Game Theory in Management

Pity the poor Project Managment proprietor. Earlier this week, Pro Football Hall of Fame running back Emmitt Smith had this to say, about a proposed rule that would penalize a running back for hitting a tackler with the top of the helmet:

“I disagree with the rule altogether.  It doesn’t make any sense for that position.  It sounds like it’s been made up by people who have never played the game of football.’’

What does this have to do with managing a Project Management Office, or an organization of Project Controls specialists? Not much, except for the fact that, in most organizations, the answer to the question of how much project management or project controls support is needed for a given endeavor is almost never provided by people fluent in those arenas. The technicians, engineers, and programmers executing their projects’ scope often determine the level of project management and PM information systems support they need and, in my experience, they often get it wrong.

But the way they get it wrong offers a certain fascination. I first noticed a pattern of assessing PM talent demand when I was working for a rather large organization, which would periodically encounter on-the-job accidents. When these accidents occurred, upper management would respond in a very predictable way, decrying the lack of a “safety culture” within the organization, and mandating that all employees attend a series of safety presentations. A renewed focus on performing our jobs safely would then permeate the organization, even those elements that weren’t engaged in any particularly hazardous duties. As the accident rate dropped, so, too, did the overt emphasis on observing the myriad safety protocols put forward by the safety engineers. The unnecessary ones would be ignored first, with very little (if any) impact on the organization’s accident rate. Eventually, some of the basic safety principles would get worked around, and another avoidable, and yet somewhat grisly accident would occur, and the cycle would begin again.

It then struck me how similar this cycle is to the perceived demand for project management expertise. Projects would be muddling along, with the engineers wondering why they had to spend any money at all on project controls specialists or other PM-types, and making the case to save money by abandoning that area of expertise. Some projects would go on to finish successfully, reinforcing the narrative that money spent on the PMO was superfluous and a waste. But, inevitably and eventually, some project would go off the tracks, racking up huge delays and overruns, and all without the ability to have forewarned upper management that a problem even existed before it was too late to avert or mitigate the disaster. A cursory post-mortem would reveal an inadequacy in that project’s cost and schedule control systems, and the demand for those who could set up such systems would realize a dramatic increase. This would continue until a certain sense of complacency would re-enter the management culture, and the engineers would return to their questioning of the need for PM expertise.

All this time, the PMO managers, as well as those in charge of project controls organizations, were dealing with one of two problems: either the perceived demand for their organizations’ services was below supply (in which case they would scramble to find billable work for their people), or else demand far outstripped available talent, in which case their people were working overtime and experiencing burn-out. Depicted graphically, it looks something like this:

In this example, the appropriate level of expertise for this organization is 25 full-time equivalents (how does one calculate the “appropriate level”? For that, you will have to buy my recently-released, must-have book Game Theory in Management). Notice how, as complacency among the technical staff reduces their perceived level of demand to zero, the available talent lags in both amplitude and time. Because technical management is reluctant to come out against, well, reality, they will tend to not communicate their anti-PM sentiments until after these have become socialized and solidified, hence the time-lag in communicating the reduced demand to the owner of those assets. However, PM organization leaders also know that, at some point in the future, this demand curve will reverse itself, so that, even as perceived demand drops to zero, they won’t be rid of all of their talent. These managers will find a way to protect as many of their charges as they can. Note, also, that these managers won’t increase their talent pool to match the zenith of the curve, since they also know that the demand will inevitably drop once the risk of project disasters has faded from memory.

Organizations trapped in this cycle will experience myriad pathologies of business decision-making, which makes all these problems so frustrating to encounter. They are all so avoidable – but you can’t tell that to those who manifest a tendency to discount or even minimize what project management, as a discipline, can bring to the board room.

And that’s why we should pity the poor Project Management proprietor.

As my regular readers know, I regularly tear into the risk management types for the multitude of foibles that lace their approach to project management. I think they overstate their ability to provide relevant business information, their data is far, far more subjective than they are willing to admit, and that Nasim Taleb (in The Black Swan, the Impact of the Highly Improbable) was right when he asserted that overuse of the Gaussian Curve is the greatest intellectual fraud of the 20^th century. But, from a strategic management point of view, their error lies in the assumption that, given enough quantitative analysis, the best strategy can be discovered for most business situations. My take is that the best strategy is one of providing for a robust response to the unforeseeable events that will inevitably transpire on a given project, rather than attempting to calculate the likelihood and impact of such events unfolding. Is there an example out there that illustrates these rival positions?

