Methods for determining the potential sales volume of an online commerce product and apparatus thereof

Abstract

The present disclosure relates to a method and an apparatus for determining a potential sales volume of an online commerce product. According to an embodiment of the present disclosure, a method for determining a potential sales volume of an online commerce product including: determining a set of queries related to a first product; obtaining first search data of a first query included in the set of queries during a first period; determining a first exposure probability, which is a probability that the first product is exposed due to a search of the first query, using the first search data; obtaining second search data of the first query during a second period; and determining a potential sales volume of the first product during the second period using the first exposure probability and the second search data, and an apparatus thereof may be provided. In this case, the first period may be a period prior to the second period.

Description

Technical Field

[0001] This disclosure relates to a method and apparatus for determining the potential sales volume of goods in online commerce. Specifically, it relates to a method and apparatus for determining the potential sales volume by substitute for abnormal sales volume in cases of stockouts or abnormal sales of goods in online commerce. Prior Technology

[0002] Online commerce refers to a form of e-commerce where goods or services are traded via the internet. Consumers can search, compare, and purchase products through online platforms, and payment and delivery can be done contactlessly.

[0003] For online commerce, although consumer demand exists, actual sales volume may not match consumer demand. For example, when a product is out of stock, consumers may want to buy it, but actual sales volume may be zero, leading to a mismatch between sales volume and demand. Furthermore, if the price of a product or the shipping cost is higher than expected, consumers may hesitate or abandon their purchase, resulting in demand existing but sales volume low. Thirdly, if a product has many negative reviews or low ratings, even if consumers have demand, they may not buy it due to a lack of trust in the product.

[0004] For online businesses, a mismatch between consumer demand and actual sales volume can lead to operational problems. For example, the failure to sell goods despite demand can cause errors in inventory management or pricing strategies. Furthermore, if consumers cannot purchase the products they want, it can reduce their satisfaction with the online business brand and weaken its competitiveness.

[0005] Therefore, there is an urgent need for a technology that can identify whether there is a discrepancy between consumer demand and actual sales volume in online commerce, and can reflect actual consumer demand, thereby improving the efficiency of online business operations.

[0006] [Existing Technical Documents]

[0007] [Patent Literature]

[0008] (Patent Document 0001) Patent Publication No. 10-1753480 (Publication Date: 2017.07.04) Summary of the Invention

[0009] The technical problem that the invention aims to solve

[0010] The technical problem to be solved by the embodiments disclosed herein is to provide a method and apparatus for facilitating the management of out-of-stock goods inventory.

[0011] Other technical problems to be solved by the embodiments disclosed herein are to provide a method and apparatus for calculating a numerical value representing the actual demand of consumers for out-of-stock goods.

[0012] Another technical problem to be solved by the embodiments disclosed herein is to provide a method and apparatus for calculating a numerical value representing the difference between the actual sales volume of a commodity and the actual demand of consumers.

[0013] The technical problems disclosed herein are not limited to those mentioned above. Those skilled in the art can clearly understand other unmentioned technical problems from the following description.

[0014] The technical solution of the present invention is as follows:

[0015] According to an embodiment of the present invention, a method and apparatus for determining the potential sales volume of online commercial goods can be provided. The method includes: determining a query set related to a first product; obtaining first search data of the first query included in the query set within a first period; using the first search data to determine the probability that the first product will be exposed due to the first query search, i.e., a first exposure probability; obtaining second search data of the first query within a second period; and using the first exposure probability and the second search data to determine the potential sales volume of the first product within a second period. Here, the first period can be a period prior to the second period.

[0016] In one embodiment, the second period may be the period during which the first product is out of stock in online commerce.

[0017] In one embodiment, the step of obtaining the first search data may further include: obtaining the third search data of the second query included in the query set within the first period; the step of determining the first exposure probability may further include: using the third search data to determine the probability that the first product is exposed due to the second query search, i.e., the second exposure probability; the step of obtaining the second search data may further include: obtaining the fourth search data of the second query within the second period; and the step of determining the potential sales volume of the first product may include: using the first exposure probability and the second search data to determine a portion of the potential sales volume of the first product within the second period, and using the second exposure probability and the fourth search data to determine another portion of the potential sales volume of the first product within the second period.

[0018] In one embodiment, the step of determining the first exposure probability may include: obtaining the number of times the first query was searched within the first period from the first search data; obtaining the number of times the first product was exposed by the first query within the first period from the first search data; and determining the first exposure probability using the number of times the first query was searched and the number of times the first product was exposed by the first query.

[0019] In one embodiment, the step of determining the potential sales volume of the first product may include: obtaining the number of searches for the first query within the second period from the second search data, and determining the potential sales volume of the first product within the second period using the number of searches for the first query within the second period and the first exposure probability.

[0020] In one embodiment, the step of determining the potential sales volume of the first product may include: obtaining the sales volume of the first product within the first period; obtaining the number of times the first product was exposed by the first query within the first period from the first search data; determining a first correlation ratio between the number of exposures of the first product and the sales volume of the first product using the sales volume of the first product within the first period and the number of times the first product was exposed by the first query within the first period; determining the expected number of exposures of the first product within the second period using the number of searches of the first query within the second period and the first exposure probability; and determining the potential sales volume of the first product within the second period using the first correlation ratio and the expected number of exposures of the first product.

[0021] In one embodiment, the step of determining the query set related to the first product may include: obtaining a first query and a third query that caused the first product to be exposed due to a search; including the first query in the query set; and excluding the third query from the query set. In this case, the number of searches for the first query may be greater than a first threshold, and the number of searches for the third query may be less than the first threshold.

[0022] In one embodiment, the step of determining the query set related to the first product may include: obtaining a first query and a third query that caused the first product to be exposed due to a search; including the first query in the query set; and excluding the third query from the query set. In this case, the exposure probability of the first query may be greater than or equal to a second threshold, and the exposure probability of the third query may be less than the second threshold.

[0023] In one embodiment, the step of determining the query set related to the first product may include: obtaining a first query and a third query that caused the first product to be exposed due to a search; including the first query in the query set; and excluding the third query from the query set. In this case, the correlation ratio between the first query and the sales volume of the first product may be greater than a third threshold, and the correlation ratio between the third query and the sales volume of the first product may be less than the third threshold.

[0024] In one embodiment, the method may further include: determining the quantity of the first commodity received into inventory during a third period using the potential sales volume of the first commodity during the second period. In this case, the third period may be a period following the second period, and the quantity of the first commodity received into inventory may be the quantity of the first commodity received into inventory at the logistics center distributing the first commodity.

[0025] According to another embodiment of the present invention, a method and apparatus for determining the potential sales volume of online commercial goods can be provided. The method includes: determining a query set related to a first product; obtaining first search data of a first query included in the query set within a first period; inputting the first search data and the sales volume of the first product within the first period into an artificial intelligence model for predicting the potential sales volume of the first product, so that the artificial intelligence model can learn; obtaining second search data of the first query within a second period; and inputting the second search data into the learned artificial intelligence model to output the potential sales volume of the first product within the second period. In this case, the first period can be a period prior to the second period. Simple Explanation of the Diagram

[0026] Figure 1 is a structural diagram showing the overall structure of an online business product potential sales volume determination device according to an embodiment of the present disclosure.

