three

Interval and Duration

Among Chuang-tzu's many skills, he was an expert draftsman. The king asked him to draw a crab. Chuang-tzu replied that he needed five years, a country house, and twelve servants. Five years later the drawing was still not begun. “I need another five years,” said Chuang-tzu. The king granted them. At the end of these ten years, Chuang-tzu took up his brush and, in an instant, with a single stroke, he drew a crab, the most perfect crab ever seen.

—Italo Calvino¹

Despite the show of force, military officers conceded that they did not have full control of Dili (a city in East Timor) and that it would take far longer to enforce security in the rest of the territory.

“How long is a piece of string?” one colonel said when asked for a time estimate.

—Seth Mydans²

The king and the colonel, in the quotations above, face a common managerial problem: How long will it take to achieve a desired outcome? We see the same question arise is business. How long will it take for a new service to gain market share, a revised policy to be successfully implemented, or, for that matter, the housing market to turn around? The usual answer is that it is hard to say. As the colonel put it, how long is a piece of sting? If we had asked Instagram founders Kevin Systrom and Mike Krieger in 2011 how long it would take for someone to offer them a billion dollars for their photo-sharing app, I'm sure they would have replied that they had no idea. (It took a year and a half.)

So, when should the king fire Chuang-tzu and hire someone else who could get the job done faster? That depends on the king's estimate of how long the job should take. And what about the colonel? Is he incompetent? That depends on how long it should take to enforce security. In another sense, deciding when to act depends on the length of an interval. If time is limited and a lengthy task can't be shortened (or, once begun, the remaining parts put off for another time), then it's foolish to begin. It would be like trying to force an SUV into a parking space barely big enough for a bicycle. In order to plan effectively, we need to be able to estimate know how long something will take and how long events or conditions in the environment will last—what will persist for a long time and what will be over quickly.

The ability to estimate the length of intervals is at the heart of the question of timing. Yet before we can estimate an interval of time, we need to recognize that an interval is present. We can't make a prediction about something we don't know exists. This becomes even more complex because the way we describe the world often omits intervals that matter. An article published in the Wall Street Journal about how and when to end an alliance exemplifies this point:

Although the article focuses on when to end an alliance (punctuation), it nonetheless neglects to factor in other relevant intervals of time. As a result, the actual triggers for ending the alliance are left ambiguous. Here are the intervals that the description omits and that are frequently overlooked in discussions of milestones, performance metrics, or service-level agreements. Each can be critical.

• How should key milestones be defined: by the time until a specific event or condition or by the time until a specific date (five thousand miles or three months, whichever comes first)?
• How long until a milestone should be reconsidered or revised?
• Does it matter if one party to an agreement misses a milestone by a small amount of time? And what is a small amount? Two weeks may be critical for one company, but may not matter for another.
• How long before a deadline arrives does a company realize that it cannot meet it? When does a company discover that it cannot meet a deadline or that it will miss a milestone?
• How long will it take to catch up or repair the damage caused by being late? When can this interval be estimated with any degree of certainty?
• Once the answers to these questions are known, how long will it be before this information is communicated to all relevant parties? Will everyone be told at the same time? If not, how will the gap between the first and last to be notified be managed?
• Do the answers to the previous six questions depend on how long the parties have known each other?

These intervals clearly matter. I don't think they were omitted because they weren't important. I think they were omitted because it's easy to leave them out. Recall the “key in the door” experiment I described in the Introduction. Our mind can jump gaps in time without noticing that it has done so. Another reason we tend to omit critical intervals from our thinking is our assumption that high speed is an advantage.

For example, researchers believed until recently that using CT scans sooner and more often to detect early-stage lung cancer would save many more lives. But that is not what studies showed. To understand why CT scans didn't improve survival rates, we have to look at the problem more closely. According to the New York Times, more cancers were found and treated as a result of CT scans, “but the death rate was the same … [because] screening led to detection and treatment of cancers that did not need to be treated—they would not have grown enough in the person's lifetime to cause any harm. And many of the deadly cancers that were treated still ended up killing patients.”⁴ In this case, research found that the extra surgeries, prompted by additional screening, sometimes caused complications, such as blood clots or pneumonia, that were life threatening or fatal.