Of course there is! (Why else would I have posed the question?) Many senior managers today had their view on the world influenced by the television series Star Trek. The original version of the series (nicknamed “the original series,” or TOS) had an iconic episode entitled “The Doomsday Machine.” In it, the starship Enterprise comes across her wrecked sister ship Constellation, which has evidently been in a horrific battle. As Captain Kirk and a boarding party investigate, they come across the  Constellation’s lone survivor, her captain, Commodore Decker (played brilliantly, I might add, by William Windom). Decker tells of what happened to his ship and crew, how they were attacked by an automated weapon of immense size and power. Just as Decker and Doctor McCoy beam back to the Enterprise, the “planet killer” reappears and attacks Enterprise.

With their Captains on board each other’s vessels, we now see played out how their competing strategies are enacted. Commodore Decker takes command of  Enterprise in order to continue with the strategy of “hitting it (the planet killer) with full phasers, at point-blank range!” Unfortunately, this is nothing more than a slight derivative of the strategy that ended up wrecking Constellation, and killing her crew, a fact that Enterprise’s first officer, Spock, points out. Spock also provides the reason that that strategy didn’t work – “The object’s hull is solid neutronium. There is no known way of blasting through it.” I would imagine not – neutronium is the almost incomprehensibly dense metal at the core of white dwarf stars, as they continue their path towards supernova and , eventually, black holes (how the creators of the planet killer extinguished the white dwarf star, mined and shaped its nuetronium into a funnel-like shape, and integrated its propulsion and defensive systems is not addressed). Decker’s strategy did not represent a very robust response the first time, and his modifications to that strategy won’t fare any better.

Kirk, on the other hand, accepts Spock’s analysis that his ship’s traditional weapons will be of no use in their current situation, and immediately abandons the strategy of attacking with those weapons. When Decker is relieved of command of Enterprise, he steals a shuttlecraft and commits suicide by flying it into the maw of the planet killer. Enterprise detects a minute reduction in the planet killer’s power emanations, and Kirk devises a new strategy. He will rig the Constellation’s impulse engines to detonate as she similarly flies into the planet killer’s interior. Naturally, there are some tense moments as Enterprise’s transporter system goes on and off-line, but, at the last possible moment, Kirk is beamed off of Constellation as she flies into the planet killer and detonates, permanently disabling the automated antagonist.

Note that Kirk had no pre-conceived strategy; or, if he did, he abandoned it immediately. He adapted to the situation, and developed a robust response to events as they unfolded. Conversely, Decker had a set strategy, and refused to significantly modify it, even in the face of overwhelming evidence that it didn’t work, and wouldn’t work in the future with nothing more than changing the range with which he confronted his foe. Don’t get me wrong – strategizing is great. But the truly insightful manager will be willing to abandon any strategy when the situation on-the-ground (or in deep space) calls for it.

And for those critics who might be tempted to assert that few real-life management situations are sufficiently similar to 50-year-old Star Trek episodes to provide insight, I have one question:

Are you really a Klingon infiltrator?

…and they’re not necessarily evil.

In last week’s blog (“Here’s a Communication Strategy: Don’t Communicate Your Strategy!”) we covered the folly of communicating any of the particulars of the organization’s intended strategy, due to one of three outcomes:

·         It’s so general that it really doesn’t communicate anything, and so wastes your intended audience’s time,

·         It’s accurate, true, and insightful, delighting your customers and shareholders alike, but tips off the competition to what you’re about to do, and

·         It’s inaccurate and untrue, angering your customers and shareholders.

As I discuss in my recently-released, must-have book, Game Theory in Management, one very interesting game that illustrates the dangers of transmitting your strategy is Diplomacy, regarded as the board game that has triggered more fist fights amongst friends than any other.

Diplomacy was supposedly the favorite game of both JFK and Henry Kissenger. It’s played on a board that has a representation of a map of Northern Europe just prior to the First World War. Each player is assigned a nation (or a part of a nation), and ground forces that can be used in either attacking other players’ territories or defending your own. The individual players don’t take turns – all players move their markers simultaneously. But here’s the catch: prior to the players actually moving, they spend time talking amongst themselves, making deals and scheming, and forming alliances, both true and fake. When it’s time to perform the move, the players write down on a piece of paper the move they will actually make. Then, at a given signal, the papers are turned over, and the markers moved accordingly.

