five

Shape

Success is rarely a straight line.

—Tom Perkins¹

I'd like to begin the discussion of shape by inviting you to take a short walk on the beach with the sculptor Henry Moore. I've timed it: it will take less than a minute.

In the passage here, Moore was concerned with physical shapes. In this chapter, I will be concerned with temporal shapes—the various curves that describe how an event or process changes over time. Most of us spend our time thinking about substance: the needs of customers, the impact of new technology, how our product or service stacks up against a competitors, and so on. In this chapter, I would like you to acquire an artist's eye. I want you to learn to notice the temporal shape of an activity or process apart from its content, in the same way that Moore studied the shape of an egg apart from its use as a food or symbol.

Recognizing the temporal shape of a process is important in making decisions about timing, for a number of reasons. First, many timing decisions reference shape directly. Without thinking about it, we include shape in our rules of thumb: “buy low, sell high” is a common example of a heuristic that incorporates shape. Likewise, when to spend your advertising dollars depends on whether sales peak during certain times of the year or are relatively constant month to month. In both cases, a shape triggers a decision or action. Second, shape offers important clues about timing. Take the idea of a cycle. Suppose you were asked in 2010 to predict when U.S. combat operations would cease in Afghanistan. Your prediction would take into account the date of congressional and presidential elections, the time in the election cycle when politicians must defend their decisions. It would also be based on another cycle: the length of a tour of duty and the number of tours that the armed forces could ask of its soldiers without degrading its global effectiveness. Third, shapes warn us that timing matters. For example, when something rises too quickly, we know it will fall. When a bubble emerges, we expect it to burst. We just don't know when. The bubble shape reminds us to start paying attention and to consider timing.

The temporal world, like the spatial world, is filled with shapes. Some of these shapes warn us of risks; others point to opportunities. Because the best way to see temporal shapes is to plot them, I will use the words shape and curve interchangeably.

AN ABRIDGED CATALOGUE OF SHAPES

As with all of the timing elements, using shape to make decisions about timing begins with knowing what to look for. The best way to train our eyes to see shapes is by examining a few that we encounter frequently. We will focus on six main shapes.

1. Point—the absence of shape (a single point in time)
2. Line—the span that directly connects two or more points in time

   Straight. If you plot how a process changes over time and find that it doesn't change at all, the result will be straight line.

   Divided. Sometimes we divide what would otherwise be a continuous (time) line into intervals—for example, weeks into days.

3. Curve—a line that changes its slope or direction over time

   Accelerating and decelerating curves—curves that change their slope in a consistent direction, like those that describe how fast your car is going after you push the pedal to the floor or slam on the breaks

   Arc—a shape like a rainbow or the top part of a gently rolling hill

   S-curve—a curve that begins slowly, reaches a peak, and then levels off

4. Cycle—an interval of time during which events periodically repeat, visually characterized by curves that rise and fall, rise and fall
5. Helix—a corkscrew shape, like a coiled spring or the handrail of a spiral staircase
6. Tandemizing shape—a shape in which two curves move in opposite directions. Then one reverses so that the two are now moving in the same direction, that is, in tandem or in parallel with each other.

The Point

Although a point is the absence of shape, it can also be the starting point for various shapes. Figure 5.1 illustrates a shape that I call the point and fan.

At time 1, to the very left of the diagram, imagine that you are faced with several alternatives. You choose one. That choice leads to others, so at time 2, you choose again from the alternatives that are present, and so on with each successive time period. Over time, the diagram assumes the shape of a fan. The equation that describes this process is an exponential function that accelerates quickly. Because choice matters only if the alternatives you are choosing among are different, when you look three or four time periods downstream, it is hard to imagine what the world and the alternatives that you are choosing among will look like. This is especially true because others are doing the same, which can change the choices that you have available That is one of the reasons we don't think about the distant future. It is uncertain. That reality, however, can cause us to miss predictable risks.

Figure 5.1 The Point and Fan

For example, let's suppose that in fifteen years, baby boomers are fully invested in the stock market, in part to protect themselves from growing inflation. Then suppose the market plunges. Boomers, along with others, will exit in droves. They have to: they will be afraid they won't live long enough to see the market recover. Their exodus will cause the market to fall even further, thus creating a vicious cycle. I suspect that one reason this risk has not been priced into the market is that no time seems right to do so; that is, the problem lacks a timing solution.

