Game Theory in Management

As most of GTIM nation knows, my third book as well as the webinar I did with the help of ProjectManagement.com involves significant research in the area of ProjectManagement.com’s July theme of Organizational Culture (and, no, I did not bribe Cameron to set this particular theme). In going back over my notes, a couple of things popped that I hold to be rather fascinating about this topic, so I’d like to pass them along.

First off, I’d like to give a call-out to Michael Maccoby’s book The Gamesman (Simon and Schuster, 1977), which I’ve referenced often and will, no doubt, refer to in the future. To take a hydraulic press and metaphorically condense his main point, Dr. Maccoby asserted that there are four types of workers in a given organization:

• The Craftsman doesn’t really care for whom he works, but cares deeply about the output.
• The Company Man tends to take on the persona of the group or team around him.
• The Jungle Fighter gets ahead through calumny and deceit.
• The Gamesman doesn’t see his salary as a roof over his head or food on the table. Rather, he sees it (and other perks) as tokens in some immense game he’s playing. Because of this approach, the Gamesman is both far more likely to master the finer points of the industry he’s participating in, and is more willing to take risks. Generally speaking, this combination tends to make this archetype more successful than the others.

Interestingly enough, psychologist Richard Bartle performed an analysis of people who play massive multi-player online role-playing games (MMORPG), and also proposed (wouldn’t you just know it?) a four-tier archetypal pattern of those players. Bartle named his categories after the suits in a deck of cards, so:

• Hearts are those people who play in order to socialize and interact with other players.
• Spades are the explorers of the virtual worlds, digging around to find things of interest or value.
• Diamonds tend to perform the game’s stated objectives or scripted adventures, and reap the rewards for doing so.
• Clubs are there to fight.

And, of course, for those of my readers who have had their Meyers-Briggs type assessed, you are aware that they also use a four-axis evaluation method to determine your type. With all of these people who are way smarter than I am using structures based on four archetypes, who am I to differ?

In order to even address the topic of organizational culture intelligently, we must come up with some sort of basis to actually quantify it. Recall the axiom that that which cannot be measured cannot be managed. I’m going to base my four-tiered architecture on Dr. Maccoby’s assertions, and simply expand them. Here’s my thinking: if Dr. Maccoby is largely right (and I think he is), it stands to reason that organizations comprised of Craftsmen, Jungle Fighters, Company Men, and Gamesmen will tend to take on the aspects of which category holds the most personnel. For example,

• If the organization is dominated by Craftsmen, it will tend to turn out high-quality goods and services, but is vulnerable to being outperformed in market share to organizations that use newer technology to approach such Craftsmen’s quality but at lower prices.
• The organization with a plurality of Jungle Fighters will expend more energy than others harming their competition, the most obvious example being those companies that spend money on politicians willing to pass laws or regulations that raise barriers to entry in a given industry. Another sure-fire tell that you’re working for a bunch of Jungle Fighters is that the organization values loyalty above competence or even success. Craftsmen or Gamesmen trapped in such a situation will see their insightful, sincerely advanced recommendations that just happen to have even the mildest of contradicting-the-status-quo aspects viewed as treachery, or rebellion.
• Gamesmen-dominated companies are extremely dynamic and versatile, and are probably more likely to succeed, especially in newer industries. They are also more likely to fail spectacularly.

“Wait!” I can hear GTIM Nation say, “What about organizations controlled by …”  Then I hear the voices trail off as everybody realizes that Company Men can’t, by definition, dominate an organization, as they tend to assume the persona of the team around them. If you believe that your company’s CEO is, in fact, a Company Man, it simply means that there is a power behind that throne, another person (or persons) making the key decisions that are passed through such a one. And those people are absolutely not Company Men.

So, now that we have a guide to types of Organizational Cultures, what are the major implications? For starters, keep in mind that it’s entirely possible that a plurality (or even majority) of the organizations within a given industry can exhibit the same macro-culture. Easy examples include technology and dot.com orgs, which are notorious for being headed by unconventional, innovative thinkers, i.e., Gamesmen. Commodities-based companies, such as grocery stores or automobile service stations, are notable for quickly eliminating non-Craftsmen, since low quality goods or services for a given cost can’t survive much competition. Jungle Fighter-led organizations can only flourish in areas where performance is subjective, even speculative, such as media companies and political action committees.

