Jeff Bezos Philosophy: "Day 1"

On September 13, Jeff Bezos launched the “Day 1” philanthropic project . Let's dig what is behind this name.



What is the philosophy of Day 1 by Jeff Bezos?

Initially, this question appeared on 'Quora': a place to receive and share knowledge, allowing people to learn from others and better understand the world.

This is how Brian Roemmele, the founder and editor of Read Multiplex, replied:

“The outside world can push you into the state of“ Day 2 ”if you don’t become or cannot quickly take on powerful trends. If you are fighting with them, you are probably fighting with the future. Use them and catch a fair wind "
- Jeff Bezos

Day 1

A recent Securities and Exchange Commission found that Jeff Bezos's speech (EX-99.1) was interesting enough. In this very letter to Amazon 2016 shareholders, Jeff describes the philosophy that he calls “Day 1” and “Day 2”. This idea came to him at the beginning of Amazon. He worked in a building called “Day 1”, so named to remind you that the company should always be in “Day 1” mode. “Day 2” for Jeff is static or a complete change in the concept of “Day 1,” followed by irrelevance. And behind it - a painful, painful decline. Then death.

Inside the company, the idea became folklore. Amazon’s success was largely due to the Day 1 concept. She has become something of a legend.

Here is a summary of this concept:

Focus on the result, not the process

Jeff explained that as the company grows, it becomes easier to rely on the process and not on the result. In this case, the process becomes a “thing”. When this happens, companies sometimes stop looking at the result and focus on whether they followed the process correctly and not on whether the result was achieved.

Make decisions quickly

Jeff says that companies following the concept of Day 2 make really good decisions, but they are only doing it slowly. The fact is that start-ups make high-quality and fast decisions without problems, but this is difficult for large companies. “Amazon management intends to maintain a high decision-making speed. Speed ​​matters in business, and also makes things much more fun. ”

For this, the “Day 1” process has a system for employees: “don't agree, but do it”. The idea is that not everyone can come to the same solution, but even so, it is still possible to work together to achieve a common goal. He mentioned that he was not quite sure about one of the proposed series on Amazon Prime: partly because of the small interest, and also because of the business conditions put forward. He said: "I had a completely different opinion, but they wanted to risk it." So I instantly answered them: “I do not agree, but I give the green light and I hope that this series will be the most viewed of all we have done before.” "Imagine how much we would slow the decision-making cycle if the team really had to convince me, and not just gain my confidence."

Do not focus only on the company

The most eloquent understanding of the philosophy of “Day 1” is as follows: “The outside world can push you into the state of“ Day 2 ”if you do not become or cannot quickly accept powerful trends. If you are fighting with them, you are probably fighting with the future. Use them and catch a fair wind. ” These powerful trends are hard to miss, but in most cases, large companies are hard to accept. He notes that currently the whole world is involved in the most important trend - machine learning and artificial intelligence. For this reason, Echo / Alexa systems are among the fastest growing and best selling products in Amazon history. Jeff declares that “No matter how hard we try, we still hardly manage to constantly have Echo stocks in warehouses. This is a high quality problem, but a problem. We are working on it".

But that's not all. Amazon is experimenting with the delivery of drones and the grocery store Amazon Go, which uses machine vision to eliminate queues at the box office. Jeff says that “Machine learning manages our demand forecasting algorithms, product search rankings, product and transaction recommendations, merchandising, fraud detection, translation, and more; although less visible, most of the machine learning work will be just that - invisible, but improving key operations. ”

Market analysis

Jeff writes: “Here is another example: market research and customer surveys can be a model for buyers, and this is especially dangerous when you invent and develop products. 55% of beta testers said they like this feature. And this is 47% higher than the first survey. Such information is difficult to interpret, and it may inadvertently be misleading. ” This premise is to some extent influenced Steve Jobs at Apple. This certainly shows how Amazon creates the voice of the first revolution. I wrote about the thousandth army that Jeff created only for Alexa.

The manifesto of "Day 1"

The concept of "Day 1" forms the basis of the Amazon manifest. I believe that this is one of the best messages for shareholders that I have ever read (and over the decades I have read enough). Dynamic goals are not just words, but fully involved concepts used in the company on a daily basis. This is a rare look inside the mind that has been creating these concepts for twenty years.

