Shulou(Shulou.com)12/24 Report--

There are several key words in e-commerce industry: homogenization of a large number of products, fragmentation of consumption scenes, consumer loyalty and so on. In this process, how to realize the breakthrough of e-commerce channel with the help of digitalization and bring about the continuous growth of front-end business?

As a rapidly rising force in the pet food track, Zhang Zhiwei, head of the Ji Pet Information Technology Center, brought the sharing of "methodology and practice of full-link data analysis in the pet industry". From the methodology of data analysis to the analysis around business scenarios, and then back to the construction of data culture within the enterprise, sharing internal thinking and practical experience.

The following is an excerpt from Zhang Zhiwei's wonderful sharing:

(the data and PPT data in this paper are test data, for reference only)

Jijia Pet owns famous pet product brands such as Crazy Dog and Lanshi, online channels such as Max, as well as pet factories, with product layout covering pet staple food, pet snacks, pet products, etc., fully covering online platforms such as Tmall, JD.com and Douyin, as well as offline channels such as pet hospitals and pet stores, serving tens of millions of pet families.

Next, I mainly share the pet industry, mainly e-commerce, how to achieve full-link data analysis, including successful methodology and practice, is divided into four parts: theory, practice, suggestions and results.

I. theoretical chapter

Ji Jia has been using the following figure to measure the maturity of the company's data analysis application, with the horizontal axis representing difficulty and the vertical axis representing value.

Through this line, you can get the information to iterate and optimize several stages:

First, in the preliminary hindsight stage, descriptive information is presented through the data.

Second, the middle insight stage to realize the diagnostic analysis of the business.

Third, in the final foresight stage, in the most mature stage of data analysis, there are four things to know: what happened, why it happened, what happens in the future, and how it affects what happens.

This is exactly what data analysts do. Junior data analysts focus on insights and insights, intermediate analysts focus on insights and foresight, and senior analysts focus on insights and foresight. The main work of data analysts can be divided into data extraction, data visualization, index experience construction and abnormal diagnosis, experimental design and special analysis. In terms of ability, it can be divided into hard power and soft power. Hard power focuses on skills such as business knowledge, while soft power tends to logical thinking, learning and innovation.

The whole team of data analysts has nine areas of knowledge, which can also be called analytical capability model. Business domain, data exploration and so on are the abilities that business data analysts should master; analytical thinking and data visualization tend to report development engineers; data management is data warehouse engineers. Other capabilities are more representative of the advanced nature of the enterprise, the professionalism of the team, including cutting-edge strategic thinking, technical capabilities, product management and team leadership and so on.

Going back to the data analysis itself, what is the relationship between the data pipeline and the data analysis life cycle? On the left side of the figure below, from collection to ingestion, preparation, persistence, and experience, the underlying engine, ETL, data integration, data warehouse and other work are actually linked. These are the jobs of data pipelines.

On the right side of the following figure is the analysis life cycle. Middle-tier data understanding and exploration of data and planning analysis is the interpretation of the data pipeline, the upper layer is communication, analysis of problems, the bottom is the stage of interpretation and activation, the application of data analysis to the actual application scenario.

The Ji family has summarized the best practices for analyzing the life cycle, from problem understanding, understanding and exploration, model development, to interpretation, interpretation and application of results.

II. Actual combat articles

To enter the actual combat chapter, we will first share the blueprint of the BI system in the data analysis link to sort out the overall context and obtain the support of the company's resources. The BI system revolves around four aspects: business, data, assets and services, so we also promote business digitization, data capitalization, asset services and service operations.

The first is business. Conventional developing companies are generally divided into day-to-day business and strategic business. Daily business includes purchasing, production, warehousing and logistics. Each strategic business is different, based on the current development of the company. The strategy is mainly about making breakthroughs in the next three years, such as big items, Douyin, private domain and so on. Data analysis provides support for business decision-making, business analysis and business management.

The second is data. Which business systems does the data come from? What tools should be adopted? Business data sources such as MES, SRM, OMS, WMS, TMS, ERP, etc., tools include API interface, decimation platform, data connector, online reporting, Excel and so on.

The third is assets. Assets are the core part of the enterprise. The underlying infrastructure includes servers, operating systems, databases and so on. Data storage with the development of enterprises, the amount of data is getting larger and larger, when the lightweight warehouse can no longer meet the business or analysis side applications, it is necessary to start to do the integration of data warehouse and lake warehouse. Data management is relatively standardized, such as data asset portal, quality management, model management, metadata management, data security management and so on.

The fourth is service. Data assets fall on the service, and the service process is divided into analysis, scheme, product, delivery, learning, and systematic operation and maintenance.

