Summary of Streetlights and Shadows: Searching for the Keys to Adaptive Decision Making by Klein - 1st edition

What does claim 1 entail?

Claim 1 is "Teaching people procedures helps them perform tasks more skillfully". The process of transforming skills into procedures is irresistible. All we have to do is break a complex task down into steps and provide some tips about when to start and finish each step.

Procedural checklists are a shield against interruptions, reducing the chance of distraction. Procedures also help us to evaluate performance. We can see if someone knows the procedures, and is following them.

Procedures are also relevant to health care. Peter Pronovost developed a procedural checklist approach for a specific and common problem in intensive care units, line infections:

• Step 1: Wash your hands with soap

• Step 2: Clean the patient’s skin with antiseptic.

• Step 3: Cover the patient with sterile drapes.

• Step 4: Wear a sterile mask and gloves.

• Step 5: Put a sterile dressing over the line once it is inserted.

Results: Line infections dropped to zero for the next year, and the hospital saved several million dollars.

The Dreyfus model of expertise emphasizes intuition and tactic knowledge that can’t be captured in rules and procedures. People might need some rules in order to get started, but they have to move beyond rules in order to achieve mastery.

Procedures are tools. Every tool has limitations. Organizations often overstate the importance of procedures. Example: during an accident investigation, if someone finds that a procedure wasn’t followed, even if it didn’t directly cause the accident, there is a good chance that procedural violation will be trumpeted as one of the contributing factors.

By appreciating the limitations of procedures and checklists we’ll be able to make better use of them. What are those limitations?

Procedures alone aren’t sufficient

In complex settings in which we have to take the context into account, we can’t codify all the work in a set of procedures. No matter how comprehensive the procedures, people probably will run into something unexpected and will have to use their judgment. It often takes government regulation to force organizations to compile reasonably comprehensive sets of procedures, and those procedures usually have gaps.

One way to ensure that a set of procedures is sufficient is to take every eventuality into account. However, this strategy can result in the opposite problem: procedural guides that are over-sufficient and sometimes incomprehensible. The problem here is that the more comprehensive the procedures are, the more voluminous they become. And the more voluminous, the more forbidding they appear, the more work to find what is needed, and the lower the chances that anyone will try.

Why is it difficult to keep procedures updated?

Procedures are often out of date, because work practices keep evolving. This is called procedural drift. Another problem is that procedural changes that get made are often excessive. The changes, though imposed to reduce the chance of adverse consequences, may create inefficiencies in working with the new procedure.

Because procedures keep evolving, procedural guides are rarely complete.

Greg Jamieson & Chris Miller (2000) studied four petrochemical refineries in the United States and Canada to see how they managed their procedures. In none of the four cases did the workers ever completely trust the procedural guides and checklists, because they never knew how updated these guides were. Over time, some procedures had become obsolete or even counterproductive.

But there is a bigger problem than the fact that procedures are rarely sufficient and often out of date. In many cases, procedures can make performance worse, not better. They can lull us into mindlessness and complacency, and cause an erosion of expertise. In some cases, procedures can mislead us.

How can procedures lead to mindlessness and complacency?

Procedures can lull people into a passive mindset of just following the steps and not really thinking about what they are doing. When we become passive, we don’t try to improve our skills. The checklists and procedural guides can reduce our motivation to become highly skilled at a job.

Important example for forecasters. It might be cheaper to hire inexperienced forecasters and give them enough procedures to get the job done, but those who depend on accurate forecasts might pay the costs and hire experienced forecasters.

How can procedures erode expertise?

When we get comfortable with procedures, we may stop trying to develop more skills. Why bother, if procedures usually get the job done? The result may be an erosion of expertise in organizations that rely too heavily on procedures.

Research supports the idea of eroding expertise. A number of studies have shown that procedures help people handle typical tasks, but people do best in novel situations when they understand the system they need to control. People taught to understand the system develop richer mental models than people taught to follow procedures.

How do procedures mislead us?

The biggest worry is that following procedures can lead us in the wrong direction and that we won’t notice because the reliance on procedures has made us so complacent.

Example: Lia Dibello found the same thing in her observations of maintenance technicians. Experts may claim to follow procedures, but when DiBello watched them she saw them deviate from the procedures when they needed to. Typically, Experts don’t compulsively follow the procedures. If they did they wouldn’t be doing their job.

Summary: We can see that procedures are insufficient, can get in the way, can interfere with developing and applying expertise and can erode over time. Procedures work best in well-ordered situations in which we don’t have to worry about changing conditions and we don’t have to take context into account to figure out how to apply the procedures, or when to jettison them.

Unintended consequences

We would establish ‘adequate performance’ as the new ideal. It is too difficult and inefficient to continually fiddle with better procedures, which creates an endless cycle of updating. By signing on the procedural strategy we would live with ineffective designs. Then we:

• Would discourage people from using their judgment.

• Generate massive volumes of procedures, because it is too expensive to go back and cull procedures we don’t need anymore

• Would save money by retaining our current ways of doing the work. Anyone who proposed a different business or work strategy would have to ensure that it was consistent with all the existing procedures. Or else, we would just let the paperwork discourage anyone from suggesting improvements.

• We would issue procedures as a way to change behavior, even though there may be simpler and more effective strategies.

Replacement

By now it should be understandable why "Claim 1: Teaching procedures helps them perform tasks more skillfully" doesn’t always apply.

• Well ordered situations: Procedures are most useful, they can substitute for skill, not augment it.

• Complex situations: Procedures are less like to substitute for expertise and may even stifle its development.

Like all tools, procedures have strengths and weaknesses. Although I have been describing their limitations, we certainly shouldn’t discard them:

• They are training tools. They help novices get started in learning tasks.

• They are memory aids. In many jobs they help workers overcome memory slips.

• They can safeguard against interruptions. For example, pilots following a flight checklist often get interrupted, the checklist helps them carry out all the steps.

• They reduce workload and make it easier to attend to critical aspects of the task.

• They are a way to compile experience and historical information. Procedures are useful when there is a lot of turnover and few workers ever develop much skill. They help less-experienced workers do a reasonably acceptable job.

• They can help teams coordinate by imposing consistency. If the people on the team know the same procedures, they can predict one anothers next moves.

Getting procedures ‘right’ is not just a matter of getting them to be accurate or efficient or updated or covering all needed contexts, which may well be impossible and prohibitively expensive. It is also a matter of getting the organization to have the right attitude toward procedures. To put procedures into perspective, it is very important to consider the difference between directions and maps.

Teaching procedures

When we do want to teach some procedures, the typical way is to present the standard procedures and make everyone memorize them. Another way to teach procedures: set up scenarios for various kinds of challenges and let the new workers go through the scenarios. If the procedures makes sense, then workers should get to see what happens when they depart from the optimal procedures.

When procedures are taught in a scenario format, people can appreciate why the procedures were put into place and can also gain a sense of the limitations of the procedures. The scenario format seems to work better than having people memorize the details of each step. The scenarios provide a good counterpoint for learning the steps of complicated tasks.

