TiDB HTAP meets new energy vehicle companies: the application practice of real-time data analysis under the direct sales model

Whether in the stock market or the auto market, new energy vehicles have already stood at the center of the stage. Behind the explosion of new cars is the combination of product and technical capabilities of the new car-making forces, and it is a wonderful combination of digital marketing and direct sales. As early as 2021, the new car-making forces have basically completed the "original accumulation" of sales. According to the official data of various brands, the top 3 annual sales of the new car-making forces represented by "Weixiaoli" have exceeded 90,000 units, which is infinitely close to the milestone of 100,000 units.
The profit and loss line for an automobile production line to maintain normal operation is about 50,000 units per year, and the annual sales of nearly 100,000 units means that the head companies of these new car-making forces have probably crossed the product profit and loss line and can start. Realize self-development in the positive cycle of "production - sales - research and development".
Of course, the reason why "Wei Xiaoli" is called a new force in car building, in addition to its power form, is another important reason for the favor of the direct sales model.

New energy vehicles, a new direct sales model

Compared with the traditional 4S store model, there is no dealer in the direct sales model favored by the new automakers, which can help car companies reduce sales and service costs to a certain extent. Direct To Customer) to quickly gain first-hand market and user voice. Based on this, car companies can quickly adjust their strategies according to market feedback, achieve more accurate and flexible marketing, and then quickly open and occupy the market and gain hematopoietic capabilities. Obviously, on the mature, crowded and competitive track of automobiles, rapid expansion and rapid adaptability are crucial to the survival and development of new car manufacturers.
The advantages of the direct sales model are easy to understand, but this does not mean that its threshold is low. On the contrary, even for the new car-making forces born in the digital age, while direct sales have brought about a significant increase in marketing capabilities, the surge in data volume has also become a major challenge in the digitalization process, requiring newer and stronger data technology. The structure can cope with it smoothly.
So, how do the new car-making forces do a good job in marketing after direct sales and achieve brilliant results?

The data challenges behind direct sales

The characteristic of direct sales is that car companies directly control the links of product sales, service and community maintenance, and can obtain more dimensional and accurate data than the traditional model. However, mastering data is only the premise of marketing upgrades. Really powerful marketing also depends on how car companies can obtain real-time analysis capabilities for massive data.
Taking a new power car company as an example, its marketing system mainly consists of three parts: a real-time report analysis system for store managers and sales staff, a new real-time business decision-making system for C Level, and the original BI system running in parallel. The three major systems perform real-time display and analysis of information in different dimensions according to different user attributes of functions and services. The basic data includes the sales data of the national sales network, user data, community data and the vehicle itself. kind. Obviously, under the premise of ensuring real-time performance, how to collect massive data and achieve effective management and analysis is the difficulty for the entire system to maximize its effectiveness.
Although stand-alone OLTP databases including MySQL and Oracle can show good performance in classified data "storage" tasks, they cannot effectively perform real-time analysis tasks in the face of multi-source data aggregation and massive data volume. competent. Traditional solutions usually use Hadoop, Hive and other big data offline data analysis platforms, which are heavy in terms of both technical and operation and maintenance, and the analysis and processing time is often delayed for hours or even days, with poor timeliness and unable to cope with thousands of daily interest rates. changing business environment.
But all this happened before TiDB was introduced.

Construction of real-time data analysis platform

At first, the car company just deployed a set of TiDB database on the virtual machine with the mentality of technical exploration, and built a simple front end, display page and SQL statement to realize the aggregation and retrieval function of some data. After a simple trial, the TiDB running on the virtual machine showed good stability and performance. As a result, car companies have also accumulated a good impression and basic experience of TiDB. The new-generation database featuring HTAP can give full play to the advantages of a one-stack data service platform. The row storage engine TiKV + column storage engine TiFlash can help the enterprise support both OLTP and OLAP business loads on the same data source, simplifying data services architecture to improve system flexibility.

In mid-2021, the car company ushered in a sales explosion, with monthly sales exceeding 10,000 units, and the pressure on the data analysis system increased sharply. The good impression of TiDB and the realistic business pressure finally made car companies make up their minds to switch the big data system responsible for core data analysis to TiDB.
The new database runs on the physical cluster, and achieves efficient data synchronization with various classified MySQL databases through the DM data synchronization tool. In a new real-time business decision-making scenario, TiDB can support complex SQL statements, and the longest SQL can even exceed 1000 rows to build data analysis logic. In the simple query and analysis scenarios for sales personnel, TiDB shows a very fast response speed. Of course, the original BI system running in parallel can also efficiently obtain data from TiDB; both welcome the new and benefit the old. As a result, the newly deployed TiDB cluster has become a high-performance data center: it supports the C Level-oriented real-time business decision-making system and the original BI system, and provides real-time data reporting and query functions for stores and sales.
In the actual production and operation and maintenance process, TiDB shows comprehensive advantages in terms of performance, reliability, ease of maintenance, and scalability. After experiencing the convenience brought by the real-time business decision-making system, the car company also synchronized the data in the vehicle production process to the TiDB cluster. Through real-time analysis and decision-making, the entire delivery time from placing an order to picking up the car has been greatly shortened , which solves the most difficult supply chain efficiency problem in automotive manufacturing.
By deploying a new generation of TiDB distributed database, the car company used its own front-end and SQL statements to build a powerful real-time data analysis platform that can serve multiple roles, successfully supporting sales expansion and marketing capability growth in 2021 . In the past, building such a real-time data analysis platform by traditional means was not only expensive, but also extremely complicated in system implementation and post-operation and maintenance. Since TiDB is completely open source, car companies do not need to worry about being bound by a single database, but can also actively explore various data applications under the active ecological blessing.

Data-driven, insight into opportunities

Whether it is the challenge of massive data or the DTC (Direct To Consumer) model represented by direct sales, behind it are the continuous upgrading of the status of data in modern business models and the urgent needs of enterprises for new marketing models. Regardless of whether an enterprise chooses to be product-driven, marketing-driven or service-driven, its final operating mode also needs to be switched to data-driven. This is the core motivation for the digital transformation of enterprises, and it is also a firm rule for the survival and development of enterprises in the digital age.
To complete the power switching driven by numbers, the database of the enterprise must meet various needs and cope with the changing internal and external environment. TiDB with HTAP capability is the ideal choice for enterprises in this macro context. In addition, TiDB with open source features and strong scalability can also reduce the cost and risk of switching databases for enterprises, leaving enough time and space for enterprises to move.
Competition is imminent, the times are changing, and opportunities are fleeting. Only technological progress can allow enterprises to move forward calmly.