Technologists | Lu Pengjun, co-CTO of Vipshop: Technicians should not rush to transform management, get closer to business as soon as possible

Google Period 2006~2014
worth 30 million dollars

In 2006, Lu Pengjun graduated from Fudan University with a master's degree in computer science and then joined Google. Like most programmers, although he has a dream to change the world, at the stage when he just walked out of campus, the path to practice his dream was not clear. In such a state of life, instead of locking in the career track prematurely, it is better to join an enterprising and innovative technology manufacturer. There are five reasons:

First, there are so many things to learn, and the eyes are all on the shoulders of giants. Various free Tech Talks can satisfy any soul who is hungry for knowledge;

Secondly, All In business growth, and despise the underlying innovation of commercial companies, often hope that employees play the role of screws, perform their own duties and pull carts; technology companies are more inclined to cultivate all-round elites, "Google requires full stacks by default. "Engineers" are notorious, "The algorithm alone is not good, you have to be able to do engineering implementation, sometimes you have to do all the testing, monitoring, pressure testing, and going online," Lu Pengjun said.

Third, as the great wheel moves forward for a period of time, the direction of life can be established more accurately in the most cutting-edge information ocean;

Fourth, get to know the top elites in the industry. Not only do three people have my teacher, but also multiple friends and multiple paths;

Fifth, on the way to rapid iteration, large companies that encourage innovation have a great tolerance for mistakes.

The lesson that Lu Pengjun impressed the most was that he caused a loss of 30 million US dollars for the company.

At that time, he was only a T5 engineer of Google, but due to his outstanding performance, he had a lot of interaction with Google's global vice president. While he was placed high hopes, he also received huge permissions: many engineers have restricted access to the HIP core code, high Controlled user logs and bare write permissions for Google databases.

Marvel’s classic line “the stronger the ability, the greater the responsibility” is not to deceive me. Lu Pengjun made a direct change to the database at a time, causing Google’s advertising system to malfunction, and the custom blacklist for customers to place ads failed, resulting in a lot of cost consumption On the channels that customers don't want to put on, the total is as high as 30 million US dollars.

Google’s inclusiveness kept Lu Pengjun from being criticized. From the lessons learned from this experience, Lu Pengjun formed a unique concept and methodology when facing problems:

First, you can make mistakes. No one is perfect, but you need to go through the roots to find the most fundamental trigger point. After incurring losses, Lu Pengjun had to face deep torture in the case study, "Why write directly, why not go through the process?" "When the process is lazy, what are you thinking?" "Why? Don't you find someone to review it for you?"

Second, establish a process and let the mechanism ensure the correctness of things.

Third, keep reviewing the process to ensure that the problem is solved effectively, while not delaying rapid iteration, but to ensure that the process is efficient.

In the process of rapid iteration, exposing the problem is not necessarily a bad thing for the overall system. Although it has been severely tortured by the replay, this is the only so-called punishment that Lu Pengjun has received.

Regrettably, even though Google’s feast is so delicious, it has always been unable to get around a huge problem: the lack of business side in the Chinese market. As he grows older, Lu Pengjun, whose technical ability continues to rise, is becoming more and more eager to join the actual combat. Despite being repeatedly retained by Google’s senior executives, Lu Pengjun followed a number of senior friends to join Baidu in 2014, "it’s a bit late to think about it now."

Baidu period 2014~2018
Transformation management slightly earlier

With the recommendation of his senior friends, Lu Pengjun transformed into a director just one year after joining Baidu and took over the renowned Fengchao team in the Internet advertising industry.

Among the many growth hurdles for technical people, the step of transformation management is the key. Before taking over Fengchao, Lu Pengjun had almost no management experience. Leading a team of 130 people was obviously a difficult challenge. However, Lu Pengjun relied on his in-depth understanding of the business to successfully complete the handover. One year later, the number of personnel under his command has exceeded 1,000, leading Fengchao's products, technology, and operations.

Talking about the experience of management transformation, Lu Pengjun gave an unexpected suggestion: Technical people need not rush to transformation management.

Technologist: This is a rare suggestion. Can you share your specific thoughts?

Lu Pengjun: Writing code can think about the problem more thoroughly, which is very important in the tackling of some core problems. Management in the Internet age is no longer purely personnel management. More often, it is necessary to control the direction and ensure the success of the business, which is inseparable from strong judgment. After many people transfer to management, they will spend a lot of energy to coordinate resources, find ways to balance the interests of all parties, and continue to empower others. In this process, if you don't handle it well, people will stay away from the front line, and your judgment will be greatly affected over time. If you don't know what's going on on the front line and you want to direct it, it's particularly unreliable.

Technologist: Can you describe the ideal working status of technical management positions in your mind?