Why does claim 1 matter?

Claim 1 matters because it creates a dilemma for workers. Too often supervisors insist that workers follow some shallow rules that never were validated. It matters because when we emphasize procedures over skills we set a standard of mediocre performance. The standard procedures become a basis for evaluating job performance, making people even less likely to adapt or improvise and more careful to comply with the rules.

It also matters because we too often issue procedures in order to change behaviour even though there may be simpler and more effective ways to do that.

What entails the study of heuristics?

For more than 35 years, decision researchers have been demonstrating the biases that corrupt our judgments and decisions. They have been showing that our thinking processes are flawed, inaccurate and unreliable. Heuristics are automatic decision rules, these are strategies we commonly use in thinking and that can bias us. There are different kind of heuristics.

The anchoring and adjustment heuristic

When we have to make an estimate and we don’t know the answer, one strategy is to find a plausible answer and then adjust it up or down. This mental strategy is called anchoring and adjustment. The initial estimate is the anchor. By suggesting a high or a low anchor we can influence people’s judgments.

Example:

Was Mahatma Gandhi greater or less than 9 years old when he was assassinated? Pick one:

• Greater than 9 years old

• Less than 9 years old

Second group:

Was Mahatma Gandhi greater or less than 140 years old when he was assassinated?

• Greater than 140 years old

• Less than 140 years old

It is known that the anchor dramatically altered the subject guesses. The anchoring-and-adjustment heuristic affects real world judgments. Real estate agents also use anchoring and adjustment to manipulate us into bidding more than we should.

Framing heuristic

The way a question is presented can determine the answer. Framing is a cognitive heuristic in which people tend to reach conclusions based on the 'framework' within which a situation was presented

The representativeness heuristic

Example:

Linda is 31 years old, single, outspoken and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in antinuclear demonstrations. Please check off the most likely alternative:

• Linda is a bank teller

• Linda is a bank teller and is active in the feminist movement

Most subjects pick the second alternative, because it seems like a fuller and more accurate description of Linda. It seems like a better representation of the kind of person Linda, while the first alternative is statistically more likely. The representativeness heuristic is one type of heuristic that we use when making judgments. In this particular example, we estimate the likelihood of an event by comparing it to an existing prototype that already exists in our minds.

What is the heuristics-and-biases research framework?

The decision bias research started in the early 1970s. Danny Kahneman and Amos Tversky ran experiments that showed systematic flaws in the reasoning processes of their subjects. They demonstrated that people use heuristics (mental shortcuts or rules of thumb) to make quick judgments, and that these heuristics lead to predictable errors. The heuristics- and biases paradigm that Kahneman and Tversky pioneered caught on quickly. Other researchers added more and more kinds of biases. As many of 60 biases have been identified in the search literature.

Many researchers have interpreted these heuristics-and-biases studies to mean that people are defective decision makers. Fortunately, the fears of decision biases are overblown. The research doesn’t really demonstrate that we are irrational. We use effective strategies in our thinking, but these have some limitations that researchers can exploit.

Kahneman and Tversky showed that, when making decisions, people use heuristics instead of formal statistics. They described heuristics as natural reasoning strategies that usually work but aren’t perfect.

People use heuristics when making judgments. The anchoring-and-adjustment heuristic and the representativeness heuristic are strategies that generally give correct answers when we have to make estimates and we don’t know the exact answers. These reasoning strategies aren’t perfect. When they break down, we see systematic error.

What about our memories?

We may be tempted to say that our memories are flawed. We all know the frustration of failing to remember someone’s name. How damaging are these memory limitations?

Daniel Schacter has argued that the limitations of memory are actually very useful. Here are a few of the seven ‘sins of memory’that Schacter has examined:

• Transcience: In transience, we lose a memory we once had, such as a name or a telephone number. Memories degrade over time. The less a person uses a memory, the more quickly it fades. But do we really want to remember everything ? Sometimes we lose a fact that we later need, but we would do much worse if we held onto everything.

• Absent-mindedness: Sometimes we forgot to do something or to record a fact (such as where we put our car keys), most likely because our attention is directed somewhere else. Absent mindedness may result from the way we focus our attention on some cues and not others; we best remember the cues to which we have attended, and we don’t effectively store in memory the cues we have ignored. If we didn’t have this mechanism for selectively recalling the cues to which we have attended, we would be inundated with irrelevant details.

• Blocking: Sometimes called the ‘tip-of-the-tongue phenomenon, we can’t recall information we want to remember. Blocking affects our ability to remember facts and incidents. Blocking seems to result from memory inhibitions that keep the irrelevant memories bottled up. When the inhibition misfires, we get blocked.

• Persistence: We may continue to remember distressing incidents that we would prefer to forget. Although these traumatic memories can cause us unwanted pain, they also make sure we stay alert to danger, and thus they help us to ensure our survival.

Schacter’s conclusion is clear: We would be foolish to rid ourselves of any of these memory features. They are frustrating when they don’t get us what we want, but each one reflects a valuable aspect of memory.

What is meant by the 'gaze heuristic'?

We learn to use something. The gaze heuristic will lead the player to the point where the ball is going to land. For example catching a ball, the gaze heuristic works very well. It takes into account wind, resistance and the spin of the ball.

Unintended consequence

Everybody thinks that biases are always negative. But no one has ever demonstrated that our judgments would be improved if we could be de-biased No one has ever compared what we would lose versus how much we would gain by giving up any of the reasoning strategies.

The heuristics-and-biases research simply shows the systematic errors that can arise from the limitations of our reasoning strategies. Some of the so-called biases have nothing to do with flawed reasoning.

Most important is that no matter how good the reasoning strategy, if it starts with poor data, it isn’t going to produce accurate results. The rule here is that if you put garbage in, garbage will come out. The original heuristics-and-biases research simply showed that people used heuristics even when the heuristics didn’t work well.

Replacement

Claim 2 states that decision biases distort our thinking which assumes there is a right way to think. It assumes that we are capable of thinking in the ‘right’ way except when we use mental strategies (heuristics) that are susceptible to errors.

What can we do to protect ourselves from the limitations without sacrificing the strengths of our mental strategies? A few ways to cope with cognitive limitations are to build expertise, to design more naturalistic judgment strategies that can mitigate the errors, to employ better data formats, and to use choice architects.

How is expertise build?

We can help people build expertise so they can use their reasoning strategies more effectively. If we want to improve decision making, and if we are vulnerable to suggestions about misleading anchors and frames, then we should help people build expertise.

Helping people learn more accurate anchors is more likely to succeed than trying to change the way they think in order to de-bias them. Greater expertise will help people identify good anchors, which will make them less vulnerable to manipulation by anchoring and adjustment.

Put judgments into perspective

We can design methods that will reduce problems that stem from the limitations from heuristics. Take overconfidence, we can benefit from safeguards, particularly if they prepare us to adapt once we encounter trouble. Two strategies:

• Outside view: using previous projects to suggest how long task will take and how many resources they will consume.

