Keith Rabois: Project Management

Stanford course CS183B: How to start a startup . Started in 2012 under the leadership of Peter Thiel. In the fall of 2014, a new series of lectures by leading entrepreneurs and experts of Y Combinator took place:



Keith Rabois : I’m going to talk about how to conduct company operations. I looked through some of the previous lectures and conclude that you have already managed to hire truly productive people and create a product that is in demand among a certain part of the user audience.

Also, I think you managed to raise capital, and now you want to create a company. I do not think that creating a company is much more complicated than creating a product. The main reason for this is that people's behavior is often irrational. We all know that.



Someone from your parents, relatives, brother or sister, the teacher just does not think about what he is doing. Creating a company as a whole is similar to how you gather all these people, whose behavior is not entirely logical, in the same building and spend at least 12 hours a day with them. This is pretty tricky.

There are several methods for solving this problem, and some cope with this successfully, while others do not. However, this is the essence of company management.



The main task in creating a company is to build a kind of “business engine”. First, you draw his scheme, design, develop the “design”, and it looks convincing, clear and beautiful.

But when you start to make it, it turns out to be terrible, its details are kept almost with adhesive tape, and it seems that it will fall apart soon. It takes a lot of effort to ensure that he properly performed his function, so people have to work 80-100 hours a week.



This seems to be a feat only because by this time the parts of your engine have not yet been “polished”. Over time, you will want to create a “car” that will work perfectly, and you will not have to be afraid every minute that it will fall apart.

And, as we usually joke, if Martians take over eBay, then it will be 6 months before the world realizes this. Over time, you will want to achieve just such a level of business automation.

As Warren Buffet says, “you need to build a business so that even idiots can manage your company, because sooner or later they will manage it.”

That’s what is really needed - to create a “machine” that “idiots” can drive. If you are a company leader, what is your actual work?



There is only one book in which this question has been revealed. It is quite old and was written by Andy Grove in 1982. It gained popularity and sold successfully.

Andy Grove’s definition of your role in the company is as follows: maximize the organization’s productive capacity.

The state of your company is your responsibility, as the CEO, you are responsible for everything, including the activities of the vice president of your company and the success of interaction with other organizations. If you are vice president of technology, you are responsible for the effectiveness of the teams and the sales team.

This is how it is necessary to evaluate the work of people, and it is important to first of all think about the performance of the company, and not about the costs.

The old approach is characterized by neglect of the level of productivity: in this case, the emphasis is only on tracking the progress of the company. And this may seem like a pretty attractive activity. People are attracted by the prospect of managing a huge organization and allegedly incurring “responsibility” for a high level of productivity.

However, I hope that today you will be able to learn how to order cocktails, teach your secretary to answer phone calls correctly, and for $ 10 per hour act as a hospitable hostess for your employees. Well, let's look at this topic.

In the beginning, when you first create a company, absolutely everything looks messy. And actually it should be so. If everything works too coherently and predictably, this may mean that you are not developing fast enough and are making insufficiently original decisions.



Therefore, new problems should arise every day, and your main occupation should be a kind of “medical sorting of the wounded” - determining the severity of a particular problem. Someone may be bothered by something like a “cold”, and because of this, you should not particularly worry, and certainly you should not devote all your time to solving this problem.

At the same time, some problems may look like a “cold”, but if they are diagnosed incorrectly, the result can be fatal. My goal is to teach you to determine if a particular problem is a cold or a deadly disease.



One of the most important things I've learned while working at Square is the concept of "editing." Over the 14 years of personnel management, I realized that this metaphor best describes the work of the administrator. This attitude to work is quite natural, and on this basis, you and your employees can understand for themselves whether they are “editors” or “writers”.

When someone asks you to do something, thinking in this way, you ask yourself if you are a “writer” or an “editor”. This metaphor conveys the essence of your work very well, and now we will discuss specific tasks when editing.

1. Simplification

Most likely, from school you remember the situation when you handed over a sheet with the text to the teacher for verification: the first thing that the testers do is take out a red pen and begin to cross out incorrect words or sentences. Thus, the main task of the editor is to simplify the text as much as possible.

