Building an e-commerce product center from 0 to 1

This series of articles will introduce the product design solutions for the core e-commerce systems in detail, helping you to understand e-commerce products in a systematic way.

After reading this article, you will have a better understanding of the following issues:

1. What is the definition and responsibilities of a product center?
2. What are the core modules of the product center?
3. How to design a product center?

At the same time, you can also systematically grasp the design ideas of most functional modules related to the products.

Preface

The previous article "Understanding E-commerce Product Architecture in One Article" gave an overall introduction to the basic business, system processes and product architecture of e-commerce. Through three large pictures, we took a preliminary peek at the basic appearance of e-commerce, understood the core components of a complete e-commerce system, and gained a global understanding of e-commerce at the conceptual level.

Starting from this article, we will continue to explain the detailed design ideas and solutions of each core system one by one.

According to the order of responsibilities of various e-commerce systems, we know that the product system is the most basic and core system. Without products, other systems cannot operate. Therefore, this article will first bring the detailed design ideas and plans of the "Product Center".

(The image is too large to display properly, please use your computer to enlarge it)

1. Overview of Product Center

1. What is a commodity system?

As the name suggests, the commodity system is a system responsible for managing commodity-related data and capabilities, which can be summarized in the following three sentences:

1. A system responsible for creating and managing products
2. The system responsible for creating and managing categories, brands, attributes, and all other product-related data
3. A system responsible for displaying products at the front desk and providing product capabilities to related upstream and downstream systems

In the e-commerce product system, commodities are the foundation of all other related systems. In technical terms, the commodity system is the dependency of other systems: from commodity procurement to inventory management, from the creation and on-shelf management of commodities themselves to commodity pricing and front-end display, from promotional activities to order payment, from order production to fulfillment and delivery, every process in e-commerce is related to commodities.

Among all these upstream and downstream systems, commodity data must be unified, universal, and standardized. For example, commodity A in the procurement system and commodity A in the order system must be the same A. Therefore, the commodity system must have the ability to provide centralized services to a large extent. Therefore, the commodity system is also the system that needs to be designed as a centralized service with the highest priority among all systems.

This is why the title of this article is Building an e-commerce product center from 0 to 1.

2. What is in the merchandise center? What can the product center do?

You can understand it through a picture:

The middle part of the above picture is the product center. The core modules of the product center can be abstracted into four parts:

1. Product classification : used to create and manage categories, and provide classification capabilities for products and other systems
2. Product Brand : used to create and manage brands, and provide brand capabilities for products and other systems
3. Product attributes : used to create and manage attributes, and provide attribute capabilities for products and other systems
4. Commodity : used to create and manage commodities, and provide commodity capabilities for other systems

Category management, brand management, attribute management and product management exist independently in the system architecture and have one-way dependencies. They are independent system modules. Each module is responsible for its own functions and can provide services to the outside world individually or in a package. The modules are also interconnected through specific objects.

Classification, brand and attributes are the basic components of a product, and together they constitute the product.

The most basic function of the product center is to support the product query and display functions of each link in the front-end, and at the same time provide product capabilities and services to downstream systems such as the price center, marketing center, order center, fulfillment center, procurement system, inventory center, warehousing and transportation, to ensure the unified input and output of product data for all systems under the e-commerce system.

Next, we will introduce the general design solutions for the four core modules of the product center.

2. Classification Management

1. Why is classification management necessary?

(The screenshot is the JD APP classification page)

The e-commerce platform can actually be seen as a mapping of a physical supermarket on a virtual website. In the supermarket shopping scene, in order to allow consumers to quickly find the goods they want to buy, the supermarket often classifies the goods.

Different categories are placed in different areas, each area is marked with shopping guide signs, and different shelves display different goods. This is the classification of the supermarket. Classification can help consumers find the products they want more conveniently and quickly.

The SKU volume of e-commerce is thousands of times that of physical supermarkets, and it becomes more difficult to search for target products on e-commerce websites. Therefore, the classification function of supermarkets is naturally applied to e-commerce platforms.

For users : Search and category navigation are the two most commonly used ways for consumers to find products (with the deepening of the application of recommendation technology, the weight of search and category has decreased to a certain extent). Category navigation can help users quickly locate target products. The realization of category navigation function depends on the category management of products.

For the operation side : managing a large number of products also has a great cost for the platform and merchants. In addition to searching for keywords, classification management is also an important means for operators to quickly locate a certain product. Classification can improve the efficiency of managing products.

2. The underlying design concept of classification management (key points)

The core design idea of ​​classification management is: separation of front-end and back-end, and many-to-many mapping.

