Wise Words Podcast Now available on all major podcast channels.

obliquity-book-summary-john-kay

Obliquity Book Summary – John Kay

What you will learn from reading Obliquity:

– Why setting goals, objectives and measurements might not actually solve the problem.

– How we consistently make mistakes by not considering alternatives when making decisions.

– What happens when our models of the world fail.

Obliquity Book Summary

Obliquity Book Summary is all about problem solving. The main thesis is this, problems are often best solved indirectly. Iteration and experience become the best tools for problem diagnosis. That is to say, you learn about the problem in the process of solving it. Therefore, problem solving is always a learning process.

 

Our goals seem simple when broken down:

Building a cathedral seems a problem amenable to direct solution. Priests issue commissions, architects draw up plans, constructors put the stones in place.  

The problem seems to be determinate – we know the objective.  

The problem seems to be closed – we can define the possibilities open to us.  

The problem seems relatively simple – we understand the relationships between the different components. 

 

Achieving your goal isn’t everything:

The cricketer and author Ed Smith expresses it well: ‘I am not saying that personal development is more important than winning; on the contrary, I am saying that enjoying the journey of self-discovery, by removing some of the pressure and angst associated with winning at all costs, is one way of helping you to win more often.’ 

 

Pseudo-Rationality Strikes in Business

Most business environments have these characteristics: 

Decision making cannot proceed by defining objectives, analysing them into goals and subsequently or any other environment breaking them down into actions.  

No priest or politician, counsellor or manager, has the capacity to do this – and those who claim to, like Le Corbusier or Lenin, have an immense capacity to damage the complex systems they attempt to plan. 

 

You should look for oblique solutions not direct ones:

Such obliquity of approach distinguishes the genius from the merely competent, the creative problem solver from the computer.  

The computer is very good at solving the problem we have specified and asked it to solve, but less useful when we are not quite sure what the problem is. 

 

How to solve complex problems:

The oblique solution of complex problems is a matter of managing the interrelationships between the interpretation of high-level objectives, the realisation of intermediate states and goals, and the performance of basic actions.  

Such skilful interpretation is required in even the simplest problem. 

 

When the prevention is the cause:

Yellowstone forest – Extinguishing all fires made the forest more vulnerable, and the bonus culture destroyed the organisation that fostered it. 

Yellowstone blazed, and Lehman failed, because those who imposed direct solutions to problems did not appreciate the complex relationships between objectives, goals and actions.  

 

Look for options and alternatives. There are many ways to same goal!

Lindblom’s thesis was that practical decision making is necessarily oblique. Such an approach, he says, involves no sharp distinction between means and ends and drastically limits analysis by problem simplification and ignoring many potentially available options. 

The process in which well defined and prioritised objectives are broken down into specific states and actions whose progress can be monitored and measured is not the reality of how people find fulfilment in their lives, create great art, establish great societies or build good businesses. 

 

Sudoku and Bounded Problem Solving:

The methods of analysis that come naturally to us are oblique, and we do not use or enjoy direct, mechanical approaches. Anyone who buys a computer program to solve a Sudoku problem has missed the point of the game. 

If the world were like Sudoku, decision making could be tackled in an equally direct way.  

 

The characteristics of Sudoku that make such an approach possible are: 

There is one and only one solution, and when it is identified we know that we have found it.  

Objectives are clear and constant.  

The play is not influenced by the responses of others to moves that are made. Interactions with others, if they are relevant at all, are limited and controllable or predictable.  

There is a complete list of possible actions and we know that all the potential actions we consider are in fact available to us. Even if we do not know what will happen in future, we know the range of possibilities and can sensibly attach probabilities to them.  

The problem is closed.  

 The number of alternative ways of filling in the boxes, although running into many millions, is nevertheless sufficiently small that all can, at least in principle, be evaluated.  

Complexity, even if extensive, is bounded. 

 

When to Use a Direct Approach:

Where objectives are clear and simple and policy and implementation can readily be distinguished, when interactions with others are limited and predictable, when we are confident in our ability to specify completely the available options and the risks to our objectives, when we understand the systems with which we deal, when we can feel confident in our abstractions, our approach can be more direct. 

