I find it fascinating how often one of the most basic distinctions in project management information systems is overlooked, or, even worse, blurred into near-meaninglessness. I’m talking about the difference between feedback and feed-forward management information systems (MISs).
For those of you who are (understandably) coming to this topic for the first time, feedback systems report on what has occurred, based on verifiable information. The general ledger is wholly predicated on the feedback concept, as are several other systems. Advantages to feedback systems include:
· The data is (or ought to be) accurate, with little or no subjectivity present.
· …and, well, that’s about it.
Here are the drawbacks to feedback systems:
· Depending on how long it takes to gather the data, process it into information, and get it to the decision-makers, the information may become so dated as to become marginally useful.
· There’s also the problem of the point of view of those interpreting the data. Dots connected by one manager into some sort of causality loop may strike another as purely coincidental.
Then there’s the feed-forward system. Such systems seek to anticipate the future, such as the use of polls preceding an election. The advantages of such systems include:
· Obviously, a management information system that can accurately predict the future is pure gold,
· …leading to its desirability.
These very advantages lead directly to the feed-forward systems’ drawbacks, including:
· Most such systems depend on subjective data to such an extent that they are little better than reading tea leaves, or other forms of divination.
· But, since they are so desirable, promoters of unproven techniques are drawn to design them in droves.
Of course, the very nature of experience (and the function of the hippocampus, if you want to get technical) is to assess facts and past events, organize them into some sort of structure, and flip that structure forward, across the Time Now line in order to derive a useful anticipation of how the future will unfold. While this little thought exercise may be part and parcel of carrying on our lives, it tends to fail as a repeatable, teachable tactic in management science. The future simply cannot be quantified, which is why the best managers are the ones who can think on the balls of their feet, adapting to new and unforeseen circumstances with (mostly) appropriate responses. PMs who stick to their plans, even in the face of dramatically changed circumstances, are usually the ones behind the projects that end in failure, if not disaster.
Which brings us back to the types of information systems that PMs like to use to learn about what’s going on in their projects. Both earned value and critical path methodologies have a remarkable capacity for predicting future performance, based on the reliable aspects of feedback systems. EVM and CPM, when set up correctly, are capable of consistently predicting final project costs and durations within ten points, an accuracy rate that can only be dreamed of by purveyors of the general ledger, or risk analysis tools, for that matter. Since they are predicated on observed, quantified past performance, such systems have the advantages of both the feedback and feed-forward types, without their drawbacks, which is probably the main reason why accountants and risk managers tend to have a natural resentment of project controls analyst. I could possibly contribute to the lessening of this resentment, if I would just stop making fun of accountants and risk managers in this blog
I’m totally cool with this month’s ProjectManagement.com theme, that of visual project management. I’m a firm believer that, once the project controls staff has collected the relevant data, processed it into usable information via Earned Value and Critical Path methodologies, they need to deliver that information in not only a timely fashion, but in a format that’s sufficiently intuitive for the decision-makers to use. Let’s face it – for new PMs with little more business acumen than what they accidentally swerved into while they were pursuing their engineering degrees, hitting them with a PERT chart or a Cost Performance Report in Format I in their first project review meeting doesn’t really help them select the most appropriate strategies in pursuing their projects’ objectives.
That having been said, is there an opposite, equally inappropriate extreme? There’s no doubt about it. And Exhibit A has to be the so-called Stop Light Chart.
For most people, a traffic light is a very handy tool for conveying critical information in a short period of time. Red equals stop, green is okay-to-go, and yellow means clear the intersection (unless you’re like the taxi driver who took me from Logan International Airport into downtown Boston, where it means accelerate to the max, even if you are up to ¼ mile from entering the intersection). For professionals who suddenly find themselves in the role of Project Manager (hey, they don’t call it “the accidental profession” for nothin’) without ever having actually taken a management science course, the reduction of critical cost and performance information into one of these three familiar colors on a readily available report is tempting in the extreme. Green indicates good or acceptable performance, yellow means pay attention here, and red is pinned on those projects/tasks that are currently in trouble – easy, right? Well, not so fast…
On what grounds, exactly, are these categories assigned? If you have a task at the reporting level with a positive cost and positive schedule variance, green is clearly called for, as is red for negative-negative. But what if you have a positive cost, negative schedule, but the cost is way positive (>15%), and the schedule is barely negative <5%)? Is that yellow?
Here’s my heartburn: by the time the reporting system is dumbed down to one of three or four colors (blue = completed on-time), all sorts of irrelevant information becomes the basis for color determination. Before you know it, the cost determiner is a comparison of budgets and actual costs, which, of course, is irrelevant to real PMs, but not to amateurs. A chilling trend among many project organizations is to perform schedule “analysis” based on whether or not the responsible PMs believe they will attain their milestones on-time. This, dear readers, is not legitimate project management information. It is polling, masquerading as usable cost and schedule performance data.
