Cost reduction programmes rely on fast, effective decision making, which in turn relies on high quality data and a thorough understanding of the situation and the cause and effect mechanics at play. However, the modern world is a particularly complex and opaque one, and the unfortunate truth is that human beings do not cope well with complexity. The complexity and human weakness means there are many areas where no matter how much information we gather and no matter how carefully we analyse it, a strategy of predict, plan and control will fail.

Complexity

The world is complex – this may seem an obvious statement to make, but the complexity is more pervasive and constraining than is generally recognised, leading to poor investment and control decisions. In fact the complexity is such that the world is often too complex to understand. We are often very poor decision makers (better suited to the simple life of a hunter gatherer on the savannah), as amply demonstrated by the high rate of project failure, business failure and macro-level policy failure. Often our only hope to make any decision making headway is to make significant simplifications and assumptions, and then, and this is the critical bit, to recognise this uncertainty and the limits of validity of any conclusions. We will often arrive at a point where the search for the optimum solution is futile. People that believe otherwise are normally talking bullshit, or operating in an artificially simplified environment.

Analysis of business survival shows similar patterns to biological survival rates, and in short shows that despite the ability of humans and companies to act with intent, in reality it is as if they operate close to the paradigm of the agent with zero cognitive ability. I.e. very limited ability to acquire knowledge about the true impact of their strategies on others or of others on them, or to put it another way, no better than a monkey throwing darts.

A simple starting point is often taken to mean a simple solution, yet tremendous complexity can result from the simplest of starting points. Take a massively simplified business decision such as to where to pitch three ice-cream huts on a beach, or a game of chess. A chess game has 32 pieces, clear rules and completely transparent and clearly and singularly defined opposition and success criteria yet the ability of humans to understand is limited, and even the most powerful computer has yet to assess all possible combinations.

Given a real project has significantly greater number of constituent parts, has far more interactions, imperfect data, unknown influencers, and human players, reality will always be more complex that some simple model. Yet we insist on teaching students and practitioners that we act as rational agents, making optimal decisions. In fact the best we can hope to do is find reasonable solutions that satisfy certain parameters. In most real world situations, it is simply not possible to 'maximise', to find the optimum solution. This unfortunate reality is kept at a comfortable distance in a number of interesting ways.

• Grossly overestimating the accuracy and completeness of the data we use,
• Grossly overstating the accuracy of our own predictions and never looking back
• Creating the illusion of strong “control” processes

Accuracy of data and predictions

Many, if not most individuals grossly overstate the accuracy, completeness and relevance of the information they use, and simultaneously, grossly overstate the accuracy of their predictions (not surprising when 70% of the population believe they are of above average intelligence). Our inability to grasp the uncertainty in the data we use and predictions we make, whist perhaps discussed at length in the arena of social sciences and some branches of economics, does not appear to have breached the ramparts of the PPM profession to any great extent.

There is systemic optimism in financial and time data, i.e. programme cost, schedule and budget. The reason so many projects are over budget or late is often not because of poor delivery. It is down to hopelessly optimistic initial estimates, founded on a gross underestimation of the inherent complexity and requirements of the task at hand.

In the author's experience, this problem of overstated accuracy, gets worse the more senior the person, with many Board and Exec managers simply unable to cope with or accept uncertainty. One MD would finesse the inputs of a project financial evaluation to give say a 5.1% return, despite the inputs being very imprecise (cost to serve raw data was 3 years old, the cost allocation incredible crude, and the sales total and mix projections inherently uncertain). Completely false precision and accuracy, combined with tremendous detail, gave a sense of comfort and allowed him to show a strong business case. To be fair he was partly driven by an absolute fear of portraying any uncertainty to the Group CEO. Stating uncertainty was seen as being vague, incomplete and a sign of poor management. With that culture in place, it is no wonder that risks and issues were not escalated and managed, they lived in permanent crisis management and a leadership revolving door.