Back to BI practice, Jijia landed many application scenarios on the BI side, including data cockpit, industry analysis, production data analysis, sales dimension analysis, customer dimension analysis, financial dimension analysis, warehousing analysis, logistics analysis and so on.

Here are a few analytical practices that are shared in detail.

1. Supply chain management

The biggest problems encountered by each enterprise are too many stories in the supply chain, too difficult, how to do the planning system, how to keep up with production capacity, how to do production scheduling, how to do procurement, and so on. Ji Jia has done some valuable exploration in this through Guanyuan BI. The following figure shows data analysis enabling supply chain management.

The first is the market demand. Based on the data of the past three to five years, Ji Jia can make a very accurate forecast of the old products. The new product adds a lot of factors to make a reference to the range. For every point of improvement in prediction accuracy, the improvement in all aspects of the operation and sales of the whole enterprise is very great, and it can be measured directly by value.

At the bottom, TMS is the BMS system that is responsible for billing the entire order and delivery to warehousing. At present, Laiji family has achieved that each cost, whether variable cost or fixed cost, is calculated to a specific item, so this part of the freight can be accurately apportioned to each main product and gift in BI.

On the production side, BI is used to accurately calculate the cost list of the whole plant, such as large material, small material, production scheduling and labor efficiency, so that the business has data to rely on.

On the warehouse side, BI realizes inventory management, understands where each brand warehouse and channel warehouse must be controlled, and finally makes early warning and dynamic adjustment through Kanban.

2. Financial management

In financial management, we use indicators to control the overall operation of the company. Profit = sales revenue-cost. To further decompose costs and sales revenue, costs can be divided into fixed costs and variable costs. Variable costs include commodity costs, platform fees, promotion, publicity, warehousing and logistics, labor fees and so on. These need to control the marginal contribution rate, that is, marginal contribution divided by sales revenue.

With such a system, profit margins can be further managed. In Ji's best practices, 15% of the marginal contribution is an indicator or baseline for staying advanced. The marginal contribution varies from high to low, and we should think about it, such as how to reduce the cost of goods through production in the case of rising raw materials.

3. Commodity analysis

In the past, the Ji family had thousands of SKU, and everyone blindly pursued sales. But it won't last long with sales without profits. So we use the matrix analysis method. It turns out that 80% of sales come from 20% of SKU, which is a terrible figure. The flow payment can be maintained, but how to deal with the long tail payment?

Carry on the analysis based on the matrix method, for example, the profit of the Taurus product is high, the sales volume is high, when its delivery efficiency has reached its peak, we should pay attention to it, and we can no longer add it any more. The Taurus product has formed a brand mind, and it is necessary to reduce the input in an appropriate amount and maintain a high gross margin; the sales volume of the problem product is high but the profit is low. Here, we should use marginal contribution analysis to analyze how to optimize the cost. New star products, low sales but high profits, here is divided into new and old brands, new products can continue to try, but old products should pay attention to inventory; skinny dog products, if the new products are in the promotion period, it is recommended that continuous observation, old products should immediately adopt the policy of stopping sales.

This solves the problems encountered in the production process of the enterprise. finally, we reduce the number of SKU and improve the operational efficiency of the whole company.

III. Suggestions and achievements

Ji family four years of construction, through 2-3 people to achieve the current data analysis system. In a word, solve some problems first, and then solve the qualitative problems. Many times we have to take the first step not only on the basis of thinking, but also on the basis of practice.

Finally, share the achievements of Ji Jia's BI construction:

The group operates the cockpit, which is analyzed from various sales indicators, and can provide an overview of the relevant data on each channel side, brand side and launch side.

Full-link data analysis, covering the cost management of the entire operating stage, so fine that you can query the specific cost sharing of each commodity.

Commodity evaluation and analysis, with the finest granularity to the store and management analysis, centralize all the evaluation of this kind of goods in the whole network to the background database.

Lean production list, you can see large materials, small materials, cost, energy consumption and analysis efficiency and so on, providing auxiliary decision-making function for lean production.

In addition to the above BI Kanban on the PC end, Ji Jia has built a mobile terminal with the goal of "the simpler, the wiser".

Sales overview, see the sales situation of the group

Industry analysis, including internal and external industry analysis

Market insight: observe the situation of the whole industry

. Wait, plate.

Finally, to sum up, in the initial stage of data analysis, we have been standing at the foot of the mountain, so look at the mountain is a mountain, see water is water, see everything is its original appearance, it is difficult to see some essence through this phenomenon. When we are in the mature stage of data analysis, we see mountains, water and clouds, everything is the manifestation of objective laws.

Once standing at the top, each team is responsible for the enterprise, and it is necessary to make strategic support for the future business success of the enterprise, and have higher requirements for themselves and others. I hope to see you at the top of the mountain with you in the future!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.