Speaker: Li Cheng, CTO of Alibaba Group
Curated by Alibaba Global Initatives
Editor: Doreen Yin, Chandee Zhuang
CTO may not be a thinker, but he must be a practitioner.
My Experience
My experience is very simple. I have been studying in school until I was 30 before 2004. In 2004, I worked with a senior brother on an outsourcing project and made an important architecture upgrade for Taobao - replacing the original PHP / MySQL architecture with Jave EE architecture. In February 2005, I formally joined Alipay as an intern.
From 2005 to 2014, my major position was Alipay architect. In 2013, the CTO of Ant Group wanted to return to the United States. And Lucy Peng, the CEO of Ant, had a talk with me. At that time, I had no confidence in leading such a large team. What I worried about most is not to lead the technical team, but how to talk to the CEO, which is considered as part of the management job of the company.
The former CTO would quarrel with the business leaders at the company’s management meeting. I don’t think I can do it. Later, Dr. Wang Jian, the founder of Alibaba Cloud, talked to me and he said it wouldn’t be a matter. Everyone is unique, but everyone can solve problems in their own way. This gave me confidence.
So, in 2013, I began to take the whole technical team of Alipay. After one year, in 2014, the company appointed me as the CTO of Ant Group, and I was in this position from 2014 to 2019.
In 2018, the company figured out that technology will indeed play a more critical role in some businesses, requiring employees with stronger technical background. At that time, it was Ant’s international business that had a strong demand for technology. I was asked to be the COO of Ant’s international business for two years. At the end of 2019, I was transferred to the CTO of Alibaba. After another period of time, the Cainiao also needs a CTO, so I also concurrently serve as Cainiao’s CTO.
Coevolution of business and technology @ Ant
2003 Escrow service “Alipay” came into being
2005 The real-time risk control system of online payment CTU launched
2007 Distributed “service” of financial system
2010 Alibaba Micro-loan service launched
2011 Bar-code payment
2013 Yu’ebao was launched while the last minicomputer stopped service
2014 Oceanbase (Alibaba’s self-developed database) supported core transactions of the Double 11 shopping festival for the first time, with peak payments exceeding 30000 transactions / s
2015 Multi-Site High Availability came into use, MY Bank was opened and Ant Financial Cloud was launched
2016 Paytm wallet in India was rolled out, and Ant’s technology started to provide service overseas.
2017 Smile Payment was launched.
2018 Ant’s BASIC financial technology was fully open: Ant Block Chain, Financial Intelligence, Security Risk Contro, Mini-app, Dragonfly system, Financial PaaS, and Oceanbase.
I selected several milestones when I worked in Ant, which well prove the idea of “co-evolution of business and technology”. At the very beginning, Alipay was very simple, that is, to make an escrow payment platform. The technology was only needed to support realizing the business. And the biggest feature of all those projects is fast and lightweight.
In 2005, many of my colleagues were very nervous when an international payment giant entered China. Jonathan Lu was Alipay’s CEO at that time. He said, “What’s so terrible about this? They need two months to develop a new feature while we only need two days. They definitely can not compete with us.”
By 2007, the company faced a technical bifurcation: Internet payment will certainly become a very mainstream payment method in the industry. We can see the scale of the business, but we need to make a judgment - what technical architecture should be used to support it.
Alipay didn’t have strong requirements for technical architecture at first. We used whatever worked. We had a very Internet-based architecture, which was born out of Taobao, as well as a very financial architecture. In 2005, we bought a minicomputer to handle core accounts.
But in 2007, we must make a judgment. Would we be more like a bank or an Internet company in the future? Fortunately, we made the right decision. We believe that we must use Internet architecture to solve payment related problems in the future.
So I decided to re-structure the system into a future oriented architecture: distributed structure. For a whole year, we distributed all core transactions, core accounts, core members and core payments.
However, as to the problem of how to solve some core technologies, such as how to solve the core transaction in such a large distributed system - it is not easy.
In 2007, another international payment giant also wanted to solve this problem. As a result, the system crashed after two weeks and CTO left. In the process of restructuring, we were also frightened, but we finally made it.
The second milestone is in 2011, when the mobile device became more and more popular, but a very key judgment needs to be made on what kind of payment should be like in the mobile scenario. When users pay with their mobile phones, what is the way?