It is necessary to be critical, if not outright sceptical of predictions made, particularly by “experts”. Always ensure that the outputs of detailed analytical reports do not portray false precision in the results, and always focus on the confidence limits of the answer. The potential deviation from the expected result, can be just as important if not more so than the result itself. Plans made must recognise the future uncertainty, and build in strategic, tactical, and operational flexibility. This may mean being inefficient when retrospectively judged, but in practice over time being more effective. The quest for efficiency can be the enemy of effectiveness. The current focus on efficiency will inevitably lead to some complete operational and project failures that ultimately cost far more to repair than the initial slim efficiency gain. Cuts in operational flexibility, contingency capacity, assurance, training, contingency, professional support etc are all examples.

Control

Control is an enticing prospect. Who would not want to seek more control? However the typical application of control governance and processes can deplete an organisations delivery, agility, and flexibility and lead to a excessive reliance on data and metrics. As an piece in Harvard Business Review (April 2010) put it “There is a deep seated desire to quantify the world around us so that we can understand it and control it. But the world isn't behaving.”

A false sense of accuracy can give a false sense of control. When the inevitable deviation occurs the response should not necessarily be a tightened grip. In fact as we saw earlier, tighter, more frequent control can lead to false alarms and significantly reduces flexibility and innovation.

One example of this was an £800m business where the sales total for the division were sent out daily to the entire Executive Committee. But sales per day were highly variable, day to day (+/- 40%), and week to week (+/- 20%) influenced by bank holidays, school holidays, the weather, differing corporate buying profiles and changing customer mix. The daily actual was compared to the budgeted daily expectation; a budget crudely created. Therefore, large variances were reported, frequently prompting a game of email ping-pong and knee jerk responses to often non-existent problems. The more frequent reporting, instigated because of not meeting budget, created an illusion of improved business control, but was ultimately distracting and counterproductive.

This highlights the need to separate out the noise and the underlying useful data. All metrics have noise, and the more detailed and the more frequently the data is reviewed, the more likely that the effects are down to noise not underlying signal. This raises the stress levels and causes unnecessary attention to essentially random events. Creating a complex, real-time control system, whilst in theory better, can create counter-productive feedback loops and unintended consequences. Take a financial investment. Checking an investment in the FTSE 100 every minute of the last year would involve half a million checks, of which say 49% would show a drop in value, even if overall in the year the shares went up 10%.

An insightful analogy comes from WWII and the 617 'dambuster' squadron. Their Lancaster bombers were suffering substantial loses, and protective armour was being incrementally added to the planes, in response to each specific incident. However, this adding significant weight to the plane. One day the squadron leader, watching one of his planes struggle into the air, had the insight that whilst individually each additional protective measure had a rationale, the upshot was a plane that was slow, unresponsive and much more vulnerable to attack in the first place. He ordered the removal of all but the most essential protection. Losses dropped immediately. Similarly, some operational and programme management practises, when badly implemented can feel like a suffocating bureaucracy, all with good intent, but ultimately counter-productive overkill.

Conclusions

Reality is far more complex and uncertain than the typical individual, organisational culture and PPM toolset can normally deal with, and we consequently have to make significant simplifications to enable us to proceed. However, the necessary simplification may be so severe it invalidates the answers in most situations, or forces us to accept 'satisfactory' results rather than the optimised solution – better to be approximately correct than precisely wrong.

Controls put in place, particularly when based on more detailed and especially more frequent datasets and reviews can have significant negative impacts and unintended consequences.

Despite the challenges of complexity and uncertainty, it is worth noting that we still make insufficient or incorrect use of the data, tools and analysis that are available and relevant. So on the one hand we need to be far more discriminating in the analysis we undertake and our acceptance of the limitations of that analysis, and yet there are significant untapped opportunities to create tremendous value through deeper and more joined up data dives.

*Why most things fail. Paul Ormerod