Here is the Amazon Exhibit 99.1 statement for the Securities and Exchange Commission, which connects the original Exhibit 99, filed during the initial public offering in 1997:

Jeff, what does Day 2 look like ?

I was asked this question at our most recent meeting. I remind people that for several decades we have "Day 1". I worked in a building called “Day 1”, and when I moved to another building I decided to keep this name. I spent a lot of time thinking about it.

“Day 2” is static followed by irrelevance. And behind it - a painful, painful decline. Then death. That's why we always have "Day 1".

Do not hesitate - such a decline will occur as slowly as possible. A certain company can reap the benefits of Day 2 for decades, but in the end the result will be obvious.
The question is, how do we counteract Day 2? What are our tricks and tactics? How to maintain the viability of "Day 1" even in a large organization?

There is no simple answer. But there are many components, paths and traps. I don’t know the answer entirely, but I probably know the particles. Here is the starter pack needed to protect Day 1: an obsession with customers, a skeptical look at models, a huge desire to take external trends and high-speed decision making.

True customer obsession

There are many ways in which you can build a business. You can be focused on competitors, on a product, on technology, on a business model, and so on. But, in my opinion, obsession with customers is the best way to protect the viability of Day 1.

Why? There are many advantages in a customer-oriented approach, but the biggest thing is this: customers are always beautiful and surprisingly unhappy, even when they say they are happy and that the business is excellent. They want something more even when they themselves do not know about it, and your great desire to please clients will lead you to inventions. No one ever asked us to create Amazon Prime, but as it turned out, everyone wanted this service, and there are many such examples.

“Day 1” requires that you calmly experiment, take failures, plant seeds, protect seedlings, and double it all with customer satisfaction. A culture obsessed by customers creates excellent conditions for this.

Do not give in to ready-made models.

As the company begins to grow larger and more complex, there is a tendency to rely on proven models. It can manifest itself in different ways, but it is dangerous, cunning and in the style of "Day 2".

A common example is process as a model. A good process serves you so that you serve customers. But if you are not vigilant enough, the process can become an integral part. This is very common in large organizations. The process becomes a model for the result you want to achieve. You stop following the result and focus on the correctness of the process. You can often hear a young manager defending a bad result with the words: “Well, we just followed the instructions.” A more experienced leader uses this opportunity to research and improve the process. The process can not be by itself. It is always worth asking: do we own the process or does the process own us? In the company "Day 2" the second option is most possible.

Here is another example: market research and customer surveys can be a model for customers, and this is especially dangerous when you invent and develop products. “55% of beta testers said they like this feature. And this is 47% higher compared to the first poll. ” Such information is difficult to interpret, and it may inadvertently be misleading. ”

Good inventors and designers understand their customers at a deeper level. They take a lot of time and effort to develop this intuition. They study and understand a huge number of stories, and not just the average values ​​from questionnaires. They live by design.

I'm not against beta testing or polls. But you, the owners of the product or service, must understand the customer, have a vision and really love the offer. In this case, beta testing and research will help you find your blind spots. A truly incredible experience with customers begins with the heart, intuition, curiosity, play, courage and taste. Polls will not give you anything like that.

Embrace external trends

The outside world can push you into the “Day 2” state if you don’t become or cannot quickly take on powerful trends. If you are fighting with them, you are probably fighting with the future. Use them and catch a fair wind.

These trends are hard not to notice (they talk and write a lot about them), but it is difficult for big companies to accept them. Now we have an obvious trend - machine learning and artificial intelligence.

Over the past decades, computers have largely automated tasks that programmers can describe using clear rules and algorithms. Modern methods of machine learning allow us to work with tasks for which it is much more difficult to describe clear rules.

Amazon has been practicing machine learning for many years. Something to see: our autonomous drones for delivery of Prime Air; Amazon Go's grocery store, which uses machine vision to eliminate queues at the checkout; and Alexa is our cloud-based AI helper. (No matter how hard we try, we still hardly manage to constantly have Echo stocks in warehouses. This is a high quality problem, but a problem. We are working on it.)