At that time, the smart-phone industry was not as mature as today, and some mobile phones did not have cameras. We deployed many plans, e.g. QR-code payment, NFC payment as well as bluetooth payment. After many attempts, we finally figured out that the QR-code payment is the one which works for most of our users.
2013 was also my first year as Ant’s CTO, and many important things happened. We have many important projects from number 1 to number 9.
Project number 1 was MY Bank. Number 2 was Yu’ebao, a money market fund product, which helped Alipay to expand its business from just payment to the real digital inclusive financial service. Project number 3 was Huabei, a credit card like service. And of course, we had many more. Some were successful while some were not.
This process has had a great impact on the whole technical architecture. From the original technical architecture of Internet payment, it began to move in the direction of digital finance. In this process, I feel that technology can help the company realize technological revolution. That year Alipay officially began to remove “IOE”(the technological structure of IBM, Oracle and EMC), and the last minicomputer was offline.
Coevolution of business and technology @ Alibaba
1999 Alibaba was established and started its first B2B business
Alibaba had a broad range of technologies. The Alibaba Cloud played the main role in its second half of development.
2003 Taobao was launched and started the C2C business.
It is very clear that in the first stage, starting from the launch of Taobao, technology supported business development. The reason why Taobao could win, in addition to its many commercial innovations, “fast” was also a very key factor. In this stage, CTO should be able to make technical layout in advance when the business grows rapidly, so that the business growth can be sustained. In 2004, the stage when I did the outsourcing business of Taobao architecture upgrade was actually a very critical period for Taobao. The formation of Java EE architecture laid a very good foundation for the growth of Taobao.
2004 Taobao fully turned to Java to lay the foundation for high-performance websites
2005 Alipay became the 3rd-party payment platform
2006 Taobao started to provide services through the way of platform
2007 Alibaba launched P4P(Pay for Performance) system, which is considered as a milestone of Internet marketing
2008 Taobao Mall (Tmall) was launched, and the Colorful Stone Project was completed, thus the Tmall and Taobao were fully interconnected.
2009 Alibaba Cloud was founded, and the Alibaba team wrote the first line of the code of Alibaba Cloud
There was a brilliant stroke in 2009 - at that time, Alibaba dared to judge that cloud computing is the future and wrote the first line of code for the cloud. Many people inside Alibaba were not optimistic about cloud computing, but why could they make the right judgement at that time? I think Alibaba has a very good saying: “you believe and you will see”. Just like the e-commerce business, Alibaba judged that e-commerce must be the future at a very early stage. Therefore, for a long time, Alibaba firmly believes in this direction and continues to move forward. As long as the time comes, it can naturally achieve commercial results. In fact, technology is the same. Alibaba made a judgment very early: cloud computing must be the future, and all subsequent developments are pivoting around this.
In retrospect, each of the following key developments is related to cloud computing. For example, after we wrote the first line of code in 2009, we couldn’t find a single customer within the company. At this time, the CEO specified that Alibaba Small & Micro Loan business must use Alibaba Cloud in order to provide a scenario to practice and improve the Cloud service. This gave Alibaba Cloud a glimmer of hope. In 2010, Simon Hu was the head of Alibaba Small & Micro Loan and later he became the head of Alibaba Cloud.
2010 Mobile Taobao was launched and Alibaba started to open to the public
2011 Alibaba Cloud began to provide cloud computing services on a large scale
2013 Alibaba completed the De-IOE project and Alibaba Cloud became the first cloud computing company with 5000 servers cluster
2014 The big-data computing platform MaxCompute was launched
2015 The Middle Platform Strategy was initiated
2016 Taobao realized “every single user with a customized front page”, and City Brain was officially launched
2017 Damo Academy was established, Tmall Genie was introduced
2018 Alibaba announced the chip strategy by establishing T-head, a data center with 300,000 servers that was set in Zhangbei.
2019 Open the era of Alibaba Cloud Intelligence
2020 We stepped into the era of Cloud Native, Alibaba Cloud + DingTalk integration
What impressed me deeply was that in 2013 and 2014, after several years of practicing with Alibaba Small Loans, Alibaba Cloud began to gain strength in technology and big data processing technlogy gradually got mature. At that time, there was a very important “5K project”. Alibaba Cloud began to have the ability to process data in large clusters of 5000 machines. With this capability, the company made another important decision. All the data of Alibaba and Alipay should be on the big data platform of Alibaba Cloud. At that time, I was exactly the CTO of Ants Group. I strongly supported the decision. Because I knew that even I didn’t, the group would still move forward. Alibaba Cloud has thus reached a new level.