[0027] Figures 2 and 3 are example diagrams illustrating the applicable situations of some embodiments disclosed herein.

[0028] Figure 4 is a flowchart of a method for determining the potential sales volume of online business goods according to another embodiment of this disclosure.

[0029] Figure 5 is a detailed flowchart illustrating a portion of the work done in the method described with reference to Figure 2.

[0030] Figure 6 is a further detailed flowchart illustrating part of the work done with reference to Figure 2.

[0031] Figure 7 is a further detailed flowchart illustrating part of the work done with reference to Figure 2.

[0032] Figure 8 is a flowchart of a method for determining the potential sales volume of online business goods according to another embodiment of this disclosure.

[0033] Figures 9 to 12 are example diagrams illustrating the methods described with reference to Figures 2 to 8.

[0034] Figure 13 is a table illustrating in detail another embodiment of the method for determining the potential sales volume of online business goods disclosed herein.

[0035] Figure 14 is a hardware structure diagram of a computing device used in some embodiments of this disclosure. Implementation

[0036] The following describes various embodiments of the present disclosure in detail with reference to the accompanying drawings. The advantages and features of the embodiments disclosed, as well as the methods for implementing them, become apparent from the accompanying drawings and the embodiments described in detail below. However, the technical concept of the present disclosure is not limited to the following embodiments, but can be implemented in many different ways. The following embodiments are provided only to complete the technical concept of the present disclosure and to fully inform those skilled in the art of the scope of the disclosure. The technical concept of the present disclosure is defined only within the scope of the claims of the invention.

[0037] When using structural elements in the various figures, it should be noted that the same structural elements should use the same reference numerals even if they are shown in different figures. Furthermore, in describing the embodiments disclosed herein, detailed descriptions of related well-known structures or functions are omitted if it is believed that such descriptions might obscure the main points.

[0038] Unless otherwise defined, all terms used in this disclosure (including technical and scientific terms) are to be used in the sense commonly understood by one of ordinary skill in the art to which the embodiments of this disclosure pertain. Furthermore, terms defined in commonly used dictionaries should not be ideally or over-interpreted unless explicitly defined. The terms used in this disclosure are for illustrative purposes, and the embodiments of this disclosure are not limited thereto. In this disclosure, singular forms include plural forms unless specifically stated in the context.

[0039] Furthermore, when describing the structural elements of the embodiments disclosed herein, terms such as first, second, A, B, (a), and (b) may be used. These terms are used merely to distinguish the structural element from other structural elements, and do not imply that the essence, order, or sequence of the structural elements is limited by the term. When it is stated that certain structural elements are "connected," "combined," or "linked" with other structural elements, it should be understood that the structural elements may be directly connected or linked to other structural elements, but structural elements may also be "connected," "combined," or "linked" to another structural element.

[0040] First, before describing the embodiments of this disclosure, the definitions of some terms in this disclosure are described.

[0041] According to some embodiments of this disclosure, the potential sales volume (hereinafter referred to as potential sales volume) of online business goods can refer to the sales volume of goods reflecting the actual demand of online business users (hereinafter referred to as users) as consumers. Potential sales volume may differ from the actual sales volume (hereinafter referred to as actual sales volume) of online business goods. Actual sales volume refers to the quantity of the goods actually sold in online commerce. Conversely, potential sales volume can refer to the quantity of the goods that are expected to be sold without specific factors. Potential sales volume may be higher than actual sales volume. Furthermore, potential sales volume may also be lower than actual sales volume.

[0042] For example, when product A is out of stock, its actual sales volume is 0. However, if we assume product A is not out of stock, its sales volume may be greater than 0, for example, 100 units. That is, the predicted sales volume assuming product A is not out of stock is the potential sales volume of product A.

[0043] As another example, product A might be eligible for a discount promotion. In this case, the actual sales volume of product A might be 100 units. However, if we assume that product A is not eligible for a discount promotion, the sales volume of product A might be less than 100 units, such as 50 units. That is, the predicted sales volume of product A under the assumption of no discount promotion is the potential sales volume of product A.

[0044] The potential sales volume disclosed herein can be used to generate forecast data that can replace actual sales volume and can be used in online commerce.

[0045] For example, as shown in Figure 2, using the actual sales volume 20 of product A on benchmark date 10, we can determine the quantity 21 of product A entering the logistics center one day after benchmark date 11 and the quantity 22 of product A entering the logistics center two days after benchmark date 12. The higher the actual sales volume 20 of product A on benchmark date 10, the higher the consumer demand for product A can be predicted after benchmark date. Therefore, it can be predicted that the quantity of product A that needs to be shipped out of the logistics center after benchmark date will increase, and thus the quantity of product A entering the logistics center after benchmark date 10 may also increase.

[0046] Specifically, when product A is sold on benchmark date 10, product A can be delivered to the buyer after benchmark date 10. To deliver product A to the buyer, the logistics center can receive product A into its inventory. However, to ensure fast and efficient delivery, the logistics center can receive product A in advance. That is, the logistics center can determine the receipt of product A before its sale, rather than after its sale.

[0047] More specifically, the quantity of Product A received by the logistics center one day after the benchmark date (11) is not necessarily equal to the actual sales volume (20) of Product A on benchmark date (11). When determining the final quantity (21) of Product A received one day after the benchmark date (11), the quantity of Product A sold before the benchmark date and needing to be shipped out, the quantity of Product A sold on the benchmark date and needing to be shipped out, the quantity of Product A expected to be sold after the benchmark date and needing to be shipped out, and the quantity of Product A sales expected to be cancelled, etc., will be considered. Similarly, the quantity (22) of Product A received two days after the benchmark date (12) can be determined in the same way. That is, based on the actual sales volume (20) of Product A on benchmark date (10), the quantity of Product A received by the logistics center in the future can be determined.

[0048] On the other hand, as shown in Figure 3, due to a shortage of product A, the actual sales volume 20 of product A on the benchmark date 10 may be 0. In this case, if product A becomes available for sale starting from day 11 after the benchmark date, determining the quantity of product A to be received into the logistics center based on the actual sales volume 20 may lead to problems in the distribution process.

[0049] Specifically, if calculations are based on actual sales volume, since the actual sales volume of Product A on benchmark date 10 was 0, the quantity of Product A received by the logistics center after benchmark date 10 might be the same as if there was no consumer demand. However, the fact that the actual sales volume of Product A was 0 on benchmark date 10 is due to a shortage of Product A, and is unrelated to consumer demand. Therefore, underestimating the quantity of Product A received by the logistics center after benchmark date 10 could lead to an increase in sales of Product A after benchmark date 10, while Product A fails to be delivered to buyers in a timely manner.

[0050] Therefore, as shown in Figure 3, in the event of special conditions (e.g., stockout), the quantity of goods entering the logistics center can be determined based on the potential sales volume 23, which is determined assuming no special conditions exist, rather than based on the actual sales volume 20, thereby solving the problem.

[0051] As another example, if the actual sales volume of product A increases sharply compared to previous periods due to discount promotions, determining the future quantity of product A to be received by the logistics center based on the actual sales volume during the discount promotion period might lead to a surplus of product A at the logistics center. Therefore, in this case, the quantity to be received by the logistics center could be determined based on the potential sales volume of product A assuming there were no discount promotions, rather than based on the actual sales volume of product A during the discount promotion period.