These results contradicted prior advice that suggested that “more than 80 percent of lung cancer deaths could be prevented with CT scans.”⁵ That analysis—in assuming that anyone with lung cancer would die of it without treatment—failed to take into account three intervals:

Interval 1: How long a patient would live despite the fact that he or she had cancer.

Interval 2: How long it would take the cancer to kill a patient if it were left untreated. Maybe the person would die of other causes before that happened (per Interval 1).

Interval 3: How long a patient with cancer would live if the cancer were treated. Some treatments are lethal.

The initial research, which justified additional CT scans, uses the traditional perspective on timing—earlier is better. This is a variation on the “first-mover advantage” theme discussed in Chapter One. As we've seen, that predisposition has significant limitations. If the benefits of a technology have to do with timing—in this case, the assumed benefit is early detection—then it is essential to know how timing enters into the problem your technology is trying to solve. In this example, that means taking into account the three intervals I just described.

THE CHARACTERISTICS OF INTERVALS

I have presented the six timing lenses in a particular order because the use of one lens facilitates the use of the next. The sequence lens directs your attention to steps and stages; the punctuation lens to beginnings, endings, pauses, specific dates on the calendar, and so on. All of these marks in time can help you find intervals that you might otherwise overlook. There are, of course, an infinite number of intervals in any given situation—and only you can decide which ones matter. However, let me offer some suggestions about what to look for, beginning with the following list of interval characteristics:

1. Type. What kind of interval is it? There are four kinds. Make sure you look for and find all four.
2. Size—the Interval Envelope. Look for very long and very short intervals. Understand the minimum and maximum length of each.
3. Objective or subjective. There may be a difference between the objective length of an interval and how it is experienced: a second can be an eternity.
4. Contents. Look inside. Is an interval transparent? Do you know what is going on during that time period? If not, what should you assume?
5. Meaning. An interval can mean different things to different people. What kinds of interpretations are possible?
6. Number. Look for multiple intervals. As the aforementioned illustrations (triggers for disengagement and CT scans) suggest, there are more than you think.

These six characteristics are organized in a particular order. First, we need to know what we are looking for, hence the discussion of type. Then we consider the size, how big or small. Is the size fixed, and if not, how big or small could it become? When you consider size, ask yourself what matters: how long an interval is or how long it seems—that is, consider the objective and subjective size of the interval. Next, look inside: What is going on during that time period (contents), keeping in mind that different stakeholders might interpret the contents of a interval differently (meaning)? Finally, always look for more than one interval (number).

Type

There are four types of intervals. Before you decide to act, make sure you have found at least one example of each. If you haven't, you are probably missing something important that will affect the timing of your actions.

1. Time in between. How much time will elapse between event A and event B? When the twin towers of the World Trade Center were hit on 9/11, for example, one of the questions the insurance company faced was whether to consider the attack one occurrence or two in terms of coverage. For the purposes of designing an insurance policy, this highlights the importance of defining the parameters of coverage keeping “time in between” intervals in mind.
2. Time since. How much time has elapsed since event A has occurred? A common question of international diplomacy, as well as modern warfare, is how long to wait before responding to an attack or provocation. If a village is shelled, leaving hundreds dead, can you wait a year to respond? The same issue comes up in business all the time. A competitor has introduced a new product. How much time can you let go by before you counter with a new product of your own?
3. Time until. A deadline is coming up. Is there enough time to complete the work? How much time before a new rule or government regulation goes into effect? What do you need to do before that happens?
4. Duration measures how long something continues. When will the recession be over? Should you sign an office lease that will commit you for three years or five years? How long will a bottle of milk last before it goes bad?

Different parties may view the same interval differently, which is why it is important to be aware of these four types. For example, are we out of danger because a number of years have passed since 9/11 (time since)? If we think that another attack is inevitable, then the “time until” the next attack must be shrinking, as the bombing at the Boston Marathon demonstrated, so our preparedness must increase proportionately.

Size—the Interval Envelope

The Interval Envelope refers to the minimum and maximum length of the longest and shortest intervals that are important for your business. That sentence is a bit of a mouthful, so let's proceed step-by-step. First, look for extremes: intervals so small or so large that you might miss them. Then, think about how small or large these intervals could become under different circumstances, and then whether that expansion or contraction would cause problems.

