On March 23, at the "Data·Decision·Future - Taobao University Business Data Application Training Camp" hosted by Taobao University, Zhang Liaoyuan, founder of Three Squirrels, delivered a干货 sharing, introducing to participating enterprises how Three Squirrels uses data for supply chain management, marketing development, and customer service. In Zhang Liaoyuan's view, data will become an open resource available to every enterprise in the future, and the most critical thing is to use data to serve corporate culture. The following is compiled from Zhang Liaoyuan's speech: I sell nuts and snacks, so if I talk about big data, it would be impractical. Therefore, today's topic is "Small Data, Big Application." Why small data? Applying a small data set to the extreme is a competitiveness we have compared to traditional enterprises. The most important thing to understand small data is: what is small data? In my view, (small data refers to) the reviews of every user on Tmall, as well as the data provided by Tmall and Alibaba's backend, including Data Cube and Business Advisor, and data from social internet. This small data enables us to become an enterprise that can manage supply chain through data, become a company with more creative design and marketing through data, and reflect our core value through data—treating customers as masters. Through these simple data, we become a truly consumer-centric enterprise. Everyone in e-commerce knows these conventional data. Actually, I think the data we care most about is the Tmall store score. You may know that each Tmall store has three scores. When I founded Three Squirrels in 2012, I believed these three scores were a revolution for traditional business. At that time, I firmly believed that by deeply mining these three data, Three Squirrels would become a company with full data management and operation. Since the first day we went online, we have maintained industry leadership in these data. The core point of data service is to become a consumer-centric enterprise. So the first thing I want to mention is brand search terms. In the past few years, nearly 10 million consumers searched for "Three Squirrels flagship store" when shopping on Tmall. Brand search terms are also the most important KPI for our brand department. Because in my mind, only when your brand term enters the user's mind can you be considered a brand. Many of our marketing actions, concepts, and strategies revolve around whether users search for the four characters "Three Squirrels."
Not a Brand Company, but Building an Industry Chain Platform Our company has been very simple for five years, doing only two things: first, provide owners with the most cost-effective and safe products; second, bring love and happiness to owners. Therefore, we redefined the customer service role. I think our most significant change is paying close attention to user review data, and based on this logic, we develop products and take marketing actions. Many times people say we are an online retailer, a company selling nuts, but internally we position ourselves as an industry chain platform driven by data management. We do not position ourselves as an FMCG company or a brand company. In fact, we are building an industry chain platform to reshape the relationship between producers and consumers. This is a screenshot from our central quality control system. In this picture, you can see that any consumer's reviews of a product accumulate to form an index, and the quality of the producer's supply forms an index. This is what we can do now. For example, the logistics index goes without saying, and the owner's complaint resolution rate is a key focus. The company has an internal network platform where this data is published daily. Any owner's dissatisfaction should be resolved in five steps: first, apologize; second, improve; and then this cycle. Although this closed loop cannot be 100% complete, we can achieve 80%. This is a simple demonstration. Full data, meaning the capture of all data from raw materials to consumers, is a goal for the next two years. Currently, we have achieved about 60%. We have already done the producer part, but not the raw material part. However, we are piloting it. For example, we have pilots for the temperature at the origin of pecans, but not fully, and we have not captured every data point from supply to consumer. The second part is the full chain from raw materials to consumers, divided into segments, forming a form and a traceable process. Currently, we have achieved about 70%. Once this segmentation is established, any information in consumer comments, such as too many empty pecan shells, can be traced to which segment's problem. Service can also be traced to which customer service representative's issue, truly making user reviews a closed loop. We hope to become an industry community. Three Squirrels is just a platform. Three Squirrels and all producers jointly serve consumers. What will truly be realized in the future is the integration of production and consumption, just like the recent national discussion on virtual economy and real economy. In the future, these two will become one. Industry community and production-consumption integration—where producers and consumers truly become one. In our industry, we believe the C2B era will come, with truly personalized customization based on batch production.
Using Data for Product Development We say redefine marketing, which is about how our data is applied. We found that on Taobao, there is an index for spicy strips that is growing rapidly. We analyzed that 2015 was 30 times that of 2013, with a peak of 500 times. So we immediately captured this data from the Taobao backend for analysis. We wanted to analyze the users who buy these spicy strips and their perception, so we also captured reviews and found that users post on Weibo using spicy strips for self-mockery and self-deprecation. Later, we understood that this is also cute. Cuteness is not just about being adorable; being dirty can also be cute. So we saw these signals and quickly developed a product—Three Squirrels Spicy Strips. Why did this project inexplicably focus on spicy strips? Because it is a project to make a traditional product fashionable. When they showed it to me yesterday, I said you are really dirty to the extreme. This product comes in a pack of 20, with 7 states, seductive spicy, and each pack is different. This product has generated a lot of sharing. Now, a simple Baidu search for Three Squirrels Spicy Strips yields many results. There are several requirements for making traditional products fashionable: first, quality optimization and taste upgrade; second, it must have its own upgrade and traffic. If you can't make such a product, it's better not to do it. This product sold 4.81 million units on Tmall in seven months, becoming the number one spicy strip product on Tmall. Total channel sales exceeded 100 million yuan. More importantly, it brought 1.2 million topics of communication, all from just one product. So in product development, we follow this logic: develop products with a user-friendly approach.
