TW202318310A - System, method, and computer program product for managing group-purchase orders - Google Patents

Abstract

A method for managing group-purchase orders is provided and allows seller devices to establish multiple management groups. Each management group includes multiple community groups. After buyer devices send order information to a server, the server merges the order information from different community groups but belonging to the same buyer device, and merges the order information from different management groups but belonging to the same buyer device. The server also provides an option to the seller devices on a management platform to merge the order information from different buyer devices.

Description

Group purchase order management system, method and computer program product

This disclosure relates to an order management system and method for online group buying.

In recent years, online shopping methods have become more and more popular, including group shopping (referred to as group buying). This mode is that the seller gathers enough buyers to purchase or use the seller's items, coupons or services at a preferential price. . However, sellers may sell goods on multiple social platforms, how to manage orders from buyers is a topic of concern to those skilled in the art.

An embodiment of the present disclosure proposes a group buying order management system, including a server, a buyer's device, and a seller's device. The server is used to provide a management platform. The seller's device is used to establish multiple management groups through the management platform, and each management group includes multiple community groups. The buyer's devices are respectively used to join the social group according to the social account. The seller's device is used to publish the product information link in the community group, the buyer's device opens the purchase interface provided by the server through the product information link, and the buyer's device sends the order information to the server through the purchase interface. The server is used to merge order information from different community groups but belonging to the same buyer's device, and merge order information from different management groups but belonging to the same buyer's device. The server is also used to provide an option on the management platform for the seller's device to combine order information from different buyer's devices.

In some embodiments, the server is also used to establish a hierarchical group buying structure. The hierarchical group buying structure includes multiple levels, and each level includes a seller device and at least one buyer device. The server is used to collect the order information of all the buyer's devices in the hierarchy, and transmit the cash flow information or logistics information in the order information to the highest-level seller's device.

In some embodiments, the server is used to transmit the product information link generated by the highest-level seller device to the lower-level seller device.

In some embodiments, the server is also used to add a virtual robot in the community group. The virtual robot inputs multiple information of the buyer's device into the first neural network to generate a state feature vector, and sends the corresponding seller's device The information is input to the second neural network to generate an action feature vector, and the inner product of the state feature vector and the action feature vector is calculated as an output value. The server is also used to determine the reward value according to whether the seller device accepts the order, calculate the error between the reward value and the output value, and train the first neural network and the second neural network according to the backpropagation method.

In some embodiments, when the corresponding seller device receives an order, the virtual robot sets the reward value to be positive. When the seller device rejects the order, the virtual robot sets the reward value to be negative. Otherwise, calculate the reward value according to the following mathematical formula.

為動作集合中的候選動作特徵向量，候選動作特徵向量表示一個回應訊息。

為下一個狀態特徵向量，

為狀態特徵向量

與候選動作特徵向量

之間的内積。 where A is the action set,

is a candidate action feature vector in the action set, and the candidate action feature vector represents a response message.

is the next state eigenvector,

is the state eigenvector

and candidate action feature vector

Inner product between.

From another point of view, the embodiments of the present disclosure provide a method for managing group buying orders, which is executed by a server. The group purchase order management method includes: providing a management platform for multiple seller devices to establish multiple management groups, wherein each management group includes multiple community groups, and multiple buyer devices are used to join community groups according to community accounts group; provide a purchase interface, whereby the seller’s device publishes a product information link in the community group, and the buyer’s device opens the purchase interface through the product information link; and receives order information from the buyer’s device, merging information from different community groups but belonging to the same The order information of the buyer's device is merged with the order information from different management groups but belonging to the same buyer's device, and an option is provided on the management platform for the seller's device to merge order information from different buyer's devices.

From another point of view, the embodiment of the present disclosure proposes a computer program product, which is executed by a computer system to implement the above-mentioned group buying order management method.