(1) Separation of front-end and back-end

It means that the front-end classification and the back-end classification are designed into two classification systems. The back-end classification is mainly for platform management, and its main logic is structured management logic (iPhone13∈mobile phone classification, sweatpants∈clothing classification); the front-end classification is mainly for user-end use, providing operation space for operations and convenience for users to shop. Its main logic is operation and user experience (for example, the classification of sweatpants on the front-end can be "winter new products", and the main consideration is the operation effect and whether users can find products better).

(2) Many-to-many mapping

The foreground classification must be based on the background classification. When creating a foreground classification, you must first select the background classification; there is a many-to-many mapping relationship between the foreground classification and the background classification, that is, one foreground classification can select multiple background classifications, and different foreground categories can select the same background classification.

3. Backend classification design plan

Product managers often talk about "classification management", "classification design" and "classification function", which generally refer to the background classification management function, which is an internal management function of the enterprise. The background classification is the real product classification.

Generally speaking, the product classification model has been determined in the early stage of platform construction, and is only upgraded and optimized in the later stage according to the development of the business (expansion of categories). After the backend classification is created, it cannot be easily modified or deleted.

From the perspective of the interface, the background classification management mainly includes two parts of visualization functions: classification creation and classification management.

(1) Category creation

To create a new category, you need to fill in two main fields: category level and category name. The category name of the background classification is generally filled in the industry category of the physical commodity (such as mobile phones) for easy maintenance and management. The category name must be unique .

The classification level is generally designed as a three-level tree structure, with the first, second, and third levels classified in a progressive manner (too many levels will make it complicated, while too few levels will not achieve the purpose of classification. The specific design needs to be considered comprehensively based on factors such as the complexity of the platform's categories).

When creating a new category at the secondary or tertiary level, you need to select the parent category.

Both the category code and category ID can be designed to be self-generated according to certain rules (such as the category code is composed of the pinyin of the category name + the timestamp). The category code and category ID must be unique . When the category is called and passed as a parameter by other systems, it can be implemented through the category code or category ID.

(2) Classification management

Category management mainly provides functions such as query, sorting, and modification of the category itself. At the same time, it can design the function of binding/unbinding the attributes or brands corresponding to the category (the function of binding the category is also available when creating attributes or brands).

As for deleting or modifying categories, as mentioned above, it is generally not recommended to modify them (including modifying the level and modifying the parent category). When modifying, it is necessary to verify whether there is related data under the category. If there is no related data, it can be modified to a certain extent. The deletion function needs to be designed more carefully. Under normal circumstances, as long as there is related data under the category (such as related products, related front-end categories, etc.), it cannot be deleted. The deletion operation can only be performed when there is no related data under the category (it is recommended not to design the deletion function in the early stage, or to design it so that advanced permissions are required to perform the deletion operation).

4. Front-end classification design plan

(Screenshot of product classification on JD.com PC)

In terms of the division of responsibilities, the front-end classification belongs to the functional category of "website/APP operation configuration" and is not a classification function in the true sense. When maintaining the front-end classification, it needs to be attached to one or more back-end categories . When the user clicks on the category to search for products on the front-end, the system queries which products should be displayed under the front-end category through the three-level mapping relationship of [front-end category-back-end category-product].

(Category management interface example)

The design logic of the many-to-many mapping between front-end and back-end categories mainly considers operational needs and differences in users' perceptions of products. Since the traffic of category navigation is very large, operations often configure some operational categories, such as recommended categories, commonly used categories, etc. For recommended/common categories, they are essentially not directly related to product attributes. For operational choices, front-end categories can be associated with multiple different back-end categories (for example, in many e-commerce websites, the mobile phone category and mobile phone accessories category in the back-end category both correspond to mobile phones in the front-end category).

3. Brand Management

1. Why do we need brand management?

In the system architecture of large e-commerce platforms, brand management, like category management, is an independent and complete function. When many people first come into contact with the product system, they may have seen the design method of simply filling in the brand of the product when creating a product. Then why do large platforms have separate brand management?

On large e-commerce platforms, brand is also an important means for users to search and filter products . The more standardized the category, the greater the use scenario of searching for products through brand filtering (for example, when buying mobile phones or home appliances on JD.com, consumers often select the brand first).

If the brand data is not structured and is allowed to be filled in at will by operators or merchants when creating products, the same brand of products may be filled in with different brands by different businesses or different merchants. For example, merchant A fills in Xiaomi as the letters "xiaomi", merchant B fills in Xiaomi as the Chinese "小米", and merchant C even fills in "showmi" and other incorrect data. When the underlying data cannot be unified, the front-end screening or other systems that need to use the brand data will not be able to use it normally.

