Five years of technology upgrade and precipitation of Taobao shopping cart

This article will summarize the product upgrade road of Taobao shopping cart from two aspects: business development and technology precipitation.

foreword

It has been more than four years since I joined the company. Fortunately, I can participate in or be responsible for the development of the Taobao shopping cart business domain for 90% of my short career. From thinking about the difference between shopping cart and favorites, to gradually starting to have a clearer positioning of shopping carts, from passively responding to needs to actively thinking about the correct and reasonable business direction of shopping cart business, from just fulfilling Taobao business demands to providing reusable services Operational capacity. Experienced 5 big promotions with the shopping cart, experienced product upgrades large or small, and changes in the development method of carts2, tried for the ultimate user experience, explored for the incremental improvement of the platform, and used technology upgrades to help Taobao shopping carts complete At the same time of business iteration, some general capabilities have been accumulated along the way to support the rapid iteration of other businesses, and a set of development knowledge bases have been continuously improved to help more and more shopping cart development students. Then it’s time to write some words and draw some pictures to record these years, what exploration and precipitation have been done in the PD and development of Taobao shopping cart, what product upgrades have been completed in Taobao shopping cart, and those useful functions that are closely related to your daily shopping are How to iterate on-line step by step.

This article mainly summarizes the product upgrade road of Taobao shopping cart from two aspects: business development and technical precipitation.

PS Here, I will share with you the shopping carts I have experienced so far. There have been major changes in product styles, interactions, and business functions. It can also be seen from the picture that the shopping carts have been around since 2021. The changes are especially obvious and bold.

How do I understand the shopping cart

Before introducing the Taobao shopping cart business, let me briefly talk about my understanding of the shopping cart business scenario, its positioning, definition, and responsibilities.

basic function

I believe that most people who have not had an in-depth understanding of the shopping cart are the most basic operations of "adding, deleting, modifying, and checking" the user's additional purchases. So what exactly is stored in a shopping cart record in the db? As shown below:

But in fact, with the continuous strengthening of the basic functions of the shopping cart, these four words are no longer enough to summarize the responsibilities of the shopping cart, so I will summarize the basic functions of the shopping cart into the following four words based on my personal understanding: add, change, count, make up

1. Add (& check): The core of the shopping cart revolves around the user's shopping cart product data stored in the db. Since it is a user's private domain, the product data basically comes from the user's active add-on purchase. The sources of add-on purchases include: details, venues, orders, etc. Links, Maochao homepage, etc. However, in order to further shorten the user's repurchase path, Taobao shopping cart iteratively launched the "frequent purchase" function in 2021, using algorithms to calculate the repurchased products with high purchase intention of users and automatically adding them to the user's "frequent purchase" shopping cart through offline means (not It is a new attempt by Taobao shopping cart to shorten the repurchase path and improve the purchase efficiency.
2. Change: After the user adds the product to the shopping cart, because the product changes dynamically in real time, it becomes invalid, the price increases, the shopping cart capacity is not enough, etc. The user can enter the shopping cart to perform single or batch operations on the product. In fact, the launch of many functions of the shopping cart is continuously improving the efficiency of the user's use of the shopping cart, for example, the failure is more refined to the sku dimension, the user can add a purchase prompt when the capacity is full, and one can be cleaned up and added to the purchase, etc.;
3. Calculation: The so-called "calculation" refers to the calculation of the price of the items in the shopping cart. The matter of "calculation" is actually displayed to different degrees in the shopping cart. [1]. The user enters the shopping cart and can immediately see the discounted price of the single product in the product dimension in the product card; [2]. The user selects multiple products Before entering the order, you can see the final settlement price at the bottom of the shopping cart, and you can see the apportioned product discount price on a single product; [3]. Open up a discount details floating layer to calculate the entire price discount calculation process and results. Clearly presented in front of the user. Shopping carts are constantly striving for correctness and clarity in price calculations. Minato: I think that the data core of the shopping cart is the items in the user's shopping cart, so the core function is to collect orders, or to say consolidated settlement. In fact, in most scenarios of Taobao so far, the vast majority of consolidated settlements come from shopping carts, which is why about 70% of the order traffic comes from shopping carts during big promotions.

4. "Making" is mainly reflected in two points: [1] The user can select multiple products in the shopping cart to place an order together, and also supports business customization and group settlement; [2] The user continuously selects and selects products in the shopping cart. In , the progress of full-reduction order collection is constantly updated, and functions such as screening, recommendation, and order collection guidance are provided to improve the efficiency of users' order collection and settlement. In the past two years, around the proposition of "collecting orders", shopping carts have been constantly trying to transform and experience.

   Position and Responsibilities

   The above are the basic functions of the shopping cart scene. Next, let's take a look outside and think about the location of the shopping cart in the nodes of the entire shopping process, or in the basic link of Taobao. We divide the shopping process into three nodes, namely pre-purchase, in-purchase and post-purchase. As shown in the figure below, the shopping cart is an important part of the "pre-purchase" link as a "step at the door" before order placement and settlement.
   