FIG. 1 is a schematic diagram illustrating a group buying order management system according to an embodiment. Please refer to FIG. 1 , the group buying order management system includes a server 110 , multiple seller devices 121 - 123 and multiple buyer devices 171 - 173 . The seller's devices 121-123 and the buyer's devices 171-173 may be smart phones, tablet computers, personal computers, or any electronic devices that have computing power and can be connected to the Internet. The server 110 provides a management platform, and the seller's devices 121-123 can use the management platform through a browser or an application program to manage orders from the buyer's devices 171-173. For example, the management platform will provide a user interface, as shown in Figure 2, the user interface 200 allows the seller to browse current orders and view related information, but Figure 2 is only an example, and the present disclosure does not limit the user interface Content.

Each seller device 121~123 can manage its own orders through the management platform. Taking the seller device 123 as an example, the seller device 123 can establish multiple management groups 141~143. Customer groups can be established. For example, the seller may sell products from multiple countries. The seller can establish a management group for each country’s products. The management group 141 can be called "Japan Pavilion", and the management group 142 can be called "Korea Pavilion". By analogy, in some embodiments, each management group 141-143 can also be called a "store". Each management group 141-143 includes a plurality of community groups, for example, the management group 141 includes community groups 151-153, and the management group 142 includes community groups 161-163. These community groups can be LINE groups, LINE communities, Facebook fan pages, Facebook communities, other online platforms such as Shopee, etc., and this disclosure is not limited thereto. Here, there can be more than one community group for each social software, for example, there can be multiple LINE groups under the same management group. Buyer devices 171~173 can join these social groups 151~153, 161~163 according to their own social accounts. The same buyer device can join multiple social groups, for example, the buyer device 172 joins the social groups 152 and 161 , and the buyer device 173 joins the social groups 162 and 163 .

The group buying methods of different community groups are described here. If the social group is a LINE group or a LINE community, the seller device 123 can publish a product information link in the corresponding social group, and the product information link can be a message in a conversation or a picture link, etc. For example, FIG. 3 shows the interface 300 of the LINE group, and the product information link 310 is presented in the dialog in the form of a picture. The buyer's device can click the product information link 310, thereby opening the purchase interface 400 provided by the server 110 (as shown in FIG. 4, but FIG. 4 is only an example, and this disclosure does not limit the content of the purchase interface), and the buyer's device can select a product specifications and quantities, etc., the buyer’s device can send order information to the server 110 through the purchase interface 400 (for example, by pressing the confirm button). Payment amount, payment method, etc., are not limited in this disclosure.

In some embodiments, the server 110 can add a virtual robot in the LINE group, and the virtual robot can create order information according to the conversations in the social group. For example, the virtual robot can obtain the keyword of the product from the information provided by the seller's device. The keyword is, for example, the specification of the product. The virtual robot will also obtain the buyer's message and determine whether the message contains the keyword of the above specification, "+", "plus" and other symbols or words representing the purchase, as well as the quantity to be purchased, if the information is complete, the virtual robot will automatically create the order information. For example, the buyer can leave a message "size M, +1", and the virtual robot will automatically capture such a conversation and create a corresponding order, where the specification is "size M" and the quantity is 1.

If the community group is a Facebook fan page or a Facebook community, the seller's device can publish a post containing information about the product to be sold. The buyer can decide whether to buy by leaving a message, for example, leave a message "size M, +1" and so on. The server 110 can automatically retrieve the message from the buyer's device and create order information through the programming language of the webpage, such as HyperText Markup Language (HTML).

If the community group is Shopee, etc., these online platforms can provide a shopping information application program interface (application interface, API), and the server 110 can retrieve the specifications, quantity, address, etc. of the order through this API. .

However, in some situations, the buyer may give an ambiguous message, such as leaving only "+1" without deciding on the size, or the buyer may not use the usual language, such as leaving "I want to buy one piece in size M", In this case, the virtual robot can use the machine learning model to automatically generate a dialogue or respond to the message, so as to ask the buyer for some missing information. Specifically, neural network-like and reinforcement learning can be combined here. In reinforcement learning, it is generally necessary to define a state set S, an action set A, and a reward r. In some conventional techniques, a state s in the state set S and an action a in the action set are input to a neural network, so as to calculate the value of such a combination (s, a). However, in this disclosure, two neural networks are used to process state s and action a respectively. Specifically, the virtual robot can first capture past conversations between the seller's device and the buyer's device, and input the message sent by the buyer's device into the first neural network to generate a state feature vector, denoted as s. In addition, the virtual robot will also input the information sent by the seller's device into the second neural network to generate an action feature vector, denoted as a. The first neural network and the second neural network mentioned above are, for example, long short-term memory models (Long Short-Term Memory, LSTM), or other types of recurrent neural networks (recurrent neural network, RNN). The outputs of the above-mentioned first neural network and the second neural network can be combined together. In this embodiment, the inner product between the action feature vector a and the state feature vector s is calculated as the output of the entire network architecture.

