Principles P3 PDF

100) EVALUATE PEOPLE ACCURATELY, NOT “KINDLY” So...... 101) Make accurate assessments. Since truth is the foundation of excellence and people are your most important resource, make the most precise personnel evaluations possible. This accuracy takes time and considerable back-and-forth. Your assessment of how responsible parties are performing should be based not on whether they’re doing it your way but on whether they’re doing it in a good way. Speak frankly, listen with an open mind, consider the views of other believable and honest people, and try to get in synch about what’s going on with the person and why. Remember not to be overcon dent in your assessments as it’s possible you are wrong. 101a) Use evaluation tools such as performance surveys, metrics, and formal reviews to document all aspects of a person’s performance. These will help clarify assessments and communication surrounding them. 101b) Maintain “baseball cards” and/or “believability matrixes” for your people. Imagine if you had baseball cards that showed all the performance stats for your people: batting averages, home runs, errors, ERAs, win/loss records. You could see what they did well and poorly and call on the right people to play the right positions in a very transparent way. These would also simplify discussions about compensation, incentives, moving players up to rst string, or cutting them from the team. You can and should keep such records of your people. Create your baseball cards to achieve your goals of conveying what the person is like. I use ratings, forced rankings, metrics, results, and credentials. Baseball cards can be passed to potential new managers as they consider candidates for assignments.... 102) Evaluate employees with the same rigor as you evaluate job candidates. Ask yourself: “Would I hire this person knowing what I now know about them?” I nd it odd and silly that interviewers often freely and con dently criticize job candidates despite not knowing them well, yet they won’t criticize employees for similar weaknesses even though they have more evidence. That is because some people view criticism as harmful and feel less protective of an outsider than they do of a fellow employee. If you believe accuracy is best for everyone, then you should see why this is a mistake and why frank evaluations must be ongoing.... 103) Know what makes your people tick, because people are your most important resource. Develop a full pro le of each person’s values, abilities, and skills. These qualities are the real drivers of behavior, and knowing them in detail will tell you which jobs a person can and cannot do well, which ones they should avoid, and how the person should be trained. I have often seen people struggling in a job and their manager trying for months to nd the right response because the manager overlooked the person’s “package.” These pro les should change as the people change.... 104) Recognize that while most people prefer compliments over criticisms, there is nothing more valuable than accurate criticisms. While it is important to be clear about what people are doing well, there should not be a reluctance to pro le people in a way that describes their weaknesses. It is vital that you be accurate.... 105) Make this discovery process open, evolutionary, and iterative. Articulate your theory of a person’s values, abilities, and skills upfront and share this with him; listen to his and others’ response to your description; organize a plan for training and testing; and reassess your theory based on the performance you observe. Do this on an ongoing basis. After several months of discussions and real-world tests, you and he should have a pretty good idea of what he is like. Over time this exercise will crystallize suitable roles and appropriate training, or it will reveal that it’s time for the person to leave Bridgewater.... 106) Provide constant, clear, and honest feedback, and encourage discussion of this feedback. Don’t hesitate to be both critical and complimentary—and be sure to be open-minded. Training and assessing will be better if you frequently explain your observations. Providing this feedback constantly is the most e ective way to train. 106a) Put your compliments and criticisms into perspective. I nd that many people tend to blow evaluations out of proportion, so it helps to clarify that the weakness or mistake under discussion is not indicative of your total evaluation. Example: One day I told one of the new research people what a good job I thought he was doing and how strong his thinking was. It was a very positive initial evaluation. A few days later I heard him chatting away for hours about stu that wasn’t related to work, so I spoke to him about the cost to his and our development if he regularly wasted time. Afterward I learned he took away from that encounter the idea that I thought he was doing a horrible job and that he was on the brink of being red. But my comment about his need for focus had nothing to do with my overall evaluation of him. If I had explained myself when we sat down that second time, he could have better put my comments in perspective. 106b) Remember that convincing people of their strengths is generally much easier than convincing them of their weaknesses. People don’t like to face their weaknesses. At Bridgewater, because we always seek excellence, more time is spent discussing weaknesses. Similarly, problems require more time than things that are going well. Problems must be gured out and worked on, while things that are running smoothly require less attention. So we spend a lot of time focusing on people’s weaknesses and problems. This is great because we focus on improving, not celebrating how great we are, which is in fact how we get to be great. For people who don’t understand this fact, the environment can be di cult. It’s therefore important to 1) clarify and draw attention to people’s strengths and what’s being done well; and 2) constantly remind them of the healthy motive behind this process of exploring weaknesses. Aim for complete accuracy in your assessments. Don’t feel you have to nd an equal number of “good and bad” qualities in a person. Just describe the person or the circumstances as accurately as possible, celebrating what is good and noting what is bad. 106c) Encourage objective reﬂection—lots reﬂection and lots of it. 106d) Employee reviews: While feedback should be constant, reviews are periodic. The purpose of a review is to review the employee's performance and to state what the person is like as it pertains to their doing their job. A job review should have little surprises in it—this is because throughout the year, if you can’t make sense of how the person is doing their job or if you think it’s being done badly, you should probe them to seek understanding of root causes of their performance. Because it is very di cult for people to identify their own weaknesses, they need the appropriate probing (not nitpicking) of speci c cases by others to get at the truth of what they are like and how they are tting into their jobs. From examining these speci c cases and getting in synch about them, agreed- upon patterns will emerge. As successes and failures will occur in everyone (every batter strikes out a lot), in reviewing someone the goal is see the patterns and to understand the whole picture rather than to assume that one or a few failures or successes is representative of the person. You have to understand the person’s modus operandi and that to be successful, they can’t be successful in all ways—e.g., to be meticulous they might not be able to be fast (and vice versa). Steve Jobs has been criticized as being autocratic and impersonal, but his modus operandi might require him being that way, so the real choice in assessing his t for his job is to have him the way he is or not at all: that assessment must be made in the review, not just a theoretical assessment that he should do what he is doing and be less autocratic. In some cases it won’t take long to see what a person is like—e.g., it doesn’t take long to hear if a person can sing. In other cases it takes a signi cant number of samples and time to re ect on them. Over time and with a large sample size you should be able to see what people are like, and their track records (i.e., the level and the steepness up or down in the trajectories that they are responsible for, rather than the wiggles in these) paint a very clear picture of what you can expect from them. If there are performance problems, it is either because of design problems (e.g., the person has too many responsibilities) or t/abilities problems. If the problems are due to the person’s inabilities, these inabilities are either because of the person’s innate weaknesses in doing that job (e.g., with a height of 5- foot-2, the person probably shouldn’t be a center on the basketball team) or because of inadequate training to do the job. A good review, and getting in synch throughout the year, should get at these things. The goal of a review is to be clear about what the person can and can’t be trusted to do based on what the person is like. From there, “what to do about it” (i.e., how these qualities t into the job requirement) can be determined.... 107) Understand that you and the people you manage will go through a process of personal evolution. Personal evolution occurs rst by identifying your strengths and weaknesses, and then by changing your weaknesses (e.g., through training) or changing jobs to play to strengths and preferences. This process, while generally di cult for both managers and their subordinates, has made people happier and Bridgewater more successful. Remember that most people are happiest when they are improving and doing things that help them advance most rapidly, so learning your people’s weaknesses is just as valuable for them and for you as learning their strengths.... 108) Recognize that your evolution at Bridgewater should be relatively rapid and a natural consequence of discovering your strengths and weaknesses; as a result, your career path is not planned at the outset. Your career path isn’t planned because the evolutionary process is about discovering your likes and dislikes as well as your strengths and weaknesses. The best career path for anyone is based on this information. In other words, each person’s career direction will evolve di erently based on what we all learn. This process occurs by putting people into jobs that they are likely to succeed at, but that they have to stretch themselves to do well. They should be given enough freedom to learn and think for themselves while being coached so they can be taught and prevented from making unacceptable mistakes. During this process they should receive constant feedback. They should re ect on whether their problems can be resolved by additional learning or stem from innate qualities that can’t be changed. Typically it takes six to 12 months to get to know a person in a by-and-large sort of way and about 18 months to change behavior (depending on the job and the person). During this time there should be periodic mini-reviews and several major ones. Following each of these assessments, new assignments should be made to continue to train and test them. They should be tailored to what was learned about the person’s likes and dislikes and strengths and weaknesses. This is an iterative process in which these cumulative experiences of training, testing, and adjusting direct the person to ever more suitable roles and responsibilities. It bene ts the individual by providing better self- understanding and greater familiarity with various jobs at Bridgewater. This is typically both a challenging and rewarding process. When it results in a parting of ways, it’s usually because people nd they cannot be excellent and happy in any job at Bridgewater or they refuse to go through this process.... 109) Remember that the only purpose of looking at what people did is to learn what they are like. Knowing what they are like will tell you how you can expect them to handle their responsibilities in the future. Intent matters, and the same actions can stem from di erent causes. 