share_log

马云颠覆阿里:我没有答案,但我知道方向

Ma Yun disrupts Ali: I have no answers, but I know the direction

深網騰訊新聞 ·  May 8 16:12

Source: Shenzhen.com Tencent News Author: Cheng Xiaoyi

User first,$Alibaba (BABA.US)$A regulation recently announced implements the company's latest strategy and principles. According to official sources, Tmall has finally cut down the pre-sale mechanism that has been criticized for a long time, and the 618 campaign will begin directly at 8 p.m. on May 20. Those complicated offers and lengthy pre-sale times in the past will all be a thing of the past.

Twelve years ago, after Tmall first introduced the pre-sale gameplay, the entire industry quickly followed suit. Merchants can know in advance how much they want to sell to reduce inventory accumulation. The platform can also lock in orders in advance and seize the market. It sounds great, but for consumers, the experience isn't that great.

“Over the past few years, when we looked internally and reflected ourselves, we knew that Ali was behind because we had forgotten who our real customers were.” Alibaba Chairman Cai Chongxin confessed.

Who is Ali's first customer, a merchant or a user? Once upon a time, Ali paid more attention to sellers on the Alibaba platform. Cai Chongxin reflected, “Customers are people who use the Alibaba App to shop, and Ali didn't give them the best experience.”

Cai Chongxin admits that Ali's morale has not been good for the past three years. “Employees are looking for a direction, and if you can clearly communicate what that direction is, then they'll get morale back.”

From sellers to users, from new retail to AI, from diversified development to focusing on the main business, from spin-off to joint efforts. One year of Ali's transformation is also the year that founder Ma Yun returned to China. Ali has almost disrupted the important strategies of the past ten years, and Ma Yun is deeply involved in every key point.

Starting with Ma Yun's return to China, he appeared in Hangzhou on March 26 of last year. Two days later, Ali began a “1+6+N” spin-off. After only two months, the board list of directors of Ali's Big Six Group confirmed that Ma Yun indicated the direction for Taotian Group and was personally responsible for resolving any resistance to change.

Pinduoduo's market capitalization surpassed Ali's eve. Ma Yun's reply pointed out that the AI e-commerce era had just begun; on the first anniversary of Ali's transformation, Ma Yun even used a thousand-character internal letter to summarize statements about the reform.

When the founder's energy returned, the decision became more “decisive”.

At first, it was possible to “activate” the heavy and huge Ali Group and simply break up the group in a clean way; now it is also possible to resolutely abandon physical retail, All-in AI, which has been deeply cultivated for nearly 10 years, and reunite the group that has just been split up for the purpose of a new AI-centered group strategy.

It's too expensive to miss a big race track. The big model is reshaping all walks of life, and the step by step pressure from Pinduoduo and Douyin may have to make new trade-offs between Ma Yun and Ali's decision makers: if Ali wants to be great again, he must find new trends in the times.

Last year, Ma Yun asked every business segment of Ali the same question: How to understand AI e-commerce? Ou De Zhang, a former Alibaba employee, revealed on his official account, “Teacher Ma said, “I have no answers, but I have a direction; you want to give me answers.”

AI base: big model “blood transfusion” Alibaba Cloud

Ali's actions in the AI field can be described as a big deal. On the one hand, Alibaba Cloud is spearheading the price war, and on the other side, the group invests to take over the “Five Little Dragons” (that is, five large domestic model companies with a valuation of over 1 billion US dollars: Dark Side of the Moon, MiniMax, Smart Spectrum AI, 10,000 Things, and Baichuan Intelligence).

“It was too expensive to miss out on this big track (big model).” An Ali employee expressed his appreciation for “AI Lightyear.”

As Ma Yun said in an internal letter, the time span of three to five years is like a century long for the Internet field, enough for drastic changes.

An Ali employee told “AI Lightyear” that the Alidamo Institute was already doing multi-modal technology research and development as early as 2019, but it wasn't called a “big model” at the time.

“In fact, there is some internal disagreement about general artificial intelligence, and the internal situation fluctuated for a while.” The Ali employee mentioned above said that after a period of time, the initial results fell far short of everyone's expectations. “The biggest problem we faced at the time was that the investment was too high, the commercialization contribution was relatively low, and that account was not fair.”

Today, although in addition to the big domestic model “Five Little Dragons,” Ali also has self-developed AI products such as the Tongyi series (Tongyi Thousand Questions, Tongyi Zhiwen, Tongyi Spirit Code, Tongyi Wanxiang, Tongyi Hearing, and Tongyi Stars), but it seems like the drunkard didn't mean alcohol.

A number of industry insiders told “AI Lightyear” that it is more likely that Ali's big model is to promote Alibaba Cloud. “GPU with cloud is the industry standard style of play.”

“AI Lightyear” has learned that one of the conditions for Ali's investment of nearly 800 million US dollars in the Dark Side of the Moon is that the Dark Side of the Moon needs to purchase Alibaba Cloud computing resources. Previously, Volcano Engine and Alibaba Cloud were the main providers of Kimi's computing power for the Dark Side of the Moon product. After completing this year's A round of financing, Kimi's future reasoning accelerated or mainly relied on Alibaba Cloud.

In this wave of large-scale model development, compared to diverse application ecosystems, major Internet companies with more resources prefer to focus internal resources on the AI foundation. This not only continues its own advantage, but it is also very reasonable from the perspective of the business model.