Lu Pengjun: I personally hope to be a "technical engineer with management capabilities" who has both technical decision-making power and the ability to process core codes; even if the management position is transformed, a mechanism must be established to ensure that I can clearly understand what is happening on the front line , Just listening to the report is definitely not enough. In some reports, the vocabulary used is very high-level, which sounds plausible, and the methods given are also very high-level, but they can’t be implemented at all. So after listening to the report, you must think positively, dig deep into the problematic areas, and continue to learn to enrich yourself.

promote business innovation

Lu Pengjun regrets the "transformation management premature", but when it comes to "whether technical people should build business thinking as soon as possible", he gave a strong affirmative reply, "I think I left Google a bit late, just because I feel close to the business. Some night".

Perhaps because of the urgency of the latecomers, Lu Pengjun showed a 120-point passion for "how to forcefully promote business growth", which is also a trait shared by many ultra-timeline growers: clear goals, not limited to means, think If it is helpful to customers, we will try our best to push customers to try, rather than stop passively collecting and satisfying needs.

Few people know that OCPC, which changed Baidu's advertising model, originally came from Lu Pengjun's desire for the honor of "Baidu's highest award".

In 2017, Lu Pengjun felt that his grasp of winning Baidu’s highest award this year was unstable, so he held a brainstorm meeting to discuss countermeasures. A T10 engineer on the team put forward a bold hypothesis, "Paying by conversion rate can be better than paying by click. Directly create value, so it should also be able to create more substantial financial returns." Lu Pengjun's eyes lit up and he immediately decided to push forward with all his strength and go online directly.

"This matter is actually a bit unreasonable, and the opposition inside the team is also very loud. Some people even said,'It is evil to change the price without greeting the customer.' A T9 engineer left his job because of this. But what I see is , The general direction of conversion billing is definitely more effective. Perhaps the current implementation method may not be right. But if everything is done properly, communication is complete, and the system is complete, the verification cycle is too long. At that time, I had to do my best. Discuss, do."

The product team can’t clearly define conversions? The operation team visited a large number of customers to collect feedback.
Advertisers will not bury points The technical team came to the door to help customers land.
At that time, there was a third-party software called Business Communication who had a lot of communication leads, and Lu Pengjun went to talk about the acquisition by himself.
……
Lu Pengjun helped this new payment model to verify and run through in a short period of time with extreme propulsion, and now the average daily turnover of this model has exceeded 100 million yuan.

Gathering Era 2018~2020
authorization is not laissez-faire, ideological consensus is the magic weapon

In 2018, Lu Pengjun, who had a very good performance at Baidu, decided to accept new challenges and join the Huanju era. "I have been advertising for 12 years and want to change the track. In 2016, I found that entertainment query accounted for 30% of the search and continued. Growth, my intuition tells me that this industry still has huge potential."

Before Lu Pengjun joined, the CTO seat has been vacant in the public annual report of Huanju Times for 4 years. The arrival of Lu Pengjun is a long drought for this company with 60% of R&D personnel.

Lu Pengjun organized the technology system of the Huanju era into three layers: the bottom layer is the file system, operation and maintenance system, and various monitoring, the middle layer is the big data recommendation algorithm, computer vision, and the third layer is the end technology, APP. The core of them is the algorithm. Huanju has just started, and he is also his strength, so he put more energy into the building of this section; and the APP technology and live broadcast technology on the terminal did not involve much before, and Lu Pengjun authorized it to Two elderly people gathered together.

What made Lu Pengjun somewhat unexpected was: "Although it is a team of more than 1,000 people, the Huanju era is completely different from Fengchao. The Fengchao team is more focused, and it belongs to the entrepreneurial style. Start-ups must constantly try new directions. When the pressure is greatest, more than a dozen apps are developed simultaneously, and every direction can’t go wrong. It’s really difficult.”

Lu Pengjun lamented that part of the pressure comes from communicating with non-technical bosses: "Perhaps this is the common pain point of many Internet companies. Sometimes when something goes wrong, the boss may not understand and feel relieved." Of course, he did not. Did not attribute all the problems to communication, "I authorized the front and back office to two veterans, but authorization does not mean decentralization. For a long time, I ignored the control of these two sections, and I found out when I turned back. There are many problems. Authorization is not wrong, but the premise is that you have to go first and see it. When the authorized object grows up and the concept is similar to yours, you can fade out. The consensus of thought is very important."

postscript:

In 2020, Lu Pengjun chose to join Vipshop as a joint CTO.

In his view, the e-commerce field is more powerful and scope for changes in the objective world, and it can also create real value.

Due to the short arrival time, the content of the manuscript can only regrettably end in the gathering period.

Asked if there is any intention to start a business? Lu Pengjun gave a negative answer. In his opinion, entrepreneurship, partnership, and part-time work... are all methods and means. The key is to exert one's influence to help this society. Entrepreneurship is not necessarily the best model. If there is a day, if I find that there is one thing I do not do, I will regret it. I will definitely choose to start a business without hesitation, regardless of success or failure."

If the tiger does not hesitate the mountain and the man does not hesitate, there will be a period later.

This article is from the public account [Technologists], sharing the growth experience and methodology of 35-year-old + technical predecessors, the community connects thousands of CTO/CIO, welcome to communicate together~