109a) Look at patterns of behaviors and don’t read too much into any one event. Since there is no such thing as perfection, even excellent managers, companies, and decisions will have problems. It’s easy, though often not worth much, to identify and dwell on tiny mistakes. In fact, this can be a problem if you get bogged down pinpointing and analyzing an in nite number of imperfections. At the same time, minor mistakes can sometimes be manifestations of serious root causes that could cause major mistakes down the road, so they can be quite valuable to diagnose. When assessing mistakes it is important to 1) ask whether these mistakes are manifestations of something serious or unimportant and 2) re ect on the frequency of them. An excellent decision-maker and a bad decision- maker will both make mistakes. The di erence is what causes them to make mistakes and the frequency of their mistakes. There is also a di erence between “I believe you made a bad decision” and “I believe you are a bad decision-maker,” which can be ascertained only by seeing the pattern. Any one event has many di erent possible explanations, whereas a pattern of behavior can tell you a lot about root causes. There are many qualities that make up a person. To understand each requires 1) a reliable sample size and 2) getting in synch (i.e., asking the person why and giving feedback). Some qualities don’t require a large sample size—e.g., it takes only one data point to know if a person can sing—and others take multiple observations ( ve to 10). The number of observations needed to detect a pattern largely depends on how well you get in synch after each observation. A quality discussion of how and why a person behaved a certain way should help you quickly understand the larger picture. 109b) Don’t believe that being good or bad at some things means that the person is good or bad at everything. Realize that all people have strengths and weaknesses.... 110) If someone is doing their job poorly, consider whether this is due to inadequate learning (i.e., training/experience) or inadequate ability. A weakness due to a lack of experience or training or due to inadequate time can be xed. A lack of inherent ability cannot. Failing to distinguish between these causes is a common mistake among managers, because managers are often reluctant to appear unkind or judgmental by saying someone lacks ability. They also know people assessed this way tend to push back hard against accepting a permanent weakness. Managers need to get beyond this reluctance. In our diagram of thinking through the machine that will produce outcomes, think about...... 111) Remember that when it comes to assessing people, the two biggest mistakes are being overconﬁdent in your assessment and failing to get in synch on that assessment. Don’t make those mistakes. 111a) Get in synch in a non-hierarchical way regarding assessments. The greatest single discrepancy between a manager and a managee is how well each performs his job. In most organizations, evaluations run in only one direction, with the manager assessing the managee. The managee typically disagrees with the assessment, especially if it is worse than the employee’s self- assessment, because most people believe themselves to be better than they really are. Managees also have opinions of managers that in most companies they wouldn’t dare bring up, so misunderstandings and resentments fester. This perverse behavior undermines the e ectiveness of the environment and the relationships between people. It can be avoided by getting in synch in a high-quality way. 111b) Learn about your people and have them learn about you with very frank conversations about mistakes and their root causes. You need to be clear in conveying your assessments and be open-minded in listening to people’s replies. This is so they can understand your thinking and you can open-mindedly consider their perspectives. So together you can work on setting their training and career paths. Recognizing and communicating people’s weakness is one of the most di cult things managers have to do. Good managers recognize that while it is di cult in the short term, it actually makes things easier in the long term, because the costs of having people in jobs where they can’t excel are huge. Most managers at other companies dodge being as open with assessments as we insist on; more typically, managers elsewhere tend to be less frank in conveying their views, which is neither fair nor e ective.... 112) Help people through the pain that comes with exploring their weaknesses. Emotions tend to heat up during most disagreements, especially about someone’s possible weaknesses. Speak in a calm, slow, and analytical manner to facilitate communication. If you are calm and open to others’ views, they are less likely to shut down logical exchanges than if you behave emotionally. Put things in perspective by reminding them that their pain is the pain that comes with learning and personal evolution—they’re going to be in a much better place by getting to truth. Consider asking them to go away and re ect when they are calm, and have a follow-up conversation in a few days.... 113) Recognize that when you are really in synch with people about weaknesses, whether yours or theirs, they are probably true. Getting to this point is a great achievement. When you reach an agreement, it’s a good sign you’re there. This is one of the main reasons why the person being evaluated needs to be an equal participant in the process of nding truth. So when you do agree, write it down on the relevant baseball card. This information will be a critical building block for future success.... 114) Remember that you don’t need to get to the point of “beyond a shadow of a doubt” when judging people. Instead, work toward developing a mutually agreed “by-and-large” understanding of someone that has a high level of con dence behind it. When necessary, take the time to enrich this understanding. That said, you should not aim for perfect understanding. Perfect understanding isn’t possible, and trying to get it will waste time and stall progress.... 115) Understand that you should be able to learn the most about what a person is like and whether they are a “click” for the job in their ﬁrst year. You should be able to roughly assess someone’s abilities after six to 12 months of close contact and numerous tests and getting in synch about them. A more con dent assessment so that you can make a more con dent role assignment will probably take about 18 months. This timeline will of course depend on the job, the person, the amount of contact with that person, and how well you do it. As I explain in the section on design, the ratio of senior managers to junior managers as well as the ratio of managers to the number of people who work two levels below them should be small enough to ensure quality communication and mutual understanding. Generally, that ratio should not be more than 1:10, preferably more like 1:5.... 116) Continue assessing people throughout their time at Bridgewater. You will get to know them better, it will help you train and direct them, and you won’t be stuck with an obsolete picture. Most importantly, assess what your people’s core values and abilities are and make sure they complement Bridgewater’s. Since core values and abilities are more permanent than skills, they are more important to ascertain, especially at Bridgewater. As mentioned, you should be able to roughly assess people’s abilities after six to 12 months of close contact and con dently assess them after 18 months. Don’t rest with that evaluation, however. Always ask yourself if you would have hired them for that job knowing what you now know. If not, get them out of the job. 117) TRAIN AND TEST PEOPLE THROUGH EXPERIENCES So...... 118) Understand that training is really guiding the process of personal evolution. It requires the trainee to be open-minded, to suspend ego in order to nd out what he is doing well and poorly, and to decide what to do about it. It also requires the trainer to be open-minded (and to do the other things previously mentioned). It would be best if at least two believable trainers work with each trainee in order to triangulate views about what the trainee is like. As previously explained, the training should be through shared experiences like that of a ski instructor skiing with his student—i.e., it should be an apprentice relationship.... 119) Know that experience creates internalization. A huge di erence exists between memory- based “book” learning and hands-on, internalized learning. A medical student who has “learned” to perform an operation in his medical school class has not learned it in the same way as a doctor who has already conducted several operations. In the rst case, the learning is stored in the conscious mind, and the medical student draws on his memory bank to remember what he has learned. In the second case, what the doctor has learned through hands-on experience is stored in the subconscious mind and pops up without his consciously recalling it from the memory bank. People who excel at book learning tend to call up from memory what they have learned in order to follow stored instructions. Others who are better at internalized learning use the thoughts that ow from their subconscious. The experienced skier doesn’t recite instructions on how to ski and then execute them; rather, he does it well “without thinking,” in the same way he breathes without thinking. Understanding these di erences is essential. Remember that experience creates internalization. Doing things repeatedly leads to internalization, which produces a quality of understanding that is generally vastly superior to intellectualized learning.... 120) Provide constant feedback to put the learning in perspective. Most training comes from doing and getting in synch about performance. Feedback should include reviews of what is succeeding and what is not in proportion to the actual situation rather than in an attempt to balance compliments and criticisms. You are a manager, and you want your machine to function as intended. For it do so, employees must meet expectations, and only you can help them to understand where they are in relation to expectations. As strengths and weaknesses become clearer, responsibilities can be more appropriately tailored to make the machine work better and to facilitate personal evolution. The more intensely this is done, the more rapid the evolutionary process will be. So you must constantly get in synch about employee performance.... 121) Remember that everything is a case study. Think about what it is a case of and what principles apply.... 122) Teach your people to ﬁsh rather than give them ﬁsh. It is a bad sign when you tell people what they should do because that behavior typically re ects micromanagement or inability on the part of the person being managed. Instead, you should be training and testing. So give people your thoughts on how they might approach their decisions or how and why you would operate in their shoes, but don’t dictate to them. Almost all that you will be doing is constantly getting in synch about how they are doing things and exploring why.... 123) Recognize that sometimes it is better to let people make mistakes so that they can learn from them rather than tell them the better decision. However, since the connections between cause and e ect can be misunderstood, providing feedback for these people is essential to the learning process. 123a) When criticizing, try to make helpful suggestions. Your goal is to help your people understand and improve, so your suggestions are important. O ering suggestions also helps those being criticized to understand that your goal is to help them and Bridgewater, not to hurt them. 123b) Learn from success as well as from failure. Point out examples of jobs that are well done and the causes of success. This reinforces good behavior and creates role models for those who are learning.... 124) Know what types of mistakes are acceptable and unacceptable, and don’t allow the people who work for you to make the unacceptable ones. When considering what failures you are willing to allow in order to promote learning through trial-and-error, weigh the potential damage of a mistake against the bene t of incremental learning. In de ning what latitude I’m willing to give people, I say, “I’m willing to let you scratch or dent the car, but I won’t put you in a position where I think there’s a signi cant risk you could total it.”