This example has an important practical implication: do not assume that the distant future is necessarily less knowable than the more immediate future, as the point-and-fan diagram would suggest. Our deep belief in the power of choice to shape our future, which creates the widening fan, leaves the impression that the distant future is always more uncertain than the more immediate future. It may not be. We simply need to take a look and, while doing so, appreciate the irony that choice, the tool we use to control our future, creates the feeling that it is futile to know where we will end up.

Points come together to form an infinite number of shapes. Let's look at a few others.

Lines

I want to consider two kinds of lines: straight, uninterrupted lines and those that are divided.

Straight Lines

We like straight lines because they seem, well, straightforward, and hence predictable. What you see is what you get: no surprises, no deviations, no sharp turns. But is a perfectly straight line always a good thing?

We sometimes observe attempts to eliminate variation. For example, consider quality-control programs. Reducing variation can cut costs and improve quality, but it also has dangers. In the 1980s, for example, technology made it possible to track a person's attention and level of interest by measuring the size of his or her pupils. When a pupil dilates, it indicates interest; when it contracts, it signals disinterest. Market researchers measured the size of people's pupils as they watched television pilots, and advised network executives to eliminate the moments when interest flagged. But Jagdish Sheth, a professor of marketing at Emory University, pointed out that when they kept the emotional level high all the time, the pilot “failed miserably.”⁴

We should not be surprised. When a heart monitor displays a flat line, life is over. Most human activities demand some degree of variability. When the musical performances of fine musicians are examined quantitatively, they are often found to be out of sync with the metronome beat. The musicians usually catch up in a measure or two, but if they are right on the beat all the time, their performance seems dull and uninteresting.

If variation is needed to give a performance vitality, its absence suggests something artificial or fake. In 2005, MIT professor Andrew Lo was looking for warning signs that a hedge fund was vulnerable to a crisis. What he discovered was that many “hedge funds were posting returns that were too smooth to be realistic.” Digging deeper, he found that funds with hard-to-appraise, illiquid investments—like real estate or esoteric interest-rate swaps—showed returns that were particularly even. In these cases, he concluded, managers had no way to measure fluctuations and simply assumed that value was going up steadily. The problem, unfortunately, is that those are exactly the kinds of investments that can be subject to big losses in a crisis. Lo concluded that “measuring the smoothness of returns gives economists a good way to estimate the level of relatively illiquid investments in the hedge fund world … without knowing the details of the investments.”⁵

Lo discovered that straight lines can be as informative as their curvy counterparts if you take time to discover what they really mean. This principle applies to other shapes as well.

Divided Lines, Dynamic Sections

We commonly divide time into past, present, and future. When we place these divisions on a time line, the line looks like this:

By convention, at least in most English-speaking cultures, we place the events of the past on the left and those in the future on the right. This tripartite division of time is important because our most powerful timing rules, the ones we rely on to decide when to act, are triples. I illustrate these triples in Table 5.1. I've placed a YES in those cells when the window of opportunity is open and an action can be successful, a NO when it is closed, and acting is impossible or inadvisable.

As the table illustrates, decisions about timing depend on what is or is thought to be possible in the past, present, and future. If you want to understand the emotions associated with a timing decision, you need to consider all three time periods. For example, compare row 5 (NO–YES–NO) with row 7 (YES–YES–NO). The rows are highlighted. In both, the decision is the same: act now. But there is a difference. In row 5, this is your first and only chance. Doing what you want to do was impossible in the past, and it won't be possible in the future. Act now. In row 7, this is your last chance to take advantage of an opportunity that was always present. The decision is the same—seize the moment—but we feel differently about it depending on the timing triple that determines the decision.

Table 5.1 Timing Triples

Curves

I want to consider three kinds of curves: those that accelerate or decelerate, a special shape I call the dramatic arc, and the familiar S-curve.

Accelerating and Decelerating Curves

If you plot the shape of all the events and processes that matter to your organization, you will find very few straight lines. Most will be curves. One of the most important curves to watch for is one that starts slowly, but then rapidly accelerates or decelerates. An “exponential function” is a typical example. As inventor and futurist Ray Kurzweil says, such a function “starts out almost imperceptibly and then explodes with unexpected fury.”⁶

There are many examples of exponential functions. Moore's law is perhaps the most famous: the number of transistors that can be placed on an integrative circuit doubles every two years, a trend that has continued since Moore first proposed it in 1965. But there are many others. In the years leading up to 2000, the increase of venture funding for Internet start-ups had an exponential shape: a slow start, followed by a dizzying climb. Another dramatic example is George Soros's explanation of the mortgage and credit crisis of 2008 in terms of the exponential growth of the CDS (credit default swaps) market. Here's what Soros said happened. I've put the last part of the quotation in italics because it is so extraordinary.

