CN-121998680-A - Real-time analysis method for sales data of coded products

CN121998680ACN 121998680 ACN121998680 ACN 121998680ACN-121998680-A

Abstract

The application provides a real-time analysis method of coded product sales data, which comprises the steps of tracking an original ticket purchasing store code and a prize exchanging store code of a customer winning, combining other customer identity lists of the ticket purchasing in the same line to form a customer ticket purchasing social network store code list, analyzing ticket purchasing frequency change values of the customer in the original ticket purchasing store based on a ticket purchasing store code section before prize exchanging and a ticket purchasing store code section after prize exchanging, extracting ticket purchasing frequency change of the customer in the same line in a social circle layer of the customer in a store comparison file, and identifying forward cumulative trend of a predicted contribution value of a ticket exchanging store and a contribution bias value in an actual contribution value in a continuous statistical period based on the correlation of the descending amplitude of the ticket purchasing frequency of the original ticket purchasing frequency after prize exchanging compared with the appearance frequency of the ticket purchasing frequency of the original ticket purchasing store before prize exchanging and the appearance frequency of the ticket purchasing store code section after prize exchanging.

Inventors

HUANG ZHIYU
ZHAO FENG
XIONG RUI
LIU SHUNCHENG
YAN GUANGYING
SHI BIN
CHEN WENLU
HE ZHENBO

Assignees

广东彩惠智能科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260123

Claims (10)