[0052] It should be noted here that special conditions are not limited to stockouts or discounted promotions. Special conditions can include any situation where the actual sales volume of a particular product differs drastically from its historical sales trend. However, for ease of explanation, the following explanation will assume a stockout situation.

[0053] Up to this point, some terms and premises of this disclosure have been explained with reference to Figures 2 and 3. Hereinafter, the structure and operation of an online commerce goods potential sales volume determination device according to an embodiment of this disclosure will be described with reference to Figure 1. Hereinafter, the online commerce goods potential sales volume determination device will be simply referred to as the potential sales volume determination device.

[0054] The potential sales volume determination device 1 disclosed herein can receive at least one query related to a specific product from the online commerce server 2. A query can refer to text searched by a user in order to purchase a product through online commerce. A query related to a specific product is one that exposes the specific product when entered through online commerce. For example, a query related to a specific product may include the name of the specific product itself, the names of similar products, the category of the specific product, and the purpose of the specific product.

[0055] Furthermore, the potential sales volume determination device 1 can remove queries from the received queries that have values ​​such as search count, specific product exposure probability, and correlation ratio between actual sales volume of a specific product that are below a threshold. Each threshold for these parameters will be different.

[0056] Furthermore, the potential sales volume determination device 1 can also receive search data (hereinafter referred to as search data) corresponding to queries related to specific products from the online business server 2. Search data can refer to data related to searches for specific queries. For example, search data may include the number of searches for a specific query, the number of times a product is exposed due to searches for a specific query, the search time for the specific query, and user information for searching the specific query (e.g., the user's age, gender, region, and whether the user is a member of the online business server).

[0057] Furthermore, the potential sales volume determination device 1 can determine the exposure probability of a specific product resulting from a search for a specific query. For example, the exposure probability of "AAA carbonated beverage" resulting from a search for the query "beverage" can be determined using the number of searches for the query "beverage" and the number of exposures for the product "AAA carbonated beverage". Specific methods for determining the exposure probability of a specific product resulting from a search for a specific query will be explained later.

[0058] Furthermore, the potential sales volume determination device 1 can determine the potential sales volume of a specific product. Next, the potential sales volume determination device 1 can use the potential sales volume to determine the quantity of a specific product entering the logistics center 3.

[0059] In some embodiments, the potential sales volume determination device 1 may be composed of multiple computing devices. For example, the potential sales volume determination device 1 may be composed of a first computing device and a second computing device. Furthermore, the first operation of the potential sales volume determination device 1 may be executed on the first computing device, and the second operation of the potential sales volume determination device 1 may be executed on the second computing device.

[0060] It should be noted that the operation of the potential sales volume determination device 1 disclosed herein is not limited to the examples described, but may include related operations of the online business goods potential sales volume determination method (hereinafter referred to as the potential sales volume determination method) of some embodiments of this disclosure that will be described below. Furthermore, the technical ideas embodied in some embodiments of this disclosure that will be described below can be incorporated into the potential sales volume determination device 1, even without further description.

[0061] So far, a potential sales volume determination apparatus according to an embodiment of this disclosure has been described with reference to FIG1. ​​Hereinafter, methods for determining the potential sales volume of online business goods according to some embodiments of this disclosure will be described with reference to FIGS. 2 to 13. In the following description, the subject of a particular step / work may be omitted; in this case, it can be understood that the step / work is performed by the computing device constituting the potential sales volume determination apparatus.

[0062] First, referring to Figure 4, an embodiment of the method for determining the potential sales volume of online business goods disclosed herein will be described. Hereinafter, the method for determining the potential sales volume of online business goods will be simply referred to as the potential sales volume determination method.

[0063] As shown in Figure 4, a set of queries related to the first product can be determined (S10). The first product can refer to a product in online commerce whose potential sales volume needs to be determined. Furthermore, the query set for the first product can refer to at least one query related to the first product. Queries related to the first product can refer to queries in online commerce or search services related to online commerce that expose the first product during a search.

[0064] Returning to Figure 4, we can obtain the search data for the first query within the training period (S20).

[0065] Here, the first query refers to one of at least one queries included in the query set. While the following describes a method for determining the potential sales volume of a first product based on search data corresponding to the first query, in some embodiments disclosed herein, the potential sales volume of the first product can be determined based on multiple search data corresponding to multiple queries included in the query set.

[0066] For example, as shown in Figure 9, the query set may include query #1, query #2, and query #3. Furthermore, the obtained search data 100 may include search data 101 corresponding to query #1, search data 102 corresponding to query #2, and search data 103 corresponding to query #3. The potential sales volume 200 of a product can be determined based on each search data. Specifically, the potential sales volume 201 of a product based on query #1 can be determined based on the search data 101 of query #1. The potential sales volume 202 of a product based on query #2 can be determined based on the search data 102 of query #2. The potential sales volume 203 of a product based on query #3 can be determined based on the search data 103 of query #3. The potential sales volume of a product can be the sum of the potential sales volumes determined based on the search data of each query included in the query set.

[0067] For ease of explanation, the following description focuses on the method for determining the potential sales volume for a single query. However, the method described below can also be applied to multiple queries.

[0068] On the other hand, the training period refers to the period prior to the period for determining the potential sales volume of the first product. Furthermore, the training period can be a pre-defined interval within the period prior to the period for determining the potential sales volume of the first product.

[0069] For example, as shown in Figure 10, the timeframe for determining the potential sales volume of the first product can be set to the date D when the first product becomes out of stock. In this case, the training period can refer to a period of 30 days prior to the date the potential sales volume is determined. Specifically, the training period can refer to the period from D-30 to D-1.

[0070] On the other hand, search data can include the number of searches for the corresponding query, product information exposed due to the search, user information for the search query, and information about the time of the search query.

[0071] Returning to Figure 4, the probability of the first product exposure generated by the search through the first query can be determined (S30).

[0072] The probability of first product exposure resulting from a search query can be interpreted as the probability that a searcher will see the first product when the first query is searched in online commerce or related search services. Specifically, if a searcher learns about the first product in an online business through a search, the first product can be considered to have been exposed. For example, if a link to purchase the first product and an image of the first product are displayed alongside the first query search in an online business, then the first product can be considered to have been exposed due to the search query.

[0073] Returning to Figure 4, we can obtain the search data for the first query within the out-of-stock period (S40).

[0074] The out-of-stock period can refer to the period during which users of the online business are unable to purchase the first item because it is out of stock.

[0075] Returning to Figure 4, the potential sales volume of the first product during the stockout period can be determined (S50). Specifically, the potential sales volume of the first product under the assumption of no stockout can be determined using the exposure probability of the first product generated by the search through the first query and the search data of the first query during the stockout period. The higher the exposure probability of the first product generated by the search through the first query, and the more searches performed in the first query, the higher the potential sales volume of the first product is likely to be. The specific method for determining the potential sales volume will be explained later with reference to Figure 7.

[0076] Returning to Figure 4, the potential sales volume of the first product can be used to determine the quantity of goods to be put into storage for the saleable period (S60).