All of this raises the following question: Is anyone in your business watching for exponential functions and the mechanisms that produce them, such as accelerating curves and vicious cycles? If not, these shapes are likely to fall through the cracks, leaving you surprised by “sudden” risks and unprepared for fast-moving opportunities.

The Dramatic Arc

In most cases, a climax occurs not at the halfway point of an event but someplace toward the end.⁸ No one reading a novel or watching a film wants the climax in the middle, leaving the rest of the story to drag on and on. Nor do we want the climax to happen at the very last moment. Bang. It's over. Instead, we want the climax to occur toward the end, but not at the very end. It's all in the timing. I call the shape that describes the asymmetrical placement of the climax the dramatic arc, illustrated in Figure 5.2. You will often find this shape whenever there is a goal-directed process. Tension rises to a peak as you near the goal and then falls as the goal is achieved.

The dramatic arc may seem far removed from a business or practical application, but that's not the case. It comes into play anytime we manage a process that involves the build-up and release of tension, such as announcing an acquisition, or timing an exit. The surge strategy adopted by the Bush administration during the war in Iraq is an example of the dramatic arc. The timing of a surge places it after the midpoint of the war. The surge, similar to a sprint at the end of a race, not only signals an ending but also makes it inevitable by expending resources at a rate that cannot be sustained. It allows those who must defend an ending to say, “We have done everything we can.”

Figure 5.2 The Dramatic Arc

If you need to terminate a project, an investment, or any other activity and you want the process to be well received, use the dramatic arc. (Unless, of course, the situation is dire. Then simply pull the plug.) The dramatic arc satisfies our esthetic sense: it feels right. And if you see one forming, know that an ending—or at least an attempt at an ending—is in progress.

S-Curves

Another very common shape is an S-curve. When you scan your environment and look into the future, you will likely see them.

Technologies, for example, often experience an early period of rapid growth, a leveling off, and then a steady climb to saturation. Reaching saturation, I should add, can take a long time. In the case of television, which had “the shortest span between early adopters and universal penetration of any medium,”⁹ it took nearly fifty years for almost every household to have one. You can anticipate an S-curve by identifying the mechanisms that create it. Consider, for example, the rate at which fax machines were adopted. When fax machines were new, a small number of people bought an early version. Most people waited. There was no point in having one unless others had one as well. Then, at a certain point, it made sense to have one because each machine now had companions with whom it could communicate.

S-curves have a number of characteristics. In the beginning, nothing seems to be happening. Once the upturn begins, however, the question becomes How quickly will it accelerate? Researchers, such as the late sociologist Everett Rogers, who introduced the term “early adopter,” have identified a number of factors to explain the acceleration of the curve. These include the comparative advantage of the product relative to others, its compatibility with existing products, whether customers can try it out before purchasing it, and whether its benefits are immediately clear. Although such a list is useful, it is only a starting point. Each factor has its own time course. For example, if there is a benefit to a product, when will customers notice it—immediately, or only after a period of time? When will the product be judged compatible or not compatible with other products—right away, after a trial period, or some other time? That is where the timing advantage is to be found—in a more detailed consideration of the factors that give the S-curve its shape. So the next time anyone includes an S-curve in her presentation, ask her to sketch the shape of the underlying curves that produce it.

Although the S-curve is useful in tracing the evolution of innovations that have run their course, it is equally useful for anticipating the life cycle and longevity of newer products.

Cycles

Every business environment is filled with cycles. These include financial (the quarterly reporting cycle), political (election cycles), technological (adoption cycles), and economic (the business cycle). Every cycle has a number of characteristics, any one of which can be important for deciding questions of timing.

Period

A period characterizes the length of a complete cycle. For example, how long will it be from one recession to the next?

Some cycles are long, like certain weather patterns—the rise and fall of ocean temperature over decades, for example. Others are measured in hours and minutes, the twenty-four-hour news cycle being one example.

It's a useful exercise to list the cycles on which your business depends. Note their current length, and ask yourself whether they might become longer or shorter in the future. How quickly could that expansion or contraction take place, and how much advance warning would you have?