   The shopping cart carries the responsibility of linking the previous and the next in the pre-purchase purchase. Promote purchases in the pre-purchase link: For users who are hesitant to place an order, use marketing and other means to improve the user's decision-making efficiency; for users with high conversion certainty, accurate recommendations allow them to buy more; Trading experience: the accuracy of discount calculation and the clarity of the process, as well as the collection and settlement of commodities in various business forms, etc.

   definition

   Based on the above description of the basic functions and responsibilities of shopping carts, as well as my own understanding of shopping carts, I define shopping carts as: a consolidated settlement tool that provides product management, order collection and settlement capabilities, and helps consumers make order decisions ;

   cliché

   Finally, let's discuss a common question: the difference between a shopping cart and a favorite? This question should be the soul torture that every student who has developed a shopping cart will face. The first time I was asked was during a new recruit training about a month or two after I joined the company, and the second time I was asked about the defense. My answer at the time was basically purely based on my own daily experience of using hand shopping, mainly focusing on one point: the difference in the intensity of purchase intention; after making a shopping cart for a few years, will there be any new experience? Let's first list the main differences between the current two products

Business scene	capacity	price expression	marketing expression	Computational power	settlement ability	Product modification ability	product search,
shopping cart	120	On the product card is the discounted price of a single product	Various marketing atmosphere expressions, service atmosphere expressions	Dynamic calculation, support multiple discounts for multiple products (single product discount, store discount, cross-store discount	Grouping and consolidated settlement capabilities	Support product sku switching and quantity modification	Currently, it supports product search and filtering of specific products (such as cross-store full reduction, price reduction, etc.).
Favorites	3000	The coupon post price is on the product card	Basically no marketing message	The rendering expression of the back price of a single product coupon does not support the combined price calculation of multiple products.	Merchandise settlement is not supported	Not supported; no concept of sku, quantity in favorites	It has perfect and rich product filtering and searching capabilities.

I think that for a user, purchasing a product requires the following decision-making processes. As the purchasing process gradually draws to a close, the user's purchasing intention and certainty are also increasing. Different links require different tools to help users complete.

So to summarize some important differences between the two:

Favorites focus on the storage management, search and search of products. The shopping cart focuses more on the marketing expression, price calculation, and order settlement of the product;
The shopping cart is closer to placing an order, and the products added to the shopping cart by the user are relatively more willing to purchase;

Although the focus of the two scenarios is different, in fact, many product functions of Favorites are worth learning from shopping carts. It is a better user experience improvement. At the same time, let's think about it, must the shopping cart and the favorites be two completely unrelated business scenarios? Are there certain related functions that can make up for the shortcomings of the two products in terms of user experience? For example (purely yy):
Provide users with the option of adding to favorites when the purchase is full;
At the same time as the shopping cart cleaning reminder, it is recommended to guide users to add products with strong purchase intentions in the favorites;
The search of the shopping cart found that, in addition to searching out the products in the shopping cart, it recommended the products in the favorites that are related and have high purchase intentions;
Favorites provide an efficient add-on path;
......

Taobao shopping cart business development

goals and strategies

business development goals

As a business team, we should aim to win the business first and take technological breakthroughs as a means to empower the efficient development of the shopping cart business. So what is the business development goal of the shopping cart scenario? In fact, how to develop the business, the first thing to think about is, what are the roles of this domain scene and the platform as the scene? What are the current pain points and demands of these roles? For the development of shopping cart business, in general, there are currently three roles: consumers, merchants and platforms, and business parties.

[Consumer role]: that is, users who need to use Taobao shopping cart to complete the purchase process; for consumers, poor experience in using basic functions of shopping cart products and low purchase decision-making efficiency are the biggest pain points;
[Merchant and platform role]: The shopping cart, as a user's private domain, provides a small space for merchants' marketing and operations. How to promote conversion, increase users' willingness to purchase, and thus gain increments is the biggest pain point faced by this type of role;
[Business side]: Business gameplay is becoming more and more complex, resulting in complex business logic, which seriously hinders R&D efficiency and business development iteration;

Therefore, in the face of these product pain points and demands in the shopping cart domain, we summarize the business development goals of the shopping cart into three directions:
Direction 1: Experience <br>What are the aspects of the experience of using the shopping cart products? As a basic product in the ordering link, the "thing" of shopping cart management is the commodity, and the first "role" is the consumer. Then we think about the relationship between people and commodities. In the use of the entire product, people (ie, consumption) The following action lines exist for objects (ie shopping cart items): storage, browsing, management, decision-making, and settlement. Then in the above links, whether the consumer's operation is efficient is defined as a good user experience of the shopping cart product.
Direction 2: Conversion <br>As a basic product, does the shopping cart have room for conversion? Can it provide the possibility for merchants to operate their own marketing, and ultimately bring increments to the platform? This is the direction of our thinking and data mining in the development of the shopping cart business. Data analysis and mining brought us the conclusion: there are a large number of stock items in Taobao shopping cart that users add but have not been converted. Sales accounted for about 60%. Therefore, the shopping cart actually has a huge space to obtain new increments; how to enhance the attractiveness of products and reawaken user demand are the two directions for the improvement of shopping cart conversion. The increment of the scene outside the shopping cart, combined with the accurate recommendation of algorithm means, allows users to buy more and more cost-effective;
Direction 3: Efficiency <br>Whether it is experience or transformation, business development is inseparable from rapid trial and error and iteration, and ultimately, efficient R&D efficiency. The R&D efficiency is also faced with the bad business status and development environment: the business is multi-terminal and multi-platform, and the requirements are numerous and complex; in the development, the demand response is slow, the communication and coordination are many, and the client version is heavily relied on. The above points seriously hinder the speed of business iteration. It is necessary to rely on technological transformation to change the R&D model and improve R&D efficiency;