[0077] The salable period refers to the period after the stockout period when the first product can be sold online. For example, as shown in Figure 10, suppose the first product is available for sale again on the first day after the stockout date D. The quantity of the first product that arrives at the logistics center on day D+1 (hereinafter referred to as the first product arrival quantity) can be determined based on past sales volume. In this case, during the period before day D+1, the actual sales volume of the first product during that period can provide a basis for determining the first product arrival quantity. On the other hand, during the period before day D+1, such as the stockout period of the first product on day D, the potential sales volume of the first product during that period can provide a basis for determining the first product arrival quantity.

[0078] So far, referring to Figure 4, a method for determining the potential sales volume of online commercial goods according to an embodiment of this disclosure has been illustrated. According to this embodiment, past search data can be used to determine the current potential sales volume of a product. Therefore, in cases where special conditions (e.g., stockouts, discount promotions) cause a difference between the actual sales volume and the product's sales trend, the potential sales volume of the product under the assumption of no special conditions can be determined. Using this potential sales volume, it is possible to predict how the sales volume of the product will change after the special conditions disappear or change. Furthermore, using the predicted sales volume, product-related data can be generated more accurately, such as the quantity of goods subsequently to be stored in the logistics center or the additional discount policies required for the product (e.g., discount quantity, discount period, discount price), etc.

[0079] Figure 5 is a detailed flowchart illustrating the process of determining the query set using the method described with reference to Figure 4. Below, some embodiments of determining the query set will be described with reference to Figure 5.

[0080] As shown in Figure 5, multiple queries that expose the first product through search can be obtained (S11). For example, if the first product is "AAA carbonated beverage", multiple queries that expose "AAA carbonated beverage" images or links that allow the purchase of "AAA carbonated beverage" in online commerce can be obtained (e.g., beverage, carbonated beverage, AAA, BBB carbonated beverage, CCC carbonated beverage, AAA carbonated, canned beverage, etc.).

[0081] Returning to Figure 5, among the multiple queries obtained, queries with a search count less than a first threshold can be removed (S12). The search count can be the sum of search counts within a preset period or the average of the search counts.

[0082] On the other hand, the first threshold can refer to the minimum number of searches required for a specific search term to be statistically significant and to ensure representativeness of the analysis. In other words, only queries with a search volume exceeding the first threshold are considered sufficient samples and reflected in the determination of potential sales volume, thereby enabling a more accurate determination of potential sales volume.

[0083] For example, we can assume that the daily average value of the first threshold is 50 times. In this case, if the daily average number of searches for "AAA carbonated beverage" among the online business searches that can expose "AAA carbonated beverage" is 40, then the searches for "AAA carbonated beverage" will be removed and not included in the query set.

[0084] Returning to Figure 5, among the multiple queries obtained, queries with an exposure probability of the first product less than the second threshold can be removed (S13). The exposure probability of the first product can be the average of the exposure probabilities over a preset period.

[0085] On the other hand, the second threshold can refer to the minimum exposure probability of the first product that is statistically significant and ensures actual relevance. In other words, only queries with an exposure probability of the first product that reaches or exceeds the second threshold will be considered as a sufficient sample and reflected in the determination of potential sales volume, thereby enabling a more accurate determination of potential sales volume.

[0086] For example, we can assume that the average value of the second threshold is 50%. In this case, if the exposure probability of "AAA carbonated beverage" is 40% among the queries for "AAA carbonated beverage" and "BBB carbonated beverage" in online commerce, then the query for "BBB carbonated beverage" will be removed and not included in the query set.

[0087] Returning to Figure 5, among the multiple queries obtained, queries with a correlation ratio to the sales volume of the first product that is less than the third threshold can be removed (S14). The correlation ratio can refer to an indicator of the degree of increase in the sales volume of the first product when the number of exposures of the first product increases due to query searches. The correlation ratio can also be the average correlation ratio over a preset period.

[0088] On the other hand, the third threshold can refer to the minimum correlation ratio between the first search term and the first product that is statistically significant and ensures the actual relevance to the sales of the first product. In other words, only queries with a correlation ratio of more than the third threshold will be considered as sufficient samples and reflected in the determination of potential sales volume, thereby enabling a more accurate determination of potential sales volume.

[0089] For example, we can assume that the average value of the third threshold is 50%. In this case, if the correlation ratio of the sales volume of "AAA carbonated beverage" to the query of "CCC carbonated beverage" in the online business query of "AAA carbonated beverage" is 10%, then the query of "CCC carbonated beverage" will be removed and not included in the query set.

[0090] On the other hand, in one embodiment, queries with a search period shorter than a preset period can be deleted from the acquired queries. For example, the preset period could be 50% of the training period. In this case, if the training period is 4 days, queries with a search period shorter than 2 days will be removed. Specifically, assuming the training period is from D-4 to D-1, queries searched only on D-4 will be removed. Furthermore, queries searched from D-4 to D-3 will not be removed. Also, queries searched on D-4 and D-1 will not be removed. That is, whether the search period within the training period is shorter than the preset period is irrelevant to whether the search periods are consecutive.

[0091] The first to third thresholds described above (hereinafter referred to as thresholds) can be preset based on the statistical analysis results of data related to online commerce. Specifically, the number of searches, the number of exposures due to searches, and the exposure correlation ratio that have a statistically determined correlation with any sales volume of a product can be determined as the thresholds.

[0092] Subsequently, the query set can be determined by excluding the queries that have been removed. Based on the determined query set, the potential sales volume of the first product can be determined by referring to the work described in Figure 4.

[0093] For example, in online commerce, if queries for "AAA carbonated beverage" can be exposed (e.g., beverage, carbonated beverage, AAA, BBB carbonated beverage, CCC carbonated beverage, AAA carbonated, canned beverage), by removing queries for "AAA carbonated," "BBB carbonated beverage," and "CCC carbonated beverage," the potential sales volume of "AAA carbonated beverage" can be determined based on the search data corresponding to the remaining queries (e.g., beverage, carbonated beverage, AAA, canned beverage).

[0094] So far, some embodiments for determining the query set have been illustrated with reference to Figure 5. Through this embodiment, the query set can be constructed using only queries highly relevant to the product. Furthermore, by utilizing only search data from queries highly relevant to the product, rather than search data from all queries that can expose the product, the potential sales volume of the product can be determined more accurately. Moreover, using less data can further improve the speed of determining the potential sales volume of the product. Furthermore, using less data can reduce the load on the computing device used to determine the potential sales volume of the product.

[0095] Figure 6 is a detailed flowchart illustrating the process of determining the probability of a product exposure resulting from a query search, as described with reference to the method in Figure 4. Below, some embodiments for determining the probability of a product exposure resulting from a query search will be described with reference to Figure 6.

[0096] As shown in Figure 6, the search count of the first query within the training period can be obtained (S31). Specifically, the search count of the first query corresponding to each unit period constituting the training period can be obtained. Here, the unit period can refer to the smallest period unit constituting the training interval. For example, the unit period can be 2 days, 1 day, 12 hours, 6 hours, 3 hours, etc.

[0097] For a specific example, as shown in Figure 11, the training period can be set from day D-30 to day D-1. In this case, the unit period of the training period can be 1 day. Therefore, the search count of the first query for each day constituting the training period can be obtained, including the first query search count of 30 on day D-1, the first query search count of 31 on day D-2, and the first query search count of 32 on day D-30.

