We often think that collecting as much information as possible will help us make the best decisions. Sometimes this is true, but sometimes it will hinder our progress. Even at some point, this can be dangerous.
Many of the most successful people use simple, diverse decision-making methodologies to eliminate the need for deliberations in specific situations.
One possibility is to say no by default, just like Steve Jobs. Or, like Warren Buffett, reject any decision that requires a calculator or computer. Or follow the principle of first principles like Elon Musk. Amazon’s founder, Jeff Bezos, has different methods than the ones mentioned above. He will ask himself, is this a reversible or irreversible decision?
If a decision is reversible, we can make a quick decision without sufficient information; if a decision is irreversible, we better slow down the decision process to ensure that we consider the full information and be as thorough as possible. Understand the problems faced.
Bezos used this methodology to make the decision to build Amazon. He realized that if Amazon failed, he could return to his previous work, he would still learn a lot, and would not regret trying. This decision is reversible, so he took risks. This is very helpful to him, and he still plays a role when he makes a decision later.
Make decisions in uncertainty
Suppose you see a comment on the Internet and decide to try a new restaurant. Because you have never been there, you don’t know if the food will be delicious, or if the atmosphere will be very dull. However, you will use the incomplete information in the comments to make a decision, because you know that if you don’t like this restaurant, it is not a big deal.
In other cases, uncertainty is also a bit risky. You may decide to accept a specific job, but you don’t know what the company culture is like, and you don’t know how it feels about your work after the “honeymoon period” ends.
You can make reversible decisions quickly, without having to struggle to find complete information. If this decision fails, we can draw wisdom from experience with little cost. Often, it’s not worth spending time and effort collecting more information to find the perfect answer. Although your research may make your decision better, you may miss an opportunity.
However, it is important to note that reversible decision-making is not an excuse to act recklessly or not to understand the situation, but rather a belief that we should adapt our decision-making framework to the type of decision we are making. Reversible decisions do not need to be made like irreversible decisions.
The ability to make quick decisions is a competitive advantage. One of the main advantages of startups is that they can move with Velocity, while established companies typically move with Speed. The difference between the two is meaningful and often means success and failure.
Speed is measured in terms of distance over time. If we fly from New York to Los Angeles, take off from JFK airport, and circle in New York for 3 hours, our Speed is fast, but we have nothing. Speed doesn’t care if you are moving toward the goal. Velocity, on the other hand, measures the displacement that occurs over time. To get Velocity, you need to move toward your goal.
This decision-making methodology explains why startups have a better advantage than established companies when making quick decisions. This advantage is magnified by environmental factors such as speed of change. The faster the environment changes, the more advantages people who make quick decisions get, because they can learn faster.
Decision making provides us with data so that we can make better decisions about the future. The faster we loop through the OODA loop, the better. This framework is not applicable to certain situations at once; it is a methodology that needs to be an integral part of the decision-making toolkit.
The basic idea of the OODA cycle theory is that armed conflict can be seen as a cyclical process in which the opposing parties can compete with each other to complete the “observation-adjustment-decision-action”. Both sides began to observe, observe themselves, observe the environment and the enemy. Based on observations, obtain relevant external information, adjust the system in time according to perceived external threats, make response decisions, and take appropriate actions.
Through practice, we can better identify wrong decisions and make adjustments, rather than sticking to past choices because of sunk cost. Equally important, we can stop thinking of mistakes or small failures as catastrophic, and see them as pure information that will provide a reference for future decisions.
“It’s better to implement a good plan right now, than to implement a perfect plan next week.” – General George Barton
Bezos likened the decision to the door. Reversible decision making is a two-way open door. An irreversible decision is a door that allows only one direction to pass; if you go, you are trapped there. Most decisions are the former and can be reversed (even if we can never recover the time and resources invested). Passing a reversible door gives us the message that we know what is on the other side.
Bezos wrote in a shareholder letter in previous years:
Some decisions are irreversible or almost irreversible one-way doors that must be carefully and carefully negotiated, carefully, carefully, and slowly. If you go over and don’t like what you see on the other side, you can’t go back to where you were before. We can call this the first type of decision. But most decisions are not like this – they are mutable and reversible – they are two-way. If you make a second-class decision, you don’t have to endure the consequences for such a long time. You can reopen the door and go back. The second type of decision can and should be made quickly by highly discerning individuals or small teams. As organizations become larger and larger, in most decisions, including many second-type decisions, there seems to be a tendency to use the first-class decision-making process of the heavyweight. The end result of this is that the decision-making is slow, without considering risk aversion, and failing to conduct sufficient experiments, thus weakening innovation. We have to find a way to overcome this tendency.
