Source: Silicon-based Laboratory

Author: Bai Jiajia; Editor: Fisherman

Original Title: "After Going Solo, Can DingTalk Really Ride on the Back of AI?"

30 years ago, there was an article called "The Chinese Software Industry on the Back of the Bull," which talked about how practitioners in the software industry must always pay attention to the industry's trends. Otherwise, every time you think you've won, you will be thrown off the "bull's" back.

This article is still relevant to the SaaS industry today.

On the one hand, AI is believed to be destined to cause a drastic change in the mode of production and productivity, and is expected to open up the deadlock of scale profitability and customized demand.

On the other hand, as more and more large model enterprises continue to launch industry-specific large models, helping enterprises improve quality and efficiency, the anxiety of whether "AI will eat up software" has become a concern for another group of practitioners.

One consensus is that the problems in the Chinese SaaS industry are not only due to insufficient production efficiency or poor product quality. Fierce competition and a distorted ecosystem are not problems that large models can solve.

In the midst of chaos, DingTalk has become a disruptor.

"Openness has become a belief of DingTalk. The comprehensive intelligence of DingTalk cannot be achieved without the participation of its ecosystem partners. In the era of large models, the golden decade of the Chinese SaaS industry has just begun," said DingTalk CEO Ye Jun at the DingTalk Ecological Conference on August 22. Does the "AI+ecosystem" route revealed in this statement really provide a solution for the industry?

01. Can Large Models Save SaaS, or Will They Kill SaaS?

The first surprise at the ecological conference appeared when DingTalk founder Chen Hang took the stage.

Chen Hang's love for DingTalk is beyond doubt. Choosing to attend the ecological conference and serve as the opening guest on Chinese Valentine's Day carries significant implications. However, this time, he did not choose to be a defender of DingTalk, but instead, at the beginning, pointed out three "unavailable" aspects, highlighting the shortcomings of DingTalk.

After leaving DingTalk in 2021, Chen Hang once again embarked on entrepreneurship and founded a company called "Two Hydrogen and One Oxygen," helping Chinese manufacturers transform into brand merchants. His three "unavailable" aspects are the extensive government procedures, internal organizational personnel and financial recruitment, and product and customer aspects that need to be addressed during the entrepreneurial stage.

Chen Hang stated that he spent a significant amount of manpower, material resources, and financial resources at each stage, but unfortunately, none of these expenses occurred on DingTalk.

In response, Chen Hang provided a set of data. In 2021, there were 52 million small and medium-sized enterprises in China, with 13,800 new enterprises daily. The implication is that compared to the large B-end, the sea of small and medium-sized enterprises is much broader. "The internet comes from the grassroots, and serving the grassroots is the true essence of the internet," he said.

The spirit of serving the grassroots of the internet is naturally politically correct, but a reality that cannot be ignored is that the Chinese SaaS industry has been going through a "crossing period" in recent years. With the retreat of capital, the business model of attracting users at low or even no cost through capital has come to an end, and many enterprises that relied on capital infusion have to face the test of survival.

Developing business for small and micro-enterprises not only has narrow application scenarios and is difficult to be upwardly compatible, but also faces difficulties in receiving payments. At this point, it may be somewhat untimely to talk about ideals in the SaaS industry.

To some extent, Chen Hang's suggestions reveal the dilemma faced by the SaaS industry. All enterprises believe that their pain points are universal and promising, but the differences brought about by factors such as region, industry, and scale are much greater than imagined by the problem presenter, making it difficult to apply even to companies in the same industry and of the same type.

The contradiction between customer customization needs and standardized product scale profitability is the most common dilemma in the SaaS industry. This is why SaaS players are placing high hopes on AI. Functions derived from large models, such as code generation and interpretation, help enterprises deliver software at lower costs and higher efficiency.

For example, after introducing Github Copilot, Chen Hang's company's code production efficiency increased by more than 50%.

However, the subtle point is that whether it is ChatGPT or Github Copilot, they are not traditional SaaS companies. Moreover, achieving a code production efficiency increase of over 50% is an outstanding achievement that many traditional SaaS companies find difficult to match.

In other words, the "one mu and three fen" (a Chinese idiom meaning a small piece of land) of SaaS companies is likely to be divided by a few leading players in the process of "large model empowering various industries."

What does AI really bring, a warm spring or a cold winter?

Kingsoft Office CEO Zhang Qingyuan once said that the future of SaaS will inevitably be AI-driven. "Instead of being anxious, we should think about how to use large models to upgrade our products."

Zhang Qingyuan's words are not without reason, but for the vast majority of small and medium-sized enterprises in the industry, the computational cost and technical threshold brought by developing large models are "unbearable burdens."