Regarding the length of the human life cycle, the poet Mary Oliver wrote, “A lifetime isn't long enough for the beauty of this world and the responsibilities of your life.”¹⁰

Parts and Phases

Some parts of a cycle are of particular interest, as the Christmas shopping season is to retailers. In addition, some parts of a cycle may become longer or shorter than others over time. I've noticed that Christmas decorations seem to go up earlier every year, particularly if there is a downturn in the business cycle.

Think about which parts of a cycle are most important. For example, social scientists conducted a study in which they asked patients how much pain they experienced during a colonoscopy.¹¹ They found that the total amount of pain patients experienced was less important than the highest degree of pain they suffered at any point during the procedure, as well as the amount they experienced during the last three minutes. The researchers called this a “peak and end” pattern.

Sometimes a cyclical pattern will have two peaks. Researchers who study auto accidents find a two-peak pattern over a twenty-four-hour period. They see “an extremely sharp error peak at about 3 AM and an error peak about one-quarter that size at about 3 PM.”¹² This is an example of a pattern that policymakers and legislators can take into account as they are considering traffic safety laws and guidelines.

It is important to consider all the phases of a cycle. For example, if you had to guess which high school sport had the greatest number of injuries, most of us would probably guess football or hockey. But according to a thirteen-year study in 1993, girls who ran cross-country in autumn suffered more injuries on a percentage basis than any other sport at the time. The reason is that they take the summer off. “In contrast, girls' track and field, a winter and spring sport that allows ample semester time for runners to get in shape, was ranked only ninth in [the] study, with less than half the injury rate of girls' cross-country.”¹³

It is easy to miss a phase or overlook the need for an extra cycle. For example, some airports have perpendicular runways. Ordinarily, that configuration is fine, but when an incoming flight has to abort its landing, turn left or right, and circle around for another try, that flight pattern can put it dangerously close to an aircraft taking off from the runway perpendicular to its approach.

Once you have identified all the phases of a cycle, ask yourself what issues might emerge if a phase has to be omitted, postponed, or repeated.

Amplitude

Amplitude refers to the magnitude or height of a cycle, peak to trough. Is a cycle composed of a series of gentle hills and valleys, or is it made up of steep climbs followed by death-defying descents? Amplitude can be important. When the market for a product or commodity rises or falls precipitously, planning goes out the window. For instance, the headline in the New York Times in July 2009 read, “Volatile Swings in the Price of Oil Hobble Forecasting.” In 2008, Southwest Airlines reported two consecutive quarters of losses—as oil prices spiked and collapsed all within a few months. “Prices were falling faster than we could de-hedge,” said Laura Wright, the chief financial officer at Southwest Airlines, a company that bought long-term oil contracts in an attempt to insure itself against volatile prices.¹⁴

Symmetry

Is the curve of the cycle symmetrical? In other words, when a sharp downturn occurs, should you expect a sharp upturn?

According to Bill Dudley, head of U.S. economic research at Goldman Sachs, stock market cycles are asymmetrical. “The idea in the marketplace that you had to have a sharp recovery because you had a sharp decline is not borne out by the empirical evidence,” he said.¹⁵ Financial bubbles are also asymmetrical. The typical bubble has a “slow start, gradual acceleration in the boom phase, a moment of truth followed by a twilight period, and [a] catastrophic collapse.”¹⁶ If it would deflate gradually, the bubble would be less threatening.

As you scan your own environment for the shapes that matter, assume asymmetry. Ask yourself which is more likely, a curve that rises quickly and then trails off slowly, or a curve that builds slowly and then quickly declines? In probability theory and statistics, the measure of the asymmetry is called “skewness.” If you haven't found a number of positively and negatively skewed curves in your environment, you are probably missing them. Assuming symmetry is a common bias of the conventional mind, like the desire to live in a state of perpetual equilibrium; it is simply unrealistic. When things don't work, we say that someone has thrown us a curve, and most curves, I would suggest, are asymmetrical. In baseball, a pitcher's slider breaks over home plate, not halfway to it.

Smoothness

Will a cycle be smooth and rounded, like an amusement park roller coaster, or full of sharp edges, like a serrated knife? The latter is a particularly important shape.