... 125) Recognize that behavior modiﬁcation typically takes about 18 months of constant reinforcement. The rst step is intellectualizing the best way of doing things. If you’re out of shape you must understand that you are out of shape, you must want to get in shape, and you must understand the way to get in shape: “I want to be t by eating well and exercising.” Then the intellect will ght with desires and emotions. With determination, the intellect will overcome the impediments to doing what’s necessary to achieve the goal, and the desired behavior will occur. After doing that consistently for 18 months, the new behavior will be internalized.... 126) Train people; don’t rehabilitate them. Training is part of the plan to develop people’s skills and to help them evolve. Rehabilitation is the process of trying to create signi cant change in people’s values and/or abilities. Since values and abilities are di cult to change, rehabilitation typically takes too long and is too improbable to do at Bridgewater. If attempted, it is generally best directed by professionals over extended periods of time. People with inappropriate values and inadequate abilities to meet their job requirements have devastating impacts on the organization. They should be properly sorted (see the principles section on sorting). 126a) A common mistake: training and testing a poor performer to see if he or she can acquire the required skills without simultaneously trying to assess their abilities. Skills are readily testable, so they should be easy to determine. Knowing them is less important than knowing people’s abilities. That makes picking people with the right skills relatively easy. Abilities, especially right-brained abilities, are more di cult to assess. When thinking about why someone is a poor performer, openly consider whether it is a problem with their abilities. Values are the toughest and take the longest to assess.... 127) After you decide “what’s true” (i.e., after you ﬁgure out what your people are like), think carefully about “what to do about it.”As it.” mentioned before, it’s important to separate thinking about “what’s true” and thinking about “what to do about it.” Figuring out what’s true takes time —often several months lled with a large sample size. Figuring out what to do about it (i.e., designing) is much faster—typically hours or days—but it isn’t instantaneous. Too often people either jump to decisions or don’t make them. 128) SORT PEOPLE INTO OTHER JOBS AT BRIDGEWATER, OR REMOVE THEM FROM BRIDGEWATER So...... 129) When you ﬁnd that someone is not a good “click” for a job, get them out of it ASAP. If you are expecting/wishing people to be much better in the near future than they have been in the past, you are making a serious mistake—instead, sort the people. People who repeatedly operated in a certain way probably will continue to operate that way because that behavior re ects what they’re like. Since people generally change slowly (at best), you should expect slow improvement (at best), so instead of hoping for improvement, you need to sort the people or change the design to supplement them. Since changing the design to accommodate people’s weaknesses is generally a bad idea, it is generally better to sort the people. Sometimes good people “lose their boxes” because they can’t evolve into responsible parties soon enough. Either there is a problem with their qualities or it will take too long to train them well. Some of these people might be good at another position within Bridgewater. Remember that identifying failure and learning from it are part of the evolutionary process. Make sure you record the reasons on the relevant “baseball card” and think about what a good next step would be for that individual.... 130) Know that it is much worse to keep someone in a job who is not suited for it than it is to ﬁre someone. Don’t collect people. Firing people is not a big deal—certainly nowhere near as big a deal as keeping badly performing people, because keeping a person in a job they are not suited for is terrible both for the person (because it prevents personal evolution) and our community (because we all bear the consequences and it erodes meritocracy). Consider the enormous costs of not ring someone unsuited for a job: the costs of bad performance over a long time; the negative e ect on the environment; the time and e ort wasted trying to train the person; and the greater pain of separation involved with someone who’s been here awhile (say, ve years or more) compared with someone let go after just a year.... 131) When people are “without a box,” consider whether there is an open box at Bridgewater that would be a better ﬁt. If not, ﬁre them. Remember that we hire people not to ll their rst job at Bridgewater nor primarily for their skills. We are trying to select people with whom we’d like to share our lives. We expect everyone to evolve here. Because managers have a better idea of people’s strengths and weaknesses and their t within our culture than what emerges from the interview process, you have invaluable information for assessing them for another role at Bridgewater.... 132) Do not lower the bar. If a person can’t operate consistently with our requirements of excellence and radical truth and can’t get to the bar in an acceptable time frame, they have to TO PERCEIVE, DIAGNOSE, AND SOLVE PROBLEMS... 133) KNOW HOW TO PERCEIVE PROBLEMS EFFECTIVELY So...... 134) Keep in mind the 5-Step Process explained in Part 2.... 135) Recognize that perceiving problems is the ﬁrst essential step toward great management. As in nature, if you can’t see what’s happening around you, you will deteriorate and eventually die o. People who can 1) perceive problems; 2) decide what to do about them; and 3) get these things done can be great managers.... 136) Understand that problems are the fuel for improvement. Problems are like wood thrown into a locomotive engine, because burning them up—i.e., inventing and implementing solutions—propels us forward. Problems are typically manifestations of root causes, so they provide clues for getting better. Most of the movement toward excellence comes from eliminating problems by getting at their root causes and making the changes that pay o repeatedly in the future. So nding problems should get you excited because you have found an opportunity to get better.... 137) You need to be able to perceive if things are above the bar (i.e., good enough) or below the bar (i.e., not good enough), and you need to make sure your people can as well. That requires the ability to synthesize.... 138) Don’t tolerate badness. Too often I observe people who observe badness and tolerate it. Sometimes it is because they don’t have the courage to make the needed changes, and sometimes it is because they don’t know how to x it. Both are very bad. If they’re stuck, they need to seek the advice of believable people to make the needed changes, and if that doesn’t work, they need to escalate.... 139) “Taste the soup.” A good restaurateur constantly tastes the food that is coming out of his kitchen and judges it against his vision of what is excellent. A good manager needs to do the same.... 140) Have as many eyes looking for problems as possible. Encourage people to bring problems to you and look into them carefully. If everyone in your area feels responsible for the well-being of that area and feels comfortable speaking up about problems, your risks of overlooking them will be much less than if you are the only one doing this. This will help you perceive problems, gain the best ideas, and keep you and your people in synch. 140a) “Pop the cork.” It’s your responsibility to make sure that communications from your people are owing freely. 140b) Hold people accountable for raising their complaints. Ask yourself: 1) does someone think there’s something wrong; 2) did this lead to a proper discussion; and 3) if they felt raising the issue didn’t lead to the proper response, did they escalate it? That’s how it should be. 140c) The leader must encourage disagreement and be either impartial or open-minded. 140d) The people closest to certain jobs probably know them best, or at least have perspectives you need to understand, so those people are essential for creating improvement.... 141) To perceive problems, compare how the movie is unfolding relative to your script—i.e., script compare the actual operating of the machine and the outcomes it is producing to your visualization of how it should operate and the outcomes you expected. As long as you have the visualization of your expectations in mind to compare with the actual results, you will note the deviations so you can deal with them. For example, if you expect improvement to be within a speci c range...... and it ends up looking like this...... you will know you need to get at the root cause to deal with it. If you don’t, the trajectory will probably continue.... 142) Don’t use the anonymous “we” and “they,” because that masks personal responsibility—use speciﬁc names. For example, don’t say “we” or “they” handled it badly. Also avoid: “We should...” or “We are...”Who is “we”? Exactly who should, who made a mistake, or who did a great job? Use speci c names. Don’t undermine personal accountability with vagueness. When naming names, it’s also good to remind people of related principles like “mistakes are good if they result in learning.”... 143) Be very speciﬁc about problems; don’t start with generalizations. For example, don’t say, “Client advisors aren’t communicating well with the analysts.” Be speci c: name which client advisors aren’t doing this well and in which ways. Start with the speci cs and then observe patterns.... 144) Tool: Use the following tools to catch problems: issues logs, metrics, surveys, checklists, outside consultants, and internal auditors. 1) Issues log: A problem or “issue” that should be logged is easy to identify: anything that went wrong. The issues log acts like a water lter that catches garbage. By examining the garbage and determining where it came from, you can determine how to eliminate it at the source. You diagnose root causes for the issues log the same way as for a drilldown (explained below) in that the log must include a frank assessment of individual contributions to the problems alongside their strengths and weaknesses. As you come up with the changes that will reduce or eliminate the garbage, the water will become cleaner. In addition to using issues logs to catch problems, you can use them to measure the numbers and types of problems, and they can therefore be e ective metrics of performance. A common challenge to getting people to use issues logs is that they are sometimes viewed as vehicles for blaming people. You have to encourage use by making clear how necessary they are, rewarding active usage, and punishing non-use. If, for example, something goes wrong and it’s not in the issues log, the relevant people should be in big trouble. But if something goes wrong and it’s there (and, ideally, properly diagnosed), the relevant people will probably be rewarded or praised. But there must be personal accountability. 2) Metrics: Detailed metrics measure individual, group, and system performance. Make sure these metrics aren’t being “gamed” so that they cease to convey a real picture. If your metrics are good enough, you can gain such a complete and accurate view of what your people are doing and how well they are doing it that you can nearly manage via the metrics. However, don’t even think of taking the use of metrics that far! Instead, use the metrics to ask questions and explore. Remember that any single metric can mislead. You need enough evidence to establish patterns. Metrics and 360 reviews reveal patterns that make it easier to achieve agreement on employees’ strengths and weaknesses. Of course, the people providing the information for metrics must deliver accurate assessments. There are various ways to facilitate this accuracy. A reluctance to be critical can be detected by looking at the average grade each grader gives; those giving much higher average grades might be the easy graders. Similarly helpful are “forced rankings,” in which people must rank coworker performance from best to worst. Forced rankings are essentially the same thing as “grading on a curve.” Metrics that allow for independent grading across departments and/or groups are especially valuable. 