A few days ago, at the 2023 Ebang Future Retail Annual Conference, Bai Ya, founder and CEO of Youzan, delivered a speech entitled "From Digital Tools to Digital Operation System". He pointed out that the main dilemma faced by domestic enterprise service providers is that customers do not know how to use SaaS tools, have no operational experience, have low willingness to pay, and at the same time, SaaS service providers need to invest a lot of customer service costs. In such a situation, he further stated, ". .There is only one way for China's digital software, putaiAdd it, let the service cost go down, let the cost of sales go down, and let the customer get the result directly.
He notedYouzan has more than 100 billion transaction volume data every year, serving a total of 2 million shopping guides and reaching 600 million consumers, using Youzan's data capabilities and customer service experience, AI can help Youzan's customers automate the problems they face in operation, saving the cost of tool Xi.
At present, Youzan AI's work scope covers the decision-making and execution of positions such as shopping guides, store managers, regional operations and headquarters operations, and can realize automatic review and automatic reconciliation after each activity.
Tips: This article is a preliminary review draft of shorthand, and it is edited and organized by Ebang Power on the basis of not affecting the original meaning.
The following is a transcript of the speech:
I'm still a little nervous standing here, I haven't stood on a stage to share for about 4 years, and I have only shared at meetings where you like yourself in the past 4 years, and I haven't seen any ** in the past 4 years, and I haven't appeared on the stage. When the previous guests shared, I was thinking, what kind of mentality should everyone here use, what do you want to hear, Comrade Jia Penglei has not seen me for three years, I was still thinking at this moment before I took the stage, the old man (Munger) is gone, it really makes me sad.
I try not to talk about the speech like an advertisement, try to talk about it like thinking in the industry, there are people who serve merchants below, and there are more brands and retailers, so I hope retailers can hear how a service provider, a company that has served hundreds of thousands of merchants, how they consider the needs of merchants and what kind of trends they see, I also hope that merchant service providers can see how we do it, and I hope to have some inspiration for everyone.
This proposition is a bit big, and I'll try to be a little more specific:
Differences and changes in the corporate service markets of China and the United States
If we look at the enterprise service market, China has no way to see it, we can only look at the United States, when did the American market rise?Before 2000, the U.S. enterprise service market basically only had two parts: customized software and generalized software. By the time of the 2000 financial crisis, there was no growth for two years, and the reason for the lack of growth was that the company did not grow during the 2000 financial crisis, so the company did not have an IT budget, so it would not pay software vendors and consulting companies. After 2000, after the financial crisis passed, business started to improve, so we had an IT budget, and we started to grow again.
However, before SaaS enterprises and SaaS software systems, there are more generalized software and customized software. After the 2000 financial crisis, all U.S. companies only bought SaaS software, not custom software and general software. The reason is very simple, not SaaS can be hot-updated, SaaS software ** as a service, etc., in fact, it is not mainly this, there are only two main reasons: the first reason, the old traditional software and customized software have no way to connect to the Internet, can not adapt to the needs of the Internet era;Second, just after the financial crisis, all enterprises have limited budgets even if they resume growth, and they are holding their wallets to make purchases, and they find that they pay $50,000 for 5 years to buy a general software, and they only need to pay $10,000 a year to buy SaaS, so they buy SaaS. So, one is that SaaS software can be networked, and SaaS software pays every year, not a one-time payment, so SaaS starts to grow.
In the past 13 years, there has been only one best track in the entire U.S. Internet market, and this track is the SaaS track, which is the track that provides business systems and operation systems for enterprises, which is the past 13 years. You can take a look at what other companies in the United States Internet companies have grown rapidly in the past 13 yearsOnly SaaS is going up like crazy.
If you went to Silicon Valley at this time last year, 80% of the billboards in Silicon Valley were SaaS billboards, because they rose very well. Of course, because AI appeared during the Spring Festival this year, the growth of AI is even more crazy.