business development strategy

Based on the three directions of shopping cart business development, we disassemble it into two main implementation strategies: product upgrade on the consumer side & R&D efficiency improvement.

[Strategy 1] Product upgrades on the consumer side: Shopping cart product upgrades are mainly reflected in two aspects: (1) experience optimization based on the tool attributes of the shopping cart; (2) incremental conversions based on the scene characteristics of the shopping cart;
[Strategy 2] R&D efficiency improvement: The improvement of shopping cart R&D efficiency is reflected in two aspects: (1) technological transformation improves R&D efficiency; (2) closed-loop business improves the efficiency of business iteration;

Finally, I summarize the development of the shopping cart (business + development efficiency) in my understanding in the past two years into a big picture (some of which will be introduced in detail in the following chapters):

What we're talking about when we're talking about the shopping cart experience

As a basic tool product, no matter whether the KPI orientation is GMV or experience, I always believe that improving the user's daily and big promotion experience is the core of product upgrades. So, what are we talking about when we talk about shopping cart experience? What have we done?

Taobao shopping cart product usage status

Since we want to improve the user experience, the first step is to understand the user's demands and pain points. Taking a research report on the experience of shopping carts in early 2021 as an example, we will discuss the distribution of shopping cart users, usage scenarios, and pain points. Take a look at:
[Crowd distribution]: People with high purchasing power are the main users of shopping carts;
[Usage scenarios]: (Statistical data at the beginning of 21): In addition to basic functions such as collecting products and shopping cart settlement, the main scenarios for users to use shopping carts are order collection, product price comparison, and product attribute comparison. It can be seen that the user's appeals for the use of shopping carts are mainly summarized as: product management and discovery, price calculation and order placement, and order collection;
[Pain Points in Use]: (Statistical data in early 2021): To sum up, users' pain points in using shopping carts come from several aspects: (1) The efficiency of cross-store order collection during the promotion period is low; (2) Price experience: including pre-orders During the hot period, the big promotion price cannot be revealed, the dynamic calculation results cannot be completely consistent with the order placed, etc.; (3) Quick discovery of shopping cart products: search, classification and screening, etc.;

What we're talking about when we're talking about the shopping cart experience

The shopping cart is a basic tool that provides product management, order collection and settlement capabilities. It revolves around two cores: "people" and "products". The relationship between people and products can be summarized into two, that is, people's management of products and people's products. purchase settlement. Then the shopping cart product experience also revolves around these two points. The whole is shown in the figure below:

Commodity management <br>means that users search, add, delete, modify, and check commodities according to their current purchase intentions. The factors that affect the user's product management experience include: (1) Add-on card points caused by the problem of shopping cart capacity; (2) The efficiency of product search and discovery; (3) Whether the operation route of product management, that is, various additions, deletions, modifications, and checks is simple efficient;
Commodity purchase <br>means that the user continuously selects commodities in the shopping cart, calculates the price, collects the order, and then calculates the price and finally completes the purchase process of order settlement, and it is also the process of constantly making order decisions. In my opinion, there are three main things that affect the efficiency of users' decision to place an order: (1) Efficient expression of information on products; (2) Accuracy of prices and clear expression of discount calculations in the process of continuous selection of products; (3) Continuous selection of products Efficient order collection guidance in the process (including real-time update of order collection progress, screening of goods in the same grade, recommendation of goods that can be collected, and uninterrupted order collection experience);
When we clarify the goal of the shopping cart experience, the strategy and specific implementation will revolve around it.

Introduction to key experience projects

Special experience of order collection <br>What is order collection?
The so-called order collection, as the name suggests, refers to the process in which users purchase multiple items in order to achieve a certain threshold to obtain a certain discount. Both online and offline. In fact, this is a commonly used promotion method by sellers: for offline/online merchant stores, in order to allow users to buy more and make profits, a variety of discounts/full discounts/full free discounts in stores are launched. In order for consumers to reach a certain threshold , will continue to select products in the store to make up the order; then for the platform side (or for offline shopping malls), in order to obtain a higher turnover of promotional activities, it will be in the whole network/all malls (registered full discount activities) Introduce full discount activities, such as Tmall promotion, which accounts for the highest proportion of marketing methods, cross-store full discount and category coupons, all of which are cross-store collect orders.