One of the most important aspects of successful project management – if not the most important aspect – is the ability to identify what the project’s issues are, and what aren’t. A project manager’s time is limited, and a performance information system that returns false positives, or masks genuine problems in overly generalized and synthesized formats, virtually invites mis-directed energies. Worse, it helps prevent authentic systems from being introduced, since the bogus systems give the appearance of being able to keep the decision-makers apprised of the project’s goings-on.
But, hey, my Boston taxi driver got me to where I was going, apparently without noticing what the traffic signals were telling him (and, seemingly, little information coming in from an accurate visual assessment of the traffic around us), so your project should be okay.
I believe that part of the reason why risk management-types have a hard time getting the management science community at large to actually execute the types of analysis they keep harping on about is that the underlying concepts are not very intuitive. Besides their relentless abasement of the language (e.g., you will search dictionaries of the English Language in vain for the definition of “opportunity” being somehow related to “upside risk”), the risk managers’ attempts to expand the use of Gaussian curves into areas of business analysis where they truly do not belong via eat-your-peas-style hectoring are not conducive to more wide-ranging acceptance. What’s needed is a vehicle for them to reduce the results of their analyses into more intuitive formats, similar to last week’s blog’s discussion of alternate ways of portraying (valid) cost and schedule performance information. I think I may have found this vehicle – the Chernoff Face.
Herman Chernoff is an American applied mathematician, currently working at Harvard. He devised a way of conveying information in a graphic format that could be readily understood by a variety of people, with varying degrees of sophistication. This format is called the Chernoff Face. The desired information is compared to a baseline for each of its components; then, when the data deviates from that baseline, the aspects of the face change correspondingly.
Consider the following cartoon face:
We’ll use this as our baseline. Now, the risk management types categorize risk events as administrative, insurable, technical non-somethingorother, Sagittarius, and only within the last two minutes of a half. Since administrative risks tend to be low-impact, we’ll adjust the face to include smaller diameter eyes, upticked brows, and a bit of a frown, so:
When the risk assessment indicates a low-impact, administrative risk, this pensive-looking Chernoff Face does a much better job of conveying this information quickly and intuitively than some decision-tree alternative listing. However, when a serious threat appears on these guys’ radar screens, a more dramatic look is indicated. For real trouble, we should increase eye diameter, push the brows up at a more severe angle, and do something with the mouth, as in:
Aside from little-trouble-big-trouble, though, the risk managers’ main communications have to do with their frustrations with people not listening to them. For those project managers who just don’t get it, in the risk managers’ eyes, this report should do the trick:
Alternately, for those PMs who do understand risk management, and simply eschew it, this style conveys the appropriate sentiment:
…which often leads to the need to figure out how to get at the project’s charge codes anyway, so…
The downside, of course, is that, by conveying risk management data this directly, it goes a long way towards de-mystifying the nature of the analysis when decoupled from the mind-numbing statistical jargon. Well, “downside” from the risk managers’ perspective, anyway…
…and if you’re a PMO Director, you will often have less time than that.
In one of my earlier, high-falutin’ gigs, I was making presentations on portfolio cost and schedule performance to some senior-level executives. A canny assistant to one of them took me aside one day, and told me “Listen, Michael, these guys may be masters of the universe, but they have the attention spans of fourth graders. Whatever you put in front of them has to arrest their interest inside of 30 seconds, or they’re checking the view of the butterflies out the window.”
This advice reminded me of my path-not-taken career, in advertising. My parents ran an advertising agency while I was growing up, and they spent time schooling me in the industry (and were somewhat taken aback when, upon attaining my undergraduate degree, I became a site rat at the Air Force Weapons Laboratory, but that’s another story). For print ads, the rule of thumb was that you had no more than 3 seconds to attract the readers’ attention with your headline; and, if you got them that far, no more than an additional 10 seconds for them to grasp the essence of the entire ad.
Does it strike anyone else as odd that many key decisions, from which types of projects to pursue to where to spend the household’s furniture budget, are so often determined by (relatively) snap decisions, or, at least, snap persuasion? I suppose it’s just part of human nature, a part that bodes very well indeed for us Project Management types (finally!).
My regular readers are, by now, used to my carping about the intellectual vacuousness inherent in the asset managers’ (read: accountants’) automatically-accepted meme of having all of the inside information the organization needs with respect to cost, budget, and project performance. But what do they have, really? The general ledger’s two main (only?) outputs are the balance sheet, and the profit-and-loss statement. Funny thing is, only a fraction of the people who use the term “the bottom line is…” really know how to perform an advanced interpretation of the profit-and-loss statement, and even fewer non-bookkeepers can make heads or tails of the balance sheet. These report formats are literally medieval, arriving around the time of Machiavelli (a coincidence?).
Now, that’s not to say that we PM types are incapable of inducing instant MEGO (a P.J. O’Rourke-ism, short for “my eyes glaze over”) by insisting on generating our cost/schedule performance information in the 1960’s era Cost Performance Report, Format I, or, for the uninitiated, the familiar (well, to us, anyway) Gantt Charts. But, once we move away from these staid formats, the sky’s the limit. One of my favorites is the calculated Variance at Completion histogram.