In 2015, many people heard of a concept of “big middle platform and small front end”, which sounded great. This actually has accomplished one thing: all the basic technologies of Alibaba and Alipay are unified and moved to the Cloud. This is a major technological change. In order to complete this technological change, Alibaba has made a very good organization design, that is the CTO of Alibaba group at that time transferred to the same position as Alibaba Cloud’s CTO, and put technology together to strengthen Alibaba Cloud. After three years of implementation of the strategy of “big middle platform and small front end”, Alibaba Cloud’s entire technical architecture is very complete. It is how Alibaba supports the achievement of a technology strategy with the strength of the whole company.
After 2017, this stage is called “creating new business with technology”, and we began to explore whether technology itself can become a part of business. The Damo Academy was established in 2017. Because we judge that the digital economy will be inevitable in the future. And the Damo Academy should focus on the core technology required by digital economy. No matter the AI, computing or chip, they are all the thinking and layout of the core issues of digital economy.
In 2020, I am already the CTO of Alibaba Group. We now have all the core businesses operating on the cloud, including the most challenging businesses, such as search, recommendation and advertisement businesses. The Cloud Computing can support them all. I think this is a very great technological progress of the cloud and a very significant technological change of the group.
Now we have entered a new ear. We should clearly define the boundary between cloud and business technology. The boundary is the Cloud Native. Our future business applications will be all based on cloud native. It will be a new era of Cloud Native. In addition, the Cloud Computing itself is also upgrading, which is known as Cloud+DingTalk Integration.
Three kinds of technical leadership in different business stages
I was inspired by Yongfu Yu (Partner of Alibaba Group, Former CEO of UC Web and Amap) for the 3 kinds of technical leadership. He thinks that enterprise development is such a wavy development process:
- Find a direction and enter a business track;
- If this direction is judged coreectly, the enterprise will enter the stage of rapid growth;
- When the enterprise grows to a certain stage, we have to git rid of inertia and fina a new direction.
In the process of wave development, technology plays a different role in each stage. In the track entry stage, CTO should be a member of the decision making team to support the CEO, to help the team be clear of the direction and help bring the business on the right track.
Once the business is on the right track and enters a period of rapid growth, the core function of CTO is no more helping judge the direction, but how to do a good job in technology. At this time, the chief architect will become very important. Technology makes the business grow faster and accelerate the growth. The business should not be slowed down by technlogy. At the same time, organizational design is as important as technical archtecture at this stage. And you can’t wait until the business stops to judge the future. CTO should judge what will happen in the future and what the second curve of business is in advance, and design a route to the future.
I think the organization as a whole needs two kinds of leadership:
- Professional leadership: CTO, Chief Architect, Technical Director and VP;
- Management leadership in organizational design.
CTO plays different roles in different times. Sometimes he/she assume a professional role, sometimes a management role, and sometimes a strategic role.
It’s hard to say that a CTO is omnipotent. CTO must have strengths and weaknesses. Take mysefl as an example, I am positioned as an engineer. Engineering ability is my stong ability while organizational aiblity is my weak one. In this process, a core team needs very good cooperation by putting people with different leadership in different positions.
Key Responsibility of CTO
- Building up the “resonating” connection between commerce and technology
What kind of role shall CTO in a company?
I think the most basic responsibility shall be building up the connection between commerce and technology. It’s mentioned that commerce and technology evolve together and during the evolving process, there need to be a “resonating” connection between them. What is the core capability a company need to build up to realize social values as well as customer values? This is not a decision based purely on technology nor business, but a mutual decision of the core management team.
It thus becomes critical if CEO and CTO can speak the same language in their conversation. CEO together with CTO need to clarify very fundamental questions: first of all, who is our customer; second, what kind of differentiated customer value that we can provide; third, what kind of business model to be adopted to realize the customer value; and last, to adopt the business model, what kind of capability, strategy and organization that we need to build up. Only when those questions are clearly identified and answered, technology team will understand what business team is planning to do.
On every task I accept in Alibaba, I will build my understanding about it in this way. There will more questions emerging later on which need further understanding. This is always my first step, and it is the duty of CTO.
- One map, one battle, one heart – marching guided by mission.