3) Surveys (of workers and of customers).... 145) The most common reason problems aren’t perceived is what I call the “frog in the boiling water” problem. Supposedly, if you throw a frog in a pot of boiling water it will immediately jump out. But if you put a frog in room-temperature water and gradually bring the water to a boil, the frog will stay in place and boil to death. There is a strong tendency to get used to and accept very bad things that would be shocking if seen with fresh eyes.... 146) In some cases, people accept unacceptable problems because they are perceived as being too diﬃcult to ﬁx. Yet ﬁxing unacceptable problems is actually a lot easier than not ﬁxing them, because not ﬁxing them will make you miserable. They will lead to chronic unacceptable results, stress, more work, and possibly get you red. So remember one of the rst principles of management: you either have to x problems or escalate them (if need be, over and over again) if you can't x them. There is no other, or easier, alternative. 146a) Problems that have good, planned solutions are completely diﬀerent from those that don’t. The spectrum of badness versus goodness with problems looks like this: a) They’re unidenti ed (worst); b) Identi ed but without a planned solution (better); c) Identi ed with a good, planned solution (good); and d) Solved (best). However, the worst situation for morale is the second case: identi ed but without a planned solution. So it’s really important to identify which of these categories the problem belongs to. 147) DIAGNOSE TO UNDERSTAND WHAT THE PROBLEMS ARE SYMPTOMATIC OF So...... 148) Recognize that all problems are just manifestations of their root causes, so diagnose to understand what the problems are symptomatic of. Don’t deal with your problems as one-o s. They are outcomes produced by your machine, which consists of design and people. If the design is excellent and the people are excellent, the outcomes will be excellent (though not perfect). So when you have problems, your diagnosis should look at the design and the people to determine what failed you and why.... 149) Understand that diagnosis is foundational both to progress and quality relationships. An honest and collaborative exploration of problems with the people around you will give you a better understanding of why these problems occur so that they can be xed. You will also get to know each other better, be yourself, and see whether the people around you are reasonable and/or enforce their reasonableness. Further, you will help your people grow and vice versa. So, this process is not only what good management is; it is also the basis for personal and organizational evolution and the way to establish deep and meaningful relationships. Because it starts and ends with how you approach mistakes, I hope that I have conveyed why I believe this attitude about and approach to dealing with mistakes is so important.... 150) Ask the following questions when diagnosing. These questions are intended to look at the problem (i.e., the outcome that was inconsistent with the goal) as a manifestation of your “machine.” It does this rst by examining how the responsible parties imagined that the machine would have worked, then examining how it did work, and then examining the inconsistencies. If you get adept at the process, it should take 10 to 20 minutes. As previously mentioned, it should be done constantly so that you have a large sample size and no one case is a big deal. 1) Ask the person who experienced the problem: What suboptimality did you experience? 2) Ask the manager of the area: Is there a clear responsible party for the machine as a whole who can describe the machine to you and answer your questions about how the machine performed compared with expectations? Who owns this responsibility? Do not mask personal responsibility—use speciﬁc names. 3) Ask the responsible party: What is the “mental map” of how it was supposed to work? A “mental map” is essentially the visualization of what should have happened. To be practical, “mental maps” (i.e., the designs that you would have expected would have worked well) should account for the fact that people are imperfect. They should lead to success anyway. 4) Ask the owner of the responsibility: What, if anything, broke in this situation? Were there problems with the design (i.e., who is supposed to do what) or with how the people in the design behaved? Compare the mental map of “what should have happened” to “what did happen” in order to identify the gap. If the machine steps were followed, ask, “Is the machine designed well?” If not, what’s wrong with the machine? 5) Ask the people involved why they handled the issue the way they did. What are the proximate causes of the problem (e.g., “Did not do XYZ”)? They will be described using verbs—for example, “Harry did XYZ.” What are the root causes? They will be descriptions. For example: inadequate training/experience, lack of vision, lack of ability, lack of judgment, etc. In other words, root cause is not an action or a reaction—it is a reason. Be willing to touch the nerve. 6) Ask the people involved: Is this broadly consistent with prior patterns (yes/no/ unsure)? What is the systematic solution? How should the people / machines / responsibilities evolve as a result of this issue? Conﬁrm that the short-term resolution of the issue has been addressed. Determine the steps to be taken for long-term solutions and who is responsible for those steps. Speciﬁcally: a. Are there responsibilities that need either assigning or greater clariﬁcation? b. Are there machine designs that need to be reworked? c. Are there people whose ﬁt for their roles needs to be evaluated?... 151) Remember that a root cause is not an action but a reason. It is described by using adjectives rather than verbs. Keep asking “why” to get at root causes, and don’t forget to examine problems with people. In fact, since most things are done or not done because someone decided to do them or not do them a certain way, most root causes can be traced to speci c people, especially “the responsible party.” When the problem is attributable to a person, you have to ask why the person made the mistake to get at the real root cause, and you need to be as accurate in diagnosing a fault in a person as you are in diagnosing a fault in a piece of equipment. For example, a root cause discovery process might proceed like this: - “The problem was due to bad programming.” - “Why was there bad programming?” - “Because Harry programmed it badly.” - “Why did Harry program it badly?” - “Because he wasn’t well trained and because he was in a rush.” - “Why wasn’t he well trained? Did his manager know that he wasn’t well trained and let him do the job anyway, or did he not know?” Ultimately it will come down to what the people or the design is like.... 152) Identify at which step failure occurred in the 5-Step Process. If a person is chronically failing it is due to either lack of training or lack of ability. Which was it? At which of the ve steps did the person fail? Di erent steps require di erent abilities. 1. Setting goals: This requires big-picture thinking, vision, and values that are consistent with those of our community. (It is helpful to ask whether the responsible party lost sight of the goals or whether he or she set goals that are inconsistent with Bridgewater’s.) 2. Perceiving problems: This requires perception, the ability to synthesize, and an intolerance of badness (i.e., some people see badness but aren’t su ciently bothered by it to push themselves to eliminate it). Of course, having perspective (typically gained via experience) helps at all steps. 3. Diagnosis: This requires logic, assertiveness, and open-mindedness. You must be willing to have open and/or di cult discussions to get at the truth. 4. Design: This requires creativity and practical visualization. 5. Doing the tasks: This requires determination and self-discipline. If you 1) identify at which of these steps the chronic failures are occurring and 2) see which, if any, of these abilities the person is short of, you will go a long way toward diagnosing the problem.... 153) Remember that a proper diagnosis requires a quality, collaborative, and honest discussion to get at the truth. Don’t just give your verdict without exploring the mistake, because there’s a reasonably high probability that you don’t know the answer. Do not be arrogant. You might have a theory about what happened, and that theory should be explored with relevant others. If you and others are open-minded, you will almost certainly have a quality analysis that will give everyone working theories to explore or you will reach conclusions that can be used for the design phase. And if you do this whenever problems recur, you and others involved will eventually uncover the root causes.... 154) Keep in mind that diagnoses should produce outcomes. Otherwise there’s no purpose in them. The outcome might not take the form of an agreement, but at a minimum it should take the form of theories about root causes (which should be written down so you have a collection of synthesized dots to use for identifying patterns) and clarity about what should be done in the future to protect against them, or to gather information to nd out.... 155) Don’t make too much out of one “dot”—synthesize a richer picture by squeezing lots of “dots” quickly and triangulating with others. A dot is a particular outcome. When you diagnose to understand the reason it occurred, you are “squeezing” the dot. Don’t try to squeeze too much out of a single dot—it can only tell you so much. Rather, try to collect and squeeze a bunch of dots in an 80/20 way, triangulating with the dots of others, so that you can synthesize a pointillist painting of what the person is like.... 156) Maintain an emerging synthesis by diagnosing continuously —You must be able to categorize, understand, and observe the evolution of the di erent parts of your machine/system through time, and synthesize this understanding into a picture of how your machine is working and how it should be modi ed to improve. But if you don’t look into the signi cant bad outcomes as they occur, you won’t really understand what they are symptomatic of, nor will you be able to understand how things are changing through time (e.g., if they are improving or worsening).... 157) To distinguish between a capacity issue and a capability issue, imagine how the person would perform at that particular function if they had ample capacity. Think back on how they performed in similar functions when they had ample capacity.... 158) The most common reasons managers fail to produce excellent results or escalate are: a. They are too removed. b. They have problems discerning quality di erences. c. They have lost sight of how bad things have become because they have gradually gotten used to their badness (the “frog in the boiling water problem”). d. They have such high pride in their work that they can’t bear to admit they are unable to solve their own problems. e. They fear adverse consequences from admitting failure.... 159) Avoid “Monday morning quarterbacking.” That is, evaluate the merits of a past decision based on what you know now versus what you could have reasonably known at the time of the decision. Do this by asking yourself, “What should a quality person have known and done in that situation?” Also, have a deep understanding of the person who made the decision (how do they think, what type of person are they, did they learn from the situation, etc).... 