• Premortem technique. A postmortem is a procedure done in a hospital to find out why a patient died. The physician benefits, the medical community benefits, and even the friends and relatives benefit from what is learned. However, a postmortem doesn’t help the patient.

After a project has failed managers can conduct a postmortem session to figure out what went wrong. We can move the postmortem to the beginning to help us anticipate problems and to become more realistic about the challenges.

Represent data in ways that support intuition: Many people have difficulty thinking about probabilities. That doesn’t mean that their reasoning is faulty, just that they aren’t skilled at calculating and combining probabilities. Percentages are tough, because they involve ratios and are hard to visualize.

If people could see the same data in a different format, they might come up with more accurate judgments.

Example: When Tversky and Kahneman used frequency data instead of probability data, the conjunction fallacy dropped dramatically.

• Frequency data: How many of 100 participants?

• Probability data: What percentage of participants?

Gigerenzer applied the frequency format to help people make better use of base rates. Heuristics-and-bias researchers have shown that people usually ignore base rates when they make estimates.

Hoffrage, Lindsey Hertwig, and Gigerenzer showed that if base-rate data were provided as frequencies, the counselors would take the base rates into account and would make more accurate judgments.

Why the use of Choice architects?

Richard Thaler and Cass Sunstein are behavioral economists who have used the findings of the heuristics-and-biases research community to design programs that encourage people to make better decisions about things such as retirement savings. Thaler and Sunstein have designed programs to help people make good choices about their finances, their health and their children’s education. Thaler and Sunstein describe themselves as choice architects.

Why does claim 2 matter?

Claim 2 (that decision biases distort our thinking) matters because the concept of bias encourages organizations to overreact to failures by enacting excessive procedures and counter-productive restrictions. It also matters because it affects the way supervisors diagnose performance problems.

For more than 35 years, decision researchers have been demonstrating the biases that corrupt our judgments and decisions. They have been showing that our thinking processes are flawed, inaccurate and unreliable. Heuristics are automatic decision rules, these are strategies we commonly use in thinking and that can bias us. There are different kind of heuristics.

What is the essence of claim 3?

Claim 3: To make a decision, generate several options and compare them to pick the best one.

Claim 3 is taught in business schools. These leaders believe that all they need is more good options and stronger methods to select the best. Claim 3 makes a stronger statement; that when faced with a decision, we should go trough the drill of listing a set of alternatives and then contrasting them.

Five more important claims:

• When people get better at making decisions they consider more options and evaluate them more deliberately.

• Only novices rush ahead with the first option that pops into their minds.

• We make decisions by comparing options, using the same criteria to find the best one.

• Decision analysis will improve the quality of our choices.

• The closer the alternative options are matched, the harder we must work to pick the best one.

These beliefs encourage us to think carefully when making decisions. Thinking carefully means identifying a range of different options so that we don’t miss anything by being too impulsive, then setting up a common set of criteria so we evaluate each of the options using the same standards.

Sometimes we want to give more weight to the most important evaluation criteria rather than counting every dimension equally. This type of analysis can be referred to the Rational Choice method.

Another technique is to use a decision tree. Decision-tree diagrams represent different options, the probability of success with each, and the anticipated benefit if that option is successful. The result of each branch is an expected value for that option, letting the decision maker see which option to select.

The option with the highest expected value will have the greatest chance for success.

What about the disclaimer?

People and organizations rarely use formal methods of decision making, such as the Rational Choice Method. Many decision analysts have come to recognize that decision makers hardly ever set up formal evaluation methods or rate different options on the same dimensions. Look at your own life. Do you always set up matrices for choices / decisions?

Reasons why people rarely make these matrices:

1. They require too much time and effort. You can’t use them if you have less than half an hour to make the decision.

2. The formal methods require a lot of data to fill in all the cells of the decision matrix, or to estimate all the probabilities. If you want to use the methods, you have to make up ratings and probabilities that you don’t have any way to confirm.

3. It is much harder to estimate complex variables than well-ordered ones. In buying a car, we can pin down reliability records and fuel economy and crash protection ratings. On the other hands, we have trouble estimating how comfortable a car will be on long trips.

4. People are not convinced with any evidence that these analytical methods help people make better decisions, or that training in formal decision making results in better performance. Decision analysts now argue that their goal is to help people make more consistent decisions with clearer justifications, not necessarily decisions that are better.

As a result, we either ignore those features or we rate all alternatives as average, which washes out this variable from the decision.

Example of an Rational Choice Model

Evaluation dimensions:

• Price

• Reliability

• Safety

• Gas mileage

A final weakness in these analytical methods is the Zone of Indifference. When one option is clearly better than the others, we need not to do any analysis. We immediately know what to choose. The closer the options become, the more the strengths and weaknesses are balanced, the harder the choice. The hardest decisions are those that must be made when the options are just about perfectly balanced.

The analytical methods are designed to help us make the toughest choices, but once we realize we are inside the zone of Indifference we should stop right here, make an arbitrary choice and move on.

What if we don’t move on? Herbert Simon argued that any company that attempted to optimize its returns and make the best decisions would fall into a never ending quest to find the best decision. Simon coined the term satisficing to describe what we all do just about all the time.

The more choices that are given, the harder the decision. The field of decision making searches for ways to help people make better choices. Yet in study after study we find that people don’t compare any options, let alone systematically review those options’ strengths and weaknesses.

Replacement

How do we make decisions, especially under some time pressure and uncertainty?

Army Research Institute did a small experiment. After this there was evidence that with 23 years of experience you recognize situations as familiar, or typical. The patterns you recognize include typical ways to react.

Therefore, the first option the commanders thought of was usually going to get the job done. They were satisficing and their first impulses were very effective.

The answer to the second mystery where they were interested in, was how to evaluate an option if you don’t compare it to others, is that the commanders evaluated an option by imagining it. They played it through in their mind, conducting a mental simulation. If it worked, they carried it out. If it almost worked, they altered it to make it better. If they couldn’t make it work, they discarded it and looked at the next most typical action.

Klein, Calderwood and Clinton-Cirocco called this strategy a Recognition-Primed Decision model. The pattern recognition suggested an effective course of action and then the firefighters used a mental simulation to make sure it would work. This RPD strategy combines intuition with analysis. The pattern matching is the intuitive party, the mental simulation is the deliberate analysis.

Therefore, Claim 3 (to make a decision, generate several options and compare them to pick the best one) can be replaced with the RPD model: Good decision makers use their experience to recognize an effective option and evaluate it through mental simulation.

Our research shows that the more experience people have, the more likely they are to make decisions without deliberately contrasting options.

The RPD model shows how we can make good decisions –using our experience- without comparing options and without generating more than one option. The decision-analysis model, as stated in claim 3, shouldn’t be our gold standard for good decision making, because we rarely use this strategy and because it doesn’t seem to result in better decisions.

What are two systems of thinking?

Many researchers are now advocating for a dual system mode of thinking.

1. The automatic system is fast, automatic, effortless and emotional and uses tacit knowledge.

2. The reflective system is slower, consciously effortful, deliberate, logical and serial, and uses explicit knowledge.

These two systems which work in different ways, complement each other, but have different mental gears that we can use.