Most of the time, it’s a challenging task driving the whole team going forward. If you can drive a small team by good relationship, it is almost impossible to relay on relation to do so in a bigger team where people may not know each other well. Leadership plays a critical role under such circumstances, to drive the team with “one map, one battle, one heart”.
Let me share with you some cases and you will find they are challenging but there are multiple ways to deal with them.
First case, the way how I build up leadership is to set very clear targets together with the team. Back to 2013, after a long hesitation, I finally decided to accept taking the role to lead the whole ANT tech team. The biggest challenge I was facing is “TRUST”- as we were peers before and all of sudden I would become of the team leader. The most common feedback on me is
My opinion towards this situation is, trust can only be built up with accomplishing tasks together. It’s not even necessary to gain acceptance on the leader if the team can believe in mutual target and realize it together. In the year of 2013, we set the following 4 targets together:
When ANT was transferring from internet platform to digital inclusivity platform, what kind of technological infrastructure we need to build up to support? We need to re-define/design platform that can provide digital inclusivity. My professionalism in designing platform infrastructure was acknowledged by the team.
What were the core technology that we need to innovate? We identified two: to launch ocean base and to build up the platform on cloud.
Would we be able to support 30k transactions per second during the annual campaign? It was well known that we had critical testing every year: what was the peak number that we can carry? It took ANT team all the efforts to take the “testing” in the past 2 years. What happened in 2012 was a unforgettable shame to the team – system collapsed before the campaign started and it took 50 minutes to recover. In the next year of 2013, we all wanted to wash off the disgrace: if our system reached its extreme of 3000 transactions per second in 2012, then let us set a target of 10 time multiplied, i.e. 30k transactions per second!
Stability – we offer not just payment services but also started offering financial services, which means our system need to be as stable as, or even more stable than, banking system. We thus set a 4 nines target (99.99%) when it was even hard to reach 99.9%.
The above 4 targets, though very challenging, gained full commitment from the team. At year end, 3 and half of the targets had been accomplished: platform transferred with Yuebao, Huabei business models online; Oceanbase launched; 4 nines accomplished; we didn’t make 30k transactions per second, but 15k transactions per second, 5 times multiplied. I started gaining trust from the team.
When I joined Alibaba Group, the first thing I did was the same – we sit together and set up the target together. As a new leader to a new platform, it was quite hard to set a completely new vision for the team. A future plan drawn by a newly onboard CTO normally didn’t make much sense. It was more practical to set up the targets together with the team and accomplish them together. That was why I decided to make steady progress on clouding strategy, by bringing the middle office to a further stage, including AI middle office, data middle office and business middle office. Another target decided then was to digitalize the organization. Alibaba was born in digital era, but it still needs digital transformation and upgrading. Setting up the original digital business system and original organization digital system was the mutual target set forth together with the team.
As CTO, you need to involve your team to think about the future outlook, not just technological connection with current businesses. What does the future look like? What shall be the path to the future? What will be the key battles to win in the future? They are the key driving force in the hand of CTO.
- Making key decisions – removing the hurdles in the way
CTO shall be able to solve problems for the team, clearing what’s in the way and making key strategic decisions. My first key strategic decision as CTO of ANT Group was to decide whether the technological infrastructure should be internet oriented or financial service oriented. My final decision was internet oriented. The second key decision was on transformation of Oceanbase.
In the year of 2010, professor Yang Zhenkun started to build up Oceanbase with his team. Oceanbase was not created as a real database. It came into being to solve the problem of Taobao bookmarking, in which the ability to store data was more important than process data. Then it came to 2012 when Oceanbase need to make the decision – should OceanBase become a real database?
ANT team had doubt about OceanBase for a good reason – our business put a very high standard on database. The Oracle database, which was set up and constantly improved for more than 10 years by their professional team, was just able to meet the business request. Would it be possible to replace it with a database built up within 2 years by Mr. Yang’s team?
I trust Dr. Wang Jian knew the art of leadership. He made a choice by saying “Let us give OceanBase to Alipay. “This decision then became the second target of the four: let us see if OceanBase can be developed further.
How did we do back then? First of all, understand what OceanBase is capable of; Second, as it was doubted by the business team to take the whole database tasks, then let us testbed with OceanBase by diverting 1% traffic to it, so it can prove itself with the 1% scenario. OceanBase didn’t fail us by successfully holding the 1% traffic and finally completed transformation from non-relationship database to relationship database. We took another bold step forward in 2014, diverting 10% of double 11 traffic to Ocean Base and make the core accounting system accessible to Ocean Base. The accounting system had a higher request on database and that was a very risky decision to make then. But all these decisions give Ocean Base a different future.