[0098] In one embodiment, the number of first query searches within a unit period when the first product is out of stock during the training period may not be obtained. For example, as shown in Figure 11, the number of first query searches on day D-3 may not be obtained. That is, search data for the period when the first product is out of stock during the training period may be excluded from the exposure probability calculation. Since online businesses may not manually expose products when they are out of stock, excluding search data for the period when the first product is out of stock during the training period allows for a more accurate calculation of the exposure probability of the first product.

[0099] Returning to Figure 6, the number of times the first product was exposed by the first query search within the training period can be determined (S32). Specifically, the number of times the first product was exposed by the first query search corresponding to each unit period constituting the training period can be obtained.

[0100] For example, as shown in Figure 11, the number of first product exposures generated by the first query search for each day constituting the training period can be obtained, including the number of first product exposures generated by the first query search on day D-1 (40), the number of first product exposures generated by the first query search on day D-2 (41), and the number of first product exposures generated by the first query search on day D-30 (42).

[0101] In one embodiment, the number of first product exposures generated by the first query search within a unit period when the first product is out of stock during the training period may not be obtained. For example, as shown in Figure 11, the number of first product exposures generated by the first query search in D-3 when the first product is out of stock during the training period may not be obtained. The reason why the number of first product exposures generated by the first query search within a unit period when the first product is out of stock may not be obtained may be the same as the reason why the number of searches for the first query within a unit period when the first product is out of stock is not obtained.

[0102] Returning to Figure 6, the probability of the first product exposure generated by the first query search can be determined (S33). Specifically, the probability of the first product exposure generated by the first query search can be determined by using the number of first product exposures generated by the first query search and the number of searches for the first query.

[0103] In one embodiment, the probability of exposure of the first product generated by the search of the first query can be the value obtained by summing the number of exposures of the first product generated by each first query search and dividing it by the sum of the number of searches of each first query.

[0104] For example, as shown in Figure 11, the sum of the number of first product exposures generated by each first query search from D-30 to D-1, divided by the sum of the number of searches for each first query from D-30 to D-1, can be the first product exposure probability 50 generated by the first query search.

[0105] So far, some embodiments for determining the product exposure probability generated by a query search have been illustrated with reference to Figure 6. This embodiment can easily determine the product exposure probability generated by a query search by the number of searches performed and the number of times the product is exposed as a result of that query.

[0106] Figure 7 is a detailed flowchart illustrating the process of determining the potential sales volume of goods during the stockout period, using the method described with reference to Figure 4. Below, some embodiments for determining the potential sales volume of goods during the stockout period will be described with reference to Figure 7.

[0107] As shown in Figure 7, the sales volume of the first product within the training period can be obtained (S51). Specifically, the actual sales volume of the first product within each unit period constituting the training period can be obtained. For example, as shown in Figure 12, the sales volume of the first product for each day constituting the training period can be obtained, including the sales volume of the first product on day D-1 (60), the sales volume of the first product on day D-2 (61), and the sales volume of the first product on day D-30 (62).

[0108] In one embodiment, the actual sales volume of the first product during the unit period when the first product is out of stock may not be obtained. For example, as shown in Figure 12, the actual sales volume of the first product on day D-3 when the first product is out of stock may not be obtained. The actual sales volume of the first product during the out-of-stock period is due to the out-of-stock situation rather than demand for the first product. Therefore, using the actual sales volume of the first product during the out-of-stock period for the correlation ratio calculation may interfere with the accurate calculation of the correlation ratio between sales volume and exposure.

[0109] Returning to Figure 7, we can obtain the number of first product exposures generated by the first query search within the training period (S52). Specifically, we can obtain the number of first product exposures generated by the first query search corresponding to each unit period constituting the training period.

[0110] For example, as shown in Figure 12, the number of first product exposures generated by the first query search for each day constituting the training period can be obtained, including the number of first product exposures generated by the first query search on day D-1 (40), the number of first product exposures generated by the first query search on day D-2 (41), and the number of first product exposures generated by the first query search on day D-30 (42).

[0111] In one embodiment, the number of times the first product is exposed by the first query within a unit period when the first product is out of stock during the training period may not be obtained. A further description of this embodiment will be omitted.

[0112] Returning to Figure 7, the correlation ratio between the number of times the first product is exposed and the sales volume of the first product can be determined (S53). This correlation ratio can be an indicator of the degree of increase in the sales volume of the first product when the number of exposures generated by the query search increases.

[0113] In one embodiment, the correlation ratio between the first query and the sales volume of the first product can be the sum of the sales volume of the first product in each unit period of the training period, divided by the sum of the number of times the first product was exposed by the first query search in each unit period of the training period.

[0114] For example, as shown in Figure 12, the total sales volume of the first product for each period from D-30 to D-1 can be determined. Furthermore, the total number of first product exposures generated by the first query search within each period from D-30 to D-1 can be determined. Finally, by dividing the total sales volume of the first product for each period from D-30 to D-1 by the total number of first product exposures generated by the first query search within each period from D-30 to D-1, the relevance ratio 70 can be determined.

[0115] Returning to Figure 7, we can obtain the search count for the first query within the out-of-stock period (S54). It should be noted that the search count for the first query refers to the search count during the out-of-stock period. For example, as shown in Figure 12, we can obtain the search count of 33 for the first query on the out-of-stock day D.

[0116] Next, the expected number of exposures of the first product generated by the first query search during the stockout period can be determined (S55). Specifically, the expected number of exposures of the first product generated by the first query search during the stockout period can be determined using the exposure probability of the first product generated by the first query search and the number of searches of the first query during the stockout period.

[0117] In one embodiment, the estimated number of times the first product is exposed during the out-of-stock period can be the product of the number of searches for the first product during the out-of-stock period and the exposure probability of the first product generated by the first product. For example, as shown in Figure 11, the estimated number of times the first product is exposed 43 during day D can be the product of the number of searches for the first product during day D and the exposure probability 50 of the first product generated by the first product calculated from the search data from day D-1 to day D-30.

[0118] In online commerce, since the first product that is out of stock is not exposed or has less exposure compared to the product that is available for sale, when determining the potential sales volume of the first product, the estimated number of exposures of the first product generated by the first query search can be used instead of the actual number of exposures of the first product generated by the first query search.

[0119] Returning to Figure 7, the potential sales volume of the first product can be determined by using the expected number of exposures of the first product generated by the first query search during the out-of-stock period and the correlation ratio between the number of exposures of the first product and the sales volume of the first product (S56).

[0120] In one embodiment, the potential sales volume of the first product during the out-of-stock period can be the product of the expected number of times the first product is exposed by the first query search during the out-of-stock period and the correlation ratio between the first query and the sales volume of the first product.

[0121] For example, as shown in Figure 12, the potential sales volume 63 of the first product within D days can be the product of the expected number of exposures of the first product generated by the first query search within D days 43 and the correlation ratio 70 determined using search data from D-1 to D-30 days and the sales volume of the first product.

[0122] So far, referring to Figure 7, some embodiments for determining the potential sales volume of a product during the stockout period have been illustrated. This embodiment utilizes past search data to determine the correlation between search data and sales volume. Furthermore, this embodiment can utilize the correlation between search data and sales volume to determine the current potential sales volume based on current search data.