Silicon Valley venture capital agency a16z discovered in an investigation at the beginning of last year that large amounts of capital in the generative AI market eventually went to infrastructure companies such as cloud providers or third-party model providers. According to a16z estimates, 10-20% of the total revenue from generative AI goes to cloud providers.

In addition to this, startups that want to train their own big models basically spend 80% to 90% of the financing amount on cloud service providers.

This can be seen from the growth of Microsoft's cloud service Azure. In the fourth quarter of 2023, Microsoft Azure narrowed the gap with Amazon AWS, the largest cloud cluster, from 12 percentage points in 2021 to 7 percentage points in 2023. AWS's global market share in the fourth quarter of 2023 was 31%, and Azure was 24%.

CFRA analysts predict that in the first quarter of this year, AI will contribute 6 to 8 percentage points to Microsoft's smart cloud revenue growth.

In addition, ToC's Alibaba Cloud Drive is also quietly competing for the personal cloud disk market with the support of the big model. Commercial IP consultant Zhang Ji said that Alibaba Cloud Drive is very fast. Coupled with the Tongyi Understanding video-to-text service, it forms a closed loop, “an experience that perfectly surpasses Baidu Netdisk and Feishu's ingenious plan.”

Other netizens also discovered that after connecting to the Tongyi Thousand Questions model, DingTalk now sends videos on the chat interface without any active action, and can automatically enter AI analysis (DingTalk Flash).

The future of e-commerce: Ali is afraid of missing out on the next era

From shelf e-commerce to content e-commerce, Ali has already fallen into a passive state due to slow people one step at a time. How to seize opportunities in the AI era is a difficult problem that Ali must overcome if it wants to maintain its leading position in the e-commerce industry.

The e-commerce price war has continued for more than a year. Judging from the data, content e-commerce represented by Douyin is superior.

According to the latest monitoring data recently released by Nielsen IQ, when overall omni-channel sales fell 0.04% year on year last year, content e-commerce sales surged 55.9% year on year, and comprehensive e-commerce fell by a huge margin, down 5.8%.

Douyin's GMV (total product turnover) increased 277% year-on-year in 2023, and the regular customer penetration rate has surpassed Tmall and Jingdong supermarkets, second only to Taobao and Pinduoduo. Kuaishou also became the fifth e-commerce platform last year after Taobao, JD, Pinduoduo, and Douyin with total transactions exceeding trillion dollars.

The momentum of content e-commerce continued from last year to this year. During the Spring Festival from January to February 2024, the growth rate of online sales in the Chinese FMCG market reached 20.7%, and the offline market growth rate was -6.4%. Among them, the biggest increase in Douyin e-commerce was over 70%, and the increase in comprehensive e-commerce was 4.4%.

Even in overseas markets, Douyin's content e-commerce can also be used.

Earnest Analytics analyzed credit card transaction data and found that since Douyin launched the TikTok Shop in September 2023, more than 11% of American households have purchased products through the TikTok Shop.

TikTok Shop's performance even surpassed that of many social content platforms in the US. The report shows that in February of this year, TikTok Shop's sales accounted for 68.1% of social shopping GMV, the live shopping app Whatnot accounted for 31%, while the combined GMV of Instagram Checkout, Facebook Shop, and Flip App was less than 1%.

As Douyin is cheaper than Pinduoduo, there are more and more discussions. “Shenzhen Internet” discovered that Taobao's advertising content on various channels has already begun to focus on “Taobao is cheaper than live streaming shopping,” and its low price magic weapon is 1688, which has already entered the market.

The products in the first batch of 1688 stores (1688 selected Taobao store, 1688 company selected Tmall store, and 1688 industrial expert selected Tmall store) were all directly supplied by the original manufacturer in the industrial belt. They support free shipping for 1 item, 24-hour delivery, and 7-day no-reason returns.

In order to win back consumers, Taobao did not hesitate to spend 10 billion dollars in skits. Recently, Taobao released the “10 billion Taobao Drama Plan”, which includes investing 10 million dollars in customized dramas and over 1 billion in traffic support; and collaborating with popular IPs, celebrities, and leading artists to create skit content.

The goal is for Taobao content e-commerce's GMV to increase 80% year-on-year in the new fiscal year, 100% increase in user volume, and 100% year-on-year increase for anchors with monthly sales exceeding one million. Since the second half of last year, Taobao has launched more than 20 short dramas.

According to eMarketer, a world-renowned market research agency, global e-commerce growth will gradually slow down in the next three years, which means e-commerce competition will become more intense in the future. AI has become the magic weapon that almost all retail leaders hope for growth.

According to eMarketer research, nearly half of retail CFOs plan to use AI to optimize costs in 2024. On the one hand, AI has the opportunity to solve the pain points of online sales of second-hand products, such as verifying the authenticity of products, handling inventory, and simplifying processes.

On the other hand, AI allows retailers to more accurately find consumers who need products. According to Coresight Research, approximately 71% of US retail decision makers have invested in data and AI-enabled content to personalize.

Looking at home and abroad, retailers' application of AI is still optimizing past functions, and there are still no disruptive uses transformed by AI in the e-commerce field. However, the addition of AI to cloud business is still the same, and no one dares to underestimate the potential disruptive power of AI on e-commerce in the future.

edit/lambor

The translation is provided by third-party software.


The above content is for informational or educational purposes only and does not constitute any investment advice related to Futu. Although we strive to ensure the truthfulness, accuracy, and originality of all such content, we cannot guarantee it.
    Write a comment