Figure 5.3 The Sawtooth Curve

When HMOs were first introduced, for example, it was assumed that their large buying power would keep costs down. That didn't happen—in part because of what I call a sawtooth curve. Here is what I mean. When drugs, tests, or treatments are new, they are expensive. As more people use them, costs decline. But as soon as a better drug or treatment is invented, everyone wants it. The result is a series of spikes, illustrated in Figure 5.3. There is not enough time between the spikes for costs to decline significantly and stay there before the next spike begins.

In the case of the HMO, the experts failed to consider the likelihood of the sawtooth curve when they were analyzing pricing trends. Why? One reason is magnitude thinking. Price, we are told, is determined by the magnitude of supply and demand. Increase supply, and prices will decline; increase demand, and prices will rise. But, as we have seen, it is not that simple. In the case of the sawtooth curve, we don't need to know precisely how high the peaks are, or exactly how much time will elapse between them, to see the implications of this shape. We simply need to recognize that the peaks will be high and that the interval between them will be short. That is sufficient to temper the view that buying power alone will keep costs in line.

Number

How many cycles will there be in a given time period? Will one cycle of boom and bust follow another almost immediately? How many cycles is your business subject to, and what could happen if the number were increased? Multiple tours of duty for U.S. soldiers in Iraq, for example, caused fatigue and health issues.

Predictability

Some of the aforementioned characteristics are predictable, such as the length of a tour of duty in Iraq or the term of office for a state senator. Others, such as the length of the business cycle, are less predictable.

If the beginning or ending of a recurrent or cyclical phenomenon is predictable, expect company. In 2008, both Florida and Michigan jumped the gun and moved up the date of their primaries. If there is an advantage to being first, expect that people will respond by jumping in earlier than you anticipated.

Linkage Between Cycles

How does one cycle prepare us for the next? For example, how does a period of high inflation prepare us for the next time prices skyrocket?

The collapse of the World Trade Organization's “Doha Development Round” of trade negotiations in the summer of 2008 illustrates what happens when the link between cycles is not considered. The negotiations failed, in large part due to the success of the previous round of talks. The 1994 Uruguay round had stipulated that countries could convert farm quotas into normal tariffs. Concerned about a flood of imports, countries were permitted to impose short-term safeguard duties to guard against a surge. But what was created as a temporary solution became a crutch. In the Doha round, negotiators' inability to manage the issue and agree on a way to reform the safeguards was part of what sunk the talks.¹⁷

Another reason the talks collapsed was that countries from the developing world felt that the prior round of talks was biased toward the rich, so in the Doha round they wanted to redress the balance. Negotiators needed to consider how the end of one cycle (or round of talks) would influence the next. The same is true in business. Stop to consider how cycles are linked. One annual cycle of strong sales, for example, may set up an unrealistic expectation for success in the next cycle. Or, conversely, downsizing jobs in one cycle in response to a shrinking market may leave business units unprepared for the coming rebound.

Influence

Cycles also influence the shape of actions that are associated or coordinated with them. For example, what activities are sequenced or timed with the beginning or ending of the fiscal year?

A passage from the 1891 novel Main-Traveled Roads illustrates the power of a cycle to influence behavior. It describes a girl who waits for her boyfriend to visit her every Sunday, as seen through the eyes of the woman who employs her.

When we think about cycles, we think about something that reoccurs, a process that goes up and down repeatedly, like the business cycle. But as the description in the passage suggests, cycles are more complex than simply periodic reoccurrence. Therefore, when you find a cycle, ask two questions: First, what external process or schedule governs, controls, or influences the cycle? Second, how would the cycle be influenced if that process or schedule changed? For example, what would happen if the fiscal year for your organization were shortened by six months or lengthened to eighteen months? What costs or benefits would result, and how could you capture them without actually modifying the fiscal year (which I assume is impossible)?

Vicious Cycles

Vicious cycles are situations in which one problem, or attempted solution, leads to another problem, which makes the first problem worse and starts the process all over again. For example, when a company needs to exit a very large position it holds in a market, that sale may cause prices to decline, thereby forcing the company to sell even more, which causes the market to drop further, and so on and on in a vicious cycle. The skyrocketing prices of homes during the U.S. housing bubble is another example of a vicious cycle. High prices and fast turnover fueled overbuilding, which created an oversupply of homes. That oversupply, among other factors, caused prices to plummet. Many owners suddenly found themselves owing more than their homes were worth. But as foreclosures rose, banks were reluctant to lend to buyers, thereby worsening the oversupply issue, leading to additional price declines and more foreclosures.¹⁹

Most serious crises have multiple vicious cycles at play. That's one reason they are crises. Organizations should keep track of the vicious cycles they may encounter—noting their cause, expected duration, and the strategies needed to terminate them. No strategic planning exercise should be considered complete without consulting this catalogue.