According to some objective data, from the statistics of related order data of cross-store full reduction and the source of shopping carts during Tmall's Double Eleven, the consolidated settlement function of shopping carts has naturally become a tool for consumers to collect orders, which is almost a hand-made tool. Amoy the only place to collect orders. The ever-developing function of collecting orders (such as the expression of full-reduction product atmosphere, order-collection guidance, and full-reduction screening, etc.) has also become a shopping tool for users.

What are the user's demands for order collection? In fact, the iteration of the shopping cart order collection function also evolves with the changes of user demands. Of course, it is also due to the complexity of the platform marketing plan year by year. The complexity of marketing rules and gameplay continues to increase the difficulty for users to collect orders. According to the summary of users' demands for shopping cart order collection provided by 21 years of research, the pain points of users' use of shopping cart order collection are mainly:
Products that meet different cross-store full-reduction rules cannot be classified and cannot quickly find the same threshold products for order collection (in the shopping cart or outside the shopping cart);
During the warm-up period of the big promotion, the shopping cart does not display the event price, which is inconvenient to collect orders;

The development history of the shopping cart's order collection function Order collection is one of the core experiences during the shopping cart promotion period. The optimization and iteration related to order collection is also an important proposition for the development of the shopping cart business in the past two years. Here, we will summarize the pain points and demands of users. The four stages of shopping cart collection and the core problems solved in each stage. First of all, let's take a look at the process of gradually improving the efficiency of users' order collection in Taobao shopping cart so far:

[Phase 1: Real-time update of the progress of the order collection]
Ability introduction

What problem is solved: Let users no longer "blindly collect" in the process of user selection of products, help users calculate the current amount that has been reduced and the amount that still needs to be purchased from the next stall; provide users with an entrance to collect orders to help users Quickly discover the products of the same grade outside the shopping cart, reach the threshold faster, and improve the efficiency of order collection; if there are problems, the entry of order collection jumps to the acceptance page, and after adding the purchase, return to the shopping cart page to refresh, and the entire order collection link is faulty; Help users to quickly find the full and reduced items in the same gear in the shopping cart. After all, the purchase intention of the goods in the shopping cart is relatively high;

[Phase 2: Screening of full-reduced products in the same stall]
Ability introduction

In fact, the function of screening full reduction, in addition to a breakthrough in business, is also one of the typical models of technological upgrading to support business development. The technical system of the entire screening scheme also supports the implementation of subsequent projects such as NewCart, search, and warm-up state. Technical challenges and solutions can be read carefully in the "Shopping Cart Filtering Capability" chapter.
What problem was solved? Quickly find the full-reduced products in the shopping cart that meet the unified gear to make up the order;
Improve the efficiency of order collection in the shopping cart;
Existing problems: Cross-store full reduction of orders, the user's order amount is still 25% less than the next threshold, accounting for 20%. At present, there are many users who have a small amount of order and there are card points;
The user's current shopping cart is full of cross-store reduction and screening floating layers. Judging from the current order collection link, the user's moving line is faulty;
📢 Cross-store full-reduction screening is the first time that Taobao shopping cart has a product function related to "commodity screening" in recent years. Set the most reasonable screening technical capabilities so far, and will help to launch pre-sale screening, price reduction screening, frequent purchases, search and other functions in succession. For related technical breakthroughs, you can directly read the "Shopping Cart Screening Capability" chapter.
[Stage 3: Accurate recommendation and complete and uninterrupted order collection link / worry-free collection]
Capability introduction In the screening floating layer of full-reduced products in the same tier, according to the difference between the price calculation result of the product selected by the user and the amount of the next tier, the algorithm dynamically recommends the product of the corresponding amount in real time, and the user adds to the purchase (in the information flow / product Details), the shopping cart will not be refreshed, and the order result will be retained and automatically updated with the price calculation result after checking the added product.

What problems have been solved? Solve the problem of users' small amount of single card points, and help users to buy more products that they want to buy in the most favorable way;
After the user selects the recommended product, the calculation price is automatically updated, and the link to collect orders is no longer interrupted;
The existing problem path is still relatively deep;
The accuracy of the algorithm needs to be continuously upgraded and improved;
📢 It is also worth mentioning that the launch of the worry-free business is also a good example of technological breakthroughs supporting business development. The implementation of the worry-free project has actually achieved several breakthroughs (client and server):
After adding the purchase (information flow, details), the shopping cart will not be refreshed, and it will be automatically checked and participate in dynamic calculation;
A set of perceptual agreements are deposited between the details add-on and the shopping cart, and after one hop for the details add-on, return to the shopping cart to make various business customizations according to the agreement;
The overall user flow is as follows:

【Stage 4: Make orders in advance during the warm-up period】
Ability introduction

What problem was solved? Calculate the price of the official period of the big promotion in advance, and collect the order in advance;
Existing problems: The result of the current warm-up period is actually completely separated from the official period, and the user experience is also faulty;
In addition, the shopping cart price preview project is also a relatively big attempt and breakthrough made by the shopping cart in 2021. For the first time, we can see the price of the official period in advance during the warm-up period. It is worth mentioning that during the entire promotion period, the volume of user appeals was very high, and the relevant appeals that users were eager to use reached 3,900+ on the same day. Finally, during the entire Double Eleven period, the related appeals for collecting orders decreased by 34.8%. As a result, the project team has won the Group's Little Strawberry, the annual TOP list of hand Taobao experience, and the TOP list of public review of hand Taobao experience, and has passed the group's approval and submitted a national patent application.
Special Price Calculation Experience <br>One of the core functions of the shopping cart is to collect orders, but the order collection must be based on accurate or basically accurate price calculation. The price calculation here includes the single product discount rendered on the product card, as well as the store Deals, cross-store deals, and more. As we all know, the marketing rules of Double Eleven are becoming more and more complex, which also brings technical complexity to the price experience of the shopping cart. Marketing rules are becoming more and more complex, there are more and more types of discounts, and it is more difficult for consumers to understand. At this time, the shopping cart is the last line of defense before placing an order. Accurate price calculation, clear discount calculation expression, and intimate most Obtaining the entrance of the optimal solution is the core goal of the shopping cart price experience. A picture to more completely describe what the shopping cart has done to improve the price experience:

[Goal 1]: Due to the complexity of marketing and the characteristics of various scenarios, in fact, there is a certain degree of price inconsistency in all links from shopping guides, shopping carts, to ordering. The shopping cart needs to continuously improve the dynamic calculation, maintain and upload (Details) The price consistency of the next (order) tour;
Shopping cart price consistency and details: so far the shopping cart and details price experience are basically the same;
And order: So far, after excluding inconsistencies caused by some normal business logic reasons, the order price in the shopping cart is basically the same. At the same time, a picture summarizes the efforts made by the shopping cart at each stage of the price consistency experience:

Dynamic computing peaks are not degraded
The Double Eleven shopping cart in 2021 has completely realized the price calculation during the peak period without downgrading;
【Goal 2】: Offers are clearly expressed. Sometimes many customer complaints from users are not because the price is wrong, but because the user cannot understand the result of the discount calculation. Then the discount details of the shopping cart are responsible for the mission of allowing users to clearly understand the price calculation process;
Shopping cart discount details upgrade discount details undertakes the task of manual shopping cart discount calculation instructions. So far, it has undergone three relatively large upgrades. No matter in terms of development method or user experience, there has been great progress:
[Version 1: weex stage]
The initial version of the discount details adopts the weex version. The user clicks on the settlement bar to open the weex page, and the page re-sends a request to the server, requesting that the input parameters of the user's last update request are the same; as shown below:

There are several major problems with this version:
1. [Experience problem]: The user selects the product in the shopping cart, calculates the price, clicks on the discount details, the weex page, the time-consuming of the server interface rt, and the time-consuming of the weex page rendering, etc., all require the user to have a relatively obvious The waiting time is poor, and the experience is poor;
2. [Expression of inconsistency]: Because the information on the discount details page is the data obtained from the server after the user clicks, which means that the time gap and other reasons bring the inconsistency between the bottom settlement column and the discount details board information;
3. [Operation is not friendly enough]: The discount details of the old version only have static discount information. If the user wants to check the details while checking the products, it means that the details need to be opened and closed continuously, and each time it is opened, it will be a new one. requests, operations and their inconvenience;
4. [Preferential details are not rich enough]: As the name suggests;

[Version 2: Ultron Componentization Stage]
Based on the increasingly complex marketing activities and the experience of the shopping cart's own discount details, we have upgraded the discount details in the NewCart project. The main transformation points include:
1. [Ultra Componentization]: Componentize the discount details panel and display it in the form of a pop-up window;
2. [Request Mode Optimization]: All dynamic calculation requests are issued with the component information of the discount details pop-up window. The user clicks to view the details, and the pop-up window is directly rendered without additional server requests, which solves the problem of waiting for users to click and is related to dynamic calculation. Inconsistent results;
3. [Content & User Operation Optimization]: The new discount details panel displays all the checked products, and you can perform corresponding operations on these products on the details board page, and the details list is updated in real time;

[Version 3: Promotion of Offer Expression]
Version 2 basically solves the user experience problem, but in fact, for the shopping cart, as the last step before placing an order, the price-related experience is also very important. I think the price experience is mainly reflected in two aspects: the price consistency with the upper (details) and lower (order) tours, and how to clearly express the discount calculation process to users, that is, in addition to bringing users the correct price, To be able to clearly tell the user how the price came. And the details of the discounts have taken on this responsibility. Therefore, in the upgrade of the third edition, we mainly did the following things:
1 [Structure upgrade]: The preferential detail structure has been upgraded to four parts: product list, product detail expression, discount detail expression, and total; the product detail expression and discount detail expression are customizable parts of the business;
2 [Expression upgrade]: Express the common items in the shopping cart and the pre-sale deposit items separately, and at the same time express the availability and unavailability of almost all discounts;