The connection between the two systems isn’t always perfect. Consider the following problem. A bat and a ball cost $ 1,10 in total. The ball costs a dollar more than the ball. How much does the ball cost? For most of us, the immediate reaction is 10 cents. Yet the correct response is 5 cent.

Bat 1,05 Ball 0,05. Why do we make this fault?

Shane Frederick (2005) has studied this and similar problems as examples of a failure of the reflective to do its job. The initial impulse seems so obvious that we don’t think twice. Therefore we don’t engage the gear for our reflective system. It also works in the other direction. We need the automatic system to check up on the reflective system.

This two system Automatic/Reflective framework fits the RPD model. The intuitive pattern-recognition part of decision making uses the automatic system to size up situations and form initial impulses about what to do. The deliberate mental simulation uses the reflective system to assess the options suggested by the patterns.

As we saw in Chapter 5, we need both intuition and analysis—both the automatic system and reflective system. If we were just using intuition, we would be making lots of blunders. And if we just tried to use analysis, we would never get very far.

Some additional claims about decision making:

• When people get better at making decisions they consider more options and evaluate them more deliberately. Not true, as people gain experience, they consider fewer options and evaluate them using mental review rather than by comparing them using the same evaluation criteria.

• We make decisions by comparing options using the same criteria to find the best one. Not true, when we try to use the same criteria for all the options, we wind up discarding some important considerations and distorting others.

• Only novices would rush ahead with the first option that popped into their minds. Not true, novices don’t feel comfortable going with the first option. They don’t have the experience to trust the first option, or to gauge its quality except by contrasting it to others.

• Decision analysis will improve the quality of our choices. Not true, we're not aware of any evidence that decision analysis improves the quality of judgments and decisions.

• The closer the alternative options are matched, the more we must work to pick the best one. Also not true. When the options match up too closely, we are in the zone of indifference. Sometimes we would be better off just flipping a coin instead of continuing with struggling. Regardless of the tactic, when we reach a point where further analysis is unlikely to break an impasse, we should recognize the futility of endless debate.

When to compare options?

Sometimes we do face situations in which we do have to contrast different courses of action. What do we do then?

We can learn from skilled chess players, they don’t settle for the first satisfactory option. They really do want to play the best move possible. Their strategy is to conduct mental simulations of each one of the promising moves, imagining how the option would play out deeper and deeper into the future.

Skilled chess players don’t compare options using the Rational Choice Method, because there is little benefit to rating options on the same evaluation dimensions, such as controlling the center of the board. These kinds of ratings don’t capture the dynamics of the game position.

The mental-simulation strategy lets chess players appreciate how a move works for that game. The mental simulations capture the context rather than filtering it out through the rational choice method.

The Rational Choice Method might be useful when we do have to make choices and don’t have very much experience. The process of setting up a matrix can help us clarify what matters to us.

How can decisions in organizations be made?

We make many decisions in teams and organizations, rather than alone. The Rational Choice approach lets us direct the behavior of team members. Building on the RPD research, John Schmitt developed a Recognitional Planning Model to expand on the Recognition-Primed Decision model. Schmitt’s model jettisons all the busy work of three courses of action and instead encourages the leader to use intuition to identify a favored course of action.

In this chapter two systems have been distinguished: Reflective and an automatic. Claim 3 puts too much emphasis on the reflective system. That’s the one we can control better. Because we rely on both systems, we have to resolve conflicts between them. We have to gauge how much we trust our intuitions and our analyses. And that depends on how skilled we are.

Claim 3 (to make a decision, generate several options and compare them) doesn’t help anyone making better decisions. It ignores the way experienced people actually make decisions. It just gets in the way of their effective strategy.

Claim 3: To make a decision, generate several options and compare them to pick the best one.

Claim 3 is taught in business schools. These leaders believe that all they need is more good options and stronger methods to select the best. Claim 3 makes a stronger statement; that when faced with a decision, we should go trough the drill of listing a set of alternatives and then contrasting them.

Why do decision-support systems fail?

Many decision-support systems are a fail. The developers think that people will welcome decision aids that make it easier to conduct decision analyses. The developers often can’t believe that decision makers are too stupid to appreciate all the benefits of such decision aids. But decision makers are not rejecting the aids. They are rejecting the mindset that keeps churning out these kinds of systems. Decision makers are rejecting the mentality that idealizes reflective, analytical thinking and marginalizes automatic, intuitive thinking, instead of blending the two kinds of thinking.

Those who advocate for a balance between the automatic (intuitive) system and the reflective system are continually challenged to ensure that their organizations value their employees’ tacit knowledge.

Once the balance shifts, expertise may begin to wither away. We identified four approaches to decision-support systems:

1. The decision is made by the decision maker alone.

2. The decision maker is helped by a support system or an algorithm.

3. The decision-support system has the final say, with inputs from the operators.

4. The decision-support system makes the entire decision on its own.

Some risks in using decision-support systems under complex conditions.

The last two of these four approaches cede decision control to the equations. They are best suited to well-ordered domains that have enough structure to get the systems organized, enough stability to make the programming economical, with not too much complexity or inter-dependencies, little need for tacit knowledge, and clear criteria for good decisions.

Algorithms aren’t used where flexibility is demanded, because they are carefully tailored to narrow conditions. In addition, algorithms can get in the way when we need people to develop and sustain skill and expertise.

Algorithms can get in the way when we need people to develop and sustain skill and expertise. When the algorithms are making the decisions, people often stop working to get better. The algorithms can make it harder to diagnose reasons for failures. As people become more dependent on algorithms, their judgment may erode, making them depend even more on the algorithms.

Another disadvantage of using decision-support systems is that we shouldn’t replace human judgments with algorithms if we want to make discoveries. Even in fairly chaotic environments, human intuition may provide clues about the way different factors relate to one another. Replacing human judgment with algorithms reduces opportunities for discovery.

When are decision-support systems rejected?

If decision biases are pervasive and crippling, then various decision-support tools should improve performance. Decision researchers often formulate various kinds of decision support systems and tools to help people make better judgments and avoid biases.

Decision Analysis uses decision trees or influence diagram to structure the process. Techniques such as Multi-Attribute Utility Analysis should help us combine information about our preferences.

Yet people don’t use these systems. Again and again, researchers sadly report that people in the field rejected their systems.

Researchers who design decision aids don’t worry about most of these issues. Their goal is usually to help people eliminate inconsistencies. Decision analysts worry about inconsistencies because the analysts have been able to demonstrate that people aren’t as consistent as statistical formulas.

Yates et al found a problem. They found that decision aids make the decision process unnatural and difficult. The decision aids are trying to change and improve on the way we think, not to support or amplify our thinking. We are responsible for the outcomes, not the analyst or the method.

People who develop decision support systems find it convenient to simplify the complexities that surround so many difficult decisions. The developers may get the details right, but may miss which way the wind is blowing.

What is the essence of project Phoenix?

