Yu Jun, didi’s senior vice president of products: “Sometimes WHEN I get a lift from work, product managers from several major companies around me come to pick me up. Because I use my real name.”
“Sometimes WHEN I get a lift from work, product managers from several major companies around me (Didi, Baidu, Sina and netease) come to pick me up. Because I use my real name.”
That was Yu Jun, one of the earliest product managers of the Chinese Internet and a former vice president of product at Baidu, known as the “father of Tieba.” He joined Didi more than a year ago as an adviser. Now, he is Didi’s senior vice president of product.
As the idol of many Chinese Internet product managers, Yu Jun’s “product methodology” has been respected by the younger generation, and his ideas and thinking have influenced generations of product managers. His understanding of product manager and recruitment requirements are also of unique value.
In the past two years, with the increasingly mature development of mobile Internet, there have been fewer and fewer breakthrough innovations in product forms. In Internet companies, market, brand and operation are becoming more and more important. But for a truly successful or even great Internet company, product is still the most important thing.
PingWest interviewed Yu Jun to talk about what he has done in Didi over the past year, and what he thinks about products and product managers.
In this interview, Yu Jun introduced the product manager system of Didi. The system is similar to, but different from, other big companies such as Baidu and Alibaba. “The average product manager needs to stay rooted in one area,” he says. “The talented one needs to accumulate experience in different areas.”
As for Didi’s products, Yu Jun said that “site carpooling is the most important thing in the first half of this year”. According to the needs of users, his work this year is mainly to continuously improve the efficiency of the product, while the social direction of Didi, which was widely rumored before, is still in the schedule.
Yu jun used to read little leisure books, but now he has a stack of books on economics in his office, including Principles of Economics, Economic Development Theory and Statistical Psychology. This is a change for him and a requirement for Didi’s job. According to Yu Jun, Didi’s product design and business model are neither completely “free market” nor third-party platform. Nor is it a one-size-fits-all “planned economy” with corporate pricing. To this end, he wants to make up a little more common sense of economics.
Yu Jun, senior vice president of Didi Chuxing products
The following is the Q&A portion of the interview:
Q: You have previously classified product managers into three categories, ABC, according to their talent. How do you develop different product managers?
A: Category C product managers certainly still have the opportunity to change themselves. Stand in his personal point of view, can look for low iteration, cycle long product slowly accumulate. Some industries may iterate once every three or five years, like the automobile industry, and gradually accumulate, they will always have more advantages than those who have not accumulated. These industries may be step-by-step, try your best, and your value will always come out, so there will always be opportunities on an individual level.
But for the whole industry, the promotion of c-type product managers will be slow. From A company perspective, Type A product managers are hard to find, and Type B managers are not hard to find, but type C is rarely used for fast-growing products.
When I first came to Didi, the ratio of ABC product managers was 1:7:2. Now it’s about 2:7:1.
In the past year, there may have been more than 100 product managers recruited, but there were also 40 to 50 product managers who left. Among them, there was a certain amount of turnover, and I was quite satisfied with a dozen of them. But good product managers are still hard to find, which is why I put so much effort into finding just over one a month on average.
I don’t really get in the business of product managers. I let them learn from each other. What they’re interested in works best. Because we have an internal running water system, if you don’t want to do it, or prefer something else, you can switch between them.
My point is, if product managers are less talented, you put down roots in one area and you accumulate.
But a good product manager shouldn’t do that. He should move from one area to another and gain experience in different areas, whether it’s in different businesses or in different roles within the same business. I’m happy to have him in any business. As long as he is excellent, if he works here for a while, you suggest him to change to another business, and he may also grow up. Moreover, this business will definitely be created and improved after being filtered by a better person.
Q: Is empathy a gift for product managers? Can it be cultivated?
A: Empathy is as much A gift as running, drawing, or writing, no doubt about it. But because the use of empathy in products is not to be number one, like running writing, or to win a world championship or a Nobel Prize in literature, he just has to get to a point where he can use it in his work. So for most people, as long as the hard work on the line, is not the level of talent.