I believe that most project-oriented orgs can and do allow for widely varied mixes of the four archetypes, making generalizations about them difficult. However, I think it’s fairly safe to say that

• Gamesmen get the projects in the door,
• Craftsmen deliver the scope on-time, on-budget (implication: most PMs will gravitate towards this archetype),
• Company Men will mirror the behavior of whichever of the above two types are most prevalent or influential, and
• Jungle Fighters will attempt to take credit for others’ successes, and deflect responsibility for their own failures. To whatever extent the said organization adheres to a meritocratic structure will directly influence how successful and common this type becomes. (Proof: the United States Chess Federation is virtually pure meritocracy, and Jungle Fighter tactics are utterly useless there.)

As for useful tactics in surviving in these environs, I can provide, with a high level of confidence, the following guidance…

Oooops! I ran out of blog pixel ink. Tune in next week, to further the discussion on My Organization Has A Culture. Now What?

Continuing with ProjectManagement.com’s June theme of technology, I feel the need to stick with my contrarian position, since most of the time advances in technology, be it in materials science, construction methods, or information systems are met with enthusiastic acceptance. However, if the large volume of horror-laced science fiction films and novels have taught us anything, it’s that carelessly or recklessly advanced technology can have terrifying outcomes. While careless or reckless advances in Project Management technology probably won’t threaten mass extinctions or the entire population of the Earth becoming enslaved to supercomputers and their robot minions, PM advances do carry the threat of being soul-crushingly boring and terrifyingly irksome. To paraphrase Kyle Reese, “Listen. Understand. Those technology-enhanced PM experts are out there. They can't be reasoned with, they can't be bargained with...they don't feel pity or remorse or fear...and they absolutely will not stop. Ever. Until you do PM the way they want.”

Meanwhile, Back In The Project Management World…

“But Michael!” I can hear GTIM Nation exclaim, “What possible nightmare scenarios can come about from advances in PM technology?” Well, I can imagine a few, including:

• Absurdly tight and mis-applied Variance Analysis Thresholds.  As commercial off-the-shelf Critical Path and Earned Value Methodology software platforms become more capable of sharing data with other systems such as the general ledger, the dopey analysis of comparing budgets to actual costs at the line-item level will become easier to do. The whole point of a Variance Analysis Threshold (VAT) is to acknowledge that some noise exists in the raw project performance data, and to avoid raising an alarm when, in fact, no performance issue is occurring. For example, a typical research Work Package (WP) will have the following weighted milestones:
  • 15% done when the initial investigation/data pull is complete.
  • 40% of the budget is claimed when the first draft of the findings report is done.
  • 65% has been attained when the second draft of the findings report is complete.
  • 85% claimed when the report is shipped off to the customer, and
  • 100% of the budget is claimed as earned when the customer approves the report.

How did I know that the completion of the second draft was worth precisely 65%? I didn’t. It’s just an approximation that tends to reflect actual performance on these kinds of WPs. What if the exact percentage were to be 59.4, in this specific instance? Well, it’s okay, since there are no VATs set as low as 5.4% (not by adults, anyway). The system is set up to acknowledge a certain amount of imprecision in EVM and CPM performance assessment claims, and compensates for them. But not if this compare-actuals-to-budget business keeps going, no siree! The technology will allow for immediate and accurate appraisals of each and every instance, not of a Cost Variance, but of a spend variance, which is very different and very irrelevant. Before you know it, the PMs will have to fill out Variance Analysis Reports on why they bought a bottle of Dr. Pepper when the spend plan clearly called for a can of Coca-Cola. For those who think I’m exaggerating, keep your ears open during your next project review. Yeah, I’m exaggerating, but not by much.