Decision makers have been studied as carefully as weather forecasters. They make predictions every day. They make short-range predictions, long-range predictions. They are very accurate in making 1-2 days predictions. A problem is that good weather forecasters have one big drawback: They cost money.

By building in decision support systems, organizations can get away with hiring relatively inexperienced forecasters and still hope for the same level of quality. That’s why Canadian forecasters are ordered to use large-scale programs providing Numerical Weather Predictions. These are computer based systems to generate forecasts.

This system isn’t perfect. What was project Phoenix? Experiments to look if skilled forecasters did a better job on their own, so without a automated forecasting system, or that the unskilled people using an automated forecasting system scored better.

The result was that unskilled forecasters had to go through training to upgrade their skills, because that was better than relying only on the system.

Snellman (1977) used the term ‘metereological cancer’ to warn about the trend for forecasters to cease being professionals and instead rely on computer-generated information.

When decision support systems are designed by people who are insensitive to tacit knowledge, their devices and strategies are likely to interfere with expertise rather than support it.

If we try to balance the human-as-hazard model with the human-as-hero model, and to balance the automatic, intuitive system with the reflective, analytical system, we should have more of a chance to create decision support systems that will get used.

Many decision-support systems are a fail. The developers think that people will welcome decision aids that make it easier to conduct decision analyses. The developers often can’t believe that decision makers are too stupid to appreciate all the benefits of such decision aids. But decision makers are not rejecting the aids. They are rejecting the mindset that keeps churning out these kinds of systems. Decision makers are rejecting the mentality that idealizes reflective, analytical thinking and marginalizes automatic, intuitive thinking, instead of blending the two kinds of thinking.

Those who advocate for a balance between the automatic (intuitive) system and the reflective system are continually challenged to ensure that their organizations value their employees’ tacit knowledge.

What does claim 5 tell us?

Claim 5: It’s bad to jump to conclusions. It’s better to wait to see all the evidence.

Whenever you hear about a dramatic news event, track the initial account and watch as it changes over the next few days. Sometimes we don’t get the real story for months.

When a story is likely to be wrong, it clogs our minds. It gets us thinking in the wrong direction, and it makes shifting over to the truth more difficult. When we get contradictory evidence, we fixate on the first story and preserve our mistaken impressions.

What is our problem? Our fixation on our initial beliefs.

What about the disclaimer?

Claim 5 (that we shouldn’t jump to conclusions, but instead wait to see all the evidence) has two weaknesses: We can’t follow the advice to keep an open mind, and we shouldn’t follow that advice because it will make us passive.

• We can’t keep an open mind: Elstein et al (1978) studied the way physicians made diagnoses. The physicians knew they were supposed to gather all the data before starting to speculate about what was wrong with the patient, just as claim 5 states. But they couldn’t stop themselves. As they got more information, they naturally began to imagine what it meant.

• We shouldn’t keep an open mind. An open mind equates to an empty mind. How will we know when we have seen all the evidence? How long are we supposed to wait while we leave our minds on idle?

The advice is also misguided

Keeping an open mind can make it harder to figure out what is happening. To understand why, look at a study conducted by Jenny Rudolph. She studied 39 resident anesthesiologists to see how they interpreted evidence and diagnosed a problem.

Conclusion of this experiment: The one who kept an open mind, failed. The only one who succeeded had jumped to conclusions and tested them. They hadn’t fixated on their first explanation. Instead, they had used that explanation to guide the tests they had performed and the way they had searched for new information. They exemplified the strategy of ‘strong ideas, weakly held’.

What are unintended consequences?

Claim 5 (that it’s bad to jump to conclusions, and one should wait to see all the evidence) fits an orderly situation where we can use past experiences to estimate the flow of information and pick a point where it is safe to begin guessing.

The replacement for claim 5 is to speculate actively, but to test those speculations instead of getting committed to them. Rather than advising people to keep an open mind, we can encourage them to engage in a speculate and test strategy. Cohen, Freeman and Thompson have developed a training approach to support the speculate-and- test strategy, and have demonstrated its effectiveness. Claim 5 doesn’t work under conditions of complexity!

Story of Watson and Crick: The story of the triple helix shows two scientists who didn’t have an open mind. They started out with a set of beliefs, which guided their investigations. They showed active exploration and a speculate-and-test mentality.

Claim 5 warns us that we can get fixated. Its recommendation to keep an open mind usually isn’t realistic in most complex settings, but there are other ways to protect ourselves. They will be described in Chapter 17.

What does anticipatory thinking involve?

Advice: People should in complex and ambiguous settings actively speculate, instead of passively absorb data. This advice is just the opposite of claim 5. Experts distinguish themselves by their ability to anticipate what might happen next. By forming sharp expectations, experts can notice surprises more readily.

When we anticipate, we get ourselves ready for what may come next. We redirect our attention. We also can prepare for a few different possibilities. Anticipatory thinking lets us manage our attention so that we are looking in the right places to spot anomalies in case they appear. For example, driving a car depends on being able to anticipate dangers, and is a metaphor for many other activities. Good drivers aren’t waiting for data. They are using their experience to actively scan for possible threats and hazards.

Anticipatory thinking describes how we actively speculate about what might happen next. The next chapter (11) will describe how we have to actively speculate about what has just happened in order to make sense of feedback.

Why does claim 5 matter?

It matters because claim 5 can disengage people and create a passive, lazy stance of waiting to receive all the data before doing anything. It also matters because claim 5 ( it’s bad to jump to conclusion, wait to see the evidence ) encourages organizations to impose bureaucratic procedures for collecting and sorting data instead of helping people engage in anticipatory thinking.

Overall conclusion: The claim matters because the advice embodied in claim 5 make people slaves to the flow of information. We always have new information just around the corner that is about to surface. If we follow claim 5, we may never try to synthesize the data.

Claim 5: It’s bad to jump to conclusions. It’s better to wait to see all the evidence.

Whenever you hear about a dramatic news event, track the initial account and watch as it changes over the next few days. Sometimes we don’t get the real story for months.

When a story is likely to be wrong, it clogs our minds. It gets us thinking in the wrong direction, and it makes shifting over to the truth more difficult. When we get contradictory evidence, we fixate on the first story and preserve our mistaken impressions.

What is our problem? Our fixation on our initial beliefs.

Sensemaking is the attempt to understand events that have occurred and to anticipate what might happen next.

What is the essence of claim 7?

Claim 7: To make sense of a situation, we draw inferences from the data

Some people picture the mind as an assembly line. Data arrive from the senses, then get refined and processed and combined and polished and assembled until we have insights and conclusions. Claim 7 stems from the assembly line metaphor. It portrays our sensemaking as making more and more inferences from the data we receive. The data get converted into information, then get further enriched to become knowledge, and finally transformed into understanding.

This process illustrates how people and organizations cope with all the data and cues that come at them.

Sensemaking is sometimes explained using the metaphor of a picture puzzle; we treat each message and data point as a piece of a puzzle and find a way to put these pieces together.