160) Identify the principles that were violated. Identify which of these principles apply to the case at hand, review them, and see if they would have helped. Think for yourself what principles are best for handling cases like this. This will help solve not only this problem but it will also help you solve other problems like it.... 161) Remember that if you have the same people doing the same things, you should expect the same results.... 162) Use the following “drilldown” technique to gain an 80/20 understanding of a department or sub-department that is having problems. A drilldown is the process by which someone who wants to do so can gain a deep enough understanding of the problems in an area as well as the root causes, so that they can then go on to design a plan to make the department or sub-department excellent. It is not a “diagnosis,” which is done for each problem. A manager doing ongoing diagnosis will naturally understand his areas well and won’t have to do a drilldown. Drilling down is a form of probing, though it is broader and deeper. Done well, it should get you almost all the information needed to turn a department around in about ve hours of e ort. A drilldown takes place in two distinct steps: 1) listing problems and 2) listing causes/diagnosing. It is followed by 3) designing a plan. If done well, getting informed via the rst two steps typically takes about four hours (give or take an hour), with the rst step of listing the problems typically taking one to two hours and the second step of diagnosing them typically taking two to four hours, if done e ciently. It’s very important that these steps are done separately and independently. That’s because going into two or three directions at the same time causes confusion and doesn’t allow adequate discussion of each of the possible causes and solutions. Having the people from the area under scrutiny actively participate in all three steps is critical. You need to hear their descriptions and allow them to argue with you when they think you are wrong. This way you are much more likely to come up with an accurate diagnosis and a good plan. After the drilldown, you will create the plan or design, which typically takes two to three hours. So the whole process, from asking the rst question to coming up with the detailed plan, typically takes about ve to eight hours spread over three or four meetings. Then there is step four—the executing, monitoring, and modifying of the plan—which typically takes six to 12 months. Here is more detail on each of the steps: Step 1—List the problems. Don’t confuse problems with possible solutions. Sometimes problems occur for rare or insigni cant reasons because nothing is perfect. Don’t pay much attention to those. But more often than not, they are symptomatic of something malfunctioning in your machine, so it pays to investigate what that is. For example, not having enough capacity is not a “problem”; it might cause problems, but it’s not a problem. Having people work so late that they might quit, getting out reports too late, etc., might be problems that are caused by a lack of capacity. But the lack of capacity itself is not a problem. To x problems, you need to start with the speci c problems and address them one by one and come up with very speci c solutions. That’s because there are lots of ways to solve problems. The problem of people working late at night might be solved by gaining capacity, or it might be solved by shifting work to another department, or by doing less, etc. To assume that lack of capacity is the problem could lead to inferior problem-solving. So unless you keep in mind the very speci c problems, you will not be e ective at solving them. In the process of solving problems, you will often see that several problems are due to the same cause (e.g., lack of capacity, a shortage of tech resources, bad management, etc.), but that is not the same thing as starting at the more general level (like saying that bad management or lack of capacity, etc., are problems), which is why I am saying you must start with very speci c problems before making generalizations. For example, when you have a “people problem,” be speci c. Specify which people you are having what problems with and avoid the tendency of saying things like, “People in operations aren’t...” Avoid the tendency not to name names for fear of o ending. Step 2—Identify root causes. Root causes are the deep-seated reasons behind the actions that caused the problems. It is important to distinguish between proximate causes, which are super cial reasons for what happened (e.g., “I missed the train because I didn’t check the train schedule”), and root causes (e.g., “I didn’t check the schedule because I am forgetful”). Typically a proximate cause is the action that led to the problem while a root cause is the fundamental reason that action occurred. So, when diagnosing, if you are describing what happened or didn’t happen to cause the problem, you are probably describing proximate causes. When you start describing the qualities that were behind these actions, you are probably getting at the root causes. To get at the root cause, keep asking why. For example, if the problem is that people are working late and the direct cause was that there wasn’t enough capacity, then ask why there wasn’t enough capacity. Then you will get closer to the root cause. If your machine is producing outcomes that you don’t want, either the design is awed or the parts/people that you dropped into the design are malfunctioning. Most, but not all, problems happen because 1) it isn’t clear who the “responsible party” is for making sure things go well or 2) the responsible party isn’t handling his or her responsibilities well (in other words, isn’t operating according to the principles to eliminate the problem). So rst ask, “Is it clear who the responsible party is?” If not, specify that. If it is clear, then ask, “Why isn’t he or she doing a good job?” There are two possible reasons for someone doing a poor job: insu cient training or insu cient ability. Though it is essential to connect problems to the responsible parties, this can be di cult if the responsible parties don’t acknowledge their mistakes and fail to diagnose why they made the mistakes. Still, clarity about responsibility and the problems’ root causes must be achieved because otherwise there is no hope for improvement. If the responsible parties do not explicitly take responsibility for ensuring that their areas operate smoothly, their areas will not operate smoothly. An important rst step toward achieving clarity is to remove the mentality of blame and credit, because it stands in the way of accurately understanding problems, and that’s a prerequisite for producing improvements. Also, it is important not to judge too quickly what the root causes are. Instead, you should observe the patterns of problems using the issues log as a tool and discuss with the responsible parties what the root causes might be each time a problem arises. You probably won’t initially be able to come to conclusions with a high degree of con dence, because there are many possible reasons for any one problem. But over time, the problems’ patterns and causes will become clear to everyone. As mentioned, there are two possible reasons why the responsible party handled something badly: 1) the responsible party didn’t encounter this problem enough times previously to learn from it and prevent it in the future (by using the principles) or 2) the responsible party is unsuited for that job. And there are also two possible reasons the person is not suited for that job: 1) not enough experience or training and 2) lack of values and/or abilities required to do the job well. So getting at the root causes is largely a matter of guring out: 1. Who is the responsible party for what went wrong? 2. Did that person encounter the problem enough times that he or she should have either learned how not to repeat it or elevated it to someone who could have helped learn how to solve it? The conclusions could be the following: 1) If the person did encounter the problem enough times to have resolved or elevated it, then the person is not suitable for the job; 2) if the person did not encounter the problems enough times to resolve or elevate it, what are the probable root causes? The most common root causes are: 1) the person is not suitable for the job in some way (doesn’t learn from mistakes, doesn’t have a high sense of responsibility, is lazy, etc.); 2) the design of the process is awed (e.g., the person is doing things in a way that can be improved); or 3) there is no possible solution. If it’s the rst root cause, the person should have their job changed; if it’s the second, you and the person need to properly diagnose the problem and come up with a di erent process that will work; and if it’s the third, you won’t know that until you have thoroughly explored whether the process can be remediated. That second alternative of trying to nd a better process takes time and patience (involving you and the person properly diagnosing the problem and nding a di erent approach that works). Normally, this is the point at which most companies and people fail. That is because people often take the identi cation of a “mistake” as the equivalent of an accusation that they are awed (dumb, lazy, etc.), so they become defensive. If instead they view the exercise as an investigation into how the process might be awed, it’s easier to make progress. So when criticizing, it’s sometimes helpful to convey explicitly the point of the exercise: mutually diagnosing the problem and exploring the pros and cons of alternative approaches. You both need to be mindful that doing this well typically takes time and patience. One of the purposes of the brainstorming session is to do this, ideally with an agreed diagnosis resulting from it. Step 3—Create a plan (brief notes): -Look at each root cause and ask yourself what should be done about it. -Creating a plan is like writing a movie script in that you visualize who will do what through time in order to achieve the goal. -Step away from the group to re ect and work on the plan, then bring it back to the group to discuss and modify. -When developing the plan, iterate through multiple possibilities and play them out in time to help determine the best choice. -Make sure to assign who is supposed to do what with rough target dates for achieving individual tasks of the plan. Once the plan design is complete, make sure the tasks, responsible parties, and timelines are reasonable and doable. -While everyone does not need to agree with the plan, it is important that the key people agree that it will work. Step 4—Implement the plan (brief notes): - Give each person a monthly to-do list to provide clarity and transparency around responsibilities and expectations for that month. Then plot the progress in open, monthly meetings with all the relevant parties. Explicitly assess how the plan is working and deal with problems that aren’t being resolved. -Make sure to hold responsible parties accountable for target dates and develop metrics around how they are meeting their commitments. -Regularly look at that list of assigned tasks to track progress and determine if any adjustments are needed. -Create transparency around the plan by posting it publicly and reviewing it regularly with the group. This helps people see the ways in which all of the problems are being addressed and reinforces accountability. Do not exclude any relevant people from the drilldown: besides losing the bene t of their ideas, you disenfranchise these people from the game plan and reduce their sense of ownership. Remember that people tell you things they want and tend not to be self- critical. It is your job as a manager to get at truth and excellence, not to make people happy. happy. For example, the correct path might be to re some people and replace them with better people, or to put people in jobs they might not want, etc. The brainstorming session must include a discussion of people’s weaknesses and failings to get at truth and excellence. Everyone’s objective must be to get at the best answer, not the answer that will make people happy. This is especially true for managers. In the long run, the best answers will be the ones that make the people we want to be at Bridgewater happiest. 163) PUT THINGS IN PERSPECTIVE So…... 