Efforts are mainly improved from three words: wish, read and calendar.
Li, is personally experienced. Users of a product are miserable, and if you’ve used it yourself, it’s the most solid experience. So whatever product you make, use it if you can. Of course some products may be difficult to use on your own, but they are the most effective. Use them.
To read is to watch others use it. Whether it’s text, voice, pictures, video, how they express their satisfaction or dissatisfaction with a product, watching how they use it, it’s all that information. You can also do it from communication, from user feedback, from all kinds of other channels, whether it’s we media or media.
Wish, you are really willing to understand others, stand in their perspective. Some people are very cold, he may be very smart, understand everything, but he may not be willing to stand in the perspective of others to think, he thinks that I and you are not a peer, not the same level. It’s natural to be willing to put yourself in someone else’s shoes.
Basic values and ideas, the most core, or user-centered, around the user to do products.
Q: How do you tell if someone has talent as a product manager?
A: I once saw A good answer on Zhihu. I couldn’t remember the details, so I recruited him. I used to emphasize to take people, to see the logic of his answer to this question, anyway, can not make mistakes, you have to be rigorous enough.
Let me give you a very specific example. Because all the core members of Didi have to drive every month. I once went out to drive with another colleague. Every time he saw a passenger, he would ask him whether he was not satisfied with the carpool waiting at the station. One of them, he said one thing, I think this person is suitable to be a product manager.
Later, I thought about why I had this reaction at that time. In fact, there are two aspects: one is very logical; One, obviously he answer the question, he thought: why do you ask this question, what is it you want, and then he also think before he answered, not just what you answer this question, he’ll stand at god’s point of view, more would think about what other people think, will be in a big environment to this idea, the description of a jump out of the question. So in one sentence, you can see that he thinks differently.
With this way of thinking, as a product manager, basic domain knowledge can be learned anyway, and proficiency can be exercised. Therefore, THERE are two problems in Zhihu. One is that he has rigorous and impeccable logic, and he is sure that he has thought a lot about himself critically. And then what you say, you can feel if he’s putting himself in other people’s shoes, that empathy.
If there are, that must be a good seedling, call over to chat, really good, then let him pick the business, he himself is a product manager, but just graduated.
Sometimes I say the potential of this person is high or low, depending on the denominator. To be able to express empathy and logic to a certain degree, but to do it in five years is obviously not the same as doing it in one year. So some people are just junior interns, and I think it’s pretty impressive that they can answer like this, and maybe the other person answers better than him, but they’ve been doing this for five years, and they know that the younger person is better. Different denominators, same numerator, definitely pick a lower denominator.
Q: How does Didi build a product manager system?
A: About three directions.
1) Internal training and growth.
Large companies often have various internal training from low-level to high-level, encouraging them to summarize writing methodology, communicate with the outside, and review various internal cases. These routine systems are gradually established.
2) Our rank system is interesting.
We are currently at D5 for new recruits, but it is usually D6 if you are not eliminated after a year. What does that mean? Is the freshman first year anyway need not compare rank bar, and then the second year also need not compare. Basically by the third or fourth year, you have to challenge yourself and see if you can get to D7. So for the first three years you don’t have to think about what rank I am, what am I going to do for that rank, what program I’m going to take or what I’m going to study, you don’t have to think about that.
I think this system can save a lot of time. The lesson of some companies is that too many people put too much thought into getting promoted. So let’s design it so that you don’t have to think about a promotion for the first three or four years. Even if you want to do that three or four years from now, you don’t have to look line by line at what I need to prepare, what projects I need to do, what skills I need to develop to get promoted next time.
That doesn’t work. Although Didi has a line by line rule, similar to some companies, our review mechanism is essentially a public review panel. For example, if you want to apply for D7, then gather six or seven D8s to come over for review. After you finish speaking, they will also write down details.