The Helix

A helix is a three-dimensional curve shaped like a wire that is wound uniformly around a cylinder or cone. That's the dictionary definition, but ask anyone who knows what a helix is, and he will take his index finger and swirl it through the air in the shape of a corkscrew (see Figure 5.4).

Figure 5.4 The Helix

The helix is an important shape when it comes to timing, for two reasons.

First, it helps us avoid a common error associated with feedback: the failure to take account of the passage of time. The image of a corkscrew reminds us that we receive feedback, not when we first ask for it, but at some time in the future. Managers, therefore, need to be conscious of what may have changed in the interim. Will those who initially requested the feedback still be with the company? When the feedback arrives, will those who receive it be in a position to understand and use it appropriately? Finally, will the information still be relevant? Surveys, in-depth interviews, and other types of feedback can be expensive. Before seeking feedback, be sure you can answer the questions here.

The second reason the helix shape is important has to do with planning. As I mentioned in the chapter on temporal punctuation, we like closure. In a busy world, managers want issues to be resolved “once and for all.” Certain kinds of issues, however, never get resolved. They keep coming back, returning again and again at successive points in the future, thus tracing out the shape of a helix.

The key is to identify issues of this kind so that you are prepared when they return. A prime example is a dilemma. A dilemma is a choice that has three defining characteristics. First, the alternatives that define the choice are incompatible. They can't coexist. You must choose one. Second, each is indispensable. You can't summarily dismiss it because it represents something essential to you or your organization. Finally, each alternative is inviolate, meaning that it cannot be compromised or watered down. When all three characteristics are present, you face a dilemma. Whatever is not chosen is not eliminated, only repressed or sidestepped. It will therefore always resurface. So after you have chosen one alternative or course of action, ask yourself two time-related questions: (1) When will we see this issue again, and (2) What might be going on at that time that would make dealing with it particularly difficult?

Tandemizing Shapes

Thus far we have focused on single curves and shapes. To lead into the next chapter on polyphony, I will mention a pair of curves that I call tandemizing shapes. Tandemizing shapes are curves that have two phases. In the beginning, they move in opposite directions. Then one curve changes direction so that the two curves move in parallel, or in tandem. I illustrate two tandemzing curves in Figure 5.5, one positive and one negative, depending on which direction they turn.

Many risk management strategies assume the first phase: something (such as a stock) moves in one direction, while something else (a bond) moves in the opposite direction. In times of crisis, however, both can turn down at the same time—that is, move in tandem. Shapes that are negatively correlated can become positively correlated. And that shift may occur at precisely the wrong moment—that is, when their negative correlation is most needed to manage risk.

Figure 5.5 Tandemizing Shapes

We like countercyclical strategies, offsetting rates, and shapes that cancel each other out because they smooth out or eliminate variability. But no one can eliminate change. Therefore, when a risk management strategy relies on components that are assumed to move in opposite directions, look for tandemizing shapes. If you cannot envision conditions when “tandemization” will occur, then you have less protection than you need.

SHAPE-RELATED RISKS AND OPPORTUNITIES

Most shape-related risks stem from a failure to consider shape and its impact. For example, an article in Florida Today noted that the Brazilian airliner Embraer was successful because it produced just the right size plane at just the right time.²⁰ Yet what the article failed to include in its description of good timing was any mention of shape. How steep was the demand curve for that aircraft, and when was the shape or slope of that curve obvious? When in the future will that aircraft no longer be the right size, and how rapidly will demand for it decline? In the Introduction, I said that our descriptions of the world are often time impoverished: they leave out many temporal characteristics. Our explanations, after the fact, also tend to miss elements such as shape.

Of course, the sculptor Henry Moore, mentioned earlier, wouldn't be surprised. He would appreciate the difficulty of the task. The shapes that he needed were all around him. He could find and examine a shape in its entirety at his leisure. That is not the case for the shapes that we need in our work. Those shapes unfold over time and require both memory and imagination to see them, neither of which we can always count on. The best protection against missing a shape is to have encountered many examples of it in the past. To that end, I suggest that you add to and update the catalogue of shapes that I have started here. Doing so will prepare you to look for those that are critical to your work.