1. A method for real-time analysis of coded product sales data, the method comprising: tracking the original ticket purchasing store codes of the customers winning the prizes and forming a customer ticket purchasing social network store code list by combining other customer identity lists of the same-party ticket purchasing; Marking customers with inconsistent ticket store codes and original ticket store codes, associating the configuration and promotion types of original ticket store service personnel and ticket store service advantages and environmental factors to the ticket purchase social network of the customers to form store comparison files, and determining a ticket store code section before the lottery and a ticket store code section after the lottery; Based on the ticket purchasing store coding section before the lottery ticket is paid and the ticket purchasing store coding section after the lottery ticket is paid, analyzing the ticket purchasing frequency change value of the customer in the original ticket purchasing store, and extracting the ticket purchasing frequency change of the same-row ticket purchasing customers in the social circle layer of the customer in the store comparison file; Confirming that the customer shows signs of loss transferred from the original ticket store to the ticket exchange store based on the correlation of the increase in the number of occurrences of the ticket exchange store code in the ticket exchange code section after the ticket exchange; After confirming the loss sign, identifying a forward accumulated trend of the contribution deviation value of the expected contribution value and the actual contribution value of the store redemption in a continuous statistical period; and when the contribution deviation value is continuously positive and the accumulated amount exceeds the average consumption of the historical single period of the customer, judging that systematic overestimation exists, identifying a defending window period of the original store due to insufficient configuration of service personnel, taking other customer identities in the same social circle layer, which have started to reduce ticket purchase of the original store, into a chain loss range, and pushing the other customer identities to the original ticket purchase store.
2. The method for real-time analysis of encoded product sales data according to claim 1, wherein said tracking of original ticket store encoding and prize-exchange store encoding for a customer winning, in combination with other customer identity lists for a peer ticket, forms a customer ticket purchase social network store encoding list, comprising: Acquiring other ticket purchasing records in a preset time range before and after a ticket purchasing period of a customer, identifying a peer ticket purchasing customer according to a rule that the payment time difference is smaller than a preset time threshold and the ticket purchasing amount difference is smaller than a preset amount threshold, and recording the identity number and the ticket purchasing store code of the peer ticket purchasing customer to form an initial social connection record centering on the customer; Aiming at the lottery ticket exchanging behavior of the winning lottery ticket, reading a lottery ticket exchanging store code and a lottery ticket exchanging timestamp, inquiring the original ticket purchasing store code of the winning lottery ticket, extracting the identity numbers of the customers of the same party from the initial social connection record, inquiring the ticket purchasing record of the customers of the same party in a preset time period after the lottery ticket exchanging time point, and acquiring the ticket purchasing store code sequence of the customers of the same party; And constructing a data structure comprising a customer identity number, an original ticket purchasing store code, a prize exchanging store code, a peer customer identity list and a corresponding ticket purchasing store code sequence according to the initial social association record and the ticket purchasing store code sequence, and storing ticket purchasing association and prize exchanging association between customer nodes and store nodes in an adjacency list mode to form a customer ticket purchasing social network store code list.
3. The method of claim 1, wherein the marking ticket store code and the original ticket store code are inconsistent customers, associating the original ticket store service personnel configuration and the promotion type, and ticket store service advantages and environmental factors to the customer's ticket social network, forming store comparison files, and determining a pre-ticket store code section and a post-ticket store code section, comprising: Aiming at the customers of which the lottery ticket store codes are inconsistent with the original ticket store codes, acquiring the number of service personnel, the personnel skill level and the average service life of the original ticket store, and simultaneously acquiring the promotion activity type codes, the activity frequency and the discount strength of the ticket store in a preset period before lottery ticket exchange to form an original ticket store service attribute record; inquiring service personnel configuration data and environment scoring data of the lottery ticket exchange store according to the store codes in the original ticket purchase store service attribute records to obtain lottery ticket exchange store environment attribute records; Reading all relevant customer identity codes in a ticket purchasing social network of the customer by comparing the original ticket purchasing store service attribute record with the lottery ticket exchanging store environment attribute record, inquiring historical ticket purchasing frequency and amount of each relevant customer in two stores, and constructing a store comparison file; Based on the store comparison file, a pre-exchange time window and a post-exchange time window are determined by taking the exchange time as a central point, and a pre-exchange ticket store coding section and a post-exchange ticket store coding section are extracted from a ticket purchase record.
4. The method for real-time analysis of sales data of coded products according to claim 1, wherein the step of analyzing the change in ticket buying frequency of customers in the original store based on the pre-exchange ticket buying store coding section and the post-exchange ticket buying store coding section, and extracting the change in ticket buying frequency of co-operating ticket buying customers in the social circle layer of the customers in the store comparison file comprises the steps of: Counting the number of times of the original ticket purchasing store code from the encoding section of the ticket purchasing store before the lottery and the encoding section of the ticket purchasing store after the lottery, and calculating the number of times of ticket purchasing of the original ticket purchasing store monthly before the lottery and the number of times of ticket purchasing of the original ticket purchasing store monthly after the lottery to obtain a ticket purchasing frequency change value of a customer in the original ticket purchasing store; And reading a social circle member list of the customer from the store comparison file according to the ticket buying frequency change value, counting the total ticket buying times of the original ticket buying stores in the time window before and the time window after the lottery exchange for each social circle member, and calculating the ticket buying frequency change of each member.