Large models are like giant rocks rolling in from outside the field of vision. Small and medium-sized enterprises can hear the thunderous sound of the rocks rolling over the ground, but they cannot see their trajectory, let alone catch a ride on them.

Will DingTalk be the one willing to lend them a hand?

02. Why Doesn't DingTalk Dare to Talk About Money?

DingTalk CEO Ye Jun did not directly introduce the changes that occurred in DingTalk after integrating large models. Instead, shortly after the start of his speech, he mentioned an article that had been widely circulated in the SaaS venture capital circle: "China Doesn't Need SaaS."

In the article, the author stated that an industry that has been losing money for eight years is not worth investing in, claiming that "in China, if outsourcing is needed, then outsource; if customization is needed, then customize; if private deployment is needed, then privately deploy; if payment is needed, then demand payment. Not doing SaaS is a sign of respect for the Chinese market and customers."

This strongly worded article resonated with many industry practitioners and also faced opposition from many people. How can we understand the complex emotions of Chinese SaaS practitioners? The best observation angle is actually DingTalk.

In just one month, DingTalk has undergone many changes, with the most attention-grabbing being its split from Alibaba Cloud.

Although DingTalk was previously seen as a supplementary service of Alibaba Cloud, it has also reaped benefits. From the transformation of collaborative tools to integrated development platforms, a large number of government and enterprise customers, and most importantly, revenue support, these are all tangible benefits.

Now, after becoming independent, DingTalk not only has to bear the pressure of profitability but also has to tell more stories to improve its valuation performance. Regardless of how many stories about making money through DingTalk's ecosystem partners were told before, the fact facing all observers is that DingTalk is still losing money.

According to reports from "Caijing," in November 2022, DingTalk's monthly cost for audio and video technology resources was approximately 250 million yuan. More than 70% of this cost was for network bandwidth, and over 20% was for server and storage costs. Over the past three years, DingTalk's annual network bandwidth costs have ranged from 1.5 billion to 2 billion yuan, and in extreme cases, have exceeded 2 billion yuan.

In terms of performance, although Alibaba has not disclosed DingTalk's loss situation in the past two years, its profit loss in the fourth quarter of the 2020 fiscal year was as high as 3.063 billion yuan, an increase of nearly 60% compared to the same period in 2019.

In fact, in order to achieve profitability and commercialization, DingTalk has made efforts.

Starting in March 2021, after officially beginning commercial exploration, DingTalk has launched explorations in three directions:

1. Software subscriptions, including three paid software packages (Professional Edition for 9,800 yuan/year, Exclusive Edition starting from 100,000 yuan/year, and Exclusive Edition starting from 1 million yuan/year);
2. Platform application commission, where DingTalk's open platform distributes products for software partners and charges up to 15% commission;
3. Hardware interface commission, where partners develop hardware based on DingTalk's open SDK interface, and DingTalk charges a 10% commission.

Ye Jun also provided a set of data to confirm the success of this model.

As of the end of March 2023, the number of paid enterprise users of DingTalk's software reached 100,000, with small and micro-enterprises accounting for 58%, medium-sized enterprises accounting for 30%, and large enterprises accounting for 12%, and the number of paid customers is steadily increasing. DingTalk's annual recurring revenue (ARR) far exceeds the $100 million evaluation standard of a unicorn company, and the GAAP revenue from software subscriptions alone has already significantly exceeded this figure.

This data seems to indicate that, barring any unexpected circumstances, DingTalk will grow a beautiful revenue curve, just like Salesforce, which is considered the standard answer by many software entrepreneurs.

However, the problem is that the commercial ecosystem in China is completely different. Subscription is a charging model commonly used by foreign software companies. It is simple and easy to understand, and has good cash flow. However, in China, this service is seen as a "platform tax," and small and medium-sized enterprises have thin profit margins and are unwilling to be further skimmed, even if it can help improve the quality and efficiency of their businesses.

Ye Jun has also admitted that the actual proportion of enterprises purchasing professional, exclusive, and exclusive edition software on DingTalk is less than 1%.

Therefore, even with the integration of large models, DingTalk is still not making money.

At the ecological conference, Ye Jun announced the pricing of the large model call service. On top of the annual fee of 9,800 yuan for the professional edition of DingTalk, an additional 10,000 yuan can be added to obtain a quota of 200,000 large model calls; on top of the annual fee for the exclusive DingTalk, an additional 20,000 yuan can be added to obtain a quota of 450,000 large model calls.

Based on the calculation that "the cost of calling a large model once is 5 cents," these packages leave almost no profit margin for DingTalk.