Claim 7 runs into several serious difficulties. The assembly-line metaphor fits orderly situations in which you only have to handle explicit knowledge. Once you have defined all the data and described the causes, you can set up filtering rules and do the sorting.

Our expertise is as much about recognizing legitimate dots as about connecting them.

When we want to put data into perspective we need stories. Without stories, the data won’t mean very much.

In judging what counts as a dot we also have to figure out how much detail the dot has to have. It depends on what we are looking for. If forecasters work at too fine a level they’ll miss the large patterns.

There is no ‘right’ level and no standard operating definitions of a “dot”. If there is no right level, the job of designing displays for forecasters becomes tougher. Sometimes the data elements we think are dots may turn out to be irrelevant.

Where do the dots come from?

The notion of connecting dots, deriving inferences from data, conjures up an image of a children’s puzzle that displays all the dots; our task is simply to find the right connections in order to see the picture. These puzzles don’t have any shadows to complicate the task. But when we make sense of events we usually are seeing only some of the dots that complete the picture, trying to remember others, discarding ‘wannabe’ dots, and adjusting dots that have gotten distorted.

Sensemaking seems to depend heavily on seeing connections. We can’t find the connections by exhaustively linking all the elements, because we don’t know in advance which are the relevant elements, the relevant dots. And we would run into an exponential explosion if we connected every combination of dots. That’s where the intuition comes in --- To help us recognize which dots are worth connecting and to remind us which dots we need to complete the picture.

When are we finished making sense of a situation?

Claim 7: To make sense of a situation, we draw inferences of the data. This claim describes a mindless process of churning out more and more inferences. The numbers of data combinations grows exponentially as we add more data elements. We may be able to generate valuable ideas, when we randomly connect data elements and then connect the connections, expecting to emerge with some insights.

But these ideas will be buried in useless inferences. Then we will have the enormous task of sorting through all the inferences to try to find ones that matter. No, this bottom-up account of random inferencing doesn’t make sense.

As Bill Clinton was fond of saying: “That dog won’t hunt” Instead, we are directing the way we form connections, just as we’re actively searching for data. We are trying to explain what is going on around us. We are looking for stories. When we arrive at a plausible story, we’re done.

What if the stories are wrong?

Claim 7 shows people endlessly generating inferences from data. It doesn’t reflect our worries about that the data and the stories might be wrong. If you’re just looking for connections, then any connections will do. Yet an important aspect of sensemaking is sorting out which dots to believe. A lot of sensemaking is double-checking whether the data are accurate. It’s also about judging whether the story rings true.

The assembly-line model of generating inferences doesn’t capture the desperation we feel when we get lost and have to sort out where we are, where we want to be, and what data we can still trust.

If we enforce claim 7: to make sense of a situation, we draw inferences of data, illustrated by the assembly line-model, we’ll filter the data to make sure that decision makers only get pre-digested materials and highly abstracted analyses of the data. However, we use our expertise to see patterns in the data. If we disconnect people from the data, we disconnect them from their expertise.

The assembly-line model takes information search out of the hands of the decision makers. Yet that’s another aspect of their skill. Experienced decision makers know how to search for data. The assembly-line model makes it hard for decision makers to build their own mental models, because they are forced to depend on the dots and the analyses that people at lower levels, with less expertise, are using.

Experts have to build their own mental models. The assembly-line model builds from data elements dots, that have been collected. But what about events that didn’t happen?.

Experts appreciate the significance of these negative cues, but when they are given a set of recommendations and summaries, skilled decision makers don’t have any way to notice what didn’t happen.

Therefore, the assembly model, and claim 7 will disconnect skilled decision makers from their expertise. It will force them to depend on subordinates who have less experience than they do

Replacement

We aren’t passively waiting for data. We are also defining what counts as data. We are questioning stories that don’t seem plausible. We are having insights about ways to connect different types of dots. That’s different from drawing more and more inferences in the hope that eventually it will all fall into place. It depends on taking an active rather than a passive stance.

We make sense of cues and data by fitting them into frames such as stories. The frames we have learned guide our attention and shape what we notice. For example, if we go for a walk in a park, a landscape architect is aware of the way the path provides interesting views. A tree specialist is distinguishing the different species and how well each is growing, and a maintenance worker is getting a feel for whether heavy lawnmowing equipment might get stuck in the muddy ground

Same park three different perspectives, three different experiences. We make sense of cues and data by organizing them into frames, such as stories, scripts, maps and strategies. But the reverse also happens. Our frames determine what we use as data. The data/frame model of sense making shows this is the process of fitting data into a frame and fitting a frame around the data.

Conclusion

We make sense of data elements by fitting them into frames, such as stories, but the reverse also happens. Our frames determine what we use as data.

The claim matters because the claim says that all we do is draw inferences from the data or connect the dots trivializes what we do. It ignores how in complex situations we have to decide what is a dot. It ignores the tacit knowledge that it takes to figure out which dots to connect.

Also claim 7 can disconnect experts from the data patterns they have learned to detect.

Claim 7: To make sense of a situation, we draw inferences from the data

Some people picture the mind as an assembly line. Data arrive from the senses, then get refined and processed and combined and polished and assembled until we have insights and conclusions. Claim 7 stems from the assembly line metaphor. It portrays our sensemaking as making more and more inferences from the data we receive. The data get converted into information, then get further enriched to become knowledge, and finally transformed into understanding.

What is claim 8 about?

In complex situations, our attempts to make adaptations may fail if we persist in pursuing the goals we started with, if we rely too heavily on identifying and minimizing risks, and if we maintain the ground rules we set at the beginning of an activity. Adapting means revising our goals, becoming resilient to threats we can’t predict, and changing the way we work together.

To adapt, we have to learn, but we also have to unlearn.

Claim 8: The starting point for any project is to get a clear description of the goal.

People definitely want to define clear goals in advance. That way they can set up a schedule and measure if they are making good progress. The more specific the goal, the better. Man management and planning tools are built around the notion that we first define the goals and then proceed to achieve them.

Peter Drucker developed Management by Objectives for managers to help subordinates specify the objectives they want to accomplish in the coming year; then the managers can measure their subordinates’performance using yardsticks they both had accepted. Project managers focus on objectives to set up plans for how they are going to achieve the project goals.

The entire system, from personnel management to strategic vision, depends on describing explicit goals at each level.

Problem: Claim 8 runs into difficulties when we put it into practice. Claim 8 discourages us from getting started until we clearly define the goal, and it leads us to expect to accomplish that initial goal. The notion of pursuing clearly defined goals can work against us in complex situations. It doesn’t work as well in stable settings, because it runs up against goal trade-offs

What are goal trade-offs?

Claim 8 creates the illusion that we have a single goal, but most of the time we’re juggling several goals at the same time. The problem here is that we needed to work both fronts at the same time. Any time we have more than one goal, we have the potential for goal conflicts. That doesn’t mean we just pick the goal we want Instead, we have to figure out how to trade off the goals so that we can achieve enough of each of them. That’s the hard part. No matter how clearly we define our goals, we still have to wrestle with goal trade-offs.