There are two years of non-growth, that is, from 2008 to 2010, also because of the financial crisis, but after each financial crisis, SaaS will soar once, this is the American market.
China is now equivalent to the United States in 2000, so during this period the Chinese SaaS market is very miserable, of course, there are many reasons for the misery, for example, Chinese merchants will say, I use the merchant's background and merchant tools on any platform are free, why do I pay for a third-party tool?I've already given you face by paying you, why don't you kneel down and serve?Therefore, various reasons have led to the very poor SaaS market in China, and it is difficult to use it well, because there is no operation process and operation process, so it is very poor.
The negative cycle dilemma of China's corporate service providers
Say it and praise it. We started by defining ourselves as a merchant services companyWe do three main things: First, we help retailers and brands acquire new customers through content marketing on social networks, and turn these customers into fans and friends in the private domain. Second, for these fans and friends in the private domain to do repeat purchase operations, such as distribution, intelligent systems, digital systems, shopping guide management, etc., all for the repurchase and distribution of old customers, to bring new customers. Third, many things in the old system are not networked, and if the inventory, member stored value, and membership identity of the old system are not communicated, so basic software such as offline cashier, ordering, and member stored value are made.
In the past 11 years, only these things have been precipitated, usually the store has attracted some fans in the social network, or when the customer arrives at the store, the shopping guide and the customer add friends, after adding friends, you need a CRM system to analyze the characteristics of friends on WeChat or fans on Xiaohongshu, and give you some suggestions for marketing planning, and these marketing planning suggestions are finally accurately touched to fans in the form of enterprise WeChat or **message, and fan transactions may be effective in independent mini programs**, or in groups** In the transaction, it may also be more likely to be tried and repurchased in the store, so we have made a store system and an offline system for the beauty industry.
How hard is this?In my opinion, it is actually very difficult, I can give you an account, we have a total of 20,000 functions for these 7 products, each function needs 25 man-days when it is developed online, and as long as someone uses it, there will be bugs, and as long as someone uses it, there will be new needs to do fine-tuning and adjustment, and it will take 25 man-days. In today's talent market, the cost of management, administration, salary, and social security adds up to 3,000 yuan per person per day. In the past 11 years, we have spent 3 billion on research and development, and none of these 20,000 functions have been done wrong, and we need to spend 3 billion without any detours. If I don't do anything with these 20,000 functions today, only maintain bugs and do small fine-tuning, for example, you can set 3 configuration items, and you want to change them to 5, these 20,000 functions need at least 300 programmers, and you need to spend a minimum of 1 a year500 million. And in today's current situation, it is basically unlikely that there will be any investors in China who are willing to spend 3 billion yuan to burn a company. Spend 1What is the concept of 500 million-200 million R&D expenses?R&D expenses may account for 10%-15% of this company, and a company that sells systems needs more than 500 million yuan a year to make a profit.
Why do I say this?Most of you here are brands, retailers, I especially want to say a word for those who do corporate services, please still cherish the current or corporate service companies, because one less one, there will be no new corporate service companies in the next 2-3 years, they will spend hundreds of millions of dollars to make a system for you to use, a little bit to get the money back, no one votes for them, they will live very hard. Today, 50% of China's enterprise service companies, even the sales unit has no gross profit, 30% of the enterprise service company sales + service is no gross profit, more than 95% of the enterprise service company sales + service + research and development is no gross profit, today in China more than 100 million US dollars of revenue is still profitable enterprise service companies do not need a hand at all, and now in this situation, if everyone is unwilling to pay, unwilling to renew, these companies can only go, go one, one less, there will be no new addition. So, all the brands and retailers should cherish the companies that provide you with the system, they are living very, very hard today, much worse than their American counterparts.
We made 20,000 features, what is the concept of it?After we did these 20,000 functions, we found that product research and development is difficult, but the most difficult thing for enterprise service companies is not product research and development, marketing and sales, because Chinese merchants and brands are not willing to pay for IT, and what is more difficult than marketing and sales is service and operation, because this system has no professional managers, no mature talents, no mature management system, even if he buys the system, it will not be used, it is very difficult to use, and all businesses will encounter a lot of problems.