The fascinating thing about Diplomacy (both the game and, I suppose, actual diplomacy) is that it’s virtually impossible to win without engaging in deceitful strategies. If, during your pre-move communications, you are always truthful, you have no hope of winning. In this, Diplomacy is very similar to Poker. The Poker player who always bids up winning hands and always folds on poor ones will be taken to the cleaners by the other players at the table.  At some point the strategies of underselling a winning hand, or overselling a losing one (“bluffing”) are called for, if, for no other reason, so that the other players (read: competition) do not have solid basis for anticipating your next move. Generally speaking, if your competition can accurately anticipate your next move, you will lose, and lose big. Consider the amount of energy going into keeping the timing and placement of the Normandy and Inchon landings a secret, not to mention Pearl Harbor. Preventing your competition/adversaries from correctly anticipating your next move is a very large part of strategic management, indeed.

As I’ve been writing repeatedly, asset, project, and strategic management are different animals, with different goals, tactics, and information systems. If you neglect or dissemble in your asset management information system, then you are engaged in fraud, and you will probably go to jail (e.g., Enron). If you fail to set up or ignore your project management information system, you  should have never been placed in a managerial role, since you are inviting disaster (e.g., the National Ignition Facility). However, if you fail to insert some level of deceit in the strategic management information coming out of your organization, then it won’t take long for the other players at the table to come to a place where they can correctly anticipate your next strategic move. When that moment comes, your organization’s days are numbered.

Of course, the implication from the previous paragraph is that, for any software package that pretends to support the strategic management function, some level of informed deceit must be included. But, since this is never the case, the inescapable conclusion remains: none of the so-called enterprise or portfolio management systems currently available can truly fulfill those functions.

And those assertions, dear readers, are completely devoid of deceit.

A Google® search on the words “Communicate with stakeholders” returned over 8.5 million hits, and, without reading all of them, I’m fairly confident that most of these sites believe that such communications are a good thing. The problem here, as with other overly-extended project management precepts, is that the definition of “stakeholder” varies significantly depending on who you’re talking to, but it’s almost always overly broad.

According to eHow – money,

The broad definition of a “stakeholder” is anyone in a position to affect or be affected by the actions of a group or organization. [i]

Well, according to Metcalf’s Law (dealing with the ability of very small variances in distant nodes of a large network to have a massive, cascading effect on many other nodes of that network. Also known as the “Butterfly Effect”), that’s just about everybody on the planet.

To be fair, most of the scholarship on communicating with stakeholders does specify the need to tailor the message to the intended audience. But here’s the problem with that: if by “tailor” we actually mean “change,” then the message sent to the employees may be different than the message sent to customers, or even shareholders, and such changes are bound to be seen as duplicitous.

There’s also the problem of such “messages” falling into the hands of the organizations’ competitors. This threat is the primary reason why no organization’s mission statement ever contains any information that is, well, usable. It’s the same “delivering the best value to our customers while pursuing the goals of our shareholders” silliness, re-phrased in ever more convoluted syntax (and almost always in the weak passive tense, which makes me insane).

In my recently-released, must-have book, Game Theory in Management, I discuss games which have as a component communications among players. One of the most iconic is Chicken, made famous by the scene in Rebel Without a Cause where Buzz Gunderson gets stuck in his car as it careens off of a cliff. Whether both cars are headed for each other, or towards a cliff, the basic payout matrix is the same:

·         If Player A swerves and Player B does not, then Player A is considered to be cowardly (“chicken”), and Player B is considered brave.

·         And, vice-versa: if Player A does not swerve, but Player B does, A is brave, and B is chicken.

·         If both players swerve, then both are considered chicken (then why play in the first place?), and

·         If neither player swerves, then both die in a fiery car crash.

If this were represented in a payoff matrix, it would look like this:

So, just based on the rules and this payoff matrix, the only reasonable strategy to adopt would be to always swerve. But, if that’s the case, why would the game theory analyst ever even engage in a game of Chicken?