[Goal 3]: In addition to accuracy and clarity of expression, in fact, the shopping cart should always have a mission, how to help users see the optimal solution of the discount and obtain the optimal solution through various channels, and reduce some IQ tax;
Shopping cart marketing capabilities In the above-mentioned complicated preferential marketing rules, some rights and interests need to be obtained by users actively, such as commodity coupons and single-item coupons in store discounts, category coupons in cross-store discounts, and gold coins and other user rights and interests . In Taobao, there are some card coupon collection entrances including details and card coupon center, which invisibly increases some burdens for users to obtain the optimal solution. Therefore, in addition to marketing expressions, the shopping cart has also launched capabilities such as coupon collection settlement, gold rush coin collection calculation, and automatic surprise coupon collection, which helps users shorten the chain of rights acquisition and obtain the best solution for commodity purchases in the most efficient way. ;

Shopping Cart Commodity Management Experience <br> Shopping Cart Commodity Management Experience: In terms of the volume of requests for shopping carts in January, it is mainly related to the inability to add purchases, and more than half of them are caused by the full capacity of the shopping cart; In the statistics of user behavior after a failed purchase, only a small proportion of the products will be ordered through the details in daily life, about 1/3 of the products will be added to the product again, and other add-on purchase traffic will be lost. The blocking of additional purchase paths has become one of the biggest problems for users to use shopping carts. The solution to the capacity problem has been continuously optimized over the years. In summary, it has gone through the following stages:

[Phase 1: Direct expansion]
Expansion for users is undoubtedly the most direct and immediate method. From 2017 to now, Taobao shopping carts have also adopted different expansion methods:
Super shopping cart: On Double Eleven in 2017, the platform provides users with the option of using Tmall points to exchange capacity;
88VIP shopping cart expansion: 88VIP users will be eligible for expansion during Double Eleven in 2019;
Pre-sale shopping cart: A separate 40 capacity is provided for pre-sale products, which does not account for the capacity of spot products;
But in fact, users' demands for capacity will not stop, and simply expanding capacity does not fundamentally solve the problem, and is not a long-term solution.
[Phase 2: Directly solve the additional purchase card points]
At the point where the user fails to add a purchase due to full capacity, it provides more choices, selects products whose one-click has expired or has a low purchase intention to be moved into the favorites, and successfully adds products with a stronger purchase intention at the moment. In the key behavior stage of users, it can effectively undertake the failure of add-on purchases, reduce user operation costs through compound operations and other strategies, shorten the user path, and effectively restore high-value add-on products that are willing to buy, thereby restoring the loss of GMV.

[Stage 3: Stimulate users to actively clean up]
According to the data, about 50% of users will take the initiative to replace and delete products when they encounter the error message layer for additional purchases, but only 20% of them will go back to the shopping cart to delete them in batches. Although the add-on-purchase cleanup function helps users solve the problem of add-on products at the detailed add-on place, the efficiency of deleting a single product is extremely low, and it is necessary to guide users to remove and clean up batches of products at an appropriate time.

The timing of the cleanup prompt is also a matter of knowledge. In the current phase, the selection is first revealed based on the number of items in the user's shopping cart. The timing of the cleanup entry will be judged later based on factors such as the category of the product that the user placed the last order.
Optimizing the deletion/cleaning experience For users who are willing to return to the shopping cart for batch cleaning, it is also very important to improve the experience of deleting items and make it easier for users to delete more.

Shopping Cart Product Management Experience: Product Discovery Problems

Commodities are the core of the shopping cart. The first step is to help users successfully and efficiently add products based on their purchase intentions to the shopping cart. The second step is how to allow users to quickly discover/find the desired products in the shopping cart.

Quick search locates about 55 related public opinions on the shopping cart per month, ranking in the top 5 of shopping cart related public opinion issues; when users have a clear purchase intention, the search ability is actually the first weapon for users to quickly find products.

In addition to the search ability to help users quickly find related products when they have a clear purchase intention, the Taobao shopping cart also provides multiple entry points for product classification and screening, which further improves the efficiency of users in finding products. In addition, improve the reach efficiency of some low-purchase-intention products, make the products in the shopping cart "live", improve the user experience and bring about an increase in conversions.

Where is the conversion-promoting space for shopping cart scenarios?

Where do shopping cart conversions come from

As a product with a tool attribute on the basic link, is there room for conversion improvement? For shopping carts, both conversion and experience are inseparable from the two cores of "people" and "products". We analyze the conversion space of shopping cart scenarios from these two aspects.

People <br>From the perspective of consumers, improving conversions can be summed up in three things:
The additional purchase link is smooth, and the user's additional purchase intention and products with high purchase intention will not be lost;
When a user places an order in one shopping cart, how to make it easier for the user to buy/collect more at one time?
How to recall users to the shopping cart, always maintain a high level of interest, and buy more times?