• The Time-Phased Estimate to Complete.  This particular piece of analytical technology is truly pushing the envelope of management science irrelevance. As I’ve often noted in previous blogs, the Estimate at Completion (EAC) is an extremely valuable piece of project performance information, but it’s easy to accurately calculate. Unable to accept this information gem at face value, many self-identified experts call for a bottoms-up estimate of the remaining project work, an Estimate to Complete, to be created, and performance measured against it. This turns the original baseline to rubber, and had been previously considered a practice pathology to be avoided. What happens to the original baseline once a bottoms-up ETC has been produced? Surely the ETC has incorporated more recent data and circumstances that could not have been foreseen at the project’s start, so we should go with the ETC, right? But if the original baseline ought not to be turned to rubber, why do the ETC in the first place? Simply because advances in PM information system technology allows for the generation of these time-phased ETCs? If that’s the case, then what happens next month, when the system delivers another time-phased ETC? Will it be used instead of last month’s ETC, or even the original baseline? What about the ETC after that? Before you know it, the Project Office will be hip-deep in rubber baselines, with their adherents bickering about whose should be considered more accurate or authoritative.
• Broader Risk Management Plans. As the technology allows more excessive cross-hatching of basic project budget, earned, and actual costs data, risk management software will have a field day with its enhanced ability to slather irrelevant Gaussian curve-based comparisons onto the real performance data. The most insipid weapon here is the “I told you so” bomb, which works like this: the risk analysis delivers a stochastic range of bad things that might happen to which activities within the baseline. If any of those things actually happen, the risk managers will be in a position to say “See, I told you so!”, providing a veneer of legitimacy to their entirely irksome approach. Those impacted by the blast, heat, and shock of the I Told You So Bomb won’t even know it until after the unfortunate occurrence risk event actually happened, and the risk managers come along afterwards and say … (you know).

Yes, science fiction is full of horror stories of technology gone wrong, and the fields of robotics and artificial intelligence make it easy to imagine terrifying outcomes. But PM isn’t all that different. Imagine a younger Arnold Schwarzenegger coming to your project review, his glowing red killer robot eyes obscured behind sunglasses, saying in his German accent and droll delivery style “You spent $105.40 on a part that appears in your basis of estimate as costing $100. Hasta la vista, baby!”

Also, just by the way, The Terminator came back in time from 2029, which is only 10 ½ years away.

Since most (if not all) of the takes on technology in the PM universe are that it’s by and large a good thing, it falls to me (of course) to take the contrarian position. Consider the famous study that showed that, when time-saving devices and appliances were introduced into homes with electricity to run them, the amount of time housewives spent doing housework did not actually decrease. Instead, two influences kept them working at about the same pace: since cleaning rugs and clothes became easier, the expectation became that people’s rugs and clothes would be clean all the time, hence the act of cleaning carpets and clothes was more frequent. Also, since more coffee brewing, ironing, and cooking could be done within the same amount of time as before, more of that starting happening. The amount of work didn’t change, just the level of cleanliness and ready availability of unwrinkled shirts and full cups of coffee (prior to the establishment of specialty coffee shops on every city corner) did.

Meanwhile, Back In The Project Management World…

I’ve spent a lot of ink pixels grousing about irrelevant information streams that are being pushed by supposed expert-level PM associations and organizations, so I’ll give that theme a rest (for this week, anyway). But I will say that many of these “must-have” analyses would be more difficult if not for the ability of modern Critical Path Methodology (CPM) and Earned Value Methodology (EVM) software packages to share data with other systems, such as the organization’s general ledger (a rather valuable utility), and the risk management system (utterly useless). So, some of these technology advances are really useful, but not all of them. But what if the PM finds herself in a situation where such advanced CPM and EVM packages are not available, or in use, or even in a situation where (horrors!) there are no personal computers or tablets? Alternately, what if our PM is in a situation where these packages are available and in-use, but she suspects their information isn’t accurate? Like Montgomery Scott says in Star Trek III, “The more you overtake the plumbing, the easier it is to stop up the drain.” In addition to warp drive technology and, well, plumbing, the same effect can be seen in certain overly-teched Project Management Information Systems. What’s the PM to do in these circumstances?