Look for Extremes

Always look for the longest and shortest intervals that are present in the situation you need to manage. Short intervals are easy to miss. According to the Wall Street Journal, nine months before the massive March 2011 earthquake and tsunami cut off power to the Fukushima Daiichi nuclear plant, one of its reactors lost power after a subcontractor “inadvertently bumped into an auxiliary relay with his elbow. That led to a ‘momentary’ flutter that was long enough to trip a circuit breaker, shutting off the reactor's main power supply, but was too short to trigger the normal backup power supply”⁶ (italics mine). The problem might have been serious except for quick-thinking control room personnel who activated a backup generator. There was no contingency in the system that accounted for this brief interval in time.

Some intervals must remain short; therefore, we need to ask about the maximum length of the shortest interval. The implications will depend on the situation. For example, Western Staff Services, a temporary help firm, says that holiday Santas need to work short shifts, four to five hours, because “it's impossible to stay joyful” for longer.⁷ That obviously has implications for seasonal staffing in retail and entertainment venues.

In some situations, a very short time period can make an occasion more memorable. In his address to a graduating class at Penn State, the biologist Edward O. Wilson cited Salvador Dali as giving the shortest speech on record: “He said, ‘I will be so brief I have already finished,’ and he sat down.”⁸ A close second is Don DeLillo's speech upon winning the National Book Award in 1985. At the ceremony, DeLillo stood up and said, “‘I'm sorry I couldn't be here tonight, but I thank you all for coming,’ and abruptly sat down.”⁹ Of course, short can be too short. (What is the minimum length of the shortest interval?) As Leonardo da Vinci cautioned, “He who wishes to be rich in a day, will be hanged in a year.”¹⁰

In some cases, the shortest interval can be extremely short. During the summer of 2011, for example, Congress was never in recess. In fact, one day, the Senate met for a total of fifty-nine seconds. The reason was that if Congress adjourned, which it can't do for more than three days without permission from the other chamber, President Obama could make recess appointments of nominees blocked by the Senate Republicans.¹¹ With Congress always in pro forma session, Obama was unable to act.

The companion question is, what is the maximum length of the longest interval? In a situation I read about in the New York Times, a gentleman believed he was making a wise business decision, but he was surprised when a relevant interval went on far longer than he had anticipated.

When I think about the minimum length of the longest interval, I think about the interrelated economic crises that have erupted across the globe in the last decade. Some crises come and go relatively quickly, but is that really a good thing? In some cases, I find myself hoping that a crisis will last long enough for a solution to be devised. In 1999, Jeffrey Garten, then dean of Yale School of Management, published an op-ed in the New York Times that made this point. He said that the global economic crisis of 1998, marked by the bailout of Long-Term Capital Management and the Russian credit default, caused millions of people in emerging markets to suffer terribly. Yet “the crisis wasn't long enough or deep enough to result in the kinds of corrective measures that would result in a less risky global economy. Indeed, [one year later] little of a fundamental nature has changed.”¹³

What was true in 1999 is true today. If a long-term solution requires considerable time to be devised, negotiated, and then implemented, then we can only hope that the crisis lasts long enough for that to happen. The skiing and snowboarding industries rely on snowy winters. When will they be motivated to lobby for environmental reforms related to global warming? Only, I suspect, after three or four or more consecutive years of warm winters. An issue that comes and goes quickly can be ignored, particularly if it is costly to address. Hence, beware of magnitude-driven thinking: just because a problem is serious does not mean that it will be addressed. It won't be if the duration of the problem is shorter than the time needed to solve it.

As these examples suggest, look for the maximum and minimum size of the longest and shortest intervals that matter. I've arranged these combinations in a two-by-two table (Table 3.1). Think about each of the four cells (1–4). For example, look back at your day and consider something that seemed to occur at lightning speed. Then ask yourself how short or long that process could become under different conditions, and if it hit those extremes, would it pose a risk or an opportunity? Would it throw off your timing, or would it make no difference at all?

Table 3.1 The Interval Envelope

Size: Is It Fixed, or Can It Vary?

Some intervals come in fixed sizes—a year is 365 days. Others can be lengthened or shortened, expanded or compressed, depending on a host of factors.

Consider airline safety. It takes only two or three major crashes within the span of a year—the current evaluation interval—to provoke an outcry about airline safety. One could reduce the sense of crisis if one used a two-year evaluation interval rather than a one-year evaluation interval. But that is unlikely to happen, because we measure so many things on an annual basis.