5. The method of claim 1, wherein the step of confirming the customer's sign of loss of the transfer from the original ticket gate to the ticket gate based on the correlation of the decrease in ticket gate frequency after the redemption compared to the ticket frequency before the redemption and the increase in the number of occurrences of the ticket gate code in the ticket gate code section after the redemption, comprises: Calculating the reduction rate of the ticket frequency of the original ticket store before the lottery and the ticket frequency of the original ticket store after the lottery, and simultaneously counting the number of times increment of the lottery codes in the ticket code section after the lottery; constructing a coordinate point according to the descent rate and the increment of the number of times of occurrence of the lottery store codes, and judging whether the coordinate point is positioned in a preset loss judging area or not; And if a transfer relationship exists and the drop rate and the increment of the number of times of the lottery ticket exchange store code appearance meet preset conditions, confirming that the customer shows the sign of loss of transfer from the original ticket buying store to the lottery ticket exchange store.
6. The method of claim 1, further comprising extracting a customer record from the store comparison profile identifying signs of churn, analyzing the tendency of churn customers to follow transfer by opinion leader effects in the social circle layer, evaluating siphonic effects of the redemption store on peripheral stores through public praise propagation of the customer, and determining the identity of other customers in the churn layer who are likely to follow transfer.
7. The method according to claim 6, wherein the step of extracting the customer records for confirming the sign of the loss from the store comparison file, analyzing the tendency of the opinion leader of the lost customer in the social circle layer to cause other customers to follow the transfer, evaluating the siphon effect of the prize exchange store to the surrounding stores through the public praise propagation of the customer, and determining the identity of other customers in the social circle layer that may follow the transfer, comprises: Extracting a customer record of confirmed loss signs from the store comparison file, acquiring ticket purchasing history data and social network relation data of the customer, counting the proportion of ticket purchasing initiation times of the customer in a social circle layer to total ticket purchasing times, marking the customer as a potential opinion leader if the proportion exceeds a preset proportion, and recording the time of the customer to purchase tickets in a prize-exchange store for the first time as an individual transfer starting point; Searching the ticket purchasing behavior change of other members in the social circle layer after the time point according to the individual transfer starting point, counting the number of the descending ticket purchasing frequency of each member in the original ticket purchasing store and the newly-increased ticket purchasing record in the prize-exchanging store, calculating the average value of the ticket purchasing frequency change of each member, and determining the sightseeing starting point of the social circle layer; Calculating the time difference between the individual transfer start point of the opinion leader and the circle layer sightseeing start point by adopting the social circle layer sightseeing start point and the potential opinion leader mark, counting the ticket purchasing times of the opinion leader in the lottery store within the time difference, calculating the propagation frequency and combining with a preset weight coefficient to obtain the siphon intensity value of the lottery store to the surrounding stores through the opinion leader; And screening the influence range through the siphon intensity value, identifying members with continuously reduced ticket purchasing frequency after the social circle layer sightseeing start point and ticket purchasing records in the lottery store, and determining the members as other customer identities possibly following transfer.
8. The method of claim 1, wherein after the identifying the sign of the loss, identifying a forward cumulative trend of the contribution bias value of the expected contribution value and the actual contribution value of the store redemption over a continuous statistical period, comprising: After confirming the loss sign, counting the total number of ticket purchasing times and the sum of ticket purchasing amounts of the original ticket purchasing stores in the code section of the ticket purchasing stores before the prize exchange, calculating the single average ticket purchasing amount, and multiplying the single average ticket purchasing amount by the number of ticket purchasing times of the original ticket purchasing stores to obtain an expected contribution value; Counting the number of times of ticket purchase in the original ticket store in the coding section after the prize is awarded and multiplying the number by the single average ticket purchase amount to obtain an actual contribution value; Calculating a contribution deviation value according to the expected contribution value and the actual contribution value, recording the deviation value according to a preset statistical period, and identifying a forward accumulation trend if the deviation value in a plurality of continuous statistical periods is positive and the deviation accumulated value in a later period is larger than that in a previous period.
9. The method for real-time analysis of encoded product sales data according to claim 1, wherein identifying a defense window period of an original store due to insufficient configuration of service personnel comprises evaluating the attractive diffusion of a redemption store formed in a surrounding community by a public praise propaganda of a lost customer in combination with other customer identities possibly following transfer in a social circle layer of the lost customer, and determining the defense window period of the original store due to insufficient configuration of service personnel, wherein the defense window period of the original store cannot be timely saved for the lost customer.
10. The method according to claim 1, wherein when the contribution deviation value is continuously positive and the accumulated amount exceeds the average consumption of the customer historic single cycle, determining that there is a systematic overestimation, identifying a defending window period of the original store due to insufficient configuration of service personnel, taking other customer identities in the same social circle layer, which have begun to reduce the ticket purchase of the original store, into a chain loss range, and pushing the other customer identities to the original ticket purchase store, comprises: determining that a systematic overestimation exists when the contribution bias value continues to be positive and the accumulated amount exceeds the customer historical single-cycle average consumption; acquiring ticket purchasing records of the customer in a preset period, counting the amount of commodities when buying the ticket each time, and calculating a sales opportunity loss value; inquiring a scheduling record of an original store service personnel according to the sales opportunity loss value, judging that personnel configuration is insufficient, and determining a defense window period; and screening other customers with descending ticket purchasing frequency in the defense window period from the social circle layer data according to the defense window period and the systematically overestimated judgment result, and pushing the other customers to the original ticket purchasing store.