Hack Number One: The Reliable Estimate at Completion

One of the most, if not THE most, valuable pieces of PM information that can be generated by an EVM system is the accurate answer to the question “At the present rate of performance, how much will this activity, task, or project cost when it’s done?” Springboarding off of Dave Christensen’s work on Cost Performance Index (CPI) stability[i], the fact of the matter is that an EAC that’s accurate to within ten points can be calculated (once the activity/task/project has cleared the 20% complete point) by dividing the Budget at Completion (BAC) by the CPI. The fascinating aspect of this (besides its underestimation by Establishment PMs) is the fact that the formula

EAC = (BAC / CPI)

can be algebraically reduced to

EAC = Cumulative Actual Costs / % complete.

Crazy, right? But it works, and reliably so. All you need is the cumulative actual costs from the general ledger, and a reasonable estimate of the activity’s/task’s/project’s percent complete, and you’re there. Not only is it gobsmackingly easier than all that bottoms-up nonsense, it’s also provably more accurate on a consistent basis.

Hack Number Two: The Reliable Duration at Completion

Earned Value and Critical Path methodologies are more closely related than many people realize. When you place your percent complete amount into the schedule network on your CPM software, it calculates a new forecast date using a formula identical to the one mentioned above, except it uses duration rather than costs. My regular readers know where I’m going with this: to calculate a duration estimate for any activity, task, or project, all you need to do is to divide the percent complete into cumulative duration, and your answer will be accurate to within ten points! In fact, for all those organizations using some form of milestone tracking system as a cheap substitute for a real schedule, stop asking the tasks’ owners for an estimate of when they will be done, and start asking them for their percent complete. Divide that number into the cumulative duration, and compare that date to their original scheduled completion date to tell if they will be early, on time, or late. Your accuracy rate will skyrocket.

Hack Number Three: Testing What You’re Being Told About At-Completion Costs

For those technically-advanced projects where they provide a new “bottoms-up” EAC on a regular basis, these estimates are usually highly optimistic. To test them, compare the Cost Performance Index to the To-Complete Performance Index (TCPI), which will always be available from the technically-advanced cost processors. The TCPI indicates how well the project must perform in cost space in order to come in under a given parameter, usually the Budget at Completion (BAC). However, if you substitute the stupid bottoms-up EAC for the BAC in the formula, and get a number above 1.00, then you probably have a problem. The aforementioned Dr. Christensen’s research found that the CPI virtually never varies more than ten points once the activity/task/project has passed the 20% complete point. Now compare the CPI and the TCPI. If it calls for a 10% improvement in performance, you are probably being fed an optimistic EAC. If a 15% performance improvement will be required to hit the bottoms-up EAC, that’s not going to happen. If the improvement called for is 20 points or more, the people pushing the bottoms-up EAC are not only lying to your face, they think you’ve dumb enough to believe them, which is problematic all by itself.

Can a PM survive a project with just these hacks? It depends on the project, but usually not. However, if your project team is completely stalled for want of readily available cups of coffee, and there are no coffee shops around, and you’re the only person who knows how to make hot coffee and assess project performance without an electronic device, you may just be in a position to rule the world.

[i] https://www.researchgate.net/publication/237574533_Cost_performance_index_stability_fact_or_fiction

In the movie Superman (1978), Lex Luthor (Gene Hackman) has deduced that a fragment of the exploded planet Krypton has arrived on Earth in the form of a meteorite that landed in Addis Ababa. One of his two assistants, Eve Teshmacher (Valerie Perrine), has the following line: “I know I'm gonna get rapped in the mouth for this, but... So what?” Lex goes on to explain that the meteorite must have the same “specific radiation” as the doomed Krypton, and is therefore lethal to Superman, a conclusion which is somewhat unintuitive to those unfamiliar with the Superman canon. Due to the power of movie magic, however, it is both accurate and relevant (and timely, given the super-caper that Lex is planning at the moment).