Our product micro**, a total of 6000 functions, there are more than a dozen first-level function navigation, each first-level function has more than a dozen second-level function navigation, second-level function and ** function navigation, and four-level, must be used in this way so that everyone can find the relevant function. Therefore, the most difficult problem faced by all enterprise service companies is that we must first find the top company, which can afford it, is the number one leader, has the resources, and has the basic conditions, and provides them with consulting and operation on behalf of them, so as to help them do it and accumulate successful experience and best practices. Then we need to make the best practices into topics, into teaching courses, online and offline courses, to train all the merchants who buy our system, and teach them how to use it.
Even so,The brand will encounter three problems: first, there is no idea and does not know how to operate;Second, there is no manpower;Third, we're all busy. This problem cannot be solved at all, and what is even more difficult is that even if I come to do training, the company is not willing to pay a penny, he thinks that I have given you face by buying your software, and you should come to do training. My training, teaching, and services to help merchants succeed have turned into costs. So this problem is very difficult, whether it is operating a SaaS system, an HR SaaS system or a marketing SaaS system, it is very difficult, because the enterprise is not willing to pay, the enterprise does not understand, and we still have to provide services. This problem has been challenging human nature, so it has been in a negative cycle, so China's enterprise service companies have not been able to develop. These 20,000 functions are all developed after research based on the needs of merchants, but these 20,000 functions are really difficult to use here, really difficult, really very complicated, this is the status quo.
We've been trying to solve this problem, but we haven't been able to solve it because it's very difficult. We tried a lot, but we couldn't solve it.
03 aiIt is the only solution for China's enterprise service providers
Last October, a friend of mine had a failed startup, and he confided in me that he used to work in Seattle and was a very good technologist at Microsoft. Later, one year, Professor Zeng, Wang Jian, and Mr. Ma went to the United States to kneel down and beg many people to come back, entered Ali's Damo Academy, and after staying in the Damo Academy for a few years, they came out to start a business and do AIGC. They can analyze the ideas of car companies, analyze the thoughts of car buyers, and use AI to write public relations and promotion for car companies. This technique is very difficult, but it is used in the wrong place.
I started focusing on AI and intelligence in October last year, and at the end of last year I went out and stayed abroad for four months. I talked to all kinds of people who were doing AI, and I was still worried, but when I found them, how should I open my mouth?Because I'm not doing AI. As a result, when I went, the AI became very popular. Anyone I meet, I don't have to say anything, they can talk to me about AI. I was in a city basically every two or three days for the past four months, and when I came back, I started working on our AI project. We don't know if today's big models will work and produce meaningful content. Even though I tried for eight months, I don't dare to let AI do pre-sales customer service now, the reason is simple, a bot as a customer service, the customer asks if it can be discounted?He said what could be done, because he said he was responsible, so he needed to control it. It was only after eight months that we started launching various products to show you:
For example, an offline brand retailer, you have a lot of stores, and the store has a lot of shopping guides. Today, we have probably served tens of thousands of brands and more than 2 million shopping guides. We have served more than 2 million shopping guides for all brand owners. There are many common scenarios for shopping guides, such as shopping guides are very concerned about whether my sales target for this month has been achieved and how far away I am from the goal. Another example is that the shopping guide said that I want to improve my performance, what can I do? The needs of shopping guides are not very complicated, but some of the most common questions, such as how my performance is and how much bonus I can get this month. The shopping guide can directly ask the AI in the shopping guide assistant, for example, how is my performance, and what can I do if I want to improve my performance. The AI can give him some advice, such as telling him how well he is doing this month, how much he is still missing, and whether it is possible to achieve his goals at the current pace. And give him some advice, such as picking out 20 WeChat friends you can find to chat with. If there are other suggestions, the AI will also tell him, such as that there is an event in the company now, and he can participate.