But at the end of the day, it’s all about the jury: do I think he’s D7 or D6, and what percentile is he in D7? For example, I listened to his review, and I thought he was a D7 20, which means he was better than 20% of the people in D7, or if he was better than 80% of the people in D7, or maybe he was only D6, but he was better in D6, 90, 60.
And then we’re going to put together the results, and we’re going to see what fraction of the D’s he is. Basically, our product committee doesn’t have much influence on the results of the public review panel, which is basically a referendum.
Besides, because our team is not too big, there are 300 PMS in the whole company, so we have a little understanding and intersection of each other’s work. In this way, you really don’t need to spend any time in reviewing and promotion, and you still have to do your work well, and most of us can feel it.
So I think we’re saving people time with this rank system. You just do your job and you don’t have to think too much. The so-called with Baidu Ali to mark is not completely to mark, just the money can mark, just at the beginning roughly by this, but later because of more than a year, according to their circumstances also have a few changes.
3) Product concept, product gene.
I hope that helping Didi has threefold value. One is the concept of user value. In fact, the first value of Didi is to create user value, which can not guarantee your success. However, this is a basic point for all successful companies. Without this point, even if they succeed, they will not grow long. So I think one is the increasing value of the user. This is not something I brought to Didi, but I hope it will continue to strengthen.
Secondly, the concept of ownership for product managers should be advocated. You’re the owner of the product, and you own it. From a global perspective, when to compromise and when to make trade-offs. Rather than saying I only represent the users, I only represent my business, or I only represent my current level, which is wrong. You are the master of the business, how do you think about standing in the long term and balancing all aspects. This may not be an idea that every company pushes.
The third is the concept of advocating talent. I keep telling key PMS that what you achieve in the next two or three years will depend on the quality of the people under you, the people you hire, and how fast they grow, not how good you are. Emphasize that by now, everyone should be able to feel the change, will be particularly concerned about whether to recruit A talent, always thinking about when to hire more A talent. It’s a concept that values talent.
In addition to the leaders and core backbone PMS, some other PMS may wonder if a strong person will compete with me. What I would say is that the capacity of each of us is not only determined by your own capacity, but by the capacity of the people around you. If the people around you are competent, your personal productivity will increase, and if everyone else is type A, even if you are type B, your personal productivity will increase. Research and development, operations, and the ability of others in your organization will also affect you.
In fact, when the proportion of A-type talents gradually increases, you will find that the product culture and product atmosphere will be significantly improved. If you have only one Type A on your team, the atmosphere may be affected by others. But when you get to three A’s, you see A difference. If you have four or five of them, everyone else will be excited about the product.
I don’t just want to hire type A, I think TYPE B, as long as they are passionate about the product, I also like them very much. You will find that their product results are likely to be better than those of someone with A potential but low product enthusiasm who is A B+. With enthusiasm and without enthusiasm, this is very important. In fact, it doesn’t matter whether there is A or B. As long as there is product enthusiasm, product enthusiasm can be stimulated by culture, and after stimulated, everyone will grow faster. Basically, I will interview any PM above D7, and I can tell in a few words whether he is a professional manager, who regards this as an ordinary job or who really likes products.
Q: How to set the general direction of Didi’s products?
A: Everything starts from user value. Which points do users not like, and which points do users want.
There’s a lot of channels and a lot of feedback. For our own use, including product manager engineers, other senior leaders, external partners, wechat circle of friends, we can collect a lot. As long as you screen, spread out, there is always a lot of need words, and then judge, one is where the pain point, one is where the want point. And then we analyze the percentage of each point, how many people it affects, how much it affects those people, how much it affects them, and then we look at the ones that are higher. It also depends on the input of technical resources, which is relatively easy to solve. What are the long-term projects, in a line like this, list them.
This is a normal step, of course, but for a product manager, always think about where the most critical, profitable, and transformative points are. Because a tech company can’t sustain growth through regular optimization, it has to figure out what really causes qualitative change, which pain points or which new elements to introduce.