The following decisions were made in the same way. As CTO, it is critical to balance risk and stability.
Take another example, on the first day Yu’ebao was introduced we knew that it would be a success as it had a crystal-clear customer value positioning – make each dollar profitable in each account balance. What we didn’t expect was that the success came so rapidly taking all the data capacity in one month. It was not quite a challenging issue for ANT as ANT platform had been completely built up upon distributed data storage which could be extended quickly. The challenges were at our partner Tian Hong Asset, as their system was not be able to cope with this rapid growth of Yu’ebao. Database extension because the core issue to solve.
We therefore made a very important decision – migrate Tian Hong Asset’s system to cloud and re-code it with distributing infrastructure, within 3 months. This was a proof of financial clouding’s value and business related to financial clouding grew since then.
All these examples reinforce that CTO need to play the key role in making key decisions and help team deal with uncertainty.
- Cope with risk and transform risks to opportunities
There are several types of technological risks.
The first type, technological infrastructure cannot support business growth, which is an unacceptable risk. This type of risk is relatively easy to be identified. The reason why Alibaba started to “de-IOE” is because we foresee the risk that “IOE” infrastructure would not be able to support company growth and must be deprived. In ANT Group, “de-IOE” started from 2010 as the campaign made us realize the limit of original infrastructure.
The double 11 in 2010 was joked by us as “manpower cloud computing”. When the campaign started, we witnessed a crazy traffic and realized that the designed capacity limit would be broken in the coming daytime and added all the data storage into it. And it proved not enough at all. We then started “killing “services that were unnecessary and focused all the capacity on the key chain. It was always not enough and we were looking everywhere for processes that could be “killed”. At the last a few minutes of that double 11, both our core payment data storage and accounting data storage had reached the limit and would have collapsed if the campaign continued for another 10 minutes. At that moment, the team made the decision and killed the accounting database. Accounting database was very important but it could survive with a few minutes ‘interruption. The core payment database could not afford any interruption or the business process would die. We held on the double 11 of 2010 in this way.
In fact, “manpower cloud computing existed for a very long time after that double 11, with people make decision on resource allocation. This was extremely painful and pushed ANT to make the firm decision to embrace the distributed storage on cloud computing.
The second type of risk is not about technological infrastructure, but on stability. CTO need to be able to foresee this coming and CEO need to support CTO with resources allocation.
Take “527” of ANT for instance, in 2015 we realized live on distant data centers, when datacenter in South China was able to realize distributed storage exchange with data center in East China. This is the first time in financial system. We carried on network breakup practice, cutting down data centers in East China and had data centers in South China take control. It proved successful. Everyone was excited. I remember I wrote a report to Lucy Peng, CEO of ANT back then, saying it was the first time ANT system realized live on distant data centers. You could sleep well from now on as our system would be as stable as rock under any circumstances.
Then on 27th, May 2015, I received a system alert when I was driving. Normally these alerts would be taken care of by the team directly but the alert insisted beeping that day. I parked my car and found out optical fiber was broken in construction sites and system was interrupted for almost 2 hours. I then learned that “live on distant data centers “would not work and make the transit if the optical fiber was broken. I stopped sending any battle field report as a way of sharing good news with CEO ever since then.
That was a hit to ANT tech team, with clients ‘trust with ANT was shaken. Externally, people started making fun of ANT technology, and ANT tech’s image was hurt. Internally, we, including me and my team, was facing trust crisis with CEO. You were not trusted with your words as they would not believe what you said could be achieved.
Risks could be a good thing. After “527” incident, we took the following approaches: first of all, we talked with the company to initiate setting up a department taking care of technical risks. The sole responsibility of this department was to secure the bottom line of ANT technology. 10% of technical resources were allocated to this department, which meant we were willing allocate this resource to secure ANT system free of risks in the future. The “527” incident helped us win this resource. Secondly, start real battle drills. We practiced network breakup scenarios before but they were prepared practices. We started unprepared battle drills since that day, until now. Third, we decided to name the date of May 27 as ANT Group’s Tech Day, which aims to warn each technician in ANT Group of any risks, no matter how long ANT will exist.
The three things were done with a more united team spirit. Risk control in ANT Group had been significantly improved and this was what I called turning risks to opportunities.