For example, there is another scene for shopping guides. The customer came into the store to try on the clothes but didn't buy them. AI can alert the shopper to tell them which garment the customer tried on but didn't buy, and that the garment is out of stock. Shoppers can recommend similar styles to customers, or tell them there's an event going on, or apply for an in-house price. However, shoppers don't know how to contact customers. Only 1% of shopping guides know how to deal with this situation, and 99% of shopping guides don't know what to do at all. As a result, AI can advise them.
In addition to the shopping guide, there is also the store manager or the manager of the area. They need to look at the performance of the area and see how the best shoppers work, so they can teach other guides. The store manager has a lot of other work to do. For example, store managers can also talk to AI to see if their performance has been completed and whether their progress is in line with requirements. The store manager can also view the report to understand the ranking of the store's 8 shopping guides, and the salary and commission they can get this month. The store manager can see the best shopping guide through Enterprise WeChat, because for which link is done well, what products they recommend to customers, and what they say, they can list it to the store manager, and the store manager can use these to teach other shopping guides.
Most of the merchants belong to the chain retail chain of the brand, such as those with hundreds of stores. So, usually there will be an operations team at the headquarters. Operations teams at brand headquarters or retail headquarters are more in demand. They need to understand why so many people try on clothes and don't buy them, and how to optimize their marketing campaigns.
For example, if I want to develop a marketing strategy, I just need to tell the AI. For example, I want to know how I'm doing, and the AI will tell me how I'm doing, as well as tell me why it's good or why it's not. If I want to know how to double the repurchase rate of my old customers, the AI will tell me how to do it, and I can optimize those methods. Once the method is established, the AI will help me execute it and supervise the entire process. For example, if I set a target of 30 million for November, how much more I have accomplished now, what else can I do, what are some ways to improve, I would like to hear your suggestion and we can have a conversation. After the conversation, the AI can automatically complete the corresponding task and supervise me. For example, if there are still 1,000 pieces of stock that need to be emptied, I can ask the AI what can be done, and the AI will tell me to discount and notify those regular customers who need it. If you can, you can also automatically change the price and send a message to the customer. All of this can be done automatically.
The scope of work of AI includes shopping guides, store managers, regional operations, and headquarters operations. Moreover, after each activity, the AI will automatically help me review the whole process. The AI will tell me what I did well, what I didn't do well, and give some advice. The review process contains all the data and reports, and each cycle can generate a report and send it to me.
We found something particularly interesting. In today's Chinese chains, if I have 200 stores, I usually need to have at least 20 finance staff to support the reconciliation of the 200 stores. And after each event, the reconciliation work is very headache. As a result, business reconciliation is very difficult. We're doing automated reconciliation. The reconciliation needs of each enterprise are different, and the caliber of each enterprise is also different. As a result, there is no one-size-fits-all software that can be used by all businesses. However, with AI, we just need to tell it what we need and it will do the reconciliation work automatically.
To summarize the process briefly: I understand what you want, and I tell you directly how you can do it. Based on your specific situation and industry best practices, we will plan and advise you like a consultant.
But it's hard to do because I wanted to do it three years ago, but I found it couldn't. It was only after we had AI that we started experimenting. It took us 8 months to find out that this can be truly achievable. Why is it possible?The reason is simple, we have been doing it for 11 years. For the past 11 years, any customer who has bought our system has had a service called "Delivery Service". So we have a large number of successful customers who will deliver to customers who have bought software.
What does it mean to be delivered?That is, he uses the system, has a certain transaction amount, and then calls the delivery completed. All completed clients will have an internal closure report. We have a lot of closure reports. Every two or three weeks, customer success experts get together to do a workshop with the best people in the industry, and each workshop writes a manual called the Customer Success Handbook.