Meanwhile, Back In The Project Management World…

In this blog I would like to combine two themes I address on something of a regular basis, to wit:

• The PM world can be divided into two camps: Processors, who don’t really care about projects being brought in on-time, on-budget, but care a great deal about people following “proper” procedures; and Performers, who don’t really care about process, but just want to bring their projects in on-time, on-budget.
• Of the thousands and thousands of PM-based software packages out there, the ultimate test of their worth has to do with their ability to deliver information that’s (1) accurate, (2) timely, and (3) relevant. Failure on any of these points renders the whole system invalid, and, therefore, worthless.

Now, there are several organizations/associations that crank out what they perceive as usable insight in documents – procedures, directives, implementation guidance, whatever – but this so-called guidance also must answer to the standards for PM-oriented software. For example, I’m aware of a certain guidance document that demands that an ongoing comparison be made between a project’s Basis of Estimate (BoE) and the actual costs incurred, at the line-item level. This guidance document was produced by a group that claims many supposedly high-level PM professionals in its ranks, and they pretty much expect to be taken seriously. Sadly, I can’t.

One of the very first rules that Project Controls Specialists (and, one would hope, Project Managers) learn is that comparing budgets to actual costs is useless. I would go one step further and say that it’s often misleading. Consider my favorite example, one that I use when teaching basic PM.

You’re the manager of a two-month-long project to produce 2,000 widgets, with a budget of $2000 (USD). You time-phase your budget to $1000 in Month 1, and $1000 in Month 2. At the end of Month 1, your accountant comes to you and says your have spent $1100. How are you doing?

If you said “I know I’m gonna get rapped in the mouth for this, but if I don’t know how many widgets I made, why is the amount I’ve spent relevant?” go to the head of the class. The seemingly intuitive conclusion that something has gone wrong because you’ve spent $100 more than planned, is rendered completely irrelevant if you have, in fact, made, say, 1300 widgets. In basic Earned Value parlance, you are not only not in bad shape, you are actually performing well ahead, enjoying a 18% positive cost variance, as well as a 30% positive schedule variance. If you had mistakenly presumed that spending $100 more than budget was a bad thing, you would have blundered in any action taken based on that so-called analysis. And yet, here’s an organization pushing the idea that such an analysis produces usable PM information simply because it’s happening at a more granular level of detail. Guys, it’s irrelevant! And, by pushing such an assertion, it kind of makes your organizations look, well…

Who’s Behind This Massive Criminal Conspiracy?

So, who, exactly, is pushing this narrative? I seriously doubt their numbers include a significant population of Performers. Performers don’t think that way. It has to be the Processors, who churn out this stuff blithely unaware of any duty they have to relevancy, nor of the fact that not even movie magic can make the budgets versus actuals analysis reliable or insightful. As long as they can push out “guidance” that compels a certain analytical behavior, they’re happy as clams, regardless of whether or not such analysis yields any usable management information. It’s almost as if they were assigning tests to the PM world, to see if they can make its adherents accept the predicate that such useless analysis must be performed in order to lay claim to “doing” legitimate Project Management. Naturally, to call into question the basis of their assertions, much less their motives, is to incur their wrath, and invite being rapped in the mouth.

But I’ll take that chance, because much as I would hate to see California fall into the Pacific Ocean due to nuclear weapon-induced cataclysmic earthquakes, that can only happen with large doses of movie magic. Conversely, furthering the idea that comparing budgets to actuals is a key aspect of legitimate Project Management doesn’t require any movie magic at all. It just needs PM’s current practitioners to become fearful of these criminal masterminds Processor-style PMs

In previous posts I’ve made the distinction between the two types of Project Management experts one comes across these days: the Performers and the Processors. In evaluating the effects of ever-improving technology on the PM field, I think it’s prudent to break out such advances based on these two very different types of consumers.