The third type of risk, which CTO and CEO may concern most, is whether the existing business model will be subverted by a new one. This is a common concern shared both by early stage companies and established companies like Alibaba. There is a risk of being completely subverted whenever a new technology comes into existence, despite the size of companies. This was a major challenge for ANT when we were entering the mobile internet era. Bitcoin’s expansion brought us this concern about whether it would totally change the way of payment. When Libra announced itself in June 2019, the concern started to grow: will global payment system be re-defined totally by a new currency? It would be an attack by dimension reduction weapon.
CTO need to join CEO to help make the judgement at each of such moment. Each decision made upon potential future challenges is also a chance to stimulate new business opportunities. The major approach will be – testbed with these potential challenges other than just monitor them, which will request some valuable waste on company’s resources.
The fourth type of risk is what we call boil a frog in warm water. The harm technology can do on a company may not be as direct as risks. However, if company efficiency gets slow due to technological, infrastructural or organizational issues, the company will be deprived of its power of competition or innovation, which will become the most devastating risk to a company. Take the middle office of Alibaba as an example. The advantage of middle office is to eliminate repeated construction on infrastructure so as to accelerate innovation based on middle office. Alibaba’s middle office has been complete and now entered a new phase. There are middle offices for core e-commerce businesses, for digital supply chain, for functional departments, for data, for AI and a serial of technological developments. Whenever a new business model is being built up, it will collaborate with all these middle offices to gain supports. The business will not work out if any of those middle office fails. In this case, we are now pushing for upgrading our middle offices by optimizing middle offices’ deliveries, which includes plenty of changes on technical infrastructures. This is how we prevent boiling frog – consistent and systematic combing and solve the issues phase by phase.
- Organizational design and administration – balance between order maintenance and innovation.
The longer you act as a CTO, the faster you will be losing your professional competence. I find mine downgrading one level every two years. The good thing is your team will grow stronger as you will complete with the team together. The core of organizational design lies in how to balance order maintenance and innovation. You will find it conflicting as innovation normally grow from chaos. Order improves efficiency but sacrifices innovation. To balance between order and innovation is like walking on a wire.
Organization of Alibaba’s tech team has two structures. One is the management report line, which distributes by functional layers:
Front office supports businesses, serving customers; middle office plays the role as capability center, supporting front office as customers by improving front office’s efficiency and competition; back office focuses most on technology advance to secure the consistency of businesses. Every layer of this organization structure is an independent business unit. What we are trying to do currently is to set a CTO for each layer of business unit and this person will take full responsibility for this business unit. The essence of this type of managerial mechanism is to have a clear definition of each layer.
This type of organization administration gives flexibility to each business unit so each unit can have its own growth. The potential issue it creates is how to consolidate the resource for the technology to be developed at company level. To solve this problem, there is another dotted line of Technical Committee, which consists of 20 to 30 core technical groups to integrate technical plan and resource across units on common technical fields. Led by professional guidance by committee and technical groups, the organization can have integrated strategy and talent resource. These two lines will function simultaneously. The clearer the definition on layers the more important this dotted professional line becomes. Take audio/video technology for instance, it is commonly used by various business units, middle offices need audio/video technology as well and so does back office which need to constantly optimize its technology to improve competition. In this field, a technical group with professionals from front office, middle office and back office will work together led by one group leader.
How to make management decisions with this dotted organizational structure is a challenge. As CTO, I believe an open mind is very important that helps you to have faith in the judgements made by professional groups. When there are different opinions and hard to reach alignments, CTO need to make independent judgement with understanding of how-to balancing order and innovation.
This is the common issue for organizations developing to a certain scale. I will not suggest complicate the organizational structure while the scale is not big enough, which will lead to additional cost on management and communication.
- Enhancing cohesion and passing down the flame
How to enhance cohesion in an organization is most important and most challenging. This is about technical culture. From time to time, we get together to discuss technical culture in Alibaba. We had this discussion last year and another round three years ago. The slogan from the discussion 3 years ago was “technology creates new commerce” and the one before that was “technology extends limits of commerce”.
Besides having slogans, one key culture of Alibaba tech team is to be practical, have respect for facts. Let facts talk. Do not propose an opinion with not data or facts supporting. This culture makes things simple if well accepted or behaved. It is not an easy job to do.
