five seconds from start to finish:

• A large insect flew into the room (“at least the size of a silver dollar”).
• It headed upwards, towards the ceiling fan, and was promptly whacked by one of its blades.
• The insect (possibly a large beetle, but no one knows for sure) crashed to the ground.
• The family’s dog, Charlie, rushed into the room, seemingly from out of nowhere.
• Charlie retrieves the insect, and rushes outside with it via his doggy door.

The four inhabitants of the room were left looking at each other, puzzled – did what we witnessed just then really happen? And, assuming so, should we be doing anything about it right now?

Meanwhile, Back In The Project Management World…

We are quickly approaching the time when the governments of the various nation-states afflicted by the pandemic will begin to ease up on the policies working as brakes on the global macroeconomic environment. As GTIM Nation is fully aware, an additional set of characteristics that I like to point to concerning the distinction between the three types of management are:

• Asset Managers’ focus is entirely within the organization itself;
• Strategic Managers are paying attention to people outside the organization, but not its customers; specifically, the competition’s actions and strategies,
• while Project Managers are looking at all the people outside the organization, but not its competitors – PMs are about customers, both existing and potential.

This is an important distinction, in that it highlights (yet another) reason why PM ought to be considered more important than the agendas of either the Asset Managers or the Strategic Managers (but don’t tell the business school faculties!). Consider the introduction of Agile/Scrum Project Management, in 1986. Information Technology (IT) projects, while clearly needing some form of PM to help curb the relatively high number of project overruns and late deliveries, were still struggling with issues stemming from traditional Configuration Management. It was simply not feasible for most (if not all) IT projects to funnel changes to their software in-development through a formal Baseline Change Control Board, considered the best guarantor of preventing scope creep. Such boards typically met only once per month, and required attendance not only from both the project teams’ organizational management, but also from the higher levels of the customers’ representatives. If a given Baseline Change Request or Proposal (BCR or BCP) was approved, then all was well. However, if it were to be rejected, or sent back for additional information or modifications, critical time and effort would be lost, with no expedient remedy available. Oh, sure, some IT-project based orgs would allow some sort of fast track, or urgent BCP/BCR, but that didn’t really fix the main problems associated with an ossified change control process. Agile/Scrum changed all of that by assuming that changes would be not only accommodated, but pretty much expected, and on a regular basis. By assigning specific titles and roles to customer and project team representatives (e.g., “Scrum Master,” etc.) in weekly (at least) meetings, such changes could be evaluated for appropriateness and priority in a highly expedited fashion. This strategy brilliantly addresses a need for retaining certain elements of “traditional” Project Management while making realistic accommodations to the demand for spontaneously addressing customer expectations, which are almost always in a state of flux.

Now, compare and contrast the introduction of the Agile/Scrum approach to any analogous initiative from, say, our friends the Asset Managers. What would Agile/Scrum Finance and Accounting look like? What tenets of Generally Accepted Accounting Principles (GAAP) would be modified or abandoned, in order to, say, deliver the Profit and Loss Statement earlier? By definition this can’t happen, as the accounting cycle is almost always set up in such a way as to make it impossible to compress. The Asset Managers can, say, process travel reimbursements more expeditiously, but they literally cannot “close the books” sooner than the last day of the accounting period, usually monthly.

What About Ceiling-Fan-Struck Insects Being Retrieved By A Family Dog?

I was just getting back to that. Once the various government-held starter pistols that signal a reprieve from quarantine begin sounding, some things will remain relatively stable – an organization’s asset portfolio, for example. Other things can be expected to change dramatically, e.g., the customer base, which is, by definition, entirely within the purview of the PMs to retain, lose, or expand. There’s really no telling what set of circumstances will lead to an unexpected opportunity suddenly becoming available in bang-bang-bang fashion, but these things do happen. As highly structured as the Asset and Strategic Managers can be, they can’t be counted on to be as agile (get it?) as PMs to respond fast enough to secure or expand the customer base, once things get back to some semblance of normalcy. In short, when the restrictions are lifted, and the business world gets back to opportunities happening in bang-bang-bang fashion, it’s the Project Managers who will swoop in from seemingly nowhere to seize them, and then go off and work them.

Just ignore that crunching sound coming out of the back yard.

