A multi-channel adaptive preferred bank message intelligent sending method, system and device

Abstract

The application relates to the technical field of financial marketing, and discloses a multi-channel adaptive optimization bank message intelligent sending method, system and device, which has the technical scheme as follows: S1, collecting full-channel user data to form a user original channel data set; S2, through a T+1 offline batch statistical model, processing full-amount user channel data collected on the previous day on the current day, generating a standardized channel state label, and constructing a global user channel state label library; S3, receiving a message pushing task, analyzing core parameters of the task, and matching corresponding channel optimization rules; S4, calling the global user channel state label library, performing effectiveness verification on each channel according to the priority set by the channel optimization rules, and screening out the optimal available pushing channel; S5, using the optimal available channel to push the message, and recording the pushing log and the delivery state; and if all online channels are in invalid states, automatically triggering an SMS bottom pushing.

Description

Technical Field

[0001] This invention relates to the field of financial marketing technology, and more specifically, to a method, system, and apparatus for intelligent sending of bank messages with multi-channel adaptive optimization. Background Technology

[0002] With the rapid popularization of digital finance, banking services are gradually becoming online and lightweight, and business message delivery has become a core means for banks to conduct marketing promotion, risk notification, business information dissemination, and customer maintenance. Banks daily need to push various messages to a massive number of users, including loan marketing, wealth management notifications, transaction reminders, risk control warnings, bill notifications, and promotional pushes. The delivery rate, cost controllability, and user experience of these messages directly determine the efficiency of banks' online operations and the quality of their customer service.

[0003] Currently, bank messaging channels are mainly divided into two categories: online digital channels and traditional SMS channels. The digital channels include three core channels: bank-operated APP, WeChat mini-program, and WeChat official account, which have the advantages of zero cost, no blocking, precise reach, and the ability to carry rich media content. SMS is a paid channel for mobile operators, and it suffers from industry pain points such as single-message billing, susceptibility to interception by mobile phone systems and operators, strong user perception of harassment, and extremely low delivery rate of marketing messages. Currently, most bank message push systems adopt a push logic that prioritizes SMS and supplements online channels, or a fixed multi-channel parallel push mode. Regardless of whether users actively use the bank's APP, mini-program, or official account, and regardless of whether users have enabled online message push permissions, the system sends SMS by default, only selectively adding push notifications from online channels. This traditional model does not require complex user channel status analysis and has simple implementation logic, but it has serious drawbacks in large-scale message push scenarios: on the one hand, massive invalid SMS pushes will generate extremely high operating costs, resulting in a waste of financial resources; on the other hand, marketing and notification SMS messages are easily intercepted by mobile phone security systems and operator spam SMS databases, leading to message delivery failures and customer reach gaps. At the same time, repeated pushes will also cause user complaints and reduce customer experience.

[0004] To optimize message delivery, some existing technologies attempt to use a multi-channel push model, simply adding apps, official accounts, and SMS channels to improve delivery rates, but they haven't achieved intelligent channel filtering and adaptive optimization. Existing related technologies mainly fall into the following three categories: 1. Fixed multi-channel parallel push technology: The system is preset to push messages simultaneously through APP, official account and SMS, without distinguishing the user’s channel usage status and permission status. All messages are sent through all channels. Although it can improve the delivery rate to a certain extent, it will exacerbate the redundant push of channels and further increase the cost of SMS. At the same time, repeated messages from multiple channels are likely to cause user resentment.

[0005] 2. Single online channel priority push technology: Some systems simply set the APP push to be prioritized, without taking into account mini-programs and official account channels, and do not verify the user's channel switch status. If the user turns off the APP push permission, the system will still give up the push or directly jump to SMS, resulting in extremely poor channel adaptation flexibility.

[0006] 3. Static channel configuration push technology: Relying on manually preset user channel push priorities, it cannot dynamically update based on real-time user behavior and channel activity status. It also cannot adapt to changes in user channel usage habits, resulting in severely insufficient adaptability. With the continuous expansion of bank user base and a significant increase in message push frequency, the traditional fixed push, manual configuration, and multi-channel parallel message sending model can no longer meet banks' core needs for cost reduction and efficiency improvement, high delivery rates, low harassment, and intelligent adaptation.

[0007] How to accurately identify the usage status and permission status of users on various online channels, dynamically select low-cost and high-delivery digital channels, and only use SMS as a last resort when no online channels are available has become a key technical problem that urgently needs to be solved in the field of intelligent message push in the banking industry.

[0008] Existing multi-channel message push technologies used by banks have several core shortcomings in terms of user channel identification, intelligent optimization strategies, cost control, delivery rate assurance, and dynamic adaptation. These shortcomings are as follows: 1. The current technology for identifying user channel status is too simplistic, lacking a refined approach. It merely checks whether a user has registered for an app or followed a public account, failing to deeply collect and analyze core data such as user channel activity, usage frequency, online status, and push notification permissions. It cannot differentiate between user states such as "registered but not used," "occasionally used," "highly active," "push notification permissions disabled," and "channel silently inactive." This easily leads to invalid channels that have been inactive for a long time or have their message switches disabled being classified as valid, resulting in failed or invalid push notifications.