There are some Alibaba proverbs in technical team such as “think towards the future”, “believing is seeing”, which means dare to invest in the future. If you always follow others, you will be exhausted from chasing. Of course, to have a sharp future outlook and dare to invest won’t be an easy task either.
It is a big challenge to the organization to build alignments on culture and visions, especially when there are constantly new comers joining and people leaving. Only based on solid culture things will sort themselves out.
CTO may not be a thinker, but he must be a practitioner
Although the six responsibilities of CTO I mentioned are not complete and it is impossible to do everything perfect, the core thinking is that I believe that CTO or technical team needs to evolve with the business development of the company. From Ant Group to Alibaba Group, I have almost experienced all the 6 stages, which I call Six Steps of CTO:
- Define goals together with the team and accomplish some things together.
- Know more about the team and business, know where is the desitination of future, create a vision with the team to ignite everyone’s enthusiasm.
- One of the core taskes of CTO is to avoid making himself/herself the ceiling of the team, but to make himself/hersefl the springboard of the team by using right people to do right things. If CTO is the ceiling, the company’s technology will be limited in a small and narrow field forever. When the CTO becomes a springboard, the comapny can expand the boundary of its business through new employees. As a CTO, one should focus on achieving business by developing people rather than by using people.
- When everything is fine, don’t forget to repair the roof on a sunny day and always be alert to danger in times of peace. Once a crisis occurs, see it optimistically. There are opportunities behind each crisis to turn crisis into opportunity.
- We should not only focus on the present, but also lay out the future and establish technical barrier for the company. In the early stage of business development, the barrier of technology is a point. When it develops to the current scale of Alibaba, the barrier of technology is a polyhedron. There must be sufficient and multifaceted layout to support the development of a large company. One need to align with CEO to decide how much to invest in the future layout.
- Finally, talent. From generation to generation, talents are the key to the future development of the company. I remember a technical director of Alibaba Cloud once shared that what is the highest level of a company’s technology, that is, whoever can be a CTO, I think this is something I should strive for.
How to develop and cultivate CTO
Finally, the most important thing is the talent.
We talked about our hierarchical and distributed governance earlier. Each business unit should have its own CTO. How to cultivate this CTO? Basically, we let him practice in the battlefield and design the development path for him. There are also classes such as “singularity camp” to train those CTOs.
Of course, the most difficult thing is to train successors. This is actually what I considered on the first day of taking office. But there are many challenges in choosing the right people.
Here are my two experiences:
- The routine of cultivating a CTO is a “Z” style
People who become CTOs in a straight line often have problems because they lack of enough experience to gain capabilities to tackle with different type of problems. Jeff Zhang, the previous CTO of Alibaba Group and now the president of Alibaba Cloud, used to be in charge of Taobao and Alibaba B2B business. I have experience working in Ant Technology, and later in Ant International for 2 years. Then I became the CTO of Alibaba Group. It is very important to have business background so that the CTO can align with CEO in the same picture.
- Design an “L” style job structure
A CTO should put his/her foot down. After all, this is a very high position in the company. The teams are there to support you, but if you cannot see the details, you won’t know what exactly happen, which will ultimately lead to wrong decisions.
How does Alibaba solve this problem? We will require a CTO to have 2 positions: A group level CTO will at the same time be asked to take a role of BU level CTO. For example, I am the CTO of Alibaba Group as the CTO of Cainiao Logistics Network. Jeff Zhang used to be the CTO of Alibaba Group and Alibaba Cloud.
What qualities should a CTO have? Each CTO has a different style, but there are several points in common: first, be realistic and pragmatic. We require every one to comment with data. It is very important to make decision down to the earth. Second, be responsible, dare to take historical responsibility and be enterprising when making key decisions. Third, we must always be introspective and open minded, without which, it is difficult to evolve. Fourth, the CTO of a large organization and business should have an overall view, be able to design architechture, and form a big picture of all aspects and all kinds of information.
Zeng Guofan ( a Chinese statesman and military general of the late Qing dynasty) knew the way of employing people very well. We can tell his principle of choosing people from his saying that “it’s hard to observe simple and clumsy (sincere and faithful) people when they are in danger”. I especially like this sentence and make it as my signatures on DingTalk, as a requirement for myself.
So far, I feel that I have not done enough in many aspects, including my judgment of the future. CEO and CTO actually want to grow and evolve together and form a tacit understanding in their work, which is very important.
若有收获，就点个赞吧
0 人点赞