Bezos gave an example of a one-hour delivery service to those who are willing to pay extra. The service began less than four months after the idea was put forward. In 111 days, the team “built a customer-facing application, determined the location of the city warehouse, identified 25,000 projects to be sold, reserved, recruited and staffed new projects for these projects, testing, iterative, design New internal use software (warehouse management system and driver-oriented) and launched in the holiday shopping season.”
As a further guide, Bezos believes that 70% certainty is the appropriate entry point for making decisions. This means that once we get the 70% of the information we need, we take action instead of waiting longer. Making a decision with 70% certainty and then making a route correction is much more effective than waiting for 90% certainty.
In “Blink: The Power of Thinking Without Thinking,” Malcolm Gladwell explains why decisions under uncertainty are so effective. We usually think that more information leads to better decisions – if doctors recommend additional tests, we tend to believe that they will lead to better results. Gladwell disagreed with this statement: “In fact, you need to know very little to find the fundamental characteristics of a complex phenomenon. All you need is evidence of ECG, blood pressure, lung fluid and unstable angina. This is A radical statement.”
In the medical field, as in many areas, more information does not necessarily ensure improved results. To illustrate this point, Gladwell gave an example. When a person comes to the hospital, the chest will be painful from time to time. His vital signs do not show any risk factors, but his lifestyle is indeed the same. He received heart surgery two years ago. If the doctor looks at all the information available, he feels that he needs to be hospitalized. But other factors besides vital signs are not important in the short term. In the long run, he is at high risk of heart disease. Gladwell writes that other factors play a very small role in determining the current situation of men, and without them they can make an accurate diagnosis. In fact, this extra information is useless. This is harmful. It confuses the problem. When doctors try to predict a heart attack, they take too much information into account.
We can all learn from Bezos’s method, which helped him build a huge company while maintaining the rhythm of entrepreneurship. Bezos used his methodology to counter the stagnation within many large organizations. It is important to be efficient, not to follow the rules of slow decision.
Once you understand that reversible decisions are actually reversible, you can begin to see them as opportunities to increase your learning speed. At the company level, allowing employees to make reversible decisions and learning from them helps you move forward at the pace of entrepreneurship. After all, if someone is moving with Speed, you will overtake them when you move with Velocity.
This is Bezos’s decision-making methodology, which directly or indirectly promotes the formation of two pizza principles in Amazon, because small teams are more reversible in making decisions.
Two pizza principles
In the early days of Amazon, Jeff Bezos developed a rule: each internal team should be small enough that two pizzas can solve the food problem. This is not about cutting food and beverage expenses. Just like almost everything Amazon does, it focuses on two goals: efficiency and scalability. The former is obvious. A smaller team spends less time managing and keeping employees up-to-date, and spends more time on what needs to be done. But for Amazon, what really matters is the latter.
The benefit of having many small teams is that they can work together and get the company’s public resources to achieve their larger goals.
In the words of Benedict Evans, a partner at venture capital firm Andreessen Horowitz, this turned the company into a machine that makes machines.
“You can add new product lines without adding new internal structures or direct reports, and you can add them to your logistics and e-commerce platforms without having to meet and go through a series of projects and processes.” Avon “You don’t need to (in theory!) fly to Seattle, arrange a meeting to get people to support your projects in Italy, or to convince anyone to add new business to their roadmap,” South pointed out.
Amazon is good at becoming an e-commerce company that sells goods, but it is best at creating new e-commerce companies that sell new products.
The company calls this method a “flywheel”: it’s large enough to stifle a typical multinational company and use it to provide a growing boost to the entire business. The heavier the flywheel, the faster it spins, the harder it is for others to stop it.
The birth and development of AWS (formerly known as Amazon Web Services) may be the best example of this approach. This is a division of Amazon that provides cloud computing services to internal and other companies—including those that compete with Amazon in other areas (for example, Netflix and Tesco use the platform, even though Amazon also sells streaming video and groceries).
Just like many things Amazon does, it all starts with commands issued by high-level executives. Bezos ordered that each team should start working together in a structured and systematic way. If the advertising team needs some data about footwear sales to decide how best to use their resources, they can’t analyze and request them by email; they need to go to the analysis dashboard and get the information in person. If the control panel does not exist, you will need to create it. This approach needs to be covered in all aspects.
From there, the next step is well worth it – let others use the same technology that Amazon provides internally.