Fixed intervals have known risk characteristics. A fund manager whose investments underperformed in the first half of the year, for instance, may buy riskier stocks during the second half because he must, on average, outperform the market annually. Birds face a fixed interval problem every night: how much energy to use in search of food. If they haven't gotten enough food by the end of the day, they will not survive the night, but if they expend too much energy searching, they will also die. As night approaches, birds have to make a life-and-death decision. Some birds take excessive risks and perish as a result.

What Determines Interval Size?

Every profession, organization, or institution has established norms regarding how long various activities should take. It is hard to break “habits” acquired over many years, whether they are good or bad. Legal systems and bureaucracies, for example, are notoriously slow. “Congress is an incremental institution,” said Professor Julian Zelizer of Princeton University. “… That's its flaw and its virtue.”¹⁴ How long something takes is heavily influenced by tradition and culture.

Psychology is another factor that influences our expectations about interval size. If a time period is painful, we want to shorten it; if it is pleasurable, we want to extend it. Most individuals are reluctant to accept financial losses, so they take their time in doing so. The typical investor is slow to sell stocks that have lost value over time, hoping they will recover. But if a stock crashes, losing most of its value overnight, most of us want to sell immediately to avoid additional pain, thereby missing the possibility of a rebound.

When we become emotionally invested in a particular outcome, that can influence our sense of timing as well. For example, the New York Times cited a report that 63 percent of the doctors surveyed overestimated the amount of time their terminally ill patients would live by a factor of 5.3. In an article accompanying the original report, Dr. Julia L. Smith, medical director of a hospice in Rochester, wrote, “‘Those of us who know our patients longer often become attached to them … We, too, hate to admit that death is near.’”¹⁵

Finally, if you need to guess whether an interval will shorten or lengthen, and have no other information to go on, look at how the interval begins and ends. If it starts with an important question, most of us want the answer as soon as possible. If we face a serious problem, we want it solved right away. If a crime is committed, we want the criminal caught and punished immediately, and so on. In Table 3.2, I've diagrammed the forces that cause us to act to shorten intervals.

In contrast, if an interval begins and ends with a problem, many of us want to delay the second problem until the first has been solved. Like a magnet, similar beginning and end points repel; different beginning and end points attract. No doubt there are exceptions, but this is a reasonable rule of thumb. I call it the magnet rule of interval size.

Table 3.2 The End Points of an Interval Attract

Objective or Subjective

Keep in mind that the length of any interval has both a subjective and an objective dimension. Sometimes time flies; sometimes it drags. It's not how long or short an interval is but how it is perceived under varying circumstances. A customer waiting in line to make an important purchase as opposed to one waiting to make a complaint will experience the same interval of time differently.

Contents

An interval can be filled with anything or nothing. Let's first consider the case of the empty interval—a time in your business, market, or industry when nothing significant seems to be going on.¹⁶ It is important to look inside empty intervals because there is often more going on than you think, just as physicists have learned that vacuums are not really empty. Here are four reasons why an interval may appear to be empty when, in fact, it is not:

• Early stage. What is going on inside the interval may need time to develop. It may not be of sufficient size or strength to be detected or recognized, and when it is, it may be too late. Nokia may not have been worried when Steve Jobs first began moving from computing into mobile technology, but by the time the iPhone was released in 2009, the Finnish cellular company was too far behind the curve to easily catch up.
• Masking. One process or activity may mask or direct attention away from another. Firms often focus on their current customers, adding more and more sophisticated bells and whistles to existing products. When the economy is growing, they prosper. What they do not notice is the rise of low-cost innovations developed in emerging markets that will threaten their business.
• Inhibition. Forces may be present that inhibit or block what is going on. A product may be extremely useful but fail to find a market because its potential customers are currently out of work and are fighting to make ends meet.
• Missing component. An interval may appear to be empty because its contents need a catalyst. Without those enabling conditions, nothing will happen. Mobile games require high-speed networks. No one wants to wait an hour to download a free app.

In some cases, of course, an interval may be as empty as it appears. Perhaps it is August, and everyone has “gone fishing.” Some industries simply change much more slowly than others. Luggage, for example, remained almost exactly the same for decades—an interval of time when nothing was going on—until the 1970s, when Bernard Sadow invented rolling suitcases. At first, the innovation did not take off, but once it did, it was the popular choice until the sleeker design of the Travelpro Rollaboard supplanted it in the late 1980s.