Using Taobao Index to Measure TV Drama Placements Only the Taobao Index can truly measure the placement level of a drama. If the index rises at the moment of your ad placement, it proves consumers have perceived it and taken action. Their central nervous system sent a command to their fingers to buy, buy, buy. "Ode to Joy," which featured Three Squirrels ads, premiered in May last year. That month, we sold out and reached 50 million in sales. That's the effect of "Ode to Joy." We have our own standards for choosing TV dramas for placement. Are dramas by big directors necessarily good? Not necessarily. Are dramas with big stars necessarily good? Not necessarily. There are several dimensions: script 25%, director 25%, actors 25%, and the match of the broadcast platform. Match is very important. Consider whether it fits the brand's tone and characteristics, and also whether the brand's tone and the drama's tone have topic potential. This is the foundation of choosing a drama. In 2016, among the top 10 hit dramas, we were in 3 of them. Choosing a drama is a bit like venture capital; there must be a set of theories and data support behind it. When done well, it can leverage four ounces to move a thousand pounds. By the way, growing brands are particularly suitable for TV drama placements. For a mature brand, it's unnecessary because consumers already know you and will think it's an ad. But for a growing brand, placement is cost-effective. Also, don't forget to have a Tmall store, because after watching, they will buy immediately; they won't go to a supermarket to find it. This is how to use data for TV drama placement.
A Chat Company That Uses Data to Act Cute Earlier, I talked about redefining the relationship between consumers and brands. We don't treat customer service as salespeople. First, customer service representatives are chat agents. Second, they should chat well with each owner. Third, they should ideally integrate their personality with the squirrel persona. Our customer service has nine major services. Let me mention two. There is a "wake-up service," which is actually a wake-up call. Many young people can't get up in the morning. No problem, tell us, and we'll call you. We have many such services. We have established data-driven customer service assessment, evaluating some indices. For example, the number of words per reception, which refers to the average word count per customer service representative. Because we adopt a customer service model not aimed at sales, we don't pursue sales targets but focus on whether users and customer service chat comfortably. For example, customer service has a vocabulary library. They should call themselves "Mouse So-and-So"; don't say "I serve you," but "Mouse So-and-So serves you"; and try not to say "you" but "master" and other words. There are many more, such as how to say when sending logistics. We try to use spot checks. For example, if a customer service representative chats with 200 people a day, we spot-check 10% of the chat records. The evaluation standard is whether the chat has squirrel personality. Our brand positions customer service not as sales but as user maintenance. Using scenario-based drama and chat modes, each customer service representative plays a little squirrel. Over time, some consumers will really imagine the customer service on the other side as a squirrel. Our company uses the popular term "two-dimensional" (2D), but actually we turn the company into a 2.5-dimensional world, bridging virtual and reality between 2D and 3D, so our users are greatly influenced. Last year alone, we received 15 million masters, chatted with 1.5 million masters, and each customer service representative averaged 185 receptions. On average, they could act cute 90 times a day, tell about 10 jokes or meme battles a day. There are quirky customer service reps, some who use classical Chinese, some who play court dramas, etc. Maybe one day a master calls a customer service rep "heroine," then it's a big deal; we have to play the hero. Much of what I've said is based on data, corresponding to three aspects: supply chain management, marketing behavior, and service model. Every brand cannot escape these three aspects. And we are lucky to have encountered the internet and the data it provides today. Today's topic is "Small Data, Big Application." Talking about big data is too far for a nut and snack seller. Those artificial intelligence things are also far from us. We have just been doing one very simple thing for five years: using small data as well as possible to serve our consumers and our enterprise. That is our original intention. In the future, I believe that under Alibaba's big data, data will become a basic service provided to us. So data is not the core competitiveness in the future. For the real economy, Alibaba, BAT, and future data companies are like building a machine for us to use. At that time, what we need to do is run a little faster than other enterprises. I think the real strength of Three Squirrels is our cultural system. Our cultural system includes the Squirrel Moral Convention, Shennong Hall, and Shuangshu, etc. These things exist, and we put cultural values first. Our data actually serves or promotes our cultural values to build today's Three Squirrels.