Even when we have clear goals, we can still get befuddled when the goals come into conflict with each other. Goal trade-offs are very difficult for us. Physicians for example have to wrestle with conflicting goals. They want to restore a patient to health and also reduce the patient’s discomfort. Soldiers also struggle with goal trade-offs.

The dilemma of goal trade-offs shows that even clear goals can be ambiguous. And this is the easy case that involves clear goals in a stable situation. In complex situations our goals are likely to be vague and ambiguous.

What are emergent goals?

What happens when we don’t have clear goals? This is the most serious limitation of claim 8. Here we are facing what Rittel and Webber (1973,1984) described as ‘wicked problems’ in which the goals are incomplete and keep changing, as well as occasionally conflicting. Solutions to wicked problems aren’t true or false. Instead, they are judged as good or bad, because there is no way to test a solution to a wicked problem. When we are faced with a wicked problem, when the goals just aren’t clear, there aren’t any objective ways to gauge success.

Most of the problems are wicked. Taking an essay is a wicked problem, midway through the exam. The student may wonder what the teacher wants to know. Different teachers might give varying grades to the same answer, because essay exams don’t have clear criteria and goals.

Claim 8 runs into trouble in complex situations in which the wicked problems prevent us from clarifying the goals at the start. Many of the problems we face are wicked problems in which we may have to re-assess our original understanding of the goals.

The goals will become clearer as we learn more. That’s why I am calling them emergent goals

Bill Duggan, in Strategic Intuition, describes a number of other instances in which leaders changed their goals on the basis of what they learned. During the process of learning, we have to change our goals, that’s very important!

Emergent goals at the project level

Perhaps goals are murky when businesses are trying to establish new industries. If we move down to the operational level, where teams are carrying out projects, we sometimes find clear goals. But often we don’t. When we are working on a project, we want to work on well-defined assignments with clear goals. That way, we know what we must deliver, and our bosses can judge our progress

However, we don’t always have that luxury, but is this always better?

Alexander Laufer (2009) studied organizations and observed that the managers of successful projects repeatedly started engineering designs before they had finished defining their project objectives.

A study of 211 RD projects found that projects that started with well-defined business and technical goals had no greater chance of succeeding than those whose goals were still fluctuating at the start.

However, the business and technical goals for the successful projects did become better defined over the life of the project than the goals for the unsuccessful projects.

Emergent goals at the individual level

Peter Drucker initially developed the concept of Management by Objectives to help supervisors manage their subordinates. After they agreed on the subordinate’s objectives, the supervisor could assess progress toward the objectives in periodic evaluations.

What are the tools for pursuing goals?

Another drawback of management by objectives is that it calls for objectives that can be measured objectively. It is easier to measure progress if you can count accomplishments, so employees discard any important objectives that they can’t quantify, further distorting the evaluation process.

Under well-ordered conditions, we can expect that the goals be clarified at the start. But complex conditions make for wicked problems that often prevent goal clarification at the outset. Program managers suppress these complications in order to create systematic management strategies.

Common management tool: The Gantt Chart. A Gantt chart illustrates the type of tools that reinforce the managers’ rigidity in holding to the initial goals. Gantt charts define tasks, show when each task should be started and when it should be completed, and list milestones.

The limitations of Gantt Charts illustrate a basic difficulty of applying analytical methods. Even though more sophisticated scheduling tools are now available, these methods work well as long as the tasks don’t change, the schedule doesn’t change, the mission doesn’t change and personnel don’t rotate in and out. When changes in tasks, goals, personnel erode that common ground, the shorthand notations on the Gantt Chart and the layers of notes and modifications all become barriers to communication and coordinated planning. Furthermore, Gantt charts depend on accurate estimates of the duration of each task. Another limitation of objective-based tools such as Gantt charts: They contribute to goal fixation.

What is meant by goal fixation?

Sengupta, Abdel-Hamid and Van Wassenhove identified an “initial goal bias”. We can think of it as goal fixation. That is, getting stuck on the original goal. They found that managers failed to revise their targets even when those targets became obsolete.

Why do managers fail to revise their targets?

Very early in their careers, people incorporate into their mental models the notion that it’s important to meet externally set targets. This bias is often reinforced in managerial life. Revising targets is seen as an admission of failure in many companies, and managers quickly realize that their careers will fare better if they stick to and achieve initial goals. Even if that leads to a worse overall outcome.

Revising goals and objectives is difficult enough. Goal fixation makes it almost impossible for us to adapt to changing circumstances or even to use what we learn as we make sense of situations.

What is an unintended consequence?

Claim 8 states that we should define goals and objectives at the start and then simply proceed towards them. Claim 8 implies that we should not start work unless we can get a clear description of the goal. But if we insisted on clear goals, we simply couldn’t get started on most of the important projects in our lives.

Replacement

The replacement for claim 8 is that when facing wicked problems we have to re-define the goals as we try to reach them. Most of the problems people encounter have poorly defined goals. No amount of thinking and analysis will make these goals well defined. In such cases, we are going to have to figure out the goals as we go along.

Goal fixation is a big problem for most people; they won’t accept the unpredictability of events and getting ready to adapt.

Alternative approach: Management by Discovery in contrast to management by objectives. Management by Discovery (MBD) says when we face complex conditions we should expect to revise and replace goals on the basis of what we learn. MBD doesn't mean revising the plan to reach the goal. MBD means revising the goal itself.

The management by objectives approach sets the expectation that we shouldn’t begin executing any plans until we have specified the goals. Management by Discovery encourages leaders to take action rather than sit around waiting for perfect goal clarity. MBD also starts with trying to clearly define goals. The difference is that MBD expects to face wicked problems and complex conditions. Instead of reacting by tighting controls, MBD takes the opposite approach of loosening controls in order to expand the discovery process.

Five important differences between MBD versus Objective-based approaches:

1. Define the objectives at the start: With MBD we still spend time trying to clarify goals at the start, we expect to change those goals on the basis of what we learn.

2. Determine a path to the end state. There are many paths to an end state; if the goals change, so will the path. With MBD we still try to envision a way to reach the end state. That’s what the plan does. However, we are also looking for alternative paths we can use when we revise the end state.

3. Identify the tasks on that path: But if the path changes, so will the tasks. We have to be ready to add, combine, delete and reshuffle the tasks. Rather than use a Gantt Chart, we should keep track of the options so we can activate one option when another seems to get stuck.

4. Prepare a schedule showing when each task begins and ends: But if the goals change, the schedule will change. On the basis of what is learned about the connections between tasks, each sub-team can understand how to provide what they need when they need it.

5. Measure progress in moving toward the goal. But if the goals change, the progress markers will change, so we should prepare ourselves and our clients and supervisors from the start to re-negotiate the goals as well as the schedule and resources.

You don't have to see Management by Discovery and Management by Objectives as conflicting. They are two ways to achieve successful outcomes. Each begins with an attempt to define the goals. The primary difference is that in complex situations with wicked problems, MBD recognizes that the original goals may become obsolete. Therefore, the MBD mindset is to move forward even though the goals haven’t been specified in detail, and to expect to revise the goals. The MBD mindset is to look for opportunities to figure out better goals than the ones specified at the beginning.