Commodities <br>Coming back to the product point of view, there are a large number of unconverted products in the shopping cart, including normal products and about 17% of invalid products. When we analyze the relationship between the time of adding a product and its conversion rate, we will find a very interesting phenomenon. After the product is added for more than one day, the conversion rate will drop sharply. This means that as the additional purchase time increases, the user's desire to purchase the product becomes lower and lower. The reason for the decrease may be: the product gradually sinks to the bottom with the additional purchase, and the exposure efficiency begins to decrease; the price of the product or the state of the product has not changed. , or there are obvious changes but the changes have not been brought to the user; based on this, we divide the shopping cart products into three types [sleeping products]: as the additional purchase time continues to accumulate, the longer the additional purchase time, the more difficult it is to be purchased. Conversion, for products whose conversion rate is lower than a certain value, we define them as sleeping products.
[Invalid product]: After the user successfully added the purchase, the product in the shopping cart cannot continue to be operated due to the merchant taking the initiative to remove the product, re-editing and publishing the product, and other reasons. Expired products can no longer view the product details, and tick, you can only bookmark or delete the product directly.
[Commodities with high purchase intentions]: Products that have been recently added to the shopping cart, or have information such as price changes and are exposed in the shopping cart, and have high purchase intentions for users;

From the perspective of "people" and "products", the overall strategy of Taobao shopping cart conversion in the past two years is summarized as follows:

The core of improving the conversion of shopping carts is to make the shopping cart products “flow”: some of the expired products are replaced by users’ failure to add purchases, and some are converted into products with higher purchase intentions by finding the same model; for sleeping products, seize their prices. Or the key points of the state change can effectively reach the user, increase the user's desire to buy, and thus bring about an increase in conversion;

Introduction to key projects

Sleeping Goods Awakening Project <br>We illustrate the business context of the project by answering a few questions:
What are sleeping commodities? As we have already introduced before, we call those products whose conversion rate decreases with the increase of purchase time as "sleeping products";
What is wake up? wake up what? When to wake up? How to wake up? The so-called awakening will affect the change of the status of the commodity placed in the order to reach the consumer, and reactivate/awaken the desire to purchase the commodity. The core logic of the overall product is summarized as follows:

So there are two key points here: when to wake up and how to wake up, namely the wake-up strategy and the touch strategy.
[Wake-up strategy] Define the status changes that affect the decision to place orders for products: marketing activities, daily price reductions, high price power, price increases Status changes: tight inventory, replenishment, new products Merchant content changes: sharing content

[Reach strategy]: The wake-up method is from outside the station to the Taobao shopping cart, and different layers of information are designed with different means of reaching the outside of the station: desktop push message reminder;
Taobao internal reflow: Tab bubble reminder at the bottom of Taobao;
Browsing in the shopping cart: prompts the user that some items in the shopping cart have been reduced in price;
So far, the Taobao shopping cart has supported the desktop push reminder and return mechanism for daily price-cut products. More wake-up strategies for price changes and product status changes will be added in the future, as well as touch strategies such as Taobao bottom bar reminders.

Surprise Coupon Project <br>Project Background (1) From the perspective of shopping cart: Analyze the distribution of the relationship between the time of adding products to the shopping cart and the conversion rate of placing an order. The conversion rate has begun to decrease significantly; at the same time, there are a large number of products in the shopping cart stock that have been added for more than one day; this means that there is a large room for improvement in the conversion rate of these products with a longer add-on time;
(2) Starting from the ability of merchants to operate: At present, the platform is gradually cultivating the minds of users who are self-operated by merchants, and provides platform tools for merchants to operate customers, such as Taobao Live - there will be clear coupons for users who come in from shopping carts. In the face of the situation of the goods in the shopping cart and the relationship between people and goods, we hope to provide a self-operated tool for merchants for the shopping cart scene of the transaction link. Targeted marketing makes the user's hesitation point disappear, and then becomes a conversion;

The key points of the business solution are the self-operation of the merchants: the merchants select the products to register, and set the coupon information according to the user's additional purchase situation;
Access to shopping cart marketing rights: The shopping cart scenario exposes the marketing rights acquisition portal set by the merchant to consumers to encourage consumers to receive them;
Shopping cart user interest point expression: the special discount that the user has received is expressed on the shopping cart page, which improves the user's purchase intention;

Project goal: Link consumers and merchants through the platform, realize the automatic distribution of rights and interests, and maximize the conversion rate of products

business evolution

The shopping cart surprise coupon business has undergone three business model optimizations. We summarize from the "merchant" and "consumer" modules as shown below.

Precipitation on the road of business support

The innovative attempts of various business scenarios of Taobao shopping cart are inseparable from the exploration and breakthrough of technology. Of course, the development of technology must also serve the business, with the goal of winning the business first. However, the shopping cart, which is a rare scene of adding, deleting, modifying and checking all lists on the basic link, objectively has some unique technical development difficulties. Then this part will analyze the technical difficulties of shopping cart business development, and how to make breakthroughs one by one to support business development, and accumulate general scalability to improve business efficiency.