First, calculate the project’s at-completion costs. A simple but surprisingly accurate formula is:
EAC = ACWPcum / % Complete
… where EAC is the estimate at completion, ACWPcum is the cumulative amount of actual costs on the task or project, and % Complete is the estimate of the task’s percent complete. Next, subtract this figure from the budget at completion (BAC), and you have the calculated variance at completion. For the piece de resistance’, arrange these figures in a table in a spreadsheet, and sort on their values, lowest to highest. Use this to set up a histogram – color all of the overruns red, and all of the underruns green. Include this graphic in your next presentation to the executives, and the following will happen:
· Since it’s a far more intuitive format than anything the accountants have, you will surpass them easily when it comes to conveying critical cost performance information; so, naturally, they will envy you.
· The PMs of the projects on the far left side of the graphic will be instantly shown to be heading up losing efforts, and by how much. Since they will have normally escaped such scrutiny, they will not appreciate being so exposed.
· The executives, on the other hand, will greatly appreciate being so graphically alerted to the problem projects within their portfolios, and will love you for it.
However, you need to beam this information into the heads of your organization’s decision-makers quickly, which means the adoption of a more visual venue than you may be used to. Which reminds me – from the movie Star Trek II: The Wrath of Khan, two minutes and fifty seconds after Kirk’s demand quoted in this blog’s title, a bridge officer reports “Captain! The mains are back on-line!” He knows this because he is looking at … a histogram!
I own a Casio® wristwatch that is suspiciously similar in appearance to an Omega® Seamaster, the watch worn by James Bond (we see a close-up of his wristwatch’s face in the movie Goldeneye). A new Seamaster (“James Bond” version) retails for $3,900 (USD), but can be purchased “on sale” for a mere $3,100. My Casio – which is also a chronometer, by the way – arrived at my door for less than $100.
Now, I readily admit that the Seamaster has many more features than my Casio; but, really, how many current Seamaster owners really need a helium release valve? If the object of the game here is to deliver the time of day, accurately and reliably, to humans walking around on the surface of the planet, then those owners of the Omega watch who do not perform dives of excessive depth or extended periods wear them, at least in part, as a signal to others that they can afford to spend amounts that most middle-class folks would view as excessive on their personal effects. Which is absolutely fine by me, don’t misunderstand – I just think we need to be clear about the form-versus-function balance being attempted here.
The first Project Management Office I headed was part of a small niche contractor, and didn’t compete for contracts quite the same way the established mega-firms did. In this environment and during this time, having a PMO at all was something of an anomaly. Since new contracts would come this company’s way more easily than they would have in an environment of proven-excellence-just-to-submit-a-proposal, the pursuit of outstanding cost and schedule performance wasn’t on everybody’s mind, at least not all the time. The establishment of management information systems that would tip off the executives about which projects were doing okay, and which were train wrecks in the making, was deemed sufficient for their purposes.
Again, don’t misunderstand: I and my staff worked diligently, trying to get all of the decision-makers in the company to recognize the value of doing project management properly. But as we droned on about the need to set up the chart of accounts consistent with the reporting level of the Work Breakdown Structures, or how a cost variance IS NOT the difference between budgets and actuals, weeks turned to months, and months to fiscal quarters. My little PMO’s exaggerated capabilities would be featured prominently on some of the company’s proposals, as if we were one of the most central aspects to managerial decision-making. Away from the marketing spotlight, however, we were largely ignored, or tolerated with a sort-of bemused patience. Some of the PMs would readily accept any help we could give them, and these ones made the whole PMO enterprise worth pursuing. These were, unfortunately, a definite minority within the firm. From that era of figuratively beating my head against the Vice Presidents’ suite office walls, I remember now as alternately an interminable amount of time, and no time at all. The day came when the company was set to graduate from its niche, and had absolutely no record of project execution excellence to point to as they entered the realm of enhanced contract competition. I discreetly left prior to this change, and none too soon: the company, failing to make the transition, was bought out soon afterwards.
The point here is that we can’t all be the Omega Seamasters of the PMO world, and that’s okay. Sometimes it’s alright to simply look like one, while doing the best we can to deliver accurate and reliable cost and performance information, even when our chronometer-like accuracy is taken for granted or ignored. Looking back, not only do I have no regrets, I’m actually very grateful for the opportunity and the lessons I learned during this time. My little PMO did its best to deliver the critical project cost and schedule performance information that the company’s executives should have used to keep the firm viable. I’ve often pointed out that the 20% worst managers with access to 80% of the information needed to obviate a given decision will consistently out-perform the 80th percentile top managers who have access to only 20% of the information so needed. Those execs willingly chose not to avail themselves of the PMO’s information stream, so they performed like … well, you know.
So, was this a “real” PMO? You can bet the difference in wristwatch prices it was.