[0123] Figure 8 is a flowchart illustrating the sequence of a method for determining potential sales volume according to another embodiment of this disclosure. Hereinafter, an embodiment for determining potential sales volume using an artificial intelligence model will be described with reference to Figure 8. However, for ease of explanation, only the parts that differ from the embodiment described with reference to Figure 4 will be described.

[0124] First, as shown in Figure 8, the query set related to the first product can be determined (S10). Next, the search data for the training period of the first query can be obtained (S20). Then, the search data for the training period is input into the artificial intelligence model, and the artificial intelligence model can be trained (S34). Here, the artificial intelligence model can refer to an artificial intelligence model that predicts the potential sales volume of the product based on the search data.

[0125] For example, an artificial intelligence model can be based on a linear regression model. In this model, the potential sales volume of the first product during the out-of-stock period can be determined by analyzing the linear relationship between the sales volume of the first product in each unit of the training period and the search data of queries related to the first product.

[0126] As another example, an AI model can be an AI model based on the ARIMAX model. In this model, by considering both time-series factors and search data as exogenous variables, the AI ​​model can more accurately determine the potential sales volume of the first item during the stockout period.

[0127] Returning to Figure 8, after the AI ​​model has completed its learning process, search data for the first query during the stockout period can be obtained (S40). Next, the search data for the stockout period is input into the AI ​​model, which outputs the potential sales volume of the first product (S57). Then, the potential sales volume of the first product can be used to determine the quantity of the first product to be put into inventory within the salable period (S60).

[0128] Specifically, search data regarding the stockout period can be input into a pre-learned artificial intelligence model. This search data may include information on the number of searches for the first query related to the stockout period. Next, the artificial intelligence model can predict and output the potential sales volume of the first product within the stockout period based on the input search data. Then, based on the output potential sales volume of the first product, the quantity of the first product to be received into the logistics center within the product's remaining salable period after the stockout period can be determined.

[0129] So far, referring to Figure 8, an embodiment of using an artificial intelligence model to determine the potential sales volume of goods has been illustrated. This embodiment, by using an artificial intelligence model, can more accurately determine the potential sales volume of goods during their out-of-stock period. Furthermore, by more accurately determining the potential sales volume of goods during their out-of-stock period, it is possible to understand the actual consumer demand for the goods. And, based on the actual consumer demand, inventory management and discount policies can be formulated more effectively.

[0130] Figure 13 is a table illustrating an example of using multiple queries to determine the potential sales volume of a product. The following will describe an embodiment of using multiple queries to determine the potential sales volume of a product, with reference to Figure 13.

[0131] First, the out-of-stock period can be determined. For example, as shown in Figure 13, the first product was out of stock on January 4th, so January 4th was determined as the out-of-stock period.

[0132] Next, the training period can be determined. For example, as shown in Figure 13, the period from January 1st to January 3rd, i.e., the period before January 4th, is determined as the training period. In this case, the unit period is 1 day.

[0133] Next, the query set can be determined. For example, as shown in Figure 13, queries A, B, C, and D that resulted in the first product being exposed due to a search can be obtained. Next, queries that searched for more than 50% of the period (e.g., 2 days, 3 days) within the training period will be determined as the query set. Query A was searched from January 1st to January 3rd, so it can be determined as a query set. Query B was searched on both January 1st and January 3rd, so it can also be determined as a query set. Query C was searched on both January 2nd and January 3rd, so it can also be determined as a query set. Query D was searched only on January 2nd, so it will not be determined as a query set and will be removed.

[0134] Next, the search count for each query included in the query set and the number of first product impressions generated by those queries can be obtained. The search count for a query can include the search count during the training period and the search count during the out-of-stock period. Conversely, the number of first product impressions generated by the search for a query can only include the search count during the training period. For example, the search counts for queries A, B, and C from January 1st to January 4th can be obtained. Furthermore, the number of first product impressions generated by the searches for queries A, B, and C from January 1st to January 3rd can be determined.

[0135] Next, we can determine the probability of first product exposure generated by each query in the query set. For example, as shown in Figure 13, the total number of searches for query A during the training period is 900 (400+300+200=900). Furthermore, the total number of first product exposures generated by searches for query A during the training period is 710 (320+240+150=710). Therefore, the probability of first product exposure generated by searches for query A is 78.89% (710÷900 ≈ 0.7889). The probability of first product exposure generated by searches for query B, calculated in the same way, is 65.71% (460÷700 ≈ 0.6571). And the probability of first product exposure generated by searches for query C, calculated in the same way, is 73.33% (220÷300 ≈ 0.7333).

[0136] Next, the expected number of exposures of the first product generated by each query search and the total number of exposures of the first product generated by the search can be determined during the out-of-stock period. Specifically, the expected number of exposures can be determined by multiplying the exposure probability of the first product generated by each query search by the number of searches for each query during the out-of-stock period. For example, as shown in Figure 13, the expected number of exposures of the first product generated by the search for query A during the out-of-stock period can be 237 times (300 × 0.7889 ≈ 237). Furthermore, the expected number of exposures of the first product generated by the search for query B during the out-of-stock period can be 131 times (200 × 0.6571 ≈ 131). And, the expected number of exposures of the first product generated by the search for query A during the out-of-stock period is 237 times (140 × 0.7333 ≈ 110). Therefore, the total expected number of exposures of the first product generated by the search during the out-of-stock period can be 478 times (237 + 131 + 110 = 478).

[0137] Next, the correlation ratio between the total number of impressions of the first product generated by each query search and the sales volume of the first product can be determined. Specifically, the correlation ratio can be the value of dividing the sum of the product sales volume by the total number of product impressions during the training period. Here, the sum of product sales volume can refer to the sum of product sales volume within each unit period of the training period. And the total number of product impressions can refer to the sum of the total number of product impressions within each unit period of the training period.

[0138] For example, as shown in Figure 13, the total exposure count of the first product on January 1st within the training period is the sum of 320 exposures generated by search query A and 210 exposures generated by search query B, totaling 530 exposures. Furthermore, the total exposure count of the first product on January 2nd within the training period is the sum of 240 exposures generated by search query A and 80 exposures generated by search query C, totaling 320 exposures. And, the total exposure count of the first product on January 3rd within the training period could be the sum of 150 exposures generated by search query A, 250 exposures generated by search query B, and 140 exposures generated by search query C, totaling 540 exposures. Therefore, the total exposure count of the first product could be 1390 exposures.

[0139] Furthermore, the sales volume of the first product during the training period can be a total of 270 units, consisting of 100 units sold on January 1st, 90 units sold on January 2nd, and 80 units sold on January 3rd. Therefore, the correlation ratio between the number of exposures of the first product generated through search and the sales volume of the first product can be 0.1942 (270 ÷ 1390 ≈ 0.1942).

[0140] Next, by using the expected number of exposures of the first product during the stockout period and the correlation ratio between the number of exposures of the first product and its sales volume, the potential sales volume of the first product can be determined. Specifically, the potential sales volume of the first product can be the product of the expected number of exposures of the first product during the stockout period and the aforementioned correlation ratio.

[0141] For example, as shown in Figure 13, the first product is expected to be exposed 478 times during the stockout period of January 4th. Furthermore, the correlation ratio between the exposure of the first product and its sales volume can be 0.1942 as previously stated. Therefore, the potential sales volume of the first product during the stockout period of January 4th can be 92 units (478 × 0.1942 ≈ 92).