164) Go back before going forward. Before moving forward, take the time to re ect on how the machine worked. By diagnosing what went right and what went wrong (especially what went wrong), you can see how the machine is operating and how it should be improved. People who are just focused on what they should do next are overly focused on the tasks at hand and not on how the machine is working; so they don’t make sustainable progress. Go back by “telling the story” to help put things in perspective. Sometimes people have problems putting current conditions into perspective or projecting into the future. Sometimes they disagree on cause-e ect relationships, or focus on details rather than addressing the big picture. Sometimes they forget who or what caused things to go well or poorly. By asking them to “tell the story” of how we got here, or by “telling the story” yourself, you put where you are in perspective. Doing this highlights important items that were done well or poorly in relation to their consequences, draws attention to the overarching goals, and helps achieve agreement. By telling the story from the past to the present, it will help you continue it into the future (i.e., design a plan). Making a good plan involves sketching out the important events through time and thinking through the speci cs in sequence so that when you are done, the nal story is vivid and easy to visualize. Then other people can understand the plan, comment on it, and eventually believe in it. It’s also required for specifying who should do what and when. 164a) Tool: Have all new employees listen to tapes of “the story” to bring them up to date. Listen to some of the associated tapes about Bridgewater’s story. Imagine how much better informed you would be than a person who just joined Bridgewater and hadn’t listened to these stories.... 165) Understand “above the line” and “below the line” thinking and how to navigate between the two. There are di erent levels and themes going on in any one conversation. It is important to know how to navigate them. If you imagine main points and subordinate points organized in outline form, an above-the-line discussion addresses the main points. That doesn’t mean you shouldn’t reference details, because some details might be necessary to the discussion. But reference details solely for the purpose of understanding major points rather than dissecting minor points. For example, suppose your major point is: “Sally can do that job well.” In an above-the-line conversation, the discussion of her qualities would target the question of Sally’s capacity to do her job. As soon as agreement was reached on whether she could perform competently, you would pass to the next major point—such as what qualities are required for that job. In contrast, a below- the-line discussion would focus on Sally’s qualities for their own sake, without relating them to whether she can do her job well. The discussion might cover qualities that are irrelevant to the job. While both levels of discussion touch on minor points, “above the line” discourse will always move coherently from one major point to the next in much the same way as you can read an outline in order to fully understand the whole concept and reach a conclusion. You go “below the line” to the minor points only to illustrate something important about the major points and progress in an orderly and accurate way to the conclusion. Your ability to do this is partially innate but can be improved with practice. 166) DESIGN YOUR MACHINE TO ACHIEVE YOUR GOALS So...... 167) Remember: You are designing a “machine” or system that will produce outcomes. This machine will consist of distinct parts (i.e., people and other resources as well as the way they interact with each other). 167a) A short-term goal probably won’t require you to build a machine. But for an ongoing mission, you will need a well-designed and e cient machine. 167b) Beware of paying too much attention to what is coming at you and not enough attention to what your responsibilities are or how your machine should work to achieve your goals. Constantly compare your machine’s outcomes to your goals in order to re ect on how well the machine is operating. Examine both the design and how the individual parts are functioning.... 168) Don’t act before thinking. Take the time to come up with a game plan. Take at least a few hours to think through your plan. Those hours will be virtually nothing in relation to the amount of time that will be spent doing, and they will make the doing radically more e ective.... 169) The organizational design you draw up should minimize problems and maximize capitalization on opportunities. Make the design an extension of your understanding of your problems and opportunities.... 170) Put yourself in the “position of pain” for a while so that you gain a richer understanding of what you’re designing for. Temporarily insert yourself into the ow to gain a real understanding of what you are dealing with (the process ow, the type of people needed, the potential problems, etc.) and to visualize a clear picture of what will work. You can accomplish this in a number of ways (reviewing work, doing work at di erent stages in the process, etc.).... 171) Recognize that design is an iterative process; between a bad “now” and a good “then” is a “working through it” period. That “working through it” period involves trying processes and people out, seeing what goes well or poorly, learning from the iterations, and moving toward having the right people in the ideal systematic design. Even with a good future design picture in mind, it will naturally take time, testing, mistakes, and learning to get to a good “then” state.... 172) Visualize alternative machines and their outcomes, and then choose. A good designer is able to visualize the machine and its outcomes accurately, though imperfectly. First visualize the parts and their interactions, and then nd the parts to t the design. Look at all the system’s pieces and their interactions. Imagine how goals 1, 2, and 3 can be achieved. Imagine how Harry, Larry, and Sally can operate in various ways with various tools and di erent incentives and penalties in place to achieve those goals. Then imagine how the system would work di erently if you replaced Harry with George, or if it was con gured in an entirely di erent way. Do this iteratively. Think through what the products and people and nances will look like month by month (or quarter by quarter) over the next year given one system; then change the system and visualize the outcomes again. At the end of this process, your plan should look like a realistic movie script, which describes the parties and their interactions through time. Remember that everything takes longer and costs more than you plan for. Recognize that some people are relatively better or worse at visualization. Accurately assess your own abilities and those of others so you can use the most capable people to create the visualization.... 173) Think about second- and third-order consequences as well as ﬁrst-order consequences. The outcome you get as a rst-order consequence might be desirable (or undesirable), while the second- or third-order consequences could be the opposite, so focusing solely on rst-order consequences, which people tend to do, could lead to bad decision-making. Though I might not like the rst-order consequences of a rainy day, I might love the second-order consequences. So if I were in a position to choose whether or not there should be rainy days, I would need to look at the second- and third- order consequences to make the right decision. For example, for every person you plan to hire, you will have to hire more to support them. I call this “The 1.6 E ect.”... 174) Most importantly, build the organization around goals rather than tasks. As an example of building the organization around goals rather than tasks, we have traditionally had a marketing department (goal: to market) that is separate from our client service department (goal: to service clients), even though they do similar things and there would be advantages to having them work together. But because marketing and servicing clients are two distinct goals, we have a separate department for each. If they were merged, the department head, salespeople, client advisors, analysts, and others would be giving and receiving con icting feedback. If asked why clients were receiving relatively poor attention, the answer might be: “We have incentives to raise sales.” Asked why they weren’t making sales, the merged department might explain that they need to take care of their clients. Keeping the two areas separate gives each department a clear focus and the appropriate resources to achieve its goals, makes the diagnosis of resource allocations more straightforward, and reduces “job slip.” Of course, when building departments around goals, your goals have to be the right size to warrant these resources. An organization might not be big enough to warrant having a few salesmen and its own analytical group. Bridgewater has successfully evolved from a one-cell organization, in which most people were involved in everything, to the current multi- cell organization because we retained our ability to e ciently focus as the organization grew. Also, I want to make clear that temporarily sharing or rotating resources is OK, and is not the same thing as a merging of responsibilities. I will discuss merging later in this document, as well the coordination required to maintain focus in large organizations. 174a) First come up with the best workﬂow design, sketch it out in an organizational chart, visualize how the parts interact, specify what qualities are required for each job, and, only after that is done, choose the right people to ﬁll the jobs (based on how their capabilities and desires match up with the requirements). 174b) Organize departments and sub-departments around the most logical groupings. Some groups naturally gravitate toward one another. Trying to impose your own structure without acknowledging these magnetic pulls is ine ective and likely will result in a bad outcome. 174c) Make departments as self-suﬃcient as possible so that they have control over the resources they need to achieve the goals. We do this because we don’t want to create a bureaucracy that forces departments to requisition resources from a pool that lacks the focus to do the job. People sometimes argue that we should have a technology department, but I am against that because building technology is a task, not a goal in and of itself. You build technology to perform valuable tasks. If we kept the tech resources outside the department, we would have people from various departments arguing about whose project is most important in order to garner resources, which isn’t good for e ciency. The tech people would be evaluated and managed by bureaucrats rather than the people they do the work for. 174d) The eﬃciency of an organization decreases and the bureaucracy of an organization increases in direct relation to the increase in the number of people and/or the complexity of the organization.... 175) Build your organization from the top down. An organization is the opposite of a building— the foundation is at the top. The head of the organization is responsible for designing the organization and for choosing people to ll its boxes. Therefore, make sure you hire managers before their direct reports. Managers can then help design the machine and choose people who complement the machine. 