What Shapes Fill an Interval?

In order to see what is going on inside an interval, we need to be aware of the shape of the processes that might be found there. We like to think in terms of straight lines and linear processes. (As Euclid told us, a straight line is the shortest distance between two points.) Straight lines are fast and predictable. But the problem, as Spanish architect Antoni Gaudí pointed out, is that there are no straight lines in nature; nature, Gaudí said, is a “symphony of forms.”¹⁷ Take the experience of waiting. Everyone knows that waiting is frustrating. But if you graph how quickly frustration builds, the result won't be a straight line. Most of us will wait quietly for a while and then, at some point, simply run out of patience. Or think about buying a new car. As soon as you drive the car off the lot, its value drops dramatically. Economists tell us that the farther a reward is pushed into the future, the faster its value declines. Even the Swiss, masters of timing, failed to anticipate a nonlinear process. They greatly underestimated how quickly electronic watches would improve and how quickly prices would drop.¹⁸

Time is filled with all kinds of mathematical curves and shapes besides straight lines, bottlenecks, wheels, processes that speed up exponentially, and so on. We need to remember to look for unexpected shapes. We will examine this characteristic more closely in Chapter Five.

Meaning

How we label an interval can change its meaning, and hence its commercial value. The artist Christo, whose work includes wrapping the Reichstag in Berlin completely in fabric, found an interesting way to transform a means into an end. As the following passage illustrates, that shift in meaning allows Christo to better cope with the long and difficult task of gaining approval for his projects, which can take years.

Thus the lengthy steps before and after Christo's project are not wasted; they are valuable in themselves. Relabeling an interval is an underutilized strategy for solving problems and allowing people to look at things differently. For example, if a plane cannot take off due to bad weather, why not label the delay a “safety hold”? It's still a delay, but passengers will be less frustrated by the wait. In your own work, try to imagine how the means to an end could become an end in itself. Could the process used to create a product become a product in its own right?

Number

There is more than one interval involved in most complex situations. For example, investigators found that it took more than an hour for the emergency alarm to sound on the ship after the Costa Concordia hit rocks off the western coast of Italy in January 2012.²⁰ It is not clear why that delay occurred, but as a result, there was not enough time between the emergency signal and the order to abandon ship for people to react. The captain needed to consider two intervals: how long to wait after the ship ran aground to sound the emergency alarm, and then, how long after that to give the order to abandon ship. In this particular case, those two intervals were related: the longer the first, the shorter the second. Obviously, the captain didn't think about this issue, or if he did, he didn't act on that knowledge.

One reason it is important to find multiple intervals is that differences between or among them matter. The New Yorker magazine referred to one of its customers, who was living in a nursing home at the time, as “the late” reader. Still alive, “this person wrote to the magazine demanding a correction. The New Yorker, in its next issue, of course complied, inadvertently doubling the error, because the reader died over the weekend while the magazine was being printed.”²¹ An interval difference: the reader's life span was shorter than the time needed to correct the error. Bad timing.

INTERVAL AND DURATION RISKS

As we have seen with the other timing elements, mistakes occur when we miss or misread them. The reasons we fail to see and understand intervals are subtle, and the list of risks I will present is far from exhaustive, but it addresses some of the barriers to finding and understanding intervals that I have seen over and over again.

Language

One reason we miss intervals is the efficiency of our language, which allows us to capture a series of actions in a single word or phrase. On February 14, 2007, for example, an ice storm hit the Northeast. Passengers spent hours stranded on a JetBlue flight in New York's JFK airport. It couldn't take off and couldn't go back to the gate (the gate was filled). David G. Neeleman, the company's CEO at the time, said he was “humiliated and mortified” by what had occurred.²²

In order to prevent this type of occurrence, we can't simply think about canceling or not canceling a flight (a noun). We must think about sequences: a plane leaves the gate, taxis to the takeoff runway, then takes off—or not. Then we have to find the intervals within that sequence and understand the factors that influence the duration of each. The noun flight represents a very high level of abstraction. It does not reveal the more detailed sequence of intervals that a flight actually consists of, any one of which can cause trouble. We need the efficiency of nouns and verbs that stand for complex sequences of action—communication would be impossible without them—but what is lost is often exactly what we need to identify timing-related risks and manage them.