[0009] 2. Delayed updates to channel status data, failing to adapt to dynamic changes in user behavior. The existing push system lacks a systematic mechanism for collecting and periodically statistically analyzing user channel data, relying mostly on static historical data. This makes it impossible to achieve T+1 batch updates of user channel status tags. Dynamic changes in user behavior, such as app usage habits, mini-program access behavior, unfollowing / re-following official accounts, and enabling / disabling message switches, cannot be synchronized to the push system in a timely manner. This leads to a disconnect between push strategies and real-time user channel status, resulting in mismatches such as "pushing online messages to users who have disabled permissions and blindly sending SMS messages to active users."

[0010] 3. Lack of a tiered adaptive channel optimization strategy and insufficient cost control capabilities: Existing technologies lack a clear multi-channel priority intelligent filtering logic, resulting in either parallel push across all channels or fixed single-channel push, failing to achieve tiered optimization of "APP-Mini Program-Official Account-SMS". The failure to prioritize the reuse of zero-cost digital channels leads to a large amount of messages that could be delivered online being wasted on SMS, resulting in consistently high SMS costs for operators and persistently high operational costs for bank messaging.

[0011] 4. Lack of a strict online channel validity verification mechanism results in insufficient delivery rate assurance. Current technology only verifies the existence of the channel, without verifying the channel's message push status or recent activity validity. Some users may follow the official account or register for the app, but have not logged in for a long time or have turned off all message push permissions. Such channels cannot reach users at all. Current technology will still prioritize pushing online messages, ultimately leading to push failures without any fallback mechanism, resulting in message omissions and ineffective customer outreach.

[0012] 5. Rigid Push Strategy and Lack of Differentiated Backup Logic: Existing technology fails to differentiate between the two core scenarios of "no available online channels" and "online channel permissions are disabled," resulting in a simplistic backup logic. For users who have access to channels but have all messaging switches off and no active online channels, it cannot accurately trigger backup SMS pushes. Furthermore, it cannot adapt different channel priorities based on message type (marketing, risk control, notification), using the same push logic for urgent risk control messages and ordinary marketing messages, leading to extremely poor adaptability.

[0013] 6. Lack of a data closed-loop iteration mechanism hinders continuous optimization of push notification performance. The existing push system only performs message sending actions without accumulating and analyzing push results, channel delivery data, and user channel behavior data. It cannot iteratively optimize channel selection weights and user status judgment rules based on historical push data, resulting in persistent issues of invalid pushes and channel mismatch, and failing to continuously improve delivery rates and cost utilization.

[0014] Lacking lightweight batch data scheduling capabilities, the system has poor adaptability. The volume of channel behavior data of massive bank users is huge. Existing technologies do not have a lightweight T+1 offline statistical model. Real-time collection and analysis of user channel data will consume a lot of system computing power and affect the operation of core business systems. On the other hand, purely static data cannot guarantee accuracy and it is difficult to achieve efficient and accurate channel status determination in massive user scenarios. Summary of the Invention

[0015] The purpose of this invention is to provide a multi-channel adaptive optimization method, system, and device for intelligent sending of bank messages. It is designed for various online message push scenarios in banks and integrates six core mechanisms: multi-dimensional channel data collection, T+1 offline statistical modeling, hierarchical channel optimization, permission validity verification, differentiated SMS backup, and data closed-loop iteration, to achieve intelligent, low-cost, and high-efficiency delivery of bank messages.

[0016] The above-mentioned technical objective of the present invention is achieved through the following technical solution: a multi-channel adaptive optimization method for intelligent sending of bank messages, comprising the following steps: S1. Collect user data from all channels and perform standardized data processing to form the original user channel dataset through unique user identifiers; S2. Based on the user's original channel dataset, the T+1 offline batch statistical model is used to process the full amount of user channel data collected the previous day on the same day, generate standardized channel status labels, and build a global user channel status label library. S3: Receive push notification tasks, parse the core parameters of the tasks, and match the corresponding channel selection rules; S4. Retrieve the global user channel status tag library, and perform validity checks on each channel one by one according to the priority set by the channel selection rules, and select the best available push channel; S5. Use the best available channel to push messages and record push logs and delivery status; if all online channels are invalid or all channel message switches are turned off, automatically trigger SMS backup push.

[0017] As a preferred technical solution of the present invention, the user data for each channel includes behavioral data, permission data, and status data.

[0018] As a preferred technical solution of the present invention, in S2, the standardized channel status labels include: validity labels, activity labels, and permission status labels for each user corresponding to each channel.