，其中

表示經過k個訊息之後的狀態，

表示在狀態

所採取的動作，

表示執行動作

之後的獎勵值，

表示執行動作

的下一個狀態。當賣方裝置接受訂單時可以設定獎勵值為正，當賣方裝置決定拒絕訂單時可設定獎勵值為負。如果還無法確定是否要接收訂單，還需要詢問一些訊息，則可以根據以下數學式計算獎勵值。

An action set A can be established according to the messages output by the seller's device. Each action in this set A represents a response message, for example, "what is the size", "what is the quantity", "what is the color", Or do not respond, etc., this disclosure is not limited thereto. Here, a training sample is expressed as

,in

Indicates the state after k messages,

indicated in the state

actions taken,

express action

After the reward value,

express action

the next state of . The reward value can be set positive when the seller device accepts the order, and the reward value can be set negative when the seller device decides to reject the order. If you still can't determine whether to receive the order and need to ask for some information, you can calculate the reward value according to the following mathematical formula.

為動作集合A中的一個候選動作特徵向量，也表示為一個回應訊息。

表示狀態特徵向量

與候選動作特徵向量

之間的内積。 in the above mathematical formula

is a candidate action feature vector in the action set A, and is also represented as a response message.

Represents the state eigenvector

and candidate action feature vector

Inner product between.

之間的誤差(例如平方差)，並根據反向傳播(back propagation)來訓練第一神經網路與第二神經網路。如此一來，上述的增強式學習方法可以根據買方裝置習慣使用的語言適應性且自動的產生回應訊息，在適當時候提出“請問尺寸是什麼”等問句，藉此協助完成訂單。 For each training sample, after determining the reward value, the reward value and output value can be calculated

The error (such as the square difference) between them is used to train the first neural network and the second neural network according to back propagation. In this way, the above-mentioned enhanced learning method can adaptively and automatically generate a response message according to the language used by the buyer's device, and ask questions such as "what is the size?" at an appropriate time, thereby assisting in completing the order.

After the server 110 collects the order information from the buyer's device, it can be presented to the seller's device 123 through the above-mentioned user interface 200 . In some embodiments, the server 110 may also provide some options on the user interface 200 to combine part of the order information. For example, the buyer device 173 purchased different products in the community group 162 and the community group 163 at the same time, and the delivery addresses of these order information are the same. If the two orders are sent in two packages, the logistics will be increased cost. Therefore, the server 110 can combine order information from different community groups and belonging to the same buyer's device. In some embodiments, order information from different management groups can also be merged. For example, the buyer device 172 purchases different commodities in the community group 152 and the community group 161 at the same time, wherein the community group 152 and the community group 161 belong to different management groups 141 and 142 respectively, and the server 110 also These two pieces of order information can be merged. In some embodiments, order information from different buyer devices can also be combined. For example, the buyer device 171 and the buyer device 172 purchased commodities in the same community group 152, and the buyer device 171 and the buyer device 172 may belong to two families belonging to the same community or are colleagues of the same company. Two buyers may want to combine orders to save on shipping costs. Therefore, the server 110 may provide an option to the seller device 123 to combine order information from different buyer devices 171 , 172 .