 

There are more variables then you think:


The many Variables involved in creating a sustainable business:

The creation of a sustainable business – a high-level objective – calls for achieving a variety of intermediate goals profitability, good products, motivated employees, customer satisfaction.  

In turn, these goals require a series of actions – cost reduction, pricing policies, product launches. 

 

The many Variables in Education:

There are many aspects to education – vocational training, citizenship, emotional development.  

True: we need to translate the high. level objective – a fine education – into goals and actions that can guide the framers of a curriculum. 

Teaching quality is defined as the relationship of the classroom experience to the intended learning outcomes.  

If you are already shuddering at the jargon of as well you may – then you the educational administrator are premature. Good teaching really does consist of classroom experiences that generate intended learning outcomes. 

 

We over quantify everything — are we really measuring the right things?:

Kelvin’s approach leads directly to the modern curse of bogus quantification. The United Nations produces an index of human development (HDI; see Figure 8), under which countries are ranked from Iceland (at the top) to Sierra Leone (at the bottom).3 The high-level objective of human development is translated into three goals or states: longevity, educational standards and gross domestic product (GDP). 

For Berlin, there was no universal answer to the question ‘What is the nature of a good society?’  

Berlin is surely right to say that the question ‘How should I live?’, like the question  

‘What is the objective of education?’, and so many others, has more than one answer. The goals we struggle to attain are often incommensurable and even incompatible. 

 

Changing the Rules:

What is true of art is also true of other areas of human endeavour. What made Henry Ford or Walt Disney or Steve Jobs great businessmen was that they modified the rules by which their success, and the success of others in their industry, were measured. 

The criteria of achievement are constantly redefined by great achievers. 

 

Seeing the trees from the woods – understanding the bigger system:

The foresters saw the trees, but not the wood. You cannot necessarily deduce the properties of the whole by adding up the properties of the individual parts. This is true of many biological systems and of all social, political and economic systems. 

 

Franklins Gambit – We rationalise the decisions we have already made:

Franklin also knew that moral algebra was generally a rationalisation for a decision taken more obliquely.  

That is Endino why as well as Franklin’s Rule he set out what I earlier called Franklin’s Gambit – ‘so convenient a thing is it to be a reasonable creature, since it enables one to find or make a reason for everything one had a mind to do.’ 

The interview report, the loan proposal, the impact assessment, the risk evaluation, are usually exemplifications of Franklin’s Gambit rather than Franklin’s Rule. They are written to rationalise the decision that has already been made. 

They made up the numbers to support the conclusions they believed their superiors wanted to reach. This was the process I exploited, profitably, in economic consultancy. 

 

We understand problems better in Context:

Perhaps we may handle problems better if they are embedded in a social situation than in a naturalistic one, and if they involve enforcement of a rule and are not merely descriptive.  

The context of the problem matters to our solution as much as the problem itself, so that our approach is intrinsically oblique. 

 

When our Models fail – A bus Schedule:

We model the way things work example – bus time schedules. If the model is a good description of the world, their nervousness is unjustified.  

But after eleven minutes they may reasonably start to wonder whether the model describes the world well.  

Perhaps they were misinformed about the schedule: perhaps the route has changed: perhaps there is a strike at the depot or an accident on the road.  

 The list of reasons why the model might fail is long and necessarily incomplete. After some length of time – fifteen minutes, twenty, twenty-five – most people conclude that their model is not, in fact, relevant and look for another route. 

 

When our Models become uncertain:

The uncertainty about the bus’s arrival has two components.  

  1. There is risk derived from the randomness reproduced within the model: the known unknown.
  2. There is uncertainty about the appropriateness of the model as a description of the world: the unknown unknown.

The first of these components allows an objective description, the second does not: there is not, and cannot be, any analysis which shows that it is right to wait for twenty minutes but wrong to wait for twenty-five.  

When new data arrives – in this case, the information that the bus hasn’t we always have the problem of whether to treat it as new data about the parameters of the model or new data about the relevance of the model. 

