An attempt at system design in my family

Some infuriating things have happened recently, and everyone says that the root of the problem lies in the unreasonable design of the system. Deng Xiaoping, the chief architect of the reform and opening up, once said, “A good system can prevent bad people from running rampant, and a bad system can prevent good people from doing good deeds adequately, and even lead to the opposite.” However, institutional design is difficult. There was an example in our family recently. The reason for how to supervise weight loss is that Zhu Apple wants to lose weight, Li Flag said that he would not eat after 6 o’clock every night, and invited his brother and sister to supervise. So we began to discuss how to supervise and how to design a reward and punishment system. I propose that every time the mother breaks the rules and eats after six o’clock, reward the brother and sister with a game each. Both brother and sister agreed with this proposal, and Zhu Pingpin also felt that it was OK. But then we felt that there was a problem with this system: would the elder brother and younger sister seduce the mother to break the rules in order to play games? Otherwise, let’s turn it around and change it to “if mother doesn’t violate the rules, reward brother and sister for playing a game”. In this way, the brother and sister will work hard to stop the mother from breaking the rules. But then I thought again: losing weight and not eating after six o’clock is my mother’s hard work. Why can my brother and sister be rewarded for no reason? We also discussed several versions, including how to add Dad as a check and balance, reward not on a one-off but weekly basis, and more. But each version has some problems, not very good. The final result is: no supervision, no reward and punishment system, or Zhu Ping’s self-consciousness. A bad system can lead to the opposite. In the above example, our fears about the first proposal are justified: it is indeed possible that the brother and sister will seduce or allow the mother to violate the rules in order to play the game. Such things have happened in history, such as the rodent eradication campaign carried out by the French colonial government in Vietnam in 1902. Every time a rat is eliminated, a government reward can be obtained by handing in a rat tail as proof. As a result, farms dedicated to raising mice were created, specifically to cut their tails. A similar example is the campaign to eradicate cobras in India by the British colonial government. The “perverse incentive” entry on Wikipedia also lists two dozen other examples where the reward and punishment system was poorly designed, leading to negative results. A successful example of the design of the reward and punishment system in the corporate environment is the commission system of Ali B2B. A common sales incentive system is that the commission ratio of the current month depends on the sales of the current month. The higher the sales, the higher the commission ratio. Ali B2B’s reward system is: the salesperson’s performance in the current month determines his commission for the next month. When I saw this system for the first time, I was amazed, and I still have an unforgettable aftertaste every time I taste it. A successful example of institutional design at the social level is the Constitution of the United States. I used to think that the creation of the U.S. Constitution was a matter of luck, as if many people guessed the rise and fall of N stocks, and there were always a few “stock gods” who could guess the rise and fall of all stocks. Similarly, many countries try all kinds of possibilities, and there is always one or two countries that are lucky and hit a possibility that can work. I changed my mind after reading Miracle at Philadelphia and The Federalist Papers. Miracle at Philadelphia provides some historical background before addressing the Philadelphia Conference. The U.S. Constitution is iterative. Articles of Confederation was the first edition, and in less than ten years there were various problems, so by 1787 the sages felt the need to iterate. The product of that iteration is today’s U.S. Constitution. The Federalist Papers is a divine book, its Chinese name is “The Federalist Papers”. If I’m going on a three-month cruise and can only take one book, I’ll take it. After reading The Federalist Papers, it becomes clear that it is no accident that the U.S. Constitution has survived for so long, becoming the second oldest constitution in the world today (after San Marino). As can be seen from The Federalist Papers, these Founding Fathers have a profound knowledge of the successes and failures of more than two thousand years from the time of Pericles, and a deep insight into various political systems and human nature. . They made the U.S. Constitution, not out of their heads. Speaking of divine books, “Corruption and Anti-Corruption in Nazi Germany” is also a divine book. The Chinese version of this book was published by Yilin Publishing House in 2015. God is God in the Chinese version of this book is still selling. The core idea of ​​this book can be summed up in one sentence: because officials in Nazi Germany are appointed from the top down, officials at all levels are only responsible to the top, so corruption is inevitable and anti-corruption is useless. I think this principle applies to the corporate environment as well. Various problems in the company, such as “KPI orientation”, various short-sighted behaviors, inter-departmental barriers, how to avoid the “innovator’s dilemma”, etc., are rooted in the fact that positions at all levels in the company are top-level Appointed from the bottom (rather than voted or elected from the bottom up). Although some “360 EIA”, “peer feedback” and other practices can alleviate it to some extent, they cannot change the root of the problem. The beauty of simplicity It is said that Ali B2B’s reward system of “the performance of the current month determines the commission of the next month” was designed by Li Qi. Wei Zhe once said in an interview, “If you ask me Li Qi’s greatest contribution to the entire sales system of Alibaba, it is some simple and easy systems.” Simple regimes are often easier to implement, less negative, and longer lasting than complex regimes. For the problems in the current system, the common practice is to patch, when encountering a problem, apply a patch to add some conditions and cover more corner cases. But often the more patched the more vulnerabilities. For example, in our family’s example, adding Dad as a check and balance role, not on a one-time basis but on a weekly basis, these are all patching. But the effect is not good, the patch is not perfect, and the patch itself can cause new problems, and more patches need to be applied. We write programs and design software architectures, and we actually have a similar situation. Therefore, I am a big believer in “optimize for simplicity”. In the company, I don’t like to make a lot of “details” in various processes and make the system very complicated. We still have to do things with conscience. Just like in our family’s example, we did not have any supervision, reward and punishment system in the end, and we still depended on Zhu Ping’s self-consciousness. “I know it when I see it” I don’t like being asked all sorts of “fine print” questions. For example, we used to have a rule that every release was rollbackable. I think the words “rollbackable” are enough. But some people have to ask, what exactly counts as rollback? Is it necessary to have a rollback script to be able to rollback, or does it count as a document with the rollback steps? If the requirement must be able to “one-click rollback” to be considered rollback, then what is “one-click”? Does pressing the button twice count as one key? Does typing three commands count as “one key”? In addition, is there a requirement for the rollback time? My system can be rolled back, but the rollback is more complicated, and each rollback takes thirty or forty minutes. Does this count as “rollback”? For these detailed questions about rollbacks, I actually don’t want to answer them in my heart. In the final analysis, these rules are not exhaustive. Instead of constantly refining the rules, it’s better to keep things simple. Specific to each case, whether it is violated or not violated, in fact, everyone’s cognition will not be too far off. I said, next time there is a failure, is it “returnable” not done, I know it when I see it. “I know it when…

Read More →

This article is reproduced from: 88%B6%E5%BA%A6%E8%AE%BE%E8%AE%A1%E5%B0%9D%E8%AF%95/
This site is for inclusion only, and the copyright belongs to the original author.

Leave a Comment