In some embodiments, the server 110 can also establish a hierarchical group buying structure. As shown in FIG. 5, the hierarchical group buying structure 500 includes multiple levels L1~L3, wherein level L1 belongs to the highest level, and level L3 is the lowest level. . Each level includes a seller device and multiple buyer devices, and the seller device sells commodities to other buyer devices in the same level. In particular, the buyer's device can "open a group" to build the next level. At this time, the buyer's device is also the seller's device of the next level. For example, the seller device 501 sells commodities to the buyer devices 511-513 and the seller device 502, and the seller device 502 also establishes its own level L2 and has its own buyer devices 514-515. Similarly, the seller device 503 is the buyer of the seller device 502 and also establishes its own level L3. That is to say, the seller device 501 sells the commodity to the seller device 502 and other buyer devices 511-513, and the seller device 502 resells the commodity to the seller device 503 and other buyer devices 514-515, and the seller device 503 resells the commodity to The buyer installs 516~517. Through reselling, the seller's devices 502 and 503 can earn the price difference of the commodity or share some profits. In such a hierarchical group buying structure 500, each level L1~L3 has its own logistics and cash flow to be managed, but this may not be efficient. Then ship to the seller's device 503, and finally the seller's device 503 can be shipped to the buyer's devices 516~517. Alternatively, the buyer devices 516-517 must first pay the seller device 503, and the seller device 503 then pays the seller device 502, and finally the seller device 502 pays the seller device 501. Therefore, in this embodiment, after constructing such a hierarchical group buying structure 500, the server 110 will collect the order information of all buyer devices in all levels L1-L3, and send the cash flow information and/or logistics information therein to The vendor device 501 of the highest level L1. In this way, all the buyer's devices can pay directly to the seller's device 501, and the seller's device 501 can also be directly shipped to all the buyer's devices, which can save logistics and cash flow costs, procedures, and/or required time.

In some implementations, the above-mentioned product information link 310 containing pictures may be made by the seller device 501 at the highest level L1, and the server 110 may automatically transmit the product information link 310 to the lower-level seller devices 502, 503, so Firstly, the seller's devices 502 and 503 do not need to repeatedly create the product information link 310, thereby encouraging group opening.

Fig. 6 is a flow chart illustrating a method for managing group buying orders according to an embodiment. Please refer to FIG. 6 , the method is executed by the server 110 . In step 601, a management platform is provided for multiple seller devices to establish multiple management groups, wherein each management group includes multiple community groups, and multiple buyer devices are respectively used to join the community groups according to the community accounts. In step 602, a purchase interface is provided, whereby the seller's device publishes a product information link in the social group, and the buyer's device opens the purchase interface through the product information link. In step 603, receive order information from the buyer's device, merge order information from different community groups but belonging to the same buyer's device, merge order information from different management groups but belong to the same buyer's device, and provide options to the seller on the management platform Device to combine order information from different buyer devices. However, each step in FIG. 6 has been described in detail above, and will not be repeated here. It should be noted that each step in FIG. 6 can be implemented as a plurality of program codes or circuits, and the present invention is not limited thereto. In addition, the method in FIG. 6 can be used in combination with the above embodiments, or can be used alone. In other words, other steps may also be added between the steps in FIG. 6 .

From another point of view, the present invention also proposes a computer program product, which can be written by any programming language and/or platform. When the computer program product is loaded into the computer system and executed, the above-mentioned Group buying order management method.

In the above-mentioned system and method, it can assist the seller to combine order information to save logistics costs, and can also help to establish a hierarchical group buying structure to manage cash flow and logistics, and can also use artificial intelligence to automatically generate dialogues to help complete orders.

Although the present invention has been disclosed above with the embodiments, it is not intended to limit the present invention. Anyone with ordinary knowledge in the technical field may make some changes and modifications without departing from the spirit and scope of the present invention. The scope of protection of the present invention should be defined by the scope of the appended patent application.