For the Performers, this is a pretty easy analysis. You see, Performers don’t care too much about following overreaching policies or guidance documents. They really just care about delivering their projects on-time, on-budget, to the client’s satisfaction. This being the case, the PM information technology that will interest them the most involves the Project Controllers’ ability to put into their hands the timely, accurate, and relevant information needed to attain that end. Assuming my bifurcation of the PM world is accurate, that means that the most valuable tool that the Performers could have would be a kind of PM Information Nonsense Detector. Such a tool would be very useful indeed when those Processors whom have (unfortunately) attained some level of organization authority begin to harp on about how the Performers “need” to have more stakeholder engagement, or a more robust risk management plan. The PM Information Nonsense Detector could then emit one of a variety of pre-selected sounds, including:

• A ring tone, like a cell phone would make. This would give the Performing PM an opportunity to pretend to pick up the PMIND as if it was a phone, and say something like “Hey, Ted. Yeah, you’ll never guess what the so-called experts from corporate want me to do this time!”
• A pre-recorded laugh track, hopefully one that includes laughter that’s clearly derisive.
• A recording of a voice very similar to the Lost In Space’s robot’s, saying “Danger! Danger! Absurd ideas being furthered!”
• An ear-splitting siren, followed by a female voice (similar to Siri) saying “Sorry, I simply can’t suffer fools gladly.”

While all this is going on, the PMIND searches the web for specific refutations to the nonsense currently being rendered. For example, if the risk management types are pushing their agendas, the PMIND would provide valuable quotes from Douglas W. Hubbard’s excellent book The Failure of Risk Management, or maybe even past Game Theory in Management blog entries.

What If We Make Them Mad?

I’m happy to provide the idea for this device to ProjectManagement.com and my readers, since I’m confident it will make a mint. For those of you who readily recognize that innovative, technically-advanced products that make a lot of money for their manufacturers also incur the wrath of those whose approach to PM is so full of folly that it provides the impetus for the demand for such a product, do not worry. We’re talking Processors here. Let’s say that, as the demand and profit margin for the PMIND soars, its patent owner(s) receive one of those threatening letters composed of letters cut out of newspapers and magazines. The response should be something like “We are in receipt of your letter threatening us if we continue to market and sell the PMIND, thereby rendering almost all of your insipid conference paper presentations and guidance-generating activities plainly absurd. Prior to full compliance with your demands to cease and desist, we have one request: given the volume of such threatening letters we receive, could we impose on you to settle on a specific format for these missives? For example, the difference in the size of the font in the cut-out letters between the “u” and “f” in the second word of the third sentence, “suffer,” is so dramatic as to render the whole sentence – and therefore your essential message – virtually unreadable. If you could simply agree to a procedure that would stipulate a mean font size consistency in cut-out letters in threatening communications, it would be much easier to implement your demands.” I can guarantee you that the Processors behind the threats will then immediately convene to discuss your request and appoint a panel to further explore the basis for generating the procedure. This panel, after many heated arguments, will finally deliver a rough outline for the procedure, where it will go to some sort of “governing committee,” who will review the outline, and make recommendations prior to the commissioning of the actual writing committee. The whole lot of them will engage in this kind of behavior ad infinitum, never getting around to actually carrying out their threats.

I Haven’t Forgotten The Opposition!

For the Processors, have I got a technology advance for you too! This device, the Process Implementation Gadget, or PIG, will inflict irritation, frustration, or even pain on those Performers who refuse to obey your arbitrary and capricious speculations well-thought-out guidance on how PM should be done. It works like this: whenever you encounter a Performer PM, the PIG will search the General Ledger for the cost accounts she is responsible for. It then searches those projects’ stored documents for things like a risk management plan, or a Lean Six Sigma Quality Control Plan, or a Stakeholder Engagement Strategy statement, or….well, you get the picture. Should the PIG fail to find any of these documents in the record, or, having found such a document, recognizes it as mostly boilerplate, it automatically sends an e-mail to that PM’s line manager, and that person’s line manager, and so on up the organization chain, with the message that the offending PM is not in compliance with the organization’s arbitrary and capricious speculations well-thought-out guidance on how PM should be done. You’ll be able to cause these PM renegades so much grief that they will gladly engage you, and allow you access to their charge codes, just to get you to sign off on whatever waiver they need to make you go away. Everyone wins! You can then go and brag about how your honorable intervention was the proximate cause of that Performer’s subsequent success at various PM seminars!

I look forward to the day that a project kickoff meeting has both PMINDs and PIGs at the ready as the strategic approach to the given project is assessed and designed, just to see which invention carries the day. I know which one I’m rooting for.