In last week’s blog I discussed some of the considerations that PMs should take into account once the nominal business re-start mayhem begins. This week I would like to examine a couple of long-held management tropes that should definitely be avoided, both in the short term and, well, permanently, if I were to get my way. Last week we looked at some useful business structures to help place our organizations with respect to the competition along the lines of the axiom, Quality, Availability, Affordability – pick any two. But what happens if we take a look at the macro-economic restart picture from the other side, as in the characteristics of the business models of the organizations that are about to fail? I’m thinking analysis into a pair of highly suspect management science “theories” will reveal them for being the invalid ideas that they have been all along, widespread acceptance notwithstanding.

First Up: Our Friends the Asset Managers

I’ve often criticized, questioned, and even ridiculed the core belief of Asset Managers, that the point of all management is to “maximize shareholder wealth.” Adherence to this mantra is extremely common among college-level business schools; in fact, it’s usually thought to be challenged only by the most ignorant of managers or students. Its logical (if extreme) extension was perhaps best articulated by former General Motors chairman Thomas Murphy, who is credited with having said “General Motors is not in the business of making cars. It is in the business of making money.”[i] This was supposedly asserted in the 1970s, when GM was king of the automobile market. Within the decade it would find itself in serious trouble from a variety of competitors who were, apparently, more interested in making cars (superior products), confident that the profits would follow. GM would ultimately – and notoriously – need a government bailout due to its bankruptcy –yes, bankruptcy – in 2009. The bailout costs the taxpayers of the United States $11.2 billion dollars. How’s that for pursuing a business strategy predicated on “maximizing shareholder wealth?” Indeed, the landmark book Crossing the Chasm, by Geoffrey Moore (Harper Business Essentials, 1991) pretty much solidified the case that organizations only concentrate on the maximizing of shareholder wealth at a phase in their life-cycle when they are no longer concerned with attracting or retaining customers.

Apparently, none of this textbook case of a management science disaster predicated on a tightly-embraced concept has made even the slightest dent on the faculty lounges of the major business schools of the United States. So, here’s my little litmus test, to see if this axiom has any legitimacy: Of those organizations that have based their business models on the “maximizing shareholder wealth” cliché, which ones will be the most successful at the restart? Consider that such organizations will be more inclined to send revenue towards dividends, rather than paying down debt or enlarging reserves – the same reserves (or credit ceiling) needed to retain or rehire staff. Contrast this approach with the Project Management-centric organization, oriented towards attaining customer-specified goals of scope, cost, and schedule. Such organizations will be far more likely to have built a larger, more loyal customer base, the same customer base needed to return to a profitable operational status quickly. For now, we can speculate, but it won’t be long until more empirical data will be available to test the robustness of the organizations that have embraced the “maximize shareholder wealth” hypothesis.

Next: One of My Favorite Targets, the risk managers

I get the impression that if a statistically significant number of project risk analyses had included the threat of a global pandemic bringing a massive amount of disruption to the normal business cycle, we’d be hearing about it by now, even if the estimated impact was off by a large factor. The fact that this exact scenario did occur points to yet another shortfall of modern risk management theory, perhaps best articulated by Nassim Taleb in his can’t-miss book The Black Swan; The Impact of the Highly Improbable (Random House, 2007). Taleb defined Black Swan events as having three characteristics:

• Completely unexpected,
• Had a large, significant impact, and
• After the fact, people tend to convince themselves that the event could have, in fact, been predicted.

…which brings us back to how a project risk analysis is currently performed. Scenarios that include bad weather, or changes in prices for materials, parts, or specialized labor, can have their cost and/or schedule impacts guessed at estimated, budging the cost and schedule baselines maybe 5 – 20% over plan. But, as COVID 19 has shown, to bring truly massive amounts of disruption into a project plan requires a Black Swan event which, by definition, is completely unexpected. In short, having an advanced risk management (no initial caps) capability in no way adds to an organization’s ability to survive a significant macro-economic disruption. So why does anybody think it will help in managing the more mundane PM events that unfold?

Despite their widespread use, I’m convinced that management strategies based on the Asset Managers’ or risk managers’ basic tenets are flawed, and ought to be abandoned straight away. If not, then we’ll see what characteristics are shared by the organizations that survive the current stress test. My bet: the survivors will be more project/product oriented than those that don’t make it.