175a) Everyone must be overseen by a believable person who has high standards. Without this strong oversight, there is potential for inadequate quality control, inadequate training, and inadequate appreciation of excellent work. Do not “just trust” people to do their jobs well. 175b) The people at the top of each pyramid should have the skills and focus to manage their direct reports and a deep understanding of their jobs. Here’s an example of the confusion that can arise when that understanding is absent: It was proposed that the head of technology have the facilities group (the people who take care of facilities like the building, lunches, o ce supplies, etc.) report to him because both are, in a sense, “facilities” and because they have some things in common, such as the electrical supply. But the head of technology didn’t understand what the facilities people do. Having people who are responsible for the janitorial services and meals reporting to a technology manager is as inappropriate as having the technology people report to the person who is taking care of facilities. These functions, even if they’re considered “facilities” in the broadest sense, are very di erent, as are the respective skill sets. Similarly, at another time, we talked about combining folks who work on client agreements with those who do counterparty agreements under one manager. That would have been a mistake because the skills required to reach agreements with clients are very di erent from the ones required to reach agreements with counterparties. It was wrong to con ate both departments under the general heading of “agreements,” because each kind called for speci c knowledge and skills. 175c) The ratio of senior managers to junior managers and to the number of people who work two levels below should be limited, to preserve quality communication and mutual understanding. Generally, the ratio should not be more than 1:10, and preferably closer to 1:5. Of course, the appropriate ratio will vary depending on how many people your direct reports have reporting to them, the complexity of the jobs they’re doing, and the manager’s ability to handle several people or projects at once. 175d) The number of layers from top to bottom and the ratio of managers to their direct reports will limit the size of an eﬀective organization. 175e) The larger the organization, the more important are 1) information technology expertise in management and 2) cross- department communication (more on these later). 175f) Do not build the organization to ﬁt the people. Jobs are created based on the work that needs to be done, not what people want to do or what people are available. You can always search outside Bridgewater to nd the people who “click” best for a particular role.... 176) Have the clearest possible delineation of responsibilities and reporting lines. It’s required both within and between departments. Make sure reporting lines and designated responsibilities are clear. To avoid confusion, people should not report to two di erent departments. Dual reporting (reporting across department lines) causes confusion, complicates prioritization, diminishes focus on clear goals, and muddies the lines of supervision and accountability, especially when a person reports to two people in two di erent departments. When situations require dual reporting, managers need to be informed. Asking someone from another department to do a task without consulting with his or her manager is strictly prohibited (unless the request will take less than an hour or so). However, appointing co- heads of a department or a sub-department can work well if the managers are in synch and combine complementary and essential strengths to this area; dual reporting in that case can work ne if properly coordinated by the co- heads. 176a) Create an organizational chart to look like a pyramid, with straight lines down that don’t cross. A series of descending pyramids make up the whole pyramid, but the number of layers should be limited to minimize hierarchy.... 177) Constantly think about how to produce leverage. For example, to make training as easy to leverage as possible, document the most common questions and answers through audio, video, or written guidelines and then assign someone to regularly organize them into a manual. Technology can do most tasks, so think creatively about how to design tools that will provide leverage for you and the people who work for you. 177a) You should be able to delegate the details away. If you can’t, you either have problems with managing or training or you have the wrong people doing the job. The real sign of a master manager is that he doesn’t have to “do” practically anything. Of course, a great manager has to hire and oversee the people who do things; but a “supreme master” manager can even hire a person or two to do this and has achieved such leverage that things are e ortlessly running superbly. Of course, there is a continuum related to this. The main message I’m trying to convey is that managers should strive to hire, train, and oversee in a way in which others can superbly handle as much as possible on their own. Managers should view the need to get involved in the nitty-gritty themselves as a bad sign. 177b) It is far better to ﬁnd a few smart people and give them the best technology than to have a greater number of ordinary and less well- equipped people. First of all, great people and great technology are almost always a great value because their e ectiveness in enhancing the organization’s productivity can be enormous. Second, it is desirable to have smart people have the widest possible span of understanding and control because fragmented understanding and control create ine ciencies and undermine organizational cohesion. Usually it is the person’s capacity that limits the scope of his understanding and control. So the mix of really smart people operating with really great technology in a streamlined organization is optimal for organizational e ciency. 177c) Use “leveragers.” Leveragers are capable of doing a lot to get your concepts implemented. Conceptualizing and managing are most important and take only about 10% of the time needed for implementing; so if you have good leveragers, you can accomplish a lot more with relative ease.... 178) Understand the clover-leaf design. Find two or three responsible parties who have overlapping believabilities and responsibilities and who are willing to challenge and check each other. If you do this, and those people are willing to ght for what they believe is best by being open-minded and assertive at the same time, and if they escalate their disagreements and failures to you, this process will have a high probability of sorting issues that they can probably handle well from issues that you should examine and resolve with them.... 179) Don’t do work for people in another department or grab people from another department to do work for you unless you speak to the boss.... 180) Watch out for “department slip.” This happens when a support department, such as HR or Facilities, mistakes its responsibilities to provide support with a responsibility to determine how the thing they are supporting should be done. An example of this sort of mistake is if those in the Recruiting department think they should determine whom we should hire or if people in HR think they should determine what our employment policies should be. Another example would be if the Facilities group determined what facilities we should have. While support departments should know the goals of the people they’re supporting and provide feedback regarding possible choices, they are not the ones to determine the vision.... 181) Assign responsibilities based on workﬂow design and people’s abilities, not job titles. What people do should primarily be a function of the job they have, and it should be pretty obvious who should do what (if they’re suited for the job). For example, just because someone is responsible for “human resources,” “recruiting,” “legal,” “programming,” etc., doesn’t necessarily mean they are the appropriate person to do everything associated with those functions. For example, though “Human Resources” people help with hiring, ring, and providing bene ts, it would be a mistake to give them the responsibility of determining who gets hired and red and what bene ts are provided to employees. When assigning responsibilities, think about both the work ow design and a person’s abilities, not the job title.... 182) Watch out for consultant addiction. Beware of the chronic use of consultants to do work that should be done by employees.... 183) Tool: Maintain a procedures manual. This is the document in which you describe how all of the pieces of your machine work. There needs to be enough speci city so that operators of the di erent pieces of the machine can refer to the manual to help them do their job. The manual should be a living document that includes output from the issues log so that mistakes already identi ed and diagnosed aren’t repeated. It prevents forgetting previous learning and facilitates communication.... 184) Tool: Use checklists. When people are assigned tasks, it is generally desirable to have these captured on checklists so they can check o each item as it is done. If not, there is a risk that people will gradually not do the agreed tasks or there will be lack of clarity. Crossing items o a checklist will serve as a task reminder and con rmation of what has been done. 184a) Don’t confuse checklists with personal responsibility. People should be expected to do their job well, not just what is on their checklists. 184b) Remember that “systematic” doesn’t necessarily mean computerized. It might mean having people do speci ed tasks and indicate that they have done them with checklists. 184c) Use “double-do” rather than “double-check” to make sure mission-critical tasks are done correctly. When people double-check someone else’s work, there is a much lower rate of catching errors than when two parties independently do the work and the results are compared. Double- doing is having two di erent people doing the same task on the same job so that two independent answers are derived. By comparing them you will not only assure better answers but you will see the di erences in people’s performances and make much more rapid improvement. I use double-dos in critical areas such as nance, where large amounts of money are involved.... 185) Watch out for “job slip.” Job slip is when a job changes without being explicitly thought through and agreed to, generally because of changing circumstances or a temporary necessity. Job slip will generally cause bad job design. It often leads to the wrong people handling the wrong responsibilities and confusion over who is supposed to do what.... 186) Think clearly how things should go, and when they aren’t going that way, acknowledge it and investigate. First decide which issue to address rst: nding the reason the machine isn’t working well or executing the tasks required to get past the problem (in which case you need to come back to the reasons later). Either way, don’t pass the problem by without discussing the reasons. Otherwise, you will end up with job slip.... 187) Have good controls so that you are not exposed to the dishonesty of others and trust is never an issue. A higher percentage of the population than you might imagine will cheat if given an opportunity, and most people who are given the choice of being “fair” with you and taking more for themselves will choose taking more for themselves. Even a tiny amount of cheating is intolerable, so your happiness and success will depend on your controls. Security controls should be viewed as a necessary tool of our profession, not as a personal a ront to an individual’s integrity. Just as a bank teller doesn’t view a check on the money in his drawer as an indication that the bank thinks he is dishonest, everyone here should understand the need for our security controls. Explain this to your people so they see it in the proper context. Even the best controls will never be foolproof, and trustworthiness is a quality that should be appreciated. 187a) People doing auditing should report to people outside the department being audited, and auditing procedures should not be made known to those being audited. 187b) Remember: There is no sense in having laws unless you have policemen (auditors). 188) DO WHAT YOU SET OUT TO DO So…... 189) Push through! You can make great things happen, but you must MAKE great things happen. Times will come when the choice will be to plod along normally or to push through to achieve the goal. The choice should be obvious. TO MAKE DECISIONS EFFECTIVELY... 