[0142] So far, Figure 13 has illustrated an example of using multiple queries to determine the potential sales volume of a product. By using multiple queries to determine the potential sales volume of a product, the potential sales volume can be determined more accurately than by using only one query.

[0143] In some embodiments of this disclosure, a stockout period for the product is assumed, and the timeframe for determining the potential sales volume of the product is described. However, it should be noted that the timeframe for determining or calculating the potential sales volume of a product is not limited to the stockout period. In other words, the timeframe for determining the potential sales volume of any product can include both the stockout period and the period when the product is not in stock.

[0144] For example, after a product has undergone a discount promotion, its potential sales volume can be determined. After the discount promotion, by comparing the potential sales volume with the actual sales volume, it can be determined whether a further discount policy should be applied. Specifically, if the actual sales volume of a product that has not undergone a discount promotion is lower than its potential sales volume, then a further discount promotion may be applied to that product. Conversely, if the actual sales volume of a product that has undergone a discount promotion is lower than its potential sales volume, then a discount promotion may not be applied to that product.

[0145] The online commerce product potential sales volume determination method and apparatus described so far in some embodiments of this disclosure can determine the sales volume of goods under the assumption of no special conditions (e.g., product shortages, product discounts, or promotions). Therefore, the potential sales volume determination method and apparatus can minimize the distortion of product sales volume-related statistics caused by special conditions. Furthermore, the potential sales volume determination method and apparatus can more accurately generate forecast data related to product sales volume (e.g., the quantity received at logistics centers).

[0146] The method and apparatus for determining potential sales volume facilitate the development of strategies for situations without such special conditions, especially when these conditions persist for an extended period. For example, when a long-out-of-stock item is restocked, the quantity to be received at the logistics center can be accurately determined based on the potential sales volume during the out-of-stock period. Furthermore, some embodiments disclosed herein avoid the problem of insufficient demand forecasts leading to inadequate logistics center receipts when determining the quantity of goods received based on the actual sales volume of long-out-of-stock items.

[0147] As another example, when a long-term discount promotion ends, the quantity of goods entering the logistics center can be accurately determined based on the potential sales volume of the goods during the promotion period. Furthermore, some embodiments of the present invention can avoid the problem of over-inbound quantity due to predicted sufficient demand when determining the quantity of goods entering the logistics center based on the actual sales volume of long-term discount promotions. The actual sales volume may increase due to the long-term discount promotion. That is, the actual sales volume of the goods may decrease when the long-term discount promotion ends.

[0148] The effects of the technical ideas disclosed herein are not limited to those mentioned above. Those skilled in the art can clearly understand other effects not mentioned from the content of this disclosure.

[0149] So far, thumbnail generation methods for some embodiments of this disclosure have been illustrated with reference to Figures 2 to 13. Hereinafter, the hardware structure of an exemplary computing device according to some embodiments of this disclosure will be described with reference to Figure 14. The computing device may be a computing device operating the potential sales volume determination device of this disclosure.

[0150] Figure 14 is an exemplary hardware structure diagram of a computing device that can be implemented in various embodiments of this disclosure. The computing device 1000 according to this embodiment may include one or more processors 1100, a system bus 1600, a communication interface 1200, memory 1400 loading a computer program 1500 executed by the processor 1100, and storage device 1300 storing the computer program 1500. Figure 14 only shows structural elements relevant to the embodiments of this disclosure. Therefore, those skilled in the art to which this disclosure pertains will recognize that other general structural elements may be included in addition to those shown in Figure 14.

[0151] Processor 1100 can control the overall operation of various structures of computing device 1000. Processor 1100 may be configured as at least one of a central processing unit (CPU), a microprocessor unit (MPU), a microcontroller unit (MCU), a graphics processing unit (GPU), or any type of processor known in the art disclosed herein. Furthermore, processor 1100 can perform operations on at least one application or program for executing methods / operations according to embodiments of the present disclosure. Computing device 1000 may have two or more processors.

[0152] Memory 1400 can store various data, commands, and / or information. Memory 1400 can load one or more programs 1500 from storage device 1300 to perform methods / operations according to various embodiments of this disclosure. Memory 1400 can be implemented as volatile memory such as RAM, but this disclosure is not limited thereto. System bus 1600 provides communication functionality between the structural elements of computing device 1000.

[0153] Bus 1600 can be implemented as various types of buses, such as address bus, data bus, and control bus. Communication interface 1200 supports wired and wireless internet communication of computing device 1000. Communication interface 1200 can also support various communication methods other than internet communication. Therefore, communication interface 1200 may include communication modules known in the art disclosed herein. Storage device 1300 temporarily stores more than one computer program 1500. Storage device 1300 may include non-volatile memory such as flash memory, hard disk, removable magnetic disk, or any form of computer-readable recording medium known in the art to which this disclosure pertains.

[0154] Computer program 1500 may include one or more instructions for performing methods / operations according to various embodiments of the present disclosure. If computer program 1500 is loaded into memory 1400, processor 1100 executes the one or more instructions, thereby enabling the execution of methods / operations according to various embodiments of the present disclosure.

[0155] For example, computer program 1500 may include instructions to perform the following tasks: determining a query set related to a first product; obtaining first search data of the first query included in the query set within a first period; using the first search data, determining the probability that the first product will be exposed due to the first query search, i.e., a first exposure probability; obtaining second search data of the first query within a second period; and using the first exposure probability and the second search data, determining the potential sales volume of the first product within the second period.

[0156] As another example, computer program 1500 may include instructions to perform the following tasks: determining a query set related to a first product; obtaining first search data for a first query included in the query set within a first period; inputting the first search data and the sales volume of the first product within the first period into an artificial intelligence model for predicting the potential sales volume of the first product, so that the artificial intelligence model can learn; obtaining second search data for the first query within a second period; and inputting the second search data into the learned artificial intelligence model to output the potential sales volume of the first product within the second period.

[0157] Although the accompanying drawings illustrate the tasks in a specific order, this should not be construed as meaning that the tasks need to be performed in the specific order shown or sequentially, or that all the illustrated tasks need to be performed to obtain the desired result. In certain situations, multitasking and parallel processing are more advantageous. Moreover, the separation of various structures in the embodiments described above should not be construed as a necessity, but rather as meaning that the illustrated program structure elements and systems are typically integrated together as a single software product or can be packaged into multiple software products.

[0158] While embodiments of this specification have been described above with reference to the accompanying drawings, those skilled in the art will understand that these embodiments can be implemented in other specific ways without altering their technical concept or essential features. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not limiting. The scope of protection of this disclosure should be interpreted in accordance with the appended claims, and all technical ideas within the equivalent scope should be interpreted as included within the scope of the technical ideas defined by this disclosure.