110: server 121~123: The seller's device 141~143: Management group 151~153,161~163: Community groups 171~173: Buyer's device 200: user interface 300: interface 310: Product information link 400: purchase interface 500: Hierarchical group buying structure 501, 502, 503: Seller's device 511~517: Buyer's device L1~L3: level 601~603: Steps

In order to make the above-mentioned features and advantages of the present invention more comprehensible, the following specific embodiments are described in detail together with the accompanying drawings. FIG. 1 is a schematic diagram illustrating a group buying order management system according to an embodiment. FIG. 2 is a schematic diagram illustrating an interface of a management platform according to an embodiment. FIG. 3 is a schematic diagram illustrating an interface of a LINE group according to an embodiment. FIG. 4 is a schematic diagram illustrating a purchase interface according to an embodiment. Fig. 5 is a schematic diagram illustrating a hierarchical group buying structure according to an embodiment. Fig. 6 is a flow chart illustrating a method for managing group buying orders according to an embodiment.

110: server

121~123: The seller's device

141~143: Management group

151~153,161~163: Community groups

171~173: Buyer's device

Claims (7)

A group purchase order management system, comprising: a server for providing a management platform; a plurality of seller devices for establishing a plurality of management groups through the management platform, wherein each of the management groups includes a plurality of community groups; and a plurality of buyer devices respectively used to join one of the social groups according to a social account, The seller's devices are used to publish a product information link in at least one of the social groups, the buyer's devices open a purchase interface provided by the server through the product information link, and the buyer's devices use the the purchase interface sends an order information to the server, Wherein the server is used to merge the order information from different social groups but belonging to the same buyer device, and merge the order information from different management groups but belonging to the same buyer device, Wherein the server is also used to provide an option on the management platform for the seller devices to merge the order information from different buyer devices. The group purchase order management system as described in claim 1, wherein the server is also used to establish a hierarchical group purchase structure, the hierarchical group purchase structure includes a plurality of levels, and each of these levels includes one of the seller devices and at least one of those Buyer Devices, Wherein the server is used to collect the order information of all the buyer's devices in the layers, and transmit the cash flow information or logistics information in the order information to the seller's device at the highest level in the layers . The group buying order management system as described in Claim 1, wherein the server is used to transmit the product information link generated by the seller's device at the highest level to the seller's device at a lower level. The group buying order management system as described in claim 1, wherein the server is also used to add a virtual robot to one of the community groups, and the virtual robot sends a plurality of messages from one of the buyer's devices input to a first neural network to generate a state feature vector, and input a plurality of messages corresponding to the seller's device to a second neural network to generate an action feature vector, and calculate the state feature vector and the action feature The inner product of the vectors is used as an output value, The server is also used to determine a reward value according to whether the corresponding seller device accepts an order, calculate the error between the reward value and the output value, and train the first neural network and the first neural network according to the backpropagation method. Second neural network. The group buying order management system as described in Claim 4, wherein when the corresponding seller's device receives the order, the virtual robot sets the reward value to be positive, wherein when the corresponding seller's device rejects the order, the virtual robot sets the reward The value is negative, otherwise the reward value is calculated according to the following mathematical formula:

where A is an action set,

is a candidate action feature vector in the action set,

is the next state eigenvector,

is the state eigenvector

and the candidate action feature vector

Inner product between. A group purchase order management method, executed by a server, the group purchase order management method includes: Provide a management platform for multiple seller devices to establish multiple management groups, wherein each of these management groups includes multiple community groups, wherein multiple buyer devices are respectively used to join the community groups according to a community account one of the groups; providing a purchase interface, whereby the seller devices post a product information link in at least one of the social groups, and the buyer devices open the purchase interface through the product information link; and receiving an order information from the buyer devices, merging the order information from different community groups but belonging to the same buyer device, merging the order information from different management groups but belonging to the same buyer device order information, and provide an option on the management platform for the seller devices to merge the order information from different buyer devices. A computer program product executed by a computer system to perform a number of steps: Provide a management platform for multiple seller devices to establish multiple management groups, wherein each of these management groups includes multiple community groups, wherein multiple buyer devices are respectively used to join the community groups according to a community account one of the groups; providing a purchase interface, whereby the seller devices post a product information link in at least one of the social groups, and the buyer devices open the purchase interface through the product information link; and receiving an order information from the buyer devices, merging the order information from different community groups but belonging to the same buyer device, merging the order information from different management groups but belonging to the same buyer device order information, and provide an option on the management platform for the seller devices to merge the order information from different buyer devices.

Images