This is another peculiar dilemma in the Chinese SaaS industry.

Like a pendulum, one end is quality, and the other end is high price. Swinging towards quality will inevitably move away from high prices. Being close to high prices means not being of high quality. This is true for SaaS companies and their customers, as well as between SaaS companies.

Even with 17 product lines, and standing at the forefront in terms of scale and technology, DingTalk, which gives 9 yuan to partners for every 10 yuan it earns, still does not dare to talk about "money," the most common word in the business world.

03. Can Ecosystem + AI Solve the Dilemma of SaaS?

In addition to AI, another key term mentioned most frequently by DingTalk is "ecosystem."

Why is the ecosystem so important?

At a time when most enterprises are still in the early stages of understanding SaaS, there are serious communication barriers between the supply side and the demand side, which prevents the release of the demand market scale for SaaS. When the market scale cannot support the maturity of all SaaS companies, competition naturally intensifies.

In this elimination game, some SaaS vendors often take "unconventional" measures to survive, leading to the displacement of good companies by bad ones.

Specifically, when a business requires cooperation between two or more SaaS vendors, the lack of trust and service support prevents them from generating synergies, sometimes even leading to a 1+12 effect.

As a result, the already narrow survival space is further compressed, and coupled with communication barriers between the demand side and the supply side, everyone can only "go beyond their duties" and do things outside their expertise to obtain orders.

For example, CRM incorporates some OA functions, OA incorporates financial and ERP functions, HR SaaS incorporates some OA and financial functions, and vendors that previously focused on single business processes, while not yet firmly established, are rushed into the process of integration due to the unstable nature of their cooperation and competition relationships.

To address these issues, DingTalk has launched "PLG+SLG."

PLG refers to "Product + Digital Base," which reduces the R&D cost of customization demands by opening the digital base to ecosystem partners and customers. SLG refers to "Sales + Service-Driven," aiming to use the open ecosystem to drive the development of the entire To B industry chain. For DingTalk, which insists on PaaS, taking a step back to solidify the underlying value of SaaS services and promote the ecosystem to form a positive cycle.

In simple terms, DingTalk has clarified its "do's and don'ts," focusing on PaaS and opening all the capabilities of the digital base to other players, conveying a new logic to the ecosystem—as long as a business can do one aspect well and deeply, it has the space to survive.

During this process, there is no longer space for "overstepping one's duties," and the complete and mutually trusting chain effect gradually comes into play. As for the communication barriers between the demand side and the supply side, DingTalk can provide consulting services.

Starting this year, DingTalk has begun integrating HR, contracts, finance, travel, marketing, and other applications into a "management toolkit" and then promoting it to customers, which is a practical application of the ecosystem strategy.

To some extent, the Chinese SaaS industry is a united industry, and DingTalk's "ecosystem expansion" strategy actually points the way for the ideal development of the Chinese SaaS industry.

However, on the other hand, in order to truly build the "Peach Blossom Spring" of the SaaS industry, DingTalk faces a difficult problem.

Even with a team of a few hundred people, it is very difficult to sift through over 10 million applications all day long. With the integration of large models, the number of applications is bound to experience explosive growth once again. Furthermore, the quality of the toolkit depends on a significant investment in human resources due to the internal know-how of different industries.

Therefore, although the scale of the DingTalk ecosystem is large and the prospects are promising, for service providers, customers, and DingTalk itself, tapping into its potential is like finding a needle in a haystack. DingTalk is increasingly resembling the interactive interface of an "airplane cockpit," which is a manifestation of this difficulty.

Can AI solve this problem?

At the ecological conference, Ye Jun mentioned a concept. In the future, the toolbar on the left side of DingTalk may only leave a few entry points, and in most scenarios, various applications will be called using natural language input.

During the demonstration, simple commands were input using natural language, and "digital employees" and the "magic wand" were able to call various applications to complete tasks such as posting job postings, screening resumes, and meeting minutes. In other words, DingTalk's vast application reserve now has a more concise way of calling and more equal traffic entry points.

This is also the reason why DingTalk is opening its intelligent base (AI PaaS) to ecosystem partners and customers, using AI to help ecosystem partners redevelop their products.

By integrating AI with the ecosystem, the PaaS strategy has a more solid business loop, and DingTalk is able to move further away from its ambition to "become the foundation of the entire SaaS industry."

But how long can this selfless support last? How long will it take for DingTalk to have deep enough value? Beyond sentiment and ideals, how can DingTalk restore market confidence in the SaaS industry, and what is the next step for DingTalk to provide a more clear growth path?

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。