2. How to design brand management?

Brand management mainly includes two functions, namely brand creation and brand management.

(1) Brand creation

There are two main scenarios for creating brands: one is active creation on the platform side, which is generally initialized directly into the system according to each category in the early stage; the other is a more common application method for platform-based e-commerce, that is, third-party merchants who have settled in the platform apply for the brand (such as in the Alibaba JD platform, if the newly settled merchant is a new brand owned by itself or has not been included in the platform, they must first go through the brand application process, such as the early Taobao brand Han Du Yishe, etc.), and the brand will be included in the brand library after the platform review.

Brand creation generally has several key elements:

1. Brand Chinese name
2. Brand English name: recommended format requires capitalization
3. Brand alias: Brands have official names and folk names. Sometimes folk names are more well-known than official ones. 4. Brand registrant
4. Brand Registration Number
5. Brand LOGO: The image size and dimensions must be limited to facilitate system display specifications
6. Brand registration area
7. Other fields are customized according to platform characteristics or industry qualification requirements

If it is created through a third-party application, the information that needs to be filled in is the same as that created independently by the platform. However, after the brand application process is completed, it will be submitted to the brand for review (the review process design is not expanded in this article). Once the review is passed, the second merchant can directly select from the brand library when operating the same brand of products.

(2) Brand management

Brand management refers to the functions of querying, modifying, and deactivating the brand itself, that is, the basic addition, query, and modification. It should be noted that when modifying or deactivating a brand, it is also necessary to verify that there is no associated data (products, categories, etc.) under the brand, otherwise the modification or deactivation cannot be performed (modification of some information can be performed if it does not affect the display of other data).

Another function is to design the mapping relationship between brands and product categories. After selecting a category when sending products, the brands that should appear under the category will be automatically retrieved (you can also not design it. The logic of not designing it is that after selecting a category, all brands will be retrieved when selecting a brand. Both solutions have advantages and disadvantages).

4. Attribute Management

1. What are attributes? Why do we need attribute management?

What are attributes? ——Attributes refer to the unique properties of a product, such as color attributes, size attributes, specification attributes, etc.

What is the function of the attribute? ——Attributes are also a necessary and important method for classifying and aggregating products. Categories and brands can only classify a large number of products to a certain extent. For example, when users buy "home appliances" on JD.com, they can search for the products they want (such as Xiaomi TV) through categories (such as TV, air conditioner, refrigerator and washing machine...) and brands (such as Gree, Xiaomi...). However, when users want to further quickly locate the most accurate intended results (such as Xiaomi TV 55-inch 4K version), categories and brands are not enough. Therefore, if you want to achieve a more fine-grained division, you can only do so through attributes .

All properties of a product are the attributes of the product, such as screen size, resolution, energy efficiency level , etc. However, products in different categories have different attributes (for example, clothing does not have attributes such as screen size, and mobile phones do not have size attributes). As the categories become richer, the attribute data also becomes very large. In order to better manage the attributes of different types of products and have clear and simple logic when applied in the foreground, the significance of attribute management becomes very important.

2. Design ideas of attribute management

Attributes can be divided into three parts from a structural perspective: attribute group > attribute > attribute value

1. Attribute group: Categorize a large number of attributes for easy management, such as the "appearance attributes" and "performance attributes" of a mobile phone;
2. Attribute: the attribute name, which refers to the specific attribute, such as color, version, and running memory;
3. Attribute value: refers to the specific value of the attribute, such as the attribute values ​​of the color attribute are: red, yellow, green;

The above three modules constitute attribute management. Sometimes, in order to facilitate communication or unify the caliber at the business level, or for other application scenarios, the design of attribute classification will be introduced. The attributes will be classified on the attribute group (relatively abstract classification), generally divided into basic attributes, sales attributes, warehousing attributes, logistics attributes, and extended attributes (there is no standard design for attribute classification. It can be divided according to the characteristics of its own platform, or attribute classification can be designed).

The following combines attribute classification with three core modules, taking the mobile phone category as an example, and explains the various concepts of attributes through a table:

The following is a screenshot of the product details page of iPhone 13 on JD.com. On the JD.com front desk, this information is a structured field (not text or graphics), and its data comes from the attributes filled in when creating the product:

3. Design of attribute management

1. Core functions of attribute groups: creation and management (query, modification, and binding of attributes)
2. Core functions of attributes: creation and management (query, modification, binding attribute groups, binding categories)
3. Attribute value: depends on the attribute function

(1) Creation of attribute groups

(2) Attribute group management (list)

(3) Attribute group management (details)

The main function is to support binding and unbinding attributes and enable/disable control of attribute groups. For the disable operation, it is necessary to verify that there is no associated data (attributes, etc.) under the attribute group.