Question to ask: Have I overlooked the time between steps and stages because of the shorthand language we use to describe events?

Magnitude Thinking

If we ask someone the rate of turnover in a company, his or her answer will be a magnitude—the number of people entering and leaving within a given period of time. We have a concept, turnover, and we assign a number to it, a magnitude, and we think we are done. That number, however, will tell us nothing about how long those positions may be vacant, whether the company had advance warning that individuals would leave, and how long it will take to replace them. As we discussed with JetBlue's canceled flight, the concept of turnover doesn't stop us from asking these questions. But it doesn't call our attention to them either. The English language (and many others) makes it easy to summarize a complex process in a single word or magnitude. In fact, anytime someone talks about a variable, you know that there are hidden intervals involved. A variable, like turnover or the number of flights departing an airport per hour, will tell you what you need to take into account or measure. But if you want to know the risks associated with a variable, you need to look further. You need to find the intervals that make up the variable, and determine when they might become too long or too short.

Question to ask: Has magnitude thinking (defining a situation based on a number that is variable) caused me to miss relevant intervals?

Time Line Thinking

Whenever we find ourselves facing a problem, the natural next step is to search for a solution, which we hope will be found as quickly as possible. As indicated by Table 3.2, we can represent the interval between problem and solution as follows:

If the problem is a serious one, we fear that finding a solution may take a long time, which can create any number of difficulties. Time to solution, however, is not the only interval we need to be concerned about. The way to start is to give problems and solutions their own time lines, which I illustrate here:

This form of visual representation accomplishes two things. First, it counteracts the unconscious gap-closing tendency that causes us to underestimate the time needed to find a solution. Second, and just as important, it reminds us that a solution can persist long after the problem that caused it is over, in which case, the solution may become the problem.

That is exactly what happened in the stock market in 1959. A speculative fever caused trading volume to mushroom. The market couldn't handle the spike in volume, and it took weeks for some trades to clear. Extra clerks and accountants were hired, but the system was again overwhelmed a few years later. The NYSE responded by adding computers and other automation. But by 1970, the cost of those improvements exceeded the volume needed to pay for them.²³ This example is over fifty years old, but the principle (solutions may become problems) is as valid now as it was then. Take airport security. Will our great-great-grandchildren, when they are adults, still have to take their shoes off before they board a plane? It's hard to say, but what I know with certainty is that some future problems will be the result of solutions that refused to die when their time was up.

There is another advantage to putting problems and solutions on separate lines: it allows us to think about their sequence. Logically, solutions should follow the problems they are designed to solve. But sometimes we need to begin a solution before we know that the problem exists, or risk being too late. Global warming is an example. If we could find a way to make a solution profitable in itself, perhaps we could get started before it is too late.

Question to ask: Has my focus on “time to solution” caused me to miss other significant intervals?

Cat-Point Thinking

We are used to thinking about performing a category of action at a single point in time. It is therefore easy to forget about the sequences of steps—and the intervals between them—that will be involved. There is a wonderful Dilbert cartoon that illustrates this point. Dilbert is buying a chair. After he selects one, the salesman compliments him on “an excellent choice,” but tells him that his next task is to sit quietly and not ask “the one question that will kill this sale.” Of course, Dilbert asks it: “Is the chair in stock?” The salesman replies that “we don't sell chairs. We sell the hope that a chair will someday be made for you.” So Dilbert naturally asks, “How long will that take?” The salesman replies, “If I could answer that question, it would be the same as selling you an actual chair. How about if I tell you it will ship in two months, and you call and yell at me every three months for eternity?” Dilbert returns to his office, and in response to a question about whether he bought a chair, he replies, “There is no way to know.”²⁴

The point of the cartoon is profound. Actions and transactions we expect to be instantaneous, or nearly so, are in reality extended sequences filled with intervals of varying lengths. Rolled up inside the name of any activity is a set of sequences and intervals that you need to discover, like the dimensions of the world the quantum physicists tell us we don't see, if you really want to know how long something will take.

Question to ask: Have I overlooked a series of intervals by treating a situation or action as a single event?