Technological breakthroughs support business development

Technology Architecture

First, use a big picture to describe the technical architecture of Taobao shopping cart, to have a global understanding

Analysis of technical difficulties in shopping cart development

Features of shopping cart business scenarios <br>Large traffic, stability and product availability are always the prerequisites and foundations;
Business scenarios in the form of lists, including additions, deletions, revisions, and searches. Although the capacity of the shopping cart is restrained, it must continue to grow;
The call-out representative application has many dependencies (downstream of discounts, commodities, inventory, purchase restrictions, etc.);
Basic products, experience first (the user experience cannot be reduced as the complexity of the business increases);

Technical difficulties in shopping cart development <br>Based on the characteristics of shopping cart business scenarios and the experience in developing shopping carts in recent years, I think there are two biggest technical difficulties in developing Taobao shopping cart business;

The introduction of marketing shopping guide/algorithm and other scenarios into the transaction core link Every year, the purchasing power and demands of users are increasing, and the business is developing continuously. There are more and more business demands for marketing shopping guide/algorithm, but the downstream dependencies are increasing, and the core link consumes more and more. Time growth, plus the shopping cart is a list form, any complexity will increase exponentially with the increase of the product list, which is a huge challenge to stability and user experience. But no matter how the business develops, stability and user experience always come first. The final result is that the rt experience on the user side is not improved, and the business of marketing shopping guide/algorithm cannot be implemented.
So is it possible to redefine the core data and non-core data of the shopping cart? The core data maintains a relatively stable data magnitude and complexity, and non-core data is used to complete business growth; for the shopping cart scenario, the distinction between whether the core should be two dimensions, first, the data magnitude dimension, that is, shopping The cart list, arranged in order, is each pit a core data for the user to enter the shopping cart (not just paging)? Second, the richness of information, such as the basic data and marketing data of the shopping cart, or the data required for placing an order and the data of the expression of interests;

The contradiction between the shopping cart paging rendering scene and the requirement of global content perception is how to perceive and aggregate the information of all the items in the user's shopping cart in advance under the premise of shopping cart paging rendering (the user actively turns the pages) to create certain atmospheres. It is expressed that many appeals of shopping carts have always been related to it: the first version of the surprise coupon is revealed, the screening entrance is revealed by the logic, tab prompts, etc.; here the perception appeal of the global product includes two levels, the basic product Logo (id, etc.), basic information of commodity IC (commodity label, etc.), complex commodity information (preferential with shopping cart scene semantics, invalidation, etc.); different levels of dependence on shopping carts;

Business upgrades brought about by technological breakthroughs

Cart filtering capability
Before Double Eleven in 2020, the Taobao shopping cart did not have any product filtering function. Commodity screening and classification and improving discovery efficiency have always been the top-level user demands in user research reports.

Difficulty analysis paging mode: the shopping cart is a paging query, and the filtered products may be distributed on different pages (for example, in extreme cases, the filtered products are expected to be on the last page of the shopping cart paging query);
Consolidated settlement: The product selection operation under each filter "page" needs to be retained during the switching process, and the dynamic calculation results at the bottom need to be retained (for example, the products filtered under the filter floating layer are the last page in the shopping cart paging query, and Selected by the user to participate in the dynamic calculation. After closing the floating layer and returning to the product, the product is selected and retained, and the dynamic calculation result is retained);
The user's operating experience requirements, such as the bottom price cannot jump, etc.;

Technical solution selection

Scheme type	Scheme description	Advantage	disadvantage
Client Screening	The user refreshes the shopping cart, queries the first page and preloads, and renders 50 items;

Click to collect and filter, and the complete information of the remaining products in the user's shopping cart will be pulled out at one time;
The server will mark the product and other components, and the client will display/hide the product according to the component label to achieve product screening | client screening, the filter floating layer check state can be kept forever, and there will be no return to the big shopping cart to slide and settle. column jumping problem;
Click to collect orders to filter and pull back all the products, so as to avoid abnormal flow problems such as various timings caused by serial sequential requests from the client.
All products are delivered at one time, and subsequent screening can be completed by the client, and there is basically no re-call to the server;
Follow-up capabilities can be reused, such as pre-sale screening; | Some traffic, the number of loaded products is more than 40, there are problems of high rt, poor user experience and performance pressure on the server (refer to the later stress measurement traffic evaluation, peak Downgrade) (need to evaluate the rt time to see the experience problem on the terminal) |
| Server Screening | Click Screening, the server will only deliver cross-store full-reduction products;
Click to close, because you need to keep the check state, you need to pull down the complete information of other remaining products in the user's shopping cart;
Click Filter/Close again, keep the logic consistent with the client-side filtering scheme, and do client-side filtering; | Precipitation ability, subsequent filtering ability can be reused | Compared with the client-side filtering scheme, the traffic is not reduced (the performance pressure is not reduced), and then Click to close, but one more call;
When opening and closing for the first time, there will be a problem that the rt is too high and the user experience is poor; |

The final technical solution is finally combined with the specific demands of the project and we choose the method of client screening. The sequence of the solution is shown in the following figure:

The essence of client-side screening, in fact, from the perspective of the client-side, is to display/hide the components according to the screening target of each component and the current screening page, so as to realize the final product screening. For example, the main components of the current shopping cart are as follows:

Then for each filter page, the structure of each filter state is as follows:

Combined with the scene of shopping cart paging rendering product list, the client-side filtering scheme is used to realize the shopping cart filtering ability that is checked and retained in the paging scene, and has scalability in the dimensions of independent shopping cart, filtering entry method, and filtering item strategy; and supports The big promotion collection order screening, pre-sale screening, multi-tab and other services are launched;

程序员