Description

Real-time analysis method for sales data of coded products Technical Field The invention relates to the technical field of information, in particular to a real-time analysis method for sales data of coded products. Background In the field of business retail and customer management, research on how to promote customer retention and accurately predict customer behavior is critical to enterprise development. This area is directly related to the revenue growth and market competitiveness of enterprises, especially in real-time analysis of product sales, where customer behavior patterns and loyalty become central factors in determining the long-term profitability of enterprises. How to balance complex customer flow and cross-regional behavior is a current challenge. The existing method often ignores the deep influence of cross-regional interaction on customer loyalty when processing customer behavior analysis. Many schemes only focus on the data presentation of a single store and do not adequately account for the chain reaction of customers flowing between different stores. The problem is particularly prominent in the scene of inter-store interaction, especially when the real cause of customer loss cannot be accurately known by enterprises due to neglect. In particular, in the context of rewards redemption, the seemingly convenience may hide more complex customer attribution changes. What is really tricky is how to accurately capture and analyze the behavior trace of customers between different stores, especially the relationship between the first redemption store and the original ticket purchase store. This factor directly determines the accuracy of the customer's risk of loss. Due to the failure to effectively track the complete behavioral paths of customers at different stores, enterprises often fail to discover the decreasing trend of customer loyalty in time. For example, if a customer purchases a product at a store, and then chooses to pay at a store B, the store A may gradually lose the consumer contribution of the customer if the store B attracts the customer to become a frequent customer through a live event, and conversely, if the customer goes to store A to pay after purchasing a ticket at the store B, the customer is attracted by the service or environment of the store A and is changed to continue to purchase at the store A, and the store B also faces the dilemma that the customer is lost. In either case, if the original ticket store cannot detect the change, the future value of the customer is overestimated based on the past data, and an incorrect marketing strategy is formulated. For example, customers purchasing tickets at suburban betting stations are diverted from loyalty due to the convenience of redemption and additional services at the central betting stations, and such a change is difficult for the original store of the enterprise to perceive, resulting in wasted marketing resources. Further, in a chain betting network, customers may purchase tickets at small community stores and then pay and renew purchases at large store stores, which not only affects single store income, but also amplifies the customer loss risk of the whole network, because the original store cannot foresee the chain effect of loyalty transfer. If the enterprise only depends on store local data, the future value of the client can be overestimated, and the attribution ambiguity caused by cross-store flow is ignored. This risk recognition lag, which is caused by incomplete recording of the behavior trace, is a problem that needs to be solved in the current technology. Disclosure of Invention The invention provides a real-time analysis method for sales data of coded products, which mainly comprises the following steps: tracking the original ticket purchasing store codes of the customers winning the prizes and forming a customer ticket purchasing social network store code list by combining other customer identity lists of the same-party ticket purchasing; Marking customers with inconsistent ticket store codes and original ticket store codes, associating the configuration and promotion types of original ticket store service personnel and ticket store service advantages and environmental factors to the ticket purchase social network of the customers to form store comparison files, and determining a ticket store code section before the lottery and a ticket store code section after the lottery; Based on the ticket purchasing store coding section before the lottery ticket is paid and the ticket purchasing store coding section after the lottery ticket is paid, analyzing the ticket purchasing frequency change value of the customer in the original ticket purchasing store, and extracting the ticket purchasing frequency change of the same-row ticket purchasing customers in the social circle layer of the customer in the store comparison file; Confirming that the customer shows signs of loss transferred from the original ticket store to