Give feedback, too, because you’ll be inspired by your feedback, and maybe a particular detail is a key point that you wouldn’t have seen without doing so much. Of course, it’s also possible that you’re doing a lot of things and there are no key points, so normal optimizations have to be done. But always keep a part of your eye on what’s really going to cause qualitative change and what’s the next major growth point that’s going to cause qualitative change.
At present, Didi is mainly about carpooling, but it is a long-term qualitative change. It does not mean that I will change obviously from today to tomorrow. I know it will definitely change qualitatively, but it may take six months and a year to slowly optimize, and market penetration and acceptance will also take time.
Q: Does Didi collect social feedback from passengers?
A: Of course there is passenger networking. With such A large user base, someone will come up with it.
In terms of carpooling at present, there is indeed a significant proportion of social needs. Like drivers picking passengers. Some drivers are commuting anyway, conveniently brought, subsidies point gas money, not how to pick; Some drivers, such as those in Zhongguancun and West Erqi, want to pick employees from technology companies and exchange ideas.
For myself, I sometimes use hitchhiker after work and the product managers of the surrounding companies immediately pick me up because I use my real name. It’s also social.
The social convenience of hitchhiking is certainly different from that of the express or something, some of it, but not too much. Because there are other user needs to be taken care of. So socializing is done, but it’s not a big focus. There is less need for social communication on the express side. Maybe when Didi first appeared, some people drove their cars to socialize deliberately. Now sometimes I can meet some old men who feel free to come out and chat, but that is the minority.
There is another social direction. We do carpooling, and then we will consider doing targeted intra-enterprise carpooling. Carpooling is now a combination for anyone. While most efficient, some people are less willing to carpool. But, for example, when we work in the same company, whether it’s Tencent, PingWest or Didi, we tend to commute to and from the same place. Especially for these big companies, people tend to live in Huilongguan. If we are from the same company, people are willing to carpool, so the social probability will definitely rise. It’s not on the list yet, but if you do it in this direction, there will definitely be social interaction.
Naturally, passengers and drivers have a choice, and hitch is two drivers who choose each other, socializing only when they match each other. Only one side wants to socialize. Every guy wants a beautiful woman. So how can there be so many beautiful women? The beautiful woman doesn’t want to.
So social is something like that, there are some, but not the mainstream, there are some big pushes for products.
Q: What is the focus of Didi’s product in the next period of time?
A: It’s really A question of supply and demand balance.
Because our business model is so unique that other big companies don’t have this kind of problem.
It is a business destined for less supply than demand. You said that if you build a search engine, build a feed (dynamic information stream) news, even if it is to do social networking or games, or build an e-commerce platform, you will never worry about the lack of supply. Maybe a few seasonal goods and fruits will be in short supply, and the rest is as long as more and more people use it. Whether it’s social games, the more content the better, the more goods the better, their company’s focus is to find demand to find demand to find demand, supply is organized, not difficult.
But in our taxi-hailing model, the needs of users are unlimited, and the people’s needs to improve their living standards. Everyone wants to take a taxi, as soon as the income increases, more demand will be released, so the demand is increasing. But demand is certainly growing faster than the supply of vehicles and roads. So over time, supply must not keep up with demand. That’s one thing, the big trend.
And then specifically, there’s the question of time and space. The morning rush hour must be in short supply, because suddenly there are so many more people wanting a taxi; Every rainy day, every hot day, every cold day, demand goes up again, maybe 30, 50 percent. But supply, the driver will not suddenly come out, at most is more than three or five, efforts to dispatch will also come out. But demand can increase 30, 50 percent at a time.
That’s the time, and then there’s the place. At the same time, here empty empty dead, there busy busy dead. Is there a lot of people in the third ring road? But the driver has no money, he can’t live in the city center, all live outside the fifth ring road or beyond. In the morning, before the drivers come in, there are a lot of people trying to hail a cab in the city center, and the demand must be far greater than the supply. Drivers don’t want to come, I drive 20 or 30 kilometers to pick you up, supply is always less than demand.