And your job as an administrator is the same: discarding unnecessary and simplifying the work for the entire team of workers. The more you simplify your workflow, the more efficient it will be.

It is difficult for people to understand and perform a large and complex system of tasks, so you need to simplify it to 1-3 operations that they could constantly perform without thinking, even visiting friends, even in the middle of the night.

Do not think that a complicated system is normal. Many people will tell you that it is too difficult to accomplish, that it is too difficult, and you will answer them: “Well, I know that others prefer to simplify tasks, but this is not for me, we are engaged in complex business.”

This is an erroneous approach. You can change the world with just 140 letters.



You are able to create the most successful companies in history based on very simple concepts.



You will be able to sell products using less than 50 letters.



I see no reason why you can’t create a company in the same way. Therefore, you need to force yourself to simplify every task, product, marketing strategy - all that you do. Just pick up the red pen and begin to cross out the excess.

2. Clarification

The second task of editors is to ask clarifying questions. When you submit a text to someone for review, what are its actions? The reviewer usually finds any ambiguities in the text and asks if you really meant what you wrote and asks for an explanation. This is your job.

Consider this situation: you are at a meeting where people will wait for your decisions. Your real job in this situation is to get answers to a whole host of questions. These can be simple questions, for example, affecting the number of working days in a week or questions about what is the company's advantage over competitors.

Investors use a similar approach, and many investors have a billion questions that you will have to answer endlessly. Therefore, you should highlight what is most important for the company, and focus only on this. This approach will allow us to conduct a substantive dialogue with employees and make quick decisions. It will also provide an opportunity not to be distracted from your daily work - the development of the company.

Nonetheless, the most general questions will be asked, since the employees will not have any additional and unnecessary information. Learning to ask the right questions is difficult, it requires a lot of practice. But when you learn how to do it properly, according to Andy Grove’s calculations, you can increase productivity by 30-50%.

3. Distribution of resources

"Editors" think over the activities of the company - this is their constant task. For example, they can take other “editors” in the Middle East dealing with the affairs of the company in this region and send them to work in Silicon Valley, because at the moment the company is more interested in this.

Thus, the distribution of resources can be carried out "from top to bottom", and in this case, the administrator directs the resources to a specific place. In the next month, quarter (or year), the administrator says: "The Middle East is not very interesting anymore, and we are not doing this anymore, let's find another area of ​​activity." Or resources can be distributed “from the bottom up,” as in the case when journalists themselves are looking for stories to publish.

The people you work with must take the initiative themselves. For example, a reporter who works with Google will prepare an interesting article on the basis of the information he heard on the air, and then give it to the editor for approval. In this case, the editor did not ask the reporter to collect information about Google. I support this approach to work.

Your goal is to use less and less “red ink” every day. Therefore, one way to assess how successfully you manage to convey to your colleagues what is important and what is not is to analyze how much “red ink” you spend every day. If, by the end of a hard day, “red ink” is spilled everywhere - it's not so bad. However, if in the next month you correct even more, and after three - even more than that, then this is a bad sign. Therefore, you can evaluate the effectiveness of your work based on how much “red ink” you spend.

4. Working with principles

Another important task of the editor, which is not so easy to accomplish for a large team of people: the company must act in the same style. If you read articles on The Economist, then you won’t think that they were written by different people. You can open any article on The Economist, and its text will look as if it was written by the author of all other publications.

Ideally, the information on your site, PR information, text on the packaging of your products, etc. should look like they were made by the same person. This is incredibly hard to do. Initially, you will want to prepare all the information yourself, and in the first stages this is normal for the founder of the company. However, over time you will change your mind. You will need to train employees to analyze the characteristics of the company’s style when creating a product.

For example, one page can be very different from another, and you begin to think about the possible reasons for this. Perhaps the text is improperly written? Perhaps some of their company managers do not fully understand the style in which information about the company should be drawn up.