This manual is detailed to how to choose the location of a bakery, how large, how many people to recruit, how to set the price, and what kind of online operation work to do, including how coupons can be issued, including how members can be divided, including whether to remember whose birthday is the cake when remembering members, whether to count what kind of words to be able to give the person who buys the cake so that he is willing to say whose birthday it is today, whether it is the lunar calendar or the solar calendar, how many people are still in their family, and what specific actions to do. Our manuals are very, very detailed. There are also a large number of customer manuals for customer service training. We often encounter merchants who call customer service, for three or four hours, asking me what I should do, what I can do, how to set it up, and teach them to do it by hand. In the past 11 years, we have accumulated a large number of databases and knowledge bases. We put a large number of knowledge bases and databases into the model, and then train the output results of the model.
These massive databases, knowledge bases, are not dead dataBecause I have more than a year100 million transaction volume, the participating merchants are closebillion, served10,000 shopping guides, reach100 million consumers. All data is updated in real-time on a weekly basis. That's why you tell me what you want to do, I have your situation, my peers' situation, I suggest you do it. Once you've optimized the movements, we're going to perform all the actions.
It's very simple for us to perform all the actions, and it's a matter of base. The hardest part is the process in the middle, which is the core thing that has been done for the past 8 months or so. That's our answer. What to do with the system of enterprise service in China, don't make old software, pile up so many functions customers will not use, you still have to engage in services, he will not pay. It's a dead end. Chinese software can only add artificial intelligence to the old software system, so that the customer does not have to learn or think at all, but only needs to tell you what he wants to do, and you can help them do automated execution. This is all our thinking, we hope that you can solve problems without Xi, and you can be inspired and grow in the process of using it.
Because of the changes over the past year, I feel like we're not a software company anymore, we're an operating company. Because we're operating an automated system. All of our work is to run the system. Customer success staff spend the next half of their time maintaining the system's output more accurately and effectively. Customers just need to use this system. Here's how we solve the problem.
Therefore, China's digital software can only go this way. Do you know why the United States has such a large SaaS market of more than 100 billion US dollars, and half of these SaaS companies are engaged in means of production, that is, tax filing, etc. China's enterprise service software hardly does the means of production. Because the means of production are in the hands of state-owned enterprises. China's enterprise service software and private enterprises are making production tools. Today, we are talking about enterprise digitalization, we are talking about retail digitalization, and the service providers of retail digitalization are not valuable, because digitalization is only a production tool, and production tools are not worth anything at all. Chinese companies are also reluctant to pay. So there is only one way for China's digital software, putaiAdd it, let the service cost go down, let the cost of sales go down, and let the customer get the result directly.
What do we do digital software like?In the past, when you used the so-called stand-alone version of the software, it was a bit like holding a notebook and keeping accounts. With the stand-alone version of the software, it started to be like running. With digital software, it looks like I'm going to ride a bike. But there are two problems with bicycles: running is free, bicycles have to buy bicycles, and you have to learn to ride them. So it's hard to get digital software in China. But with artificial intelligence, it's a self-driving couch. Sit on, you say go** and go. You don't need to learn, and it's very simple. Even in the future, it may not be paid upfront, but based on results.
So, my answer is that only artificial intelligence can save all the digital software service providers in China. Otherwise, everyone will not be able to survive in the future. Thank you!
About this meeting::
The 2023 Ebang Future Retail Conference is one of the theme conferences of the 2023 Ebang Annual Meeting, which was held on November 29 at Pudong Shangri-La Hotel, Shanghai.
The 2023 Ebang Annual Meeting is hosted by Ebang Power, co-organized by Horseshoe Club and Ebang Think Tank, with the theme of "Perfect Strategy for the Shift Period". For the first time, the agenda of the annual meeting covers the whole cycle of enterprise digitalization: domestic e-commerce, brand globalization, and industrial digitalization. Through the traction of the theme of "Perfect Strategy", 64 heavyweight guests conducted in-depth research on 54 hot topics, actively sought responses to a series of key issues in 2023, and jointly explored the way of growth based on short-term reality and adhering to long-termism.