When you know that a particular shape exists, you can use that knowledge to improve decisions, processes, and predictions. Consider a circadian rhythm, which refers to the body's rhythm of waking and sleeping in sync with the sun's twenty-four-hour cycle of rising and setting. As Abigail Zuger wrote in the New York Times, it is possible to picture a day when health care takes the circadian rhythms of disease into account. “Asthma testing would take place in the middle of the night for the most telling results… Cardiac patients would have life-threatening clots in their coronary arteries dissolved in the late afternoon, when the body's clotting system is at its weakest.”²¹

Another way to use shape to manage risk and spot opportunities is to ask more precise questions. Consider the bubble. The questions people tend to ask are the same: Is there or is there not a bubble forming, and if there is, when will it burst? That's where the discussion usually ends. Instead, ask about shape (as well as rate, interval, and sequence). If the bubble does exist, will it pop, or will it deflate slowly, and if the latter, over what period of time? Also ask about mechanism. Is a vicious cycle involved? For example, if you are surrounded by houses that have been foreclosed, your own home will lose value, which will cause your neighbor's to lose value, which will cause your home to lose even more value, and so on. We say that we value transparency and perfect information, but if everyone sees the same shape at the same time and interprets it in the same way, expect herd behavior. In the case of a bubble, that can mean a panicked race for an exit before the bubble bursts. So remember to ask about the shape of the information curve: When will increasing numbers of individuals know certain facts and then act on them? Finally, ask yourself how much advance warning you are likely to have before a known bubble busts.

Seeking answers to these questions—even just raising them—will better prepare you for what might occur, particularly if you have researched past bubbles in your industry and looked for the answers to these questions in the historical record. Don't let fragments of information go to waste. We know, for example, that when an economic recovery begins and inflation returns, the Federal Reserve will be to slow to raise interest rates for fear of stopping the recovery in its tracks. Unless you plot two curves, one that shows the rising rate of inflation and the second that shows the predictable lagged response of the Fed, information about the lag will get lost.

Remember the mantra: before precision, pattern. If you find yourself sketching a curve that has an odd but important shape, you may want to name it, as I did with the tandemizing shape or sawtooth curve. This will help you remember it.

Finally, be aware of your own biases. We all seek the power and adrenaline rush of the decisive moment. When unfavorable conditions are present, we want them to end, period. Over time, we come to realize that we need to consider more time-extended shapes. For example, “more doctors and patients are starting to view cancer as a chronic illness—something to be treated, not cured.”²² This is an important historical shift with implications for how patients are treated. We want our lines short and, of course, straight. When we find a curve, we want it to be symmetrical and smooth with no gaps, discontinuities, or sharp edges. As a result, we are likely to be late in discovering shapes like the sawtooth curve and in understanding its implications. When something changes in a positive direction, we want it to continue, as was the case with housing prices: always moving up. When we can't answer a question—for example, when will a bubble burst—with precision, we move on and forget to ask other questions, such as those I mentioned earlier. The best protection in any business venture is to list the shapes that you assume will be present and then challenge them, keeping in mind your biases. For example, if you expect sales to ramp up smoothly, look for times when sales will jump and then fall back before advancing again. If you expect a project to end, expect some parts to continue on, even unofficially, like embers in a fire that takes a long time to die out.

There are many shapes that were not included in this chapter and others that were mentioned only in passing. Space was limited, and some shapes, like bubbles and cliffs (a precipitous drop or change in a prior condition) tend to dominate public discussion because of the fear and anxiety associated with them. I counted seventeen shapes that Henry Moore was interested in as he walked on the beach—eggs, shells, nuts, plums, and so on. I suspect that temporal shapes are just as numerous and diverse as spatial ones. We just need to look for them.

THE TEMPORAL IMAGINATION

I have focused on shapes that are easy to visualize, such as points, lines, and cycles. As I have indicated, there are many others. I use a shape from Laurence Sterne's novel Tristram Shandy, which he created to visually represent the narrative arc of his book, to remind myself to look for a variety of shapes, many of which do not have names (see Figure 5.6). One purpose of Sterne's novel, according to Howard Anderson, the editor of the Norton Critical Edition, was to reveal “the infinite ways in which conventional ideas of all sorts have handcuffed our minds and imaginations.”²³ That is also an aim of this book.

Figure 5.6 Narrative Arc of the Novel Tristram Shandy