But a lot of machine learning is not obvious. Machine learning controls our demand forecasting algorithms, product search rankings, product and transaction recommendations, merchandising, fraud detection, translation, and more; though less noticeable, but most of the work of machine learning will be just that - imperceptible, but improving key operations.

With AWS (Amazon Web Services) cloud platform, we reduce costs and other machine learning and AI barriers so that all organizations and companies can take advantage of advanced technologies.

Using pre-built versions of popular depth learning systems that work according to the P2-computing model (optimized for such a workload), customers are already developing powerful systems in various fields: from early detection of the disease to increased yields. We also made higher-level Amazon services more accessible in a convenient form. Amazon Lex (which is inside Alexa), Amazon Polly and Amazon Rekognition eliminate the hard work of recognizing natural language, generating speech and analyzing images. Access to them is provided through a simple API call, which does not require special knowledge in machine learning. Watch out for this area. Further more.

Fast decision making

Day 2 companies make really good decisions, but they do it slowly. To preserve the energy and dynamics of Day 1, you must be able to combine high speed and quality in solutions. This is easy for startups, unlike large companies. The Amazon team is committed to supporting high-speed decision making. Speed ​​matters in business, and also makes the atmosphere much more fun. We do not know all the answers, but here are some of our thoughts.

First , never use a universal decision-making process. Many solutions have two sides. For these solutions, you can use a lightweight process. What happens if you make a mistake? I spoke about this in my last year’s address.

Secondly, most decisions should be made with approximately 70% of the information you would like to have. If you wait up to 90%, most likely you will be slow. In addition, in any case, you should be able to quickly recognize and correct the wrong decisions. If you can easily correct the direction of the decision, then the errors will not be as costly as you think (but the delay is always expensive).

Thirdly, use the principle of "do not agree, but do." He will save a lot of time. If you have conviction in a certain direction, but there is no unanimous decision, it would be better to say: “Listen, I understand that we have not come to a unanimous decision, but maybe we will risk it? Do not agree and do? ". At this stage, no one knows the correct answer, so you will surely quickly hear a positive answer.

It does not work unilaterally. If you are a boss, you also need to adhere to this principle. I always disagree, but I do. We recently gave the green light to a show from Amazon Studios. I told my team that I am not sure that it will be interesting, that it will be easy to release and that the business conditions are not very profitable, and that we still have many offers. They had a completely different look, and they wanted to take a chance. I immediately answered them: “I do not agree, but I give the green light and I hope that this series will be the most viewed of all that we have done before.” Imagine how much we would slow the decision-making cycle if the team really had to convince me, and not just gain my trust.

Note that in this example I did not think: “Well, guys are completely wrong and do not catch the essence, and it is not worth my efforts.” This is the most real difference of opinion, a frank expression of my point of view, a chance for the team to weigh my opinion and go their way. Given that the team has already taken 11 Emmys, 6 Golden Globes and 3 Oscars, I'm glad they let me in the room!

Fourth , recognize the real inconsistency problems as early as possible and immediately raise them to a new level. Sometimes teams have different goals and completely different views. They are not consistent. No discussions and no meetings will solve this deep misunderstanding of each other. Without escalation, the default dispute resolution mechanism in this scenario will be exhausted. The final decision is made by one who has higher stamina.

Over the years at Amazon, I have seen many examples of true disagreement. When we invited third-party vendors who had to compete with us directly in our own products, this became especially clear. Many smart Amazon employees simply disagreed with the chosen direction. A big decision entailed a bunch of small ones, many of which had to be raised to a higher level of senior management.

“You completely exhausted me” is a terrible decision-making process. It is slow and energy depriving. It is much better to raise the issue to another level.

So, did you stop on the quality of decision making or thought about speed? Do global trends fill your sails with a fair wind? Are you a victim of models or models work for you? And, most importantly, do you make your customers happy? You can have the scale and capabilities of a large company, and the spirit and heart of a small one. The choice is yours.
Many thanks to each client for allowing you to provide services, our shareholders for support, and each employee for their hard work, ingenuity and passion.

As always, I enclose a copy of the original letter of 1997. We still have "Day 1".

Regards,
Jeff.

---