Intervals That Are Too Long or Short

There are risks associated with intervals that are either too long or too short. One way to check for those risks is to fill in all four cells of the Interval Envelope—that is, try to estimate the maximum and minimum length of the longest and shortest intervals that are involved in the work you are doing, and identify the risks if those intervals were to occur. In the 2002 World Cup, for instance, the United States upset Portugal, 3–2. One reason Portugal's coach Antonio Oliveira gave for the defeat was that “with our players in European leagues, we only had two weeks to prepare.”²⁵

Be alert for horizon errors, which occur when you don't look far enough back into the past or far enough forward into the future. Scott McCleskey (head of compliance at Moody's from April 2006 to September 2008) noted that “once Moody's issues these ratings [‘on thousands of municipal bonds’], it rarely reviews them again—leaving them fallow, sometimes for decades.”²⁶ Aging financial infrastructure poses the same risk of horizon errors as aging physical infrastructure: after a while, it may not be safe.

Questions to ask: Do the longest or shortest intervals in a series present risks? Have I considered what will happen if the shortest interval in a sequence is shorter than I anticipated, or the longest interval, longer?

Elasticity

Is the size of an interval fixed, or can it be made longer or shorter as needed? Term limits, for example, which serve to standardize the length of an interval, help insure accountability. But they have significant risks as well—namely, the loss of expertise, which is particularly important when the issues to be decided are complicated and the consequences for mistakes long lasting. In a business situation, is a deadline a true deadline, or can it be adjusted, giving those working on a task a longer or shorter time to accomplish it?

Question to ask: Have I considered whether an interval is fixed or elastic, and does the distinction matter?

Norms

We often establish norms about how long an interval should last. When these norms are violated, people notice. In February 2000, for example, after just two years as president of Brown University, E. Gordon Gee surprised the Brown campus and the higher education community nationwide when he abruptly resigned to become chancellor of Vanderbilt University. “Two years is just too short, and I admit that up front,” Gee said, “this is a fast-forward world we're in right now, in which there's a lot of change.”²⁷

When you violate a norm associated with an interval, be prepared to explain yourself or have your actions questioned. I suspect that if Gee also chose to leave his Vanderbilt job in just a year or two (he did not; he stayed until 2007), he would have difficulty repairing his reputation no matter what explanation he provided.

Question to ask: What will happen if the length of an interval is longer or shorter than expected?

Missing the Location of an Interval

When an interval occurs can be just as important as how long it is. For example, Tony Gwynn of the San Diego Padres made the following comments in 1999: “What if it is the last year of your contract and you feel you're not playing well? With the kind of money out there, could Creatine or Andro or steroids make a difference in your play? And if you only take them one year to prolong your career, would it really be dangerous to your health?”²⁸ If we know when a risk is present, we are more able to avoid or mitigate it. Were baseball to focus its enforcement on the times when players are most likely to use banned substances, those efforts might be more effective. When players begin, resume, or permanently discontinue drug use is probably not random.

Question to ask: Have I considered where an interval occurs within the larger sequence of events?

Content Risks

You can make an error because there is more (or less) going on during an interval than you realize. Some intervals are truly empty: nothing is going on, as was the case in the luggage industry before Sadow. But sometimes there is a lot happening, such as when banks systematically lower their debt levels right before quarterly reports, and then undo those actions at the start of the next quarter, which has led the SEC to consider new rules to deter banks from “dressing up their books.”²⁹ It is also possible for the contents of one interval to mask the contents of another, such as when important nonpolitical news doesn't get reported during a hotly contested presidential election.

Questions to ask: What, if anything, is truly happening within a given interval? Is it exactly what it seems or something else altogether?

Misinterpreting Intervals

It is common to misread how others will interpret an interval. Back in 1993, for example, a federal judge awarded $11.5 million (at the time a record) to a former General Electric employee who exposed defense contract fraud at the company. GE said that the whistleblower, Chester Walsh, delayed reporting the fraud so that he could collect the large sum. Walsh countered that “he needed that much time [more than four years] to build his case against his employer.”³⁰ After you locate an interval that is critical for the success of your work, ask yourself how others might interpret it and whether there is a risk that they could interpret it in the wrong way. According to a study by the National Marriage Project in 2001, women tend to view living together as a step toward marriage, whereas men reportedly view it as trying out a relationship.³¹

Questions to ask: Am I interpreting the contents and meaning of an interval correctly? Might others interpret the same interval of time differently?