(4) Creation of attributes and attribute values

There are also two scenarios for creating attributes and attribute values: one is active creation on the platform side, and the other is a more common application method for platform-based e-commerce. That is, when a third-party merchant creates a product, if the attributes in the attribute library are not enough, you can directly add attributes on the product creation interface.

Creating attributes on the platform side

Focus on the associated categories: attributes must be associated with categories. When creating a new product, the first step is to select the category. When you select the attributes later, the system can query the corresponding attributes based on the selected category (for example, under the mobile phone category, attributes such as clothing size will not be displayed). Create attributes/attribute values ​​when sending products to third parties.

Taobao third-party backend adds attribute/attribute value function

Add attribute/attribute value function to JD.com's third-party backend

The above is the design introduction of the attribute management related functional modules. The interface of attribute management ( binding category , binding attribute group) is similar to that of attribute group.

Deleting and unbinding attributes are the same as attribute groups. You also need to verify whether there is any associated data.

5. Product Management

The product management module is mainly responsible for creating and managing products and providing product capabilities for other systems. It is the core module of the product center. Product management also relies on three modules: classification, brand, and attribute.

Taking the creation of a product as an example, among the fields that need to be filled in when creating a product, the most important one is to select the category, brand, and attributes. It can be said that the product is a module built on the category module, brand module and attribute module.

1. Two concepts - SPU & SKU

Before explaining the product module in detail, it is necessary to explain the two terms "SPU and SKU". Not only e-commerce platforms, but most systems involving transactions may involve SPU and SKU. I believe most people have come into contact with these two names. The following is Baidu Encyclopedia's explanation of SPU and SKU:

SPU (Standard Product Unit)——Standardized product unit. It is the smallest unit of commodity information aggregation and a collection of reusable, easily searchable standardized information that describes the characteristics of a product.

In layman's terms, products with the same attribute values ​​and characteristics can be called an SPU SKU (stock keeping unit) - SKU is the unit of measurement for inventory in and out, which can be in units of pieces, boxes, pallets, etc.; SKU is a product that can be sold.

In layman's terms, SPU positions a product and SKU positions a commodity. For example, in the mobile phone category, the iPhone 13 is an SPU, and the iPhone 13 256G black is a SKU (JD’s SKU will also add attributes such as the public version). Generally, the basic attributes can be regarded as the attributes of the spu, such as the screen size, CPU model, pixels and other basic attributes of the mobile phone; sales attributes and warehousing and logistics attributes are stored on the sku, such as color, version and other sales attributes.

2. Create a product

The product creation function looks relatively simple from the interface, which is just to fill in a bunch of fields. However, depending on the size, complexity, vertical expertise and other factors of the platform, it can be abstracted into two design solutions:

Solution 1: Create multiple SKUs at the same time and generate related SPUs simultaneously . The overall solution is to create SKUs directly. Maintaining multiple different attributes means producing and maintaining multiple SKUs. This solution is suitable for most 2C comprehensive e-commerce platforms (such as Taobao and JD POP create products in this way).

Solution 2: Create SPU first, and then create SKU based on SPU . The overall solution is that the platform's master data team is responsible for maintaining SPU, and merchants (including self-operated and POP) operate (or purchase and sell) to maintain SKU based on SPU. When creating SKU, first select SPU (the basic attributes in SPU have been maintained by the data team), maintain sales attributes, logistics attributes, etc. based on SPU, and then generate SKU. This solution is suitable for highly professional vertical B2B industries, such as automobiles, pharmaceuticals, etc.

The reason for the two directions is that the merchants on vertical B2B platforms (traditional industries and older business owners) have limited operational capabilities and a much higher error rate in maintaining product attributes than merchants on 2C platforms. The platform has high requirements for the structural management of products. In order to avoid the same product being maintained by different merchants with multiple different attributes (such as the tread width and size of car tires), the platform often chooses to have a dedicated data team maintain the basic attributes of the product, that is, maintain the SPU.

Moreover, B2B vertical e-commerce has fewer categories, relatively small SKU volumes, a high degree of category standardization, and is highly feasible for unified maintenance by the platform.

However, for comprehensive e-commerce platforms with tens of thousands of categories, it is not realistic to rely on the platform's data team to maintain them uniformly. Or for non-standard categories like clothing, there is a low demand for structured management of goods. Therefore, there are differences in the design direction of comprehensive platforms (JD.com, Taobao) and vertical platforms.