[0159] 1: Potential Sales Volume Determination Device 2: Online Business Server 3: Logistics Center 10, 11, 12: Unit Duration 20, 60, 61, 62: Sales volume of the product 21, 22: Quantity of goods received into the logistics center 23, 63, 200, 201, 202, 203: Potential sales volume of the product 30, 31, 32, 33: Number of searches for the query 40, 41, 42: Number of product impressions generated by searches 43: Expected number of times products generated from search results will be displayed. 50: Probability of product exposure generated by search 70: The correlation ratio between product exposure frequency and product sales volume. 100, 101, 102, 103: Search data S10, S11, S12, S13, S14, S20, S30, S31, S32, S33, S34, S40, S50, S51, S52, S53, S54, S55, S56, S57, S60: Steps

Claims

1. A method for determining the potential sales volume of online commercial goods, executed by a computing device, characterized in that it includes: The steps to determine the query set associated with the first product; The steps to obtain the first search data of the first query included in the query set within the first time period; The steps include: determining a first exposure probability using the first search data, the first exposure probability being the probability that the first product is exposed due to the first query search; obtaining second search data for the first query within a second period; and determining the potential sales volume of the first product within the second period using the first exposure probability and the second search data, the first period being a period prior to the second period, wherein the potential sales volume is the predicted sales volume of the first product within the second period. The step of determining the potential sales volume of the first product includes: obtaining the number of searches for the first query within the second period from the second search data; and determining the potential sales volume of the first product within the second period using the number of searches for the first query within the second period and the first exposure probability. The step of determining the query set related to the first product includes: obtaining a first query and a third query that expose the first product due to a search; and including the first query in the query set while excluding the third query from the query set, wherein the correlation ratio between the first query and the sales volume of the first product is greater than a third threshold, and the correlation ratio between the third query and the sales volume of the first product is less than the third threshold.

2. The method for determining the potential sales volume of online commercial goods according to claim 1, characterized in that the second period is the period during which the first commodity is out of stock in the online commercial transaction.

3. The method for determining the potential sales volume of online commercial goods according to claim 1, characterized in that the step of obtaining the first search data further includes: The method further includes the step of obtaining third search data for the second query included in the query set within the first period, and the method also includes the step of determining a second exposure probability using the third search data, the second exposure probability being the probability that the first product is exposed due to the second query search. The step of obtaining the second search data further includes the step of obtaining a fourth search data for the second query within the second period. The step of determining the potential sales volume of the first product includes the steps of determining a portion of the potential sales volume of the first product within the second period using the first exposure probability and the second search data, and determining another portion of the potential sales volume of the first product within the second period using the second exposure probability and the fourth search data.

4. The method for determining the potential sales volume of online commercial goods according to claim 1, characterized in that the step of determining the first exposure probability includes: The step of obtaining the number of searches for the first query within the first time period from the first search data; The steps are as follows: obtaining the number of times the first product was exposed by the first query within a first period from the first search data; and determining the first exposure probability by using the number of searches of the first query and the number of times the first product was exposed by the first query.

5. The method for determining the potential sales volume of online commercial goods according to claim 1, characterized in that the step of determining the potential sales volume of the first product includes: The steps to obtain the sales volume of the first product within the first period; The steps include: obtaining the number of times the first product was exposed by the first query within a first period from the first search data; determining a first correlation ratio between the number of times the first product was exposed and the sales volume of the first product within the first period using the sales volume of the first product and the number of times the first product was exposed by the first query within the first period; determining the expected number of times the first product was exposed within the second period using the number of searches for the first query and the first exposure probability; and determining the potential sales volume of the first product within the second period using the first correlation ratio and the expected number of times the first product was exposed.

6. The method for determining the potential sales volume of online commercial goods according to claim 1, characterized in that the step of determining the query set related to the first product includes: The steps to obtain the first and third queries that caused the first product to be exposed due to the search; The steps include including the first query in the query set and excluding the third query from the query set, wherein the number of searches for the first query is above a first threshold and the number of searches for the third query is less than the first threshold.

7. The method for determining the potential sales volume of online commercial goods according to claim 1, characterized in that the step of determining the query set related to the first product includes: The steps to obtain the first and third queries that caused the first product to be exposed due to the search; The steps include including the first query in the query set and excluding the third query from the query set, wherein the probability of the first product exposure generated by the search of the first query is above a second threshold, and the probability of the first product exposure generated by the search of the third query is less than the second threshold.

8. The method for determining the potential sales volume of online commercial goods according to claim 1, characterized in that it further includes: The step of determining the quantity of the first commodity to be received into inventory in a third period using the potential sales volume of the first commodity within the second period, wherein the third period is the period following the second period, and the quantity of the first commodity to be received into inventory is the quantity of the first commodity to be received into inventory at the logistics center distributing the first commodity.

9. A method for determining the potential sales volume of online commercial goods, executed by a computing device, characterized in that it includes: The steps to determine the query set associated with the first product; The steps to obtain the first search data of the first query included in the query set within the first time period; The steps include: inputting the first search data and the sales volume of the first product within the first period into an artificial intelligence model for predicting the potential sales volume of the first product, so that the artificial intelligence model can learn; obtaining the second search data of the first query within a second period; and inputting the second search data into the learned artificial intelligence model to output the potential sales volume of the first product within the second period, wherein the first period is the period prior to the second period, and the potential sales volume is the predicted sales volume of the first product within the second period; wherein the step of determining the query set related to the first product includes: obtaining the first query and the third query that caused the first product to be exposed due to the search. The steps include including the first query in the query set and excluding the third query from the query set, wherein the correlation ratio between the first query and the sales volume of the first product is above a third threshold and the correlation ratio between the third query and the sales volume of the first product is less than the third threshold.

10. A device for determining the potential sales volume of online commercial goods, characterized in that it comprises: Communication interface; The system includes a memory that loads a computer program and one or more processors that run the computer program, which includes instructions for performing the following tasks: determining a query set related to a first product; obtaining first search data for the first query included in the query set within a first period; determining a first exposure probability using the first search data, the first exposure probability being the probability that the first product is exposed due to the first query search; obtaining second search data for the first query within a second period; and determining the potential sales volume of the first product within the second period using the first exposure probability and the second search data, the first period being a period prior to the second period, wherein the potential sales volume is the predicted sales volume of the first product within the second period, wherein determining the potential sales volume of the first product includes: obtaining the number of searches for the first query within the second period from the second search data; and determining the potential sales volume of the first product within the second period using the number of searches for the first query within the second period and the first exposure probability, wherein determining the query set related to the first product includes: The task involves obtaining a first query and a third query that cause the first product to be exposed due to a search; and including the first query in the query set while excluding the third query from the query set, wherein the correlation ratio between the first query and the sales volume of the first product is greater than a third threshold, and the correlation ratio between the third query and the sales volume of the first product is less than the third threshold.

11. A device for determining the potential sales volume of online commercial goods, characterized in that it comprises: Communication interface; The system includes a memory that loads a computer program, and one or more processors that run the computer program, which includes instructions for performing the following tasks: determining a query set related to a first product; obtaining first search data for a first query included in the query set within a first period; inputting the first search data and the sales volume of the first product within the first period into an artificial intelligence model for predicting the potential sales volume of the first product, so that the artificial intelligence model can learn; obtaining second search data for the first query within a second period; and inputting the second search data into the learned artificial intelligence model to output the potential sales volume of the first product within the second period, wherein the first period is a period prior to the second period, and the potential sales volume is the predicted sales volume of the first product within the second period, wherein determining the query set related to the first product includes: obtaining a first query and a third query that cause the first product to be exposed due to a search; The operation of including the first query in the query set and excluding the third query from the query set, wherein the correlation ratio between the first query and the sales volume of the first product is above a third threshold and the correlation ratio between the third query and the sales volume of the first product is less than the third threshold.

Images