INTERVAL AND DURATION OPTIONS AND OPPORTUNITIES

Paying attention to intervals—their size, sequence, and location—can enable you to design a more effective course of action. For example, NASCAR used a sequence of intervals to solve a difficult timing problem. Suspecting that various teams were modifying racecars in ways that were prohibited, NASCAR officials wanted to cut down on the cheating and faced a question of timing: When should they issue a warning? On the one hand, if the warning were issued too early, it wouldn't be taken seriously. Why now? someone would ask. Perhaps the threat that it warns against won't materialize. On the other hand, a warning that is issued too late is, well, too late. As conventionally understood, this is a cat-point problem. At what point in time should a warning (a category of action) be issued? But that is not how NASCAR approached the problem. Its solution was to use a sequence of intervals.³²

1. NASCAR issued a warning during the previous season that “penalties would escalate if infractions continued.”
2. Three weeks before teams arrived at Daytona for Speedweeks, teams were warned again.
3. The Sunday before the Daytona 500, scheduled for February 18, 2007, “NASCAR suspended five crew chiefs and a team vice president, and assessed fines and subtracted valuable race points from five drivers and teams before Sunday's season-opening Daytona 500.”
4. The penalties were timed so that they began at the start of the new season, thus completing the cycle.

There were two warnings before violators were punished, and the intervals between them were of decreasing size. Warnings were issued about a year, a month, and a week in advance. Penalties were timed to coincide with the start of the new season, thereby using a punctuation mark. Thus NASCAR's solution did not rely on a single warning, because deciding how early to warn drivers and to have that warning be effective was an impossible problem to solve. It would be attempting to solve a timing problem by reference to a single magnitude—that is, using an interval of a given size. Instead, NASCAR moved from a point solution, selecting one point in time, to a path solution, one that involved a sequence of intervals.

Finding intervals leads to better solutions. For example, research shows that cancer patients don't like waiting at the hospital for blood test results in order to know whether they are strong enough to receive chemotherapy. The conventional solution would be to reduce waiting time. Can the results of the test be obtained more quickly? If not, can patients be distracted by TV or engaged in other activities?

The Cancer Center at Sloan-Kettering in New York asked the design firm IDEO to investigate. The result was to offer some patients the blood tests the day before they were scheduled for chemotherapy. The Center found that many patients actually preferred making two trips rather than having testing done at the last minute.³³ The key was where in the sequence of events (testing–results–chemo) the important waiting interval was located, rather than its size. Size suggests a magnitude solution: when an interval is too long, shorten it. But in this case, the better solution was to relocate the critical interval (the wait from test to result) to the previous day. If the results were favorable, patients wouldn't have to wait for chemotherapy when they arrived the next day. Notice that this redesign works because of what does not happen over time. If the patient's test results varied minute by minute, the two-trip solution would not work.

Think about what makes the NASCAR and cancer examples similar despite their obvious differences. In both cases, the solutions involved seeing a sequence of intervals, stretching over months rather than weeks for NASCAR, and days not just minutes for the Cancer Center. That is why it is important to search for intervals: they can suggest new and better solutions when conventional ones prove inadequate.

THE TEMPORAL IMAGINATION

The comedian Jonathan Winters recalls a conversation he had with a fellow tourist when he was visiting the Temple of Athena in Greece.

Winters reminds us to think about what we should expect with the passage of time. We expect parts to wear out, what is broken to be fixed, memories to fade, a jazzy new phrase to become a cliché, and leaders to retire or be ousted. When a process that should take a long time doesn't, we know something is wrong, as when a widow remarries the day after she buries her husband. She hasn't allowed enough time, a long enough “decent” interval, to separate these two events.

Sometimes the fact that nothing changes with the passage of time, a true empty interval, should raise a warning flag. As we know, there were not enough lifeboats on the Titanic when it sank. But the ship was in full compliance with all marine laws. The difficulty was that the British Board of Trade “hadn't updated its regulations for nearly 20 years,” and “it had been 40 years since the last serious loss of life at sea.”³⁵ Regulations were simply out-of-date. I suggest we time-stamp all policies, rules, and regulations related to safety, or any other critical matter, with dates of origin and expiration so that we know when they need to be revisited, revised, or revoked.