Because of the pre-game communications. Remember, based on the eHow—Money definition, your Chicken opponent is certainly a “stakeholder.” If, prior to actually getting into your car and speeding off to the critical decision point, you were to have a talk with your opponent, and convince him that you are either extremely brave (or crazy), and absolutely will not swerve, then you have increased the odds of selecting the no-swerve strategy, and living past then next ten minutes.  Conversely, if, after your interaction with your opponent, you come away convinced that he is brave enough (or crazy enough) to not engage in the swerve strategy, you would be well-advised to update your approach accordingly.

Which brings us back to why would anyone in the business world would communicate a strategy that’s worth hearing. If the message being delivered is inconsistent across the projected audiences, then the strategist will be seen as duplicitous. If the strategy being communicated is accurate and provides actual insight, then the competitors will inevitably catch wind of it, and use it to your strategic disadvantage. In attempting to communicate your strategy, you are either wasting people’s time, lying to them, or giving your competition an advantage. So, here’s my recommendation for communicating your strategy:

Don’t communicate your strategy.

“Plans are nothing. Planning is everything.”

--Dwight D. Eisenhower

Strategic management is an odd animal, similar to those creatures that are actually combinations of known animals that appear in classic mythology, like the griffin. This makes writing insightful generalities about strategic management very tricky, since any guidance on how to arrive at the best strategy for a given management scenario is going to (obviously) depend heavily on the scenario. That having been said, there are, actually, some valuable (if conversely axiomatic and counter-intuitive) methods for developing a winning strategy, in virtually any situation. And it has to do (wouldn’t you just know it?) with management information systems.

Management writers (like me) love using war analogies. They just seem to fit so well – instead of people’s lives and property being in the hazard, though, in business we see people’s jobs (or even careers) and corporate assets being arrayed against each other, and the manner in which the winners win offers a certain fascination. And the wartime analogy I wish to invoke to gain insight into how one arrives at a winning strategy has to do with the legendary Battle of Midway.

In mid-1942, the United States had suffered a string of defeats at the hands of the Imperial Japanese Navy (IJN), with the only high-point being a tactical draw in the Battle of the Coral Sea, in May. By the time of that battle, U.S. Naval Intelligence had become aware that the IJN had plans to launch a massive attack against a target known only by code-breakers as “A.F.”  Suspecting that “A.F.” might be Midway Island, the USN arranged for radio operators on Midway to transmit in the open that their water distilling plant was having operations difficulties. Sure enough, a few days later the IJN passed along a message that “A.F.” was short of drinking water.

And the target wasn’t the only thing the USN knew. Long story short, by the time Chester Nimitz began to plan his strategy, he was in possession of the approximate number of ships and planes arrayed against him, their timetable – just about the IJN’s entire order of battle. Conversely, since the Japanese scouting submarines did not arrive at their stations until after the American fleet had sailed from Pearl Harbor, Admiral Chuichi Nagumo had absolutely no idea of the number of fleet units, their timetables, or (most importantly) their whereabouts leading up to June 4, 1942. In reality, the Japanese force was considerably more potent than their American counterparts. They had more ships, of superior firepower and capability. Their aircraft performed better than their American counterparts, and their crews were exceptionally trained. But, by the end of June 5, the Japanese fleet had been shattered, with a four-to-one loss ratio in aircraft carriers alone.

What was the difference in Nimitz’s and Nagumo’s strategies? I believe that the American victory at Midway was due to two distinct advantages:

·         The aforementioned information advantage. Nimitz knew almost everything about Nagumo’s order of battle before the shooting even started, while Nagumo knew almost nothing of the American fleet’s disposition until dive bombers were tearing up his flight decks, and

·         The strategic approach to the battle. Nagumo had a very rigid set of tactical goals, to be executed on a specific schedule. Nimitz essentially sent Admiral Raymond Spruance, the tactical commander, to a position to the North and East of Midway (“Point Luck”) with the American fleet, to simply respond to whatever the Japanese did. Spruance had greater latitude of action, meaning that he was far more likely to come up with a robust response to whatever happened in the battle area.

And those, I believe, are the two keys to developing a winning strategy in the business world. Ensure your organization’s management information systems produce timely, accurate, and, most of all, relevant information when forming a strategic approach. And, whatever you do, do NOT initiate a rigid set of tactics that must be executed on a given timetable. Instead, allow your managers on the project to develop a robust response to whatever happens to them as the project team executes its scope.