Over time, you will need to solve this problem, but for now, you should start by designing everything so that each page is composed in a single style. This is quite difficult to implement, and in almost every company there are elements that are knocked out of the general concept.

Apple is known for performing poorly, even during the time of Steve Jobs. If you ask any employee of this company about the design of the page where the ads for the recruitment are placed: do they look like created by Apple, then the answer in all cases will be “No”. Thus, it is impossible to complete this task 100%. However, you should try to get as close to this level.

5. Delegation

The next thing that causes difficulties is the assignment of work to other employees. Again, this problem can easily be explained using the “editing” metaphor: most of the work is done by “writers”, not “editors”. And in fact, in your company, you should not do most of the work.

You relieve yourself of the obligation to carry out the work, entrusting it to other people. The problem in this case is that you have a great responsibility. There are no excuses for the CEO, the founder of the company. He cannot say that this or that department or employee made a mistake, and thereby relieve himself of responsibility.

As a founder, you are responsible for every little thing, especially when mistakes are made. And the question arises, how to simultaneously entrust the work to others and not to give up responsibility? This is a very difficult task. You can go to two dangerous extremes: either entrust almost all the work to employees and relieve yourself of responsibility, or control every minor process. I want to share with you a couple of tricks to solve this problem.



The first trick, taken from Andy Grove’s book, is called “employee experience assessment”. This is a beautiful definition of a simple question for an employee: "did he have to do this work before?" And the more often he had to do the same job in the past, the more trust he should be given.

And vice versa - the less employees are familiar with the work that you give them, the more you need to instruct them and constantly monitor them. This is the most basic approach, and it should be kept in service.

An interesting point and at the same time very important is that the CEO should not apply only one management style - each employee requires a personal approach. Therefore, the work of one person needs to be controlled to a greater extent, if he does not have the proper experience, while another person can be assigned more work to do independently, if he already knows a lot.

Everything is in order if you ask questions regarding the performance of work to employees, and half of them ask for more control over their actions, and the other half is ready to take more responsibility. This is a normal situation and there is no problem. At the beginning of my activity, I did not understand this at all.

I was baffled by the situation when people conducted surveys among employees at my request and returned with similar conflicting results. Later, I finally realized that maybe I had organized the work correctly. Then I began to explain to other people that this is a normal situation for the company.



A more delicate question, to which I managed to find an answer, touches on the problem of decision making. I borrowed this table from Peter Thiel , and it helped me a lot in my life. At least with the help of her, he taught me at least something.

Here you choose the level of confidence in your decision in the range between "the highest level" and "the lowest." Sometimes you are sure of the fallacy of your decision, and sometimes you really wouldn’t act in one way or another, but you don’t know if this decision is right or not.

And there is also a range of severity of consequences. There are situations when the wrong decision can lead to disastrous consequences for the company. And vice versa - there are situations when the result of an error does not greatly affect the success of the company.

And here I believe that if the consequences of making the wrong decision in a certain situation are insignificant, and at the same time you are not sure of the correctness of your solution to the problem, then it is definitely worth delegating this work to an employee. At the same time, he should be allowed to work completely independently, allowing him to make mistakes and gain experience.

On the other hand, if the consequences of a wrong decision are significant, and you are completely confident in the correctness of your judgment, then you should not allow your colleague to make a mistake. You are solely responsible for this error, so if the problem is very serious, you cannot allow this to happen.

And the best solution in this situation will be a dialogue and explanation to a colleague of your thoughts. It is always easier to avoid any explanation, and in our world explaining something to people is a very difficult job, but it's worth a try.

One very, very talented employee worked with me on LinkedIn, who was very annoyed if I did not agree with his opinion on a particular issue. Therefore, I had to spend a lot of time explaining to him why I made this or that decision.

His universal answer, if I could not completely convince him of anything, was the phrase: "OK, you are the boss." For me it sounded like I was abusing my powers. Each time he said this, I realized that I was walking on thin ice and could run into big trouble if I did it too often. You need to keep track of how often you do this.