[i] Wikipedia contributors. (2019, May 27). Thomas Murphy (chairman). In Wikipedia, The Free Encyclopedia. Retrieved 18:53 MDT, May 4, 2020, from https://en.wikipedia.org/w/index.php?title=Thomas_Murphy_(chairman)&oldid=899030673

Sooner or later, our projects are going to start back up, and some organizations will conduct this restart better than others. How will they do that, exactly? Well, here are a few points to consider when the commercial environment makes solid moves back towards normalcy.

First, let’s consider the implications that come with a blank slate, or the resetting of the business environment. Recall the oft-stated GTIM axiom, of Quality, Availability, Affordability: pick any two. When the massive economic timeout was called, which two of these three did your organization pursue most energetically? It’s important, because:

• If your organization’s product or service was a quality leader, and was priced that way, then availability was probably okay, i.e., no significant wait times. When the planet’s economies come back on-line, this aspect will be stressed, as pent-up demand will challenge your previous levels of availability. When that happens, management will need to make a decision, whether to hedge on some aspects of quality, raise prices, or be prepared to arrange for wait lists. Keep in mind, any of these has the potential to reduce the organization’s market share; but, since availability was previously a strength, then wait lists are (generally speaking) the biggest potential threat to a successful restart.
• On the other hand, if your organization’s product or service was readily available and affordable, but wasn’t necessarily known for its rare quality, then your ability to attract your competitors’ pre-timeout customers will be predicated on your ability to convince them that your organization’s version of the good or service is “good enough” for the job. This also assumes that such an organization is in a position to regain its potentially furloughed staff, since, generally speaking, those companies that concentrate on affordable and available goods and services do not require a preponderance of highly-skilled staff, which makes them vulnerable to higher turnover rates, which carry with them their own difficulties.
• Finally, high-quality and readily available goods or services already had waiting lists prior to the macro timeout, and restarting activities will probably focus on reconciling the backlog. However, while this is going on, the competitors with the previously-addressed characteristics will be seeking to gain market share by picking off those customers who have the budget, but not the time.

Now let’s shift gears again, and consider the Corner Cube hypothesis. Harkening back to another oft-stated ideas, that there are three distinct and different types of management, to wit:

• Strategic Management is focused on market share,
• Asset Management wants to “maximize shareholder wealth,” and
• Project Management seeks to deliver scope on-time, on-budget, for what are essentially customer-specified parameters.

A little mental exercise here: which of these are the goals of start-up companies? Do entrepreneurs tend to base their decisions on, say, how much return they can expect from their investments? Or do they not rather invest large amounts of their own time and talents in an effort to attract and satisfy customers? Of course customer satisfaction, i.e. Project Management, is the priority, as laid out in my very first published (and peer reviewed) article, Managing to the Corner Cube; Three Dimensional Management in a Three Dimensional World (https://www.pmi.org/learning/library/three-dimensional-management-world-5329). In fact, successful start-ups will generally follow this progression:

• Do whatever it takes to attract and satisfy customers in the organization’s field of endeavor.
• Only after a strong customer base has been attained will deliberate efforts at increasing market share commence, meaning that Strategic Management will take precedence over Asset Management. When I say “deliberate efforts,” it’s because the establishment of a strong customer base will usually automatically increase or establish some level of market share, and in a way that successful Asset Management, well, won’t.
• Remember the axiom that the point of all management is to “maximize shareholder wealth?” Well, the Corner Cube hypothesis refutes this directly, and here’s Exhibit A: only after a reliable customer base has been established AND an acceptable market share secured will the successful manager even think about increasing return on investment.

Soooo, what does this mean in re-start space? Even those organizations that have met their goals of customer base, market share, and return on investment will easily find themselves having to approach the re-start as if they were brand-new, meaning Project Management re-emerges as the most important, most immediate goal of the re-starting organization.

But then, that’s been the real goal this whole time, hasn’t it?