Those inconspicuous beginnings gave birth to a beast. The business currently accounts for 10% of Amazon’s total revenue, so much so that financial regulations force the company to report it as a top-level department of its own: Amazon divides its company into “US and Canada” “international” and ” AWS”.
AWS is so large that it can be compared to Amazon’s branch offices in other regions; as big as Netflix, a company that accounts for one-third of Internet traffic in North America is just another customer.
By 2016, the company released a “snowmobile”, a truck used to move data. Companies that work with AWS provide a lot of information, and sometimes the Internet simply can’t. So now, if you want to upload a lot of data to Amazon’s cloud, the company will drive a truck to your office, fill it up, and then drive it back. If you need to upload 100 gigabytes of video, that is, about 5 million 4k movies with surround sound, you’ve found that there’s no faster way to travel on the highway at 120 kilometers per hour.
When AWS saw Amazon open its internal technology to external customers, another part of the company did the same thing on Amazon’s website.
The Amazon Marketplace was launched in 2000 to allow third-party sellers to sell their products on the site. Over the years, this feature has continued to expand, making Amazon a “department store” – the only place to buy an existing product on the Internet.
The Marketplace is better than the two pizza rules, allowing Amazon to expand into new areas without the need to hire any additional staff.
Amazon has a wide variety of goods sold, and its internal computer scientists face a problem. “E-commerce companies like Amazon handle billions of orders every year,” a research team at Amazon wrote. “However, these orders account for only a small portion of all reasonable orders.” Solution? Training artificial intelligence is purely to generate plausible fake orders and better guess how to market new products.
Amazon reported that its revenue from the Marketplace accounted for about 20% of the company’s total revenue. However, this indicator only calculates the fees paid by third-party sellers to the company, and underestimates the huge scale of the business. “The market now accounts for about half of Amazon’s total sales,” Evans of Andreessen Horowitz estimates. “In other words, Marketplace means Amazon’s share of e-commerce (but by the way, it doesn’t price itself) It reports twice the revenue share.”
As a result, Amazon is becoming less and less like a large retailer like Tesco or Wal-Mart, squatting on the edge of the city, stifling local commercial streets, and more like a shopping mall: independent retailers can exist and even maintain a clean life. But the premise is that they have a place in the shopping center, and they always remember that the real money earner is the landlord.
Since 2014, Amazon has added a third flywheel to its business: artificial intelligence. The company has always been in the industry’s leading position, most notably its neural network-based recommendation algorithm. However, until recently, this method was still untargeted, segmented, and almost world-class. (Think about what you bought on Amazon last time and recommend it to you in a few weeks: “Do you like duvets? Why not buy 10 more?”)
The situation changed when the company decided to build hardware that would become Echo. In Amazon’s classic model, it starts from the end, then works backwards from there, writing a “press release” for future concept products, and then trying to figure out what expertise to develop or acquire to achieve this goal. . Need a personal assistant? Acquisition of Cambridge-based True Knowledge. Need far-field speech recognition, let Echo hear the voice of people on the other side of the room? Start solving this problem now, because no one really solves this problem.
Institutionally, most members of the Alexa Artificial Intelligence team are still under AWS, using their infrastructure and providing another portion of digital services to third parties who want to build voice control in their devices. But the scale of artificial intelligence is unique. Of course, the value of the data is that the more people use Echo, the more voice samples they need to train, so the better Echo. In addition, machine learning technology is so basic and versatile that every advancement in Amazon will generate a rebound in the entire business, improve efficiency, open up new areas, and propose further research directions.
Amazon’s weakness
But nothing is eternal, and Amazon has its weaknesses. For example, the two pizza principles may be a good strategy for building an infinitely expanding company, but it does not bring a pleasant, stress-free work environment.
For a long time, Amazon has been criticizing the treatment of warehouse workers: Like many companies in the industry, huge valuations and high-tech ambitions go hand in hand with low-income, low-skilled jobs.
Amazon differs from companies such as Deliveroo, Apple, and Facebook in that there are almost as many complaints about highly skilled employees working at headquarters.
A report in the New York Times in 2015 stated that Amazon employees were crying at their desks and suffering from near-collapse pressure. The rapid flow of its employees is legendary, and insiders describe a scenario where someone leaves and others have to rewrite all their code so that people still there can understand it – but when the rewrite is complete, The rewritten staff also left and needed someone to start the process again.
But from the first day, Jeff Bezos has been at the top of the food chain, directly controlling the $740 billion (530 billion euros) business, and almost no other company’s boss can match it.