In reality, even on a comprehensive platform (such as JD.com), different categories will have different design methods. Some categories have vertical depth, and the platform also maintains the SPU and the merchants create the SKU.

(1) Create a sku design plan, explained in direction 1

This method is quite common. Students who have seen the backends of mainstream e-commerce platforms such as Taobao and JD.com should have a certain impression of it. The logic is to directly create sku and generate corresponding spu at the same time as sku, as shown in the following figure of Taobao merchant backend:

(Screenshot of creating a product in Taobao’s third-party backend - filling in sales information)

As can be seen from the figure, multiple SKUs with the same basic attributes correspond to the same SPU. The color attribute + flight type attribute determines a SKU. The price and inventory are stored on the SKU (even if the same price and inventory are filled in batches, they are stored according to the SKU dimension at the bottom of the data; and displayed in the SPU dimension at the front end (better experience). Switching sales attributes means switching different SKUs under the same SPU dimension:

The above is the product details page of iPhone13 on Tmall PC. It can be seen that iPhone13+Black+128G and iPhone13+Pink+128G are two independent SKUs (different skuids), and the SPU corresponding to the two skus is the same (spuid is the same)

The above is the core design logic of Solution 1: that is, when creating a product, maintaining different sales attributes means maintaining multiple different SKUs, but ultimately generating the same SPU synchronously.

After understanding the design logic, it is very simple to design functions. Whether it is a self-operated background or a third-party pop background, the interface for creating products is similar, and can be divided into 7 steps according to the process:

Step 1: Select a category (product classification). Note that when selecting a category, the background classification is called (the classification management service is used here):

(Picture: Taobao merchant backend - create new product - select category)

After selecting a category, the system automatically loads the corresponding brand data under that category (for some categories, you need to select the brand when filling in the product information).

The following steps are relatively simple. You just need to fill in the corresponding fields. I will not go into details here. Here is a screenshot of the background for reference:

(The picture shows a screenshot of a Taobao merchant posting a product on the backend)

The above is the design plan for creating products. The difference between Plan 2 and Plan 1 is: when creating sku, it is necessary to create sku based on spu, that is, spu is created and maintained by the internal data operation and maintenance team. When operating and third-party merchants create sku, they need to select spu first (if not, they need to submit an application). The specific interface is not much different from Plan 1, and this article will not expand on it.

3. Other modules of product management

After the product is created, it needs to be reviewed before it can be put on the shelf and displayed on the front desk. This involves the design of the product's state machine, but it is also relatively simple and will not be explained in detail in this article.

Finally, a note about price : I don’t know if anyone will have any questions, but the product center does not mention price at all. In the minds of many people, products and prices are one and the same, and the price must be filled in when creating a product. In fact, this is a misunderstanding.

In the early stages of the business, when there is only a retail price, you can directly fill in simple prices such as the retail price and the strike-through price. As the business develops, the business model becomes more and more complex. The method of directly filling in the price when creating a product will soon fail to meet the needs of the business.

For example, JD.com later developed the enterprise purchasing business, and also developed various user prices (plus price, student price, new user price, etc.), and added promotional prices. At this time, the product system could no longer support various forms of price demands, so in the system architecture, products and prices are two completely independent and separate systems.

VI. Conclusion

Finally, let me answer the questions at the beginning of the article:

1. What are the responsibilities of a product center?

1. Responsible for creating and managing products
2. Responsible for creating and managing categories, brands, attributes, and all other product-related data
3. Responsible for the display of front-end products and providing product capabilities for related upstream and downstream systems

2. What are the core modules of the product center? —— Four core modules

1. Product classification: used to create and manage categories, and provide classification capabilities for products and other systems
2. Product Brand: used to create and manage brands, and provide brand capabilities for products and other systems
3. Product attributes: used to create and manage attributes, and provide attribute capabilities for products and other systems
4. Commodity: used to create and manage commodities, and provide commodity capabilities for other systems

3. How to design each module of the product center? —— One picture to understand the core points of each module

As a dependency of other upstream and downstream systems, the product center is one of the systems that interacts most with other systems. Therefore, for the product center, reasonable architecture design and good scalability are of paramount importance. Laying a solid foundation in the early stage can avoid repeated reconstruction of the system in the later stage as the business develops.

The above is all the content of the product center. It mainly introduces the concepts and responsibilities of each module of the product center, as well as the core design ideas of each module. It is recommended that while reading the article, you can also experience different platforms, analyze and summarize more based on the article, and then come to your own comprehensive understanding.

Author: Dougson Salted Fish King

Source: E-commerce Product Manager from 0 to 1