In some cities, there are stages where supply and demand balance is a little bit, but this is phased. In the long term, Didi’s model is different from other models, and it is very special.
We’re going to have to spend a lot of time trying to solve this problem. Because you have less supply than demand, so many people, different social environment, they don’t understand. There are probably a lot of people who can’t get a taxi every day, and they will definitely scold you. While there are some people who understand this model, there are always others who don’t understand and aren’t satisfied.
It’s a lot of pressure. We’re very cautious. To solve this balance, supply should be improved, including the efficiency of supply, the efficiency of order delivery, and the increase of carpooling is also a kind of efficiency, anyway, all kinds of ways to improve efficiency.
But the efficiency is gradually improved, and so many people need to take a taxi on the demand side, what needs are met? Regulars, for example, are regular cabs, membership; Or are people just standing in line; Or through a certain dynamic price adjustment, price leverage; Or through some combination, some people can be preferential treatment, such as the old and the sick.
There’s a lot of pressure on Didi. Should dynamic price adjustment be controlled? If you have control, you have to find other ways to achieve demand and supply, one is matching, and the other is actually picking demand. There are always a lot of people can’t get a car every day, so which people can’t get a car? How to reasonably meet the needs of some people first, how to achieve both fair and reasonable, and economic efficiency?
I think I can do this for many years, and the more mature the business, the more energy I have to spend on it. Big data, ethics and economic research, we have found many economists and psychologists to study and explore. Therefore, the complexity of our business is very interesting. Many overseas Chinese professors are very interested in this business model and are willing to participate. Because this is a big challenge for them. There’s never been anything this interesting.
Specifically, carpooling is an obvious direction at present. Because there are many aspects of certainty, and then there is the exploration of pricing. Another example is queuing. At present, Zhongguancun, International Trade and other places are piloting it. In the past, when there are many people coming off work in the evening, they still wait for the bus, which is uncertain. But standing in line now, telling you how many people are in front of you and where you are in line, will soon be rolled out nationwide. But the whole line is not ok, some people may be in a hurry, you have to give him an exit, in what way to identify, this is enough exploration.
Q: What passengers care most about is price. How does Didi make dynamic price adjustment?
A: Our business model is non-standard. That’s part of the business model. Only a fraction of what you really offer to passengers is perceived as standard. For example, the App experience, after-sales customer service, the efficiency of sending orders.
But a lot of time, in the middle of the car, including the driver to pick up, phone communication, driving stability and service attitude, experience is determined by these. Every driver is different, and the same driver is not in a good mood when he makes every order. This is particularly easy to conflict, each single is different, there will always be a probability of conflict, how do you go to judge fair.
Understanding fairness involves an infinite number of positions, so why do we have economists and psychologists in the business, because it’s interesting and complicated. The fairness of the rules must be gradually improved to keep the business stable. Some disputed machines and labor are not clear judgment, is the platform to bear this part of the cost, such as empty compensation.
As for dynamic markup, we have established a cap mechanism, but the price lever can identify some passengers who are really anxious. A passenger complained that he could not get a taxi when he was really in a hurry. He said, “Don’t give me so many choices, I only have one choice, as long as I can get a taxi, I will add 100 yuan.
How to meet the needs of the passengers most. This can be done for many years, with the support of all aspects, no matter technical products, or economics, psychology, or big data.
We had a big economist come in and ask whether Didi’s model is a platform or an enterprise. Platforms, like some e-commerce businesses, do not set prices. Anyway, countless sellers and buyers negotiate a price in full competition, but enterprises set their own prices.
After all the math, we have a unique model. Oneself price efficiency is high, but once oneself price, affirmation both sides offend. Passengers always hope that the price is as low as possible, no matter what price you set, he will think you set too high, is a profiteer; The driver will always want the price as high as possible, no matter what price you set, you must be low, favoring passengers. So the tradeoffs here are important.
Maybe there will be a combination model, some of which will be free pricing, but the mainstream will probably still have to come up with a negotiated price in the middle.