190) RECOGNIZE THE POWER OF KNOWING HOW TO DEAL WITH NOT KNOWING So...... 191) Recognize that your goal is to come up with the best answer, that the probability of your having it is small, and that even if you have it, you can’t be conﬁdent that you do have it unless you have other believable people test you.... 192) Understand that the ability to deal with not knowing is far more powerful than knowing. That is because there’s way more that we don’t know than what we could possibly ever know. 192a) Embrace the power of asking: “What don’t I know, and what should I do about it?” Generally you should nd believable people and ask their advice, remembering that you are looking to understand their reasoning rather than get their conclusions. 192b) Finding the path to success is at least as dependent on coming up with the right questions as coming up with answers. Successful people are great at asking the important questions and then nding the answers. When faced with a problem, they rst ask themselves if they know all the important questions about it; they are objective in assessing the probability that they have the answers; and they are good at open-mindedly seeking believable people to ask.... 193) Remember that your goal is to ﬁnd the best answer, not to give the best one you have. The answer doesn’t have to be in your head; you can look outside of you. In life the goal is for you to do the right thing, considering the probability that you might be wrong. So it is invaluable to know what you don’t know so that you can gure out a way to nd out and/ or to get help from others.... 194) While everyone has the right to have questions and theories, only believable people have the right to have opinions. If you can’t successfully ski down a di cult slope, you shouldn’t tell others how to do it, though you can ask questions about it and even express your views about possible ways if you make clear that you are unsure.... 195) Constantly worry about what you are missing. Even if you acknowledge you are a “dumb shit” and are following the principles and are designing around your weaknesses, understand that you still might be missing things. You will get better and be safer this way. 195a) Successful people ask for the criticism of others and consider its merit. 195b) Triangulate your view. Never make any important decisions without asking at least three believable people. Don’t ask them for their conclusions or just do what they tell you to do. Understand, visualize, and assess their reasoning to see if it makes sense to you. Ask them to probe your own reasoning. That’s critical to your learning as well as to your successful handling of your responsibilities. 196) MAKE ALL DECISIONS LOGICALLY, AS EXPECTED VALUE CALCULATIONS So...... 197) Considering both the probabilities and the payoﬀs of the consequences, make sure that the probability of the unacceptable (i.e., the risk of ruin) is nil. 197a) The cost of a bad decision is equal to or greater than the reward of a good decision, so knowing what you don’t know is at least as valuable as knowing. 197b) Recognize opportunities where there isn’t much to lose and a lot to gain, even if the probability of the gain happening is low. It is a reality that there are always multiple possibilities and nothing is certain. All decisions are therefore risk/reward bets. Know how to pursue fabulous risk/ reward ratios that have a huge upside and very little downside, albeit a small probability of happening. My life has been lled with these. 197c) Understand how valuable it is to raise the probability that your decision will be right by accurately assessing the probability of your being right. I often observe people giving opinions as soon as they have them, which seems at about the point that they think there’s more than a 50% chance of them being right. Often they don’t pay any attention to the value of raising the probability of being right (e.g., from 51% to 85%) by re ecting harder on whether the answer is right and doing the investigations and double-checking with others to make sure that the answer is right. Remember that, in an expected value sense, raising the probability of being right (e.g., from 51% to 85%) can be worth more than just going from probably wrong (e.g., 45%) to probably being right (e.g., 51%) because we are all playing probabilities. Think about the e ects of altering the probabilities of achieving must-dos: if you have a 51% probability of handling a “must-do” correctly, it means that only a bit more than half of your must-dos will be done appropriately, whereas an 85% probability of handling a decision well means that only 15% of the must-dos will be handled badly. 197d) Don’t bet too much on anything. Make 15 or more good, uncorrelated bets. 198) REMEMBER THE 80/20 RULE, AND KNOW WHAT THE KEY 20% IS So...... 199) Distinguish the important things from the unimportant things and deal with the important things ﬁrst. 199a) Don’t be a perfectionist, because perfectionists often spend too much time on little di erences at the margins at the expense of other big, important things. Be an e ective imperfectionist. Solutions that broadly work well (e.g., how people should contact each other in the event of crises) are generally better than highly specialized solutions (e.g., how each person should contact each other in the event of every conceivable crisis), especially in the early stages of a plan. There generally isn’t much gained by lots of detail relative to a good broad solution. Complicated procedures are tough to remember, and it takes a lot of time to make such detailed plans (so they might not even be ready when needed). 199b) Since 80% of the juice can be gotten with the ﬁrst 20% of the squeezing, there are relatively few (typically less than ﬁve) important things to consider in making a decision. For each of them, the marginal gains of studying them past a certain point are limited. 199c) Watch out for “detail anxiety,” i.e., worrying inappropriately about unimportant, small things. 199d) Don’t mistake small things for unimportant things, because some small things can be very important (e.g., hugging a loved one).... 200) Think about the appropriate time to make a decision in light of the marginal gains made by acquiring additional information versus the marginal costs of postponing the decision. There are some decisions that are best made after acquiring more information, and some that are best made sooner rather than later. The later a decision is made, the more informed it can be; however, making it later can also have adverse consequences (e.g., postponing progress). Understanding the trade-o between the marginal gains of acquiring the extra information against the marginal costs of postponing a decision is an important factor in the timing and preparation of decision-making.... 201) Make sure all the “must do’s” are above the bar before you do anything else. First, distinguish between your “must do’s” and your “like to do’s”. Don’t overlook any “must do’s,” and don’t mistakenly slip the “like to do’s” onto the list. Then, get all the “must do’s” above the bar. Then get all the “must do’s” excellent. If you have time, turn to the “like to do’s” and try to get them above the bar. Only if you have time (though you certainly will not if you are thinking broadly), turn toward making things perfect. Chances are, you won’t have to deal with the unimportant things, which is better than not having time to deal with the important things. I often hear people say, "Wouldn’t it be good to do this or that,” referring to nice-to-do’s rather than must-do’s that have to be above the bar. Chances are, they are being distracted from far more important things that need to be done well.... 202) Remember that the best choices are the ones with more pros than cons, not those that don’t have any cons. Watch out for people who tend to argue against something because they can ﬁnd something wrong with it without properly weighing all the pros against the cons. Such people tend to be poor decision-makers.... 203) Watch out for unproductively identifying possibilities without assigning them probabilities, because it screws up prioritization. You can recognize this with phrases like “It’s possible that...” then going on to say something that’s improbable and/or unimportant, rather than something like, “I think there’s a good chance that...” followed by something that’s important or probable. Almost anything is possible. All possibilities must be looked at in terms of their likelihoods and prioritized.... 204) Understand the concept and use the phrase “by and large.” Too often I hear discussions fail to progress when a statement is made and the person to whom it is made replies, “Not always,” leading to a discussion of the exceptions rather than the rule. For example, a statement like “The people in the XYZ Department are working too many hours” might lead to a response like “Not all of them are; Sally and Bill are working normal hours,” which could lead to a discussion of whether Sally and Bill are working too long, which derails the discussion. Because nothing is 100% true, conversations can get o track if they turn to whether exceptions exist, which is especially foolish if both parties agree that the statement is by and large true. To avoid this problem, the person making such statements might use the term “by and large,” like “By and large, the people in the XYZ Department are working too many hours.” People hearing that should consider whether it is a “by and large” statement and treat it accordingly. 204a) When you ask someone whether something is true and they tell you that “It’s not totally true,” it’s probably true enough. 205) SYNTHESIZE So…... 206) Understand and connect the dots. To do this well, you have to synthesize what is going on. Usually it takes diagnosing a few (e.g., ve or so) dots of the same type to get at the true root cause so that you can see how the machine should be modi ed to produce better outcomes. For example, one type of outcome involves someone, let’s call him Harry, handling a type of responsibility (entering an order). You will need at least a few experiences to learn about Harry doing this. It will pay for you to understand Harry and his handling of orders and have him understand you by looking objectively at the outcomes and by getting in synch, especially about the bad outcomes. The quality of your understanding of your machine and its constituent parts will depend on how well you diagnose and process the important outcomes. If you don’t do this continuously and you don’t synthesize well, you will fail. This isn’t easy. See how the dots connect through time. This requires collecting, analyzing, and sorting lots of di erent types, and it ain’t easy for most folks. Imagine a day in which eight outcomes occur. Some are good, some bad. Let’s represent this day as follows, with each type of event represented by a letter and the quality of the outcome represented by its height. In order to see the day this way, you must categorize outcomes by type and quality, which will require synthesizing a “by and large” assessment of each. If you didn’t examine the bad outcomes as they occurred, you couldn’t understand what they are symptomatic of. Keep in mind our example is a relatively simple one: only eight occurrences over one day. Now let’s look at what a month looks like. Confusing, eh? Some people are much better at this than others. In order to understand how your machine is working to achieve your goals, you have to perceive change over time, charting improvement vs. deterioration. The chart below plots just the type X dots, which you can see improving. As mentioned in the section on diagnosis, you must categorize, understand, and observe the evolution of the di erent parts of your machine through time, and synthesize this understanding into a picture of how your machine is working and how it should be improved. People who do this well are rare and essential. As with most abilities, synthesizing well is partially innate and partially learned through practice.... 207) Understand what an acceptable rate of improvement is, and that it is the level and not the rate of change that matters most. I often hear people say, “It’s getting better,” as though that is good enough when “it” is both below that bar and improving at an inadequate rate.

Document Details

Tags

Related

Summary

Full Transcript

Upgrade to continue