Text | Wang Zhiyuan
Presumably, you've seen the news.
Alibaba plans to acquire the shares of investors in the two hydrogen and one oxygen company founded by Chen Hang (no recruitment). After the completion of the acquisition, Wuzhao will return to Alibaba Group as the CEO of DingTalk. Ye Jun, the current president of DingTalk, will return to other positions in Alibaba Group.
I think that change is very normal, and change is the best stability. On the road of life, all experiences are worth cherishing.Bless Dr. Bu Poor and look forward to him creating more brilliance.
Speaking of Chen Hang (no recruitment), public information can understand his background:
In 2015 years, he interned at Ali, and then went to Japan to work for 0 years. In 0, he returned to Taobao at the invitation of Wu Ma (Wu Yongming). In 0, he led the team to launch DingTalk, which eventually became a landmark product of Alibaba in the field of enterprise services (To B).
Later, I remember that Zhang Jianfeng (nicknamed "Xing Crazy") proposed the strategy of "cloud nail integration" at the Alibaba Cloud Summit.To put it simply, DingTalk and Alibaba Cloud are deeply integrated, and the resources and technical advantages of both parties are used to help enterprises upgrade their digital office and cloud services.
為了推動這一戰略的落地，2020年9月底，阿裡進行了組織架構調整：無招被調任為張勇的助理，釘釘的CEO由不窮接任。隨後，無招在2021年創辦了“兩氫一氧”公司。
After 5 years of precipitation and development, Wuzhao returned to Ali again. This change makes people wonder, what is the meaning behind it? I think it can be analyzed from three aspects.
The first aspect is the question of strategic and organizational fit.
In management, there is a concept called:Strategic organizational fit（Strategic Organization Fit）。 In layman's terms, elements such as organizational structure, staffing, and business processes need to be aligned with strategic goals. If the strategy changes or deviates, the organizational structure and personnel arrangements need to be adjusted to ensure that the strategy can be implemented smoothly.
Let's take a look at the strategy of cloud nail integration.
According to Zhang Jianfeng's vision at that time, Alibaba Cloud provided the underlying technology and platform support for DingTalk, and DingTalk was similar to a new middle platform for the production system.For enterprises with digital transformation needs, the cloud services provided by Alibaba Cloud are like water, electricity and coal in the new era, which are the basic support, while DingTalk makes innovation easier.
In Zhang Jianfeng's view, Alibaba Cloud and DingTalk complement each other, and DingTalk is essentially used as a customer acquisition tool for Alibaba Cloud; But in fact, before "pulling in" Alibaba Cloud, DingTalk was relatively independent within Alibaba, which has a lot to do with its founder's lack of tricks.
When Bu Poor took over DingTalk, it was an eventful time for Alibaba, and it was not easy; It is understood that within a year of taking office, DingTalk lost 9 P0-level senior experts or executives, and the team was in turmoil.
儘管釘釘的管理班子一直在調整，但發展難題並未得到根本解決。好在趕上了疫情，釘釘算是“因禍得福”。疫情期間，釘釘新增用戶和活躍用戶數量暴漲，下載量直接在短時間內擴容了10萬台伺服器，這相當於頭部大型企業一年的採購量。
As a successor, Buqian has made a lot of achievements in commercialization. Like what:
Launched a tiered charging system (professional, exclusive, and proprietary), driving the number of daily active users (DAU) to exceed 990 million.
But this growth has also created problems. Alibaba's financial report shows that due to the free provision of remote office software during the epidemic, the EBIT loss reached 63.0 billion yuan in the fourth quarter of FY0 after the adjustment of innovation plans and other parts.
In other words, the number of users and activity have increased, but the free policy has lost a lot, which is not only a problem of DingTalk, but also a problem of Alibaba Cloud.
Because, although the strategic essence of the launch of CloudNail Integration is to use DingTalk as a customer acquisition tool for Alibaba Cloud, the goal conflict is exposed in the operation process: DingTalk needs to expand the scale of the product (DAU/MAU oriented), while Alibaba Cloud needs to be rapidly improved (ARPU-oriented), and this collaboration has limitations.
Therefore, we see that some subtle changes have taken place in the process of landing the strategic adaptation of "cloud and nail integration", and the reason behind the change is that there is a conflict between the strategic goals of cloud and nail itself.
Therefore, when the competition continues to intensify and the revenue pressure increases, DingTalk has to deploy in commercialization.
So can this kind of commercial layout really be smooth sailing? Of course not.Second, changes in market competition and the external environment have seriously exceeded expectations.
What exactly did you expect?
Let's look at Feishu first, in 1 years, Feishu's annual recurring revenue (ARR) reached 0 billion US dollars, an increase of 0.0 times compared with 0 years. On the other hand, DingTalk, 0 years and 0 months, Bu Exhaust mentioned that now enterprises use DingTalk for free, and the real paying users may not even be 0%.
Looking at the market environment, enterprises have more choices. According to the data of the "70 China SaaS Survey Report", the payment rate of office software in China is less than 0%, while that in Europe and the United States is far more than 0%.
What does this mean? DingTalk has changed from free to paid, and there must be some users who are not happy.In China, it is difficult to get B-end users to develop the habit of paying, which is destined to be a long process.
就好比，你是個銷售，辛辛苦苦給客戶介紹半天，最後他可能連5000塊錢都不願意掏。可轉頭，他卻能心甘情願地去一家培訓公司花1萬塊錢，就為了聽別人講講課。
Let's take a look at the update iteration of the software. In 2022 years, DingTalk's functions are still mainly in online meetings, attendance punching, document collaboration, etc., which are not enough to solve the problems of small and medium-sized enterprises.