This is not a platform where competition can be fully free, where countless drivers and passengers have enough time to choose between each other. It is a temporary market in a short time, in a small geographical area. You send out an order, 100 people and 30 cars within three miles, and you need a deal in minutes. After a few minutes, the bus leaves, and the passengers give up or find another mode of transportation. So deals have to be made within minutes, and with every minute of delay, drivers lose money and passengers lose too.
It is not like most commodities, buy a day early or buy a day late, in fact, there is no big difference, the goods will always be there, the buyer is not so anxious. But taxis are something that people expect to make a deal in minutes.
It is also possible to provide a platform for both parties to negotiate fully. That’s what hitchhiking is all about. Both parties can see each other when choosing. It is to choose a person now, still can choose price later, each party offers price freely, but this efficiency can be very low. There is a small percentage of people who are very Internet savvy who feel very good about it and can choose the lowest price or the best car for the same price.
For more than 90 percent of users, a quick taxi ride is the number one request. You can only recommend this time, this place, the same price by combining the calculation, considering a little supply and demand, recommend to him.
We’re also exploring the possibility of giving you a free market under certain conditions. Also some scenarios, give a suggested price, allowing for appropriate adjustments. But it has to be initiated by the passenger, because the passenger pays, and the driver causes problems.
None of this has come to a conclusion yet, just in this direction, and we continue to explore. In most cases, you can only give one price to improve efficiency. Early business is to grab single mode, order out drivers grab. Later, IT was found that this efficiency was too low, so it was still efficient to send orders. Although sending orders certainly caused a small part of harm, the efficiency was high in most cases, which was beneficial to both drivers and passengers.
Q: Can Didi, which aims to improve efficiency, build emotional connections with users like some other Internet products?
A: It must be, because I understand user value as money, time, body and emotion. Especially the emotion is more complex, more potential space. We used to be really bad at this, but Uber did a pretty good job.
Now the foundation is still strengthening the construction, you want to do emotional connection, but you can not be divorced from the user, is not really the user when the same thing. Whether in terms of product efficiency, customer service, or product functions and concepts, it is natural to make a solid foundation of this value, and then to do emotional connection. If the fundamentals of your product are flawed, it doesn’t matter if you bring in a branding guru.
At present, we are indeed weak in the aspect of emotional connection. I cannot answer this question well and I am not good at it. But if my colleagues in the company have good ideas in the future, I think I can understand them. What I can do is, in terms of product features, make users feel more satisfied, more personalized.
For example, in addition to one to five stars, different products have different evaluation methods. There are some experiments going on, more human, more lively ones, not just to get data efficiently. In addition, we are also considering, for example, a passenger says I don’t like this driver, he can choose to take a taxi in the future to shield.
Because of professional characteristics, some drivers often drive in a range because they live near, in fact, the probability of repeated receipt of a passenger is not small, that passengers can be shielded if they do not like it. Some drivers are also unhappy with the behavior of a large number of passengers, so they could theoretically block passengers. We will explore this and finally speak with data or market opinion.
If you give both parties the right to block their dissatisfaction, I think that’s definitely a benefit.
You can also do more personalization, you like the driver to call or not to call, there are really two kinds of people, and taxi prefer air conditioning or not air conditioning. Passengers who don’t like to talk to the driver or don’t want music in the car are already in the demand pool.
If you do more of these things, it will be much easier for you to make emotional connections between products and users. In the first half of the year, we also made a baby car, with a baby seat, a small amount. But passengers with young children do feel different when it comes to getting out. There’s also a wheelchair bus, which is smaller, but it’s a great direction. I think there’s a lot more to explore.
20 yuan limited time grab tickets: the last 34 concessions!
HAY for 3,000 people a day! The first tech carnival explosion
Oct 22, Beijing Jiuxianqiao 751 East District story
Hang out with mystery giants, trendy brands, and cool gamers
Open 3000 square meters of infinite wonderland
