Click to read the original article for details Digitalization, of course, requires data. So where does data come from? Some businesses are naturally online, such as platform e-commerce, O2O, and community group buying. Being online generates data. However, traditional enterprises are primarily offline, and in most cases, users are not online, and it will be difficult for them to be online for a long time in the future. So how can traditional manufacturers achieve marketing digitalization? Some have proposed the goal of "five online" aspects: employees online, products online, users online, marketing online, and transactions online. This is an ideal goal, not reality. If users are not online, the other online aspects lose their meaning. For traditional manufacturers, getting users online through offline channels may take a long time. But marketing digitalization cannot wait. Look at the difficulty of B2B adoption. After six or seven years, the user online rate is still very low, which shows that digitalization for traditional manufacturers is definitely different from digitalization for internet companies. The 2C digitalization model does not work in the 2B field. Even Hema Fresh, which does well in new retail for offline stores, only has about 60% of orders online. This article aims to convey: how can traditional manufacturers achieve marketing digitalization when user online rates are low? -01- Where does data come from? Data sources are diverse; it doesn't necessarily require users to be online, nor does it require the manufacturer's own users to be online. From the perspective of data sources, there are several pathways:
First, use technical means (e.g., web scraping) to capture data from the internet.
Digital R&D and digital communication typically do not use the online data of one's own users but rather use technical means to scrape data from the internet. There is a vast amount of information on the internet. To obtain valuable information (including data), there is a technology called web scraping, which is a program that automatically captures information from the World Wide Web according to certain rules. The search engines we commonly use are general-purpose crawlers, but they are often only used for keyword retrieval. Focused crawlers can target specific web resources and can capture information (data) based on semantic information (rather than keywords).
Second, internal staff capture data manually or through technical means.
For example, in the process of deep distribution, how to obtain terminal display (2B) data? In the past, it relied on manual store visits and then entering data into the system. Now, with AI object recognition technology, it is possible to take photos at the terminal and then use AI object recognition to "calculate" the terminal display, including both your own and competitors' displays. If the capture interval is appropriate, it can even "calculate" terminal sales volume.
Third, offline transactions followed by users going online.
The most common method now is one product, one code. After purchase, users scan the QR code to go online. This is not an online transaction scenario but rather transaction first, then online. The one-product-one-code data capture method is most suitable for promotion by traditional manufacturers. It can achieve B-end online (e.g., box codes, B codes) and C-end online (C codes). Through technical means, BC-linked data can also be obtained. BC-linked data is the most important data for traditional manufacturers' digitalization.
Fourth, purchase data from e-commerce, new retail, O2O, community group buying, KA platforms, and digital system companies.
E-commerce, new retail, O2O, and community group buying are inherently online transactions, and KA also generates data at the point of payment. As long as you cooperate with them, you can obtain data depending on the depth of cooperation, or you can pay to purchase data. However, in general, obtaining data on your own products is not a problem, but obtaining competitor information should be "desensitized data." If it is not desensitized, it usually violates rules. Now, unless it is a customized system, digital systems are basically SaaS or PaaS, and these systems also have a large amount of data that can be purchased. -02- How to digitalize without big data? In the first lecture, a key concept for traditional manufacturers' digitalization was mentioned: BC-linked data. Many people have noticed this concept and asked about its specific meaning. In recent days, I have discussed this concept with several people, and many strongly agree with it. When talking about data, the general objects are B-end and C-end, i.e., 2B data and 2C data. But traditional manufacturers reach C-end through offline channels, so it must be BC integration. This concept has been repeatedly mentioned in the context of three-dimensional connection. Since it is BC integration, the data is also BC-linked. BC-linked data means not only knowing that C-end purchased the product but also knowing at which B-end the C-end made the purchase. In other words, it means not only knowing that B-end sold the product but also knowing to whom (C-end) it was sold. In this way, B-end data and C-end data are interrelated. BC-linked data has great utility and serves several important functions: First, the distribution of products in the channel and the speed of circulation are fully transparent. This provides great convenience for manufacturers' channel management. Second, BC linkage provides space for manufacturers to digitalize when channel data is insufficient. I have always emphasized that digitalization for traditional manufacturers is definitely not like e-commerce's full data or big data, as users are relatively less online. This is a basic reality that marketing digitalization must acknowledge. Therefore, when data is insufficient, marketing digitalization must have a leverage effect, using small data to move big markets. The basic principle is: Use C-end digitalization to leverage the massive existing offline (B-end) system. The application of BC-linked digitalization will be the focus of traditional manufacturers' digitalization. In the future, we will propose relevant digitalization strategies and policies that will involve BC-linked data. -03- What does "online" mean? Being online generates data. That's right! So what does "online" mean? The general understanding of "online" refers to the "five online" aspects, with the core being users online. User online is easily misunderstood as online transactions. In fact, in many cases, we are already online without realizing it.
First, when the phone is on, it is online; even when off, it is online.
When the phone is on, even if not in use, it is already online, and data is generated without the user's awareness. Similarly, being off is also online. Don't think that when the phone is off, there is no signal; if there is a signal, it is online. Being online does not necessarily mean a person is online; more often, it is online from a technical perspective. Being online generates data.
Second, usage means online.
Many smart home appliances and 3C products are now interconnected; usage means online. For example, Xiaomi's various home appliances, such as door locks, air conditioners, rice cookers, and speakers, generate online data as soon as they are used. Similarly, not using them also generates data.
Third, perception means online.
If you have sensors installed at home, such as cameras, as long as they are turned on, they are technically online.
Fourth, behavior means online.
There is a lot of information and many apps on phones. Authorized apps on the phone may capture data from the phone. Posting a moment on WeChat could potentially be captured by some crawler technology, and your social relationships could be captured.
Fifth, reading means online.
Content online, typically Toutiao, Douyin, Kuaishou, and live streaming; opening, browsing, liking, and commenting all generate online data. The platform's algorithmic recommendations are all calculated based on data.
Sixth, socializing means online.
WeChat, QQ, blogs, Facebook, and DingTalk; as long as you socialize online, online data is generated. -04- Transaction data and behavioral data When users are online, not only do transactions generate data, but all behaviors also generate data. On Alibaba, Pinduoduo, Meituan, etc., as soon as you enter the app or mini-program, all behaviors, including clicks, dwell time, shopping cart, transactions, and payments, generate data. Platforms use this data to create user profiles and then push targeted content. Marketing digitalization is not narrowly about obtaining data from online transactions but broadly about deriving business opportunities from online data, understanding users, and gaining insights into them. The soul of digitalization is user profiling, which relies not only on transaction data but also on behavioral data. For example, repeatedly browsing similar products without placing an order is behavioral data. How to interpret and use behavior depends on the user profiling tool; different tools interpret differently, and the marketing strategies adopted also differ.