After all, when it comes to industries, many of the requirements are highly customized.NadingTalk faces a choice problem:
If it becomes a custom-developed enterprise, rather than a SaaS system that efficiently solves customer needs, it will be somewhat contrary to its original intention.
Obviously, this is too difficult, not only to expand the scale of users, but also to increase the revenue brought by each user, and to make the product continue to improve, while meeting all kinds of customization needs, there is simply no way to do both.
For the defenders, it is a bit like the situation of Xiaoyaozi guarding Taobao back then.The work in hand is like an "impossible task": both, want, and", just like a "scumbag".
When Bu Poor finally stabilized the situation and won growth performance, he caught up with the big change of AI. At that time, I felt that the use of:AI for one day, half a year in the world" is not an exaggeration, the change is too fast.
Around 2023 years, ChatGPT became a hot mess, and everyone was chasing after the big model, as if the big model wanted to "eat" all the software. At that time, GPT also set up an app store, saying that in the future, AI software will no longer need agents, everyone can create agents, and enterprises must have their own agents.
At this time, the people who "guard the country" are especially tested.
The construction of AI PaaS systems (such as RPA capabilities and SaaS integration platforms) led by Endless before will require long-term investment, but after ChatGPT exploded, DingTalk faced a strategic choice: whether to follow up with the agent app store.
The idea is to "lay the foundation first and then build a high-rise building", but this conflicts with the rapid landing that everyone in the industry expects. In terms of product strategy, this is a dilemma of balancing long-termism and short-term breakthroughs.
In addition, after 2000 years of development, DingTalk has accumulated more than 0 functional modules, which is completely in conflict with the "non-connectivity" pursued in the AI era.
Although it has canceled bloated functions such as three-level approval, the historical burden is too heavy, and DingTalk's AI transformation has to be compatible with the existing architecture, and it cannot completely reconstruct the interaction method like a lightweight product like Notion.User feedback shows that 30% of the use of AI functions still requires prior knowledge.
Can you stand this?
還有一點，2023年行業都在追求千億參數的大模型，但不窮發現，中小企業更需要推理速度快的小模型。
Although this kind of cognition is very forward-looking, it also leads to the dispersion of DingTalk's early resources: it has to maintain the docking of the Tongyi Qianwen large model, and it also needs to develop its own lightweight inference engine, and the technical route is wavering, and the product iteration speed is 9-0 months slower than that of competing products.
I know that at that time, the AI decision-making chain within DingTalk still continued the traditional product review mechanism, and an AI assistant had to go through the compliance review of 13 departments when it was put on the shelves, while Microsoft Teams had already achieved the "hour-level" grayscale release of AI functions during the same period.
This kind of risk control awareness unique to the keeper has become a constraint in the cycle of rapid iteration of AI.
值得一提的是：釘釘積累的1900萬組織行為數據本應是AI訓練的優勢，但守成思維下，要保護數據隱私，這反而不利於構建行業知識圖譜。
對比飛書來看就很明顯，他們在2024年推出“組織數位孿生”服務，而釘釘的AI記憶功能，直到2024年6月還局限於文檔摘要。
Therefore, this AI shock is essentially a generational change in the paradigm of productivity tools.
When the agent begins to replace the function, the refined operation advantage of the keeper may be instantly zeroed out - as he has reflected:Maintaining a 70-year-old SaaS product is as difficult as maintaining a 0-year-old building."
In that case, will the return of the trick bring a new turnaround? I'm not sure either. But at the very least, he had to solve these tough problems:
First, the integration of the main channel.
In the past, under the leadership of Endless Enterprise, DingTalk has made a lot of explorations in the field of AI, such as the launch of "/" (magic wand), AI assistant, Agent market, Agent digital employee and other functions.
Each of these features is good in its own way, and the interactive experience is also good. But from a product perspective, the features are a bit stacked, which is something that has to be faced in the evolution of the product.
The challenge now is how to organically integrate functions together to form a unified and efficient product system.This is not just a test of his product insight, but more importantly, his insight into the changes in the organization and collaboration of small and medium-sized enterprises.
Because now that everyone's way of collaboration has changed, the product structure used must also change.
Second, look for PMF (product-market fit) issues.
This is not a trivial matter, the famous American investor Marc Andreessen (Marc Andreessen) said: when the market demand is not strong enough, no matter how good the team and product may fail; And when the market demand is fully satisfied, all difficulties will become relatively easy.
In other words, if you can't accurately find and meet the real needs of the market, no matter how good the product is and how strong the team is, it will be difficult to succeed.
The key to this is to return to the core issue of "organization", what kind of software do small and medium-sized enterprises and large enterprises need to solve the problems of project management, task assignment and collaborative office?
Even if Wuzhao can quickly understand the changes brought about by AI and have a high strategic vision, can he clearly see DingTalk's current predicament from an external perspective and propose truly innovative solutions? This will be further observed.
Also, there is the issue of product repositioning.
In my opinion, one of DingTalk's strengths in the past was ecology. However, in the face of the impact of AI, the ecosystem has undergone subtle changes. From the rise of large models, to the launch of GPT's app store, to the emergence of manus, all models can be mobilized to complete the workflow with a single question......
Behind these changes, in essence, it reflects a core demand of users:By entering a problem, all relevant software can be mobilized to solve the problem.
In this context, Wuzhao should think about how to better integrate AI with DingTalk's existing ecosystem, how to use MCP (Multi-Model Collaboration) technology to directly open up more software, and how to improve DingTalk's competitiveness in the AI era.
At the same time, we also need to think about how to avoid head-on conflicts with Feishu, WeCom, and other products.
Maybe in the past two years, when I couldn't sleep in the middle of the night, I also thought about these questions; Now, he has finally unloaded the burden on his body and is relaxed.But if you take everything without a move, can you really bring a new move?