 

Models fail when things they fail to take into account happen:

The risk models that financial institutions use ensure that it is very unlikely that these institutions will fail for the reasons that are incorporated into these models. That does not mean that they will not fail, only that if they fail it will be for other reasons. 

 

When computers get it wrong – The Sat Nav Problem:

Computers are replacing ticket sellers. In 2009, I asked the Transport for London website how to get from Paddington Station to Hyde Park Gardens.  

It told me to take a bus in the opposite direction, and then retrace the route of the bus on foot to Paddington Station. When I arrived there, I should walk directly to Hyde Park Gardens.  

There is an absurd logic to this proposal. The route is the quickest way to get from one location to the other by London Transport. It is, of course, quicker still not to use the bus at all.  

The computer’s ludicrously oblique solution is the direct answer to a badly formulated version of the problem. 

 

Remember results change the lens people view the action:

The qualities of charisma, boldness and vision that had once won plaudits for Barnevik were now portrayed as arrogance, aggression, resistance to criticism.  

Of course, neither ABB’s organisation nor Barnevik’s personality had changed: what had changed was the company’s results, and hence the lens through which the organisation and the personality were viewed.” 

The mistake is to make inferences about the relationships between outcomes and processes when we cannot observe and do not understand the processes themselves. 

 

Hedgehogs, foxes and avoiding credit for your decisions:

Hedgehogs are people who know the answers. Foxes know the limitations of their knowledge. Hedgehogs create headlines for journalists, and their confident certainties attract the attention of politicians and business leaders. 

Machiavelli understood that to be an effective decision maker it was wise not to seek public credit for the success of your decisions. Yet another of obliquity’s many paradoxes. 

 

A Great Joke about Green Lumber Fallacy:

A well-known joke tells of the economist in the wilderness who, when he sees a bear approaching, pulls out his computer and begins to calculate an optimal strategy.  

His colleague, appalled, says: ‘We don’t have time for that!’ ‘Don’t worry,’ replies the economist smugly, ‘the bear has to work out an optimal strategy too.’  

Behind the joke lies a deeply serious point. The bear gains a decisive advantage by not suffering the illusion that the approach based on calculation might work. 

 

The Illusion of purposeful design:

For years I struggled with the idea that if profit could not be the defining purpose of a corporation, there must be something else that was its defining purpose. If business did not maximise profit, what did it maximise?  

I was making the same mistake as those victims of the teleological fallacy who struggled for centuries with questions like ‘What is a tiger for?’ Tigers, we now understand, are not the product of any purposive design.  

Tigers are the creature you would design if you were more skilful and knowledgeable than you could ever be, to do the sorts of things that tigers do. But that is not how they came into existence. 

 

Remember, Irrationality is an evolutionary tool:

This conception of irrationality doesn’t mean that the subjects solve problems badly, just that they don’t solve them the way the experimenters think they should. But we get angry and reject outcomes we think are unfair for good evolutionary reasons. 

 

Revisiting Franklins Gambit:

Franklin’s Gambit is perhaps the most common fault in decision making – and particularly in public decision making – today. 

There is an appearance of describing objectives, evaluating options, reviewing evidence. But it is a sham. 

The objectives are dictated by the conclusions, the options presented so as to make the favoured course look attractive, the data selected to favour the required result. 

Real alternatives are not assessed rigorously: policy-based evidence supplants evidence-based policy.

 

Real Problem Solving:

The oblique, unaccustomed, perspective was how Brunelleschi cracked the egg, built the dome of Santa Maria del Fiore and discovered how to represent perspective.  

Many great achievements are of this kind. Alexander Graham Bell’s invention of the telephone, like Akio Morita’s creation of the Sony Walkman and Steve Jobs’s reinterpretation of Morita’s idea in the iPod, were solutions to problems people did not know they had. 

Iteration and experience lead us to the best principles of analysis. In obliquity we learn about the structure of a problem by the process of solving it.  

Klein’s paramedics and firefighters became competent by learning the rules and became good through practice.