As we approach the peak of the COVID-19 pandemic, part and parcel of the post-event analysis will include attempts to answer the question: how much did this cost us, and how much time did it remove from our productivity? Whether by nation, macro-organization, corporation, small business, or from the point of view of specific industries or families, the upcoming weeks and months will, no doubt, see many attempts at quantifying these impacts, usually in order to ridicule political opponents, but more importantly to ascertain those business strategies that are most effective at making the analyses’ target group more economically resistant to similar future occurrences. Unfortunately, both purposes of analysis invite quackery, and in massive proportions. Those who are supposedly looking to advance management science can be counted on to advance their own particular pet theories (although establishing how risk management [no initial caps] can help insulate organizations against pandemic shutdowns would be quite the reach, I fully expect something to be pushed that way). And yet, this serious question needs a serious treatment, regardless of charlatans muddying the waters. How can this be done, precisely? As fate would have it, I have some experience in this type of scenario, and the good news is that the solution is far simpler and easier to access than almost all of the alternatives often pursued, and that optimal approach comes from basic PM.

First, let’s dispense with the intuitive but nevertheless incorrect notion that estimators can deliver the information being sought here. They can’t, as is easily observed by considering the common practice among projects to derive a “bottoms-up” estimate at completion, or EAC. Back when I was working to attain my estimator’s certification I was taught that there were three types of estimates:

• Order or Magnitude, or “ballpark” estimate, derived by comparing your intended project to similar work done previously, and accounting for differences in size, function, geographic and inflation. The bases of comparison are published by organizations whose data has been shown to be consistently reliable; however, these estimates’ accuracy is rated between 45% and 60% of final costs.
• A Budget Estimate lists the intended resources by type (Heavy Equipment, Labor [by type], Overhead, Hotel Load, etc.), based on the most recent information for each category. Perhaps the most common type of estimate, these are rated at between 30% and 45% accuracy.
• A Detailed Estimate is usually produced by a professional estimator, using off-the-shelf software for that purpose, and is so detailed that it can be handed off to the procurement specialist to begin buying the labor and materials needed to work the project. This type of estimate is accurate to within 15% and 30%.

Keep that last accuracy bracket in mind as we proceed with this analysis – a professional estimator, using OTS estimating software, can be expected to be up to thirty points off of the real answer, and even the optimal scenario is 15% accuracy.

However, what’s being estimated here isn’t a project. It’s the impact of a Black Swan event (as described in the excellent book by the same name by Nassim Taleb) across a broad spectrum of organizations and projects. There is simply no reliable way of knowing what a “final impact” of any given occurrence, much less one with the massive, far-reaching impact of a pandemic, could have on any given subject. The trucking industry is being hard-pressed by the unavailability of the facilities that drivers need, and that can perhaps be quantified. But what about the impact of lower fuel prices? Yes, hotels are largely slowing down or shutting, but demand from on-line retailers is jumping. How to cross-evaluate these factors, and dozens, if not hundreds more like them? It’s simply impossible to do. Facilities such as restaurants can compare their pre-virus numbers to post-virus versions, but since when does any business, let alone an entire industry, experience level revenue figures across weeks, let alone months?

“Alright, Michael” I can hear GTIM Nation say, “how does one calculate the impact of a shutdown, as your title implies?” Here’s how it’s done, at least in project space.

Watch your Cost Performance Index (CPI) and Schedule Performance Index (SPI), at the cumulative level. For those new to these information streams, these are derived so:

CPI = BCWPcum / ACWPcum

SPI = BCWPcum / BCWScum

where BCWPcum is the cumulative Earned Value, ACWPcum is the cumulative actual costs, and BCWScum is the cumulative time-phased budget. What you will see happen is a sub-1.00 spike in both indices (1.00 is performance exactly as planned; below that number indicates trouble, for both indices) downward in the next couple of months, most likely more pronounced in SPI than CPI. Then, over the following few reporting cycles, you should see these figures climb their way back to where they were before, the speed with which they do so determined by the robustness of the subject organization. How accurate is this method? Studies have shown it’s good to within ten points. This method of impact quantification is not only accurate, but it’s best simply because it’s not predicated on the thousands upon thousands of parameters which, even if they were identifiable, cannot be reliably captured.

If you have no Earned Value Management System in place, two things: are you really doing Project Management at all? Secondly, well, I can’t help you.

Also, stay safe out there.