When it comes to wicked problems, the standard objectives-based tools, such as Gantt Charts, may not be useful.

One strategy is to compile a “goal field” and use it to track what you have learned. You can share this goal field with your subordinates to help calibrate your discoveries. You can ask them to keep their own goal fields so that you can review their progress.

A goal field represents the primary goals, and how they relate to one another. The goals usually won’t form a neat and simple hierarchy, so it is important to show the interconnections and to highlight areas in which the goals might come into conflict.

Why does claim 8 matter?

It matters because claim 8 prevents people from adapting. MBD isn’t a choice. It is what we have to do in unpredictable circumstances when we are faced with ambiguous goals.

It matters because too many leaders and managers have one reaction to adversity: Increasing the controls.

Many people try to tame turbulence, complexity by redoubling their efforts to specify the goals and details. This works well for managers early in their careers, but when they are assigned to programs with wicked problems, this can be the wrong approach.

Problem: Most management tools and training programs thinks that goals are stable, but also people are highly vulnerable to getting fixated on goals.

In complex situations, our attempts to make adaptations may fail if we persist in pursuing the goals we started with, if we rely too heavily on identifying and minimizing risks, and if we maintain the ground rules we set at the beginning of an activity. Adapting means revising our goals, becoming resilient to threats we can’t predict, and changing the way we work together.

To adapt, we have to learn, but we also have to unlearn.

Claim 8: The starting point for any project is to get a clear description of the goal.

What is claim 10 about?

Common ground is the knowledge, beliefs and history we share that let us coordinate smoothly with one another. In a relay race, the runners know what to expect from one another. They are on the same wavelength. Shared beliefs let us make accurate assumptions about other people. More important, they let us make accurate predictions about what others are going to do. Teams do better when the members can predict one another’s actions more accurately.

We need to predict in order to coordinate with others, even if the others are complete strangers.

Claim 10: Leaders can create common ground by assigning roles and setting ground rules in advance.

Claim 10 asserts that we can create common ground. Good leaders prepare the members of their team to build and use predictions about one another. They carefully and clearly explain what each person should do. They explain how the team is going to work. They try to eliminate any ambiguities that may create confusion. This process will strengthen team coordination, because coordinations depends on interpredictability.

What does it actually mean? “create common ground?” Before companies try to create common ground, they should see how it works and why it fails.

What about the disclaimer?

Common ground is never as good as we think it is, and it is always eroding. We can’t prevent its erosion, especially under complex and changing conditions. Runners on a relay team find it easy to maintain common ground because there is very little variation in their actions.

When goals, roles, and abilities change, we can’t count on sustaining common ground.

Claim 10 asserts that giving clear instructions can provide common ground. And it’s true that clear instructions at the start are one basis for common ground. Certainly, clear instructions are better than vague ones, clear procedures are better than confusing ones. But giving clear instructions and having standardized procedures at the start can’t protect us.

The problem is that there is no magic formula to guarantee common ground. Some leaders are better than others at strengthening common ground at the outset, but that can’t prevent its erosion during dynamic and unexpected events.

There are also examples that are involved in breakdowns of common ground. Then, people didn’t get the information that they needed from others in their organization who had that information.

More worrisome are the common ground breakdowns in which team members hold mistaken beliefs about what others know. Common ground is always imperfect and continually eroding. Some of the reasons for this erosion are that we aren’t very good at taking the perspective of others, we get tripped up by ambiguous words, and we have different life expectancies that color our understanding.

When is attention self-centered?

We aren’t always able to take someone else’s perspective. We may assume that if we hear or see something, then others will see and hear it and understand it the way we do. We don’t appreciate that they may be focusing on something else because their different goals and priorities affect how they look and listen.

What is the ambiguity of words?

Even if we try to take another person’s perspective, we can’t be sure we mean the same thing by the same words. Risk is a word that can lead to confusion.

Teamwork doesn’t mean the same thing all over the world. Western managers can’t simply announce a set of best practices for the multinational team to follow and assume that they have created common ground. Another word that creates confusion is ‘leadership’.

We all have different experiences with leadership problems. Even when people within the same organization agree that they have a leadership problem, they are likely to mean different things, some remembering incidents, other recalling unfair acts.

One antidote to the slipperiness of words is to make better use of examples. This method, defining by example, can reduce the ambiguity of words.

Different Life experiences: Common ground also erodes because we have different life experiences that affect how we see things. Teams that have worked together for a while have shared working experiences and precedents about how to handle situation. That’s why bringing in new team members creates risks. They don’t have the shared experiences.

Claim 10 (that leaders can create common ground by assigning roles and setting ground rules in advance) is necessary but not sufficient. It suggests that we can improve common ground by preparing a team more carefully. It certainly makes sense to prepare a team by explaining everyone’s roles and functions, going over the ground rules, and describing the goals. But claim 10 goes too far in suggesting that we can rest easy once we have completed the preparation. But common ground is never perfect and is always eroding.

Common ground is much easier to establish in well-ordered stable situations. When a situation changes rapidly, we can’t count on common ground.

The replacement for claim 10 is that all team members are responsible for continually monitoring common ground for breakdowns and repairing common ground when necessary.

Once we have finished preparing the team we should still be on the lookout for breakdowns. We should continually monitor common ground, staying alert for early signs of confusion. That requirement holds for the team’s leader and also for everyone on the team.

In complex situations, the original plans, goals, roles are likely to change, degrading the team’s common ground. Instead of trying to increase control, degrading the team’s common ground. Instead of trying to increase control, we should expect common ground to erode and then repair it on the fly.

When people agree to be members of a team, they take on a responsibility to fix any important common ground breakdowns they spot, in order to increase predictability and team coordination.

If team members shirk that responsibility, the results can be tragic. Good teams are proactive, alerting everyone to unexpected events that might catch people by surprise. Common ground typically suffers during transitions.

Why does claim 10 matter?

It matters because we take common ground so much for granted that we don’t appreciate how fragile it is and how easily disrupted.

It can give teams a false sense of security. It matters also because many leaders like to issue directives and then consider their work done. They blame subordinates for anything that might go wrong instead of accepting their responsibility for monitoring and repairing the common ground in their organizations.

It matters because performance depends heavily on common ground. Most of the team failures involve breakdowns in common ground. Team coordination depends on predictability, without adequate common ground, we can’t reliably predict one another’s actions.

Common ground is the knowledge, beliefs and history we share that let us coordinate smoothly with one another. In a relay race, the runners know what to expect from one another. They are on the same wavelength. Shared beliefs let us make accurate assumptions about other people. More important, they let us make accurate predictions about what others are going to do. Teams do better when the members can predict one another’s actions more accurately.

We need to predict in order to coordinate with others, even if the others are complete strangers.

Claim 10: Leaders can create common ground by assigning roles and setting ground rules in advance.