[0019] As a preferred embodiment of the present invention, in S2, the application process of the T+1 offline batch statistical model is as follows: S21. Set activity thresholds, validity thresholds, and permission compliance thresholds for each type of channel; S22. For each user and each channel, calculate the channel effectiveness score separately: ; in: The channel activity score is the effectiveness score. Score the effectiveness of channel authorization. Score for message switch permissions. The score is based on the timeliness of recent use. , , , Preset weight parameters; S23. Based on the channel effectiveness score and the preset label classification rules, assign labels to the status of each channel; S24. Update the standardized channel status tags of the previous day on the same day to form a global user channel status tag library.

[0020] As a preferred embodiment of the present invention, in S23, the labels include: valid and available, conditionally available, invalid and unavailable, and fully expired.

[0021] As a preferred embodiment of the present invention, S3 includes: S31. Analyze the core parameters of the message push task; S32. Match the preset differentiated channel selection rules according to the core parameters; the channel selection rules include the channel priority weights configured for different message types.

[0022] As a preferred technical solution of the present invention, S31, the core parameters of the task include: target user list, message type, push priority, whether to force delivery, and message validity period; message types include: marketing, transaction reminder, risk control warning, and billing notification.

[0023] As a preferred technical solution of the present invention, after S5, S6 is also executed: summarizing the daily push results, channel delivery data, and user channel behavior change data, and iteratively optimizing the channel selection rules and selection weights.

[0024] A multi-channel adaptive optimization intelligent bank message sending system includes: Multi-source data acquisition module: Collects user data from all channels and performs standardized data processing to form the original user channel dataset through unique user identifiers; T+1 Offline Statistical Modeling Module: Based on the user's original channel dataset, the T+1 offline batch statistical model processes all user channel data collected the previous day on the same day, generates standardized channel status labels, and builds a global user channel status label library. Push task parsing module: Receives push message tasks, parses the core parameters of the tasks, and matches them with the corresponding channel selection rules; Multi-channel optimization and verification module: Retrieves the global user channel status tag library, performs validity verification on each channel one by one according to the priority set by the channel optimization rules, and selects the best available push channel; Online message distribution module: Push messages using the best available channels and record push logs and delivery status; SMS backup trigger module: used to automatically trigger SMS backup push when all online channels are in an invalid state or all channel message switches are turned off.

[0025] A multi-channel adaptive optimization intelligent bank message sending device includes: a processor and a memory, wherein the memory stores a computer program executable by the processor, and the processor implements the above-described method when executing the computer program.

[0026] In summary, this invention offers the following advantages: It constructs a multi-dimensional user channel status collection and T+1 statistical mechanism to precisely identify the activity status, usage behavior, and message push permission status of users across three major online channels: the bank's app, WeChat mini-program, and WeChat official account; it establishes a tiered channel optimization strategy, prioritizing the matching of messages to effective, active, and authorized online digital channels; and it triggers SMS channel backup sending only when the user has no available online channels or all online channel message switches are turned off, while dynamically adjusting the push strategy based on message type, forming a fully intelligent and adaptive message distribution system. It significantly reduces bank message operation costs: by maximizing the reuse of zero-cost online channels and reducing redundant SMS sending, it fundamentally reduces operator SMS costs and optimizes the cost of large-scale message push. It significantly improves the overall message delivery rate: avoiding the problem of marketing SMS being blocked by the system and operators, it leverages the advantages of online channels—no blocking and high reach—to solve the pain point of traditional SMS delivery failure. It achieves refined and differentiated user outreach: based on users' actual channel usage habits and permission status, it intelligently adapts push channels, avoiding invalid and duplicate pushes, reducing user harassment, and improving customer service experience. Building a closed-loop data iteration system: By periodically updating user channel tags on a T+1 basis and accumulating push effect data, the system continuously optimizes channel selection rules to improve long-term push accuracy and system robustness. Adapting to large-scale push scenarios with massive users: Adopting a lightweight architecture of offline T+1 statistics + real-time calls, it balances data accuracy and system operating efficiency, adapting to the business needs of banks with massive users and high-frequency message pushes.

[0027] Therefore, this invention can be widely applied to message delivery across all scenarios, including bank marketing push notifications, transaction alerts, risk warnings, bill notifications, and customer maintenance. It effectively solves the industry pain points of high cost, low delivery rate, and poor adaptability of traditional message push, providing core technical support for the digital and intelligent operation of banks. For the target users of the message to be pushed, it comprehensively collects client behavior data and permission data from their bank APP, WeChat mini-program, and WeChat official account. It generates a user channel status tag library through a T+1 offline statistical model; it verifies the activity validity, push on / off permissions, and recent usage status of each online channel in real time, and adaptively selects the best available online channel to complete the message sending according to the preset priority; if no available online channels are found after verification, it automatically triggers the SMS channel as a backup push; at the same time, it continuously iterates and optimizes the channel selection rules based on historical push data to achieve full-process adaptive intelligent push. Attached Figure Description

[0028] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0029] It is readily understood that, based on the technical solution of this invention, various embodiments of the invention can be conceived by those skilled in the art without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention. Rather, these embodiments are provided to enable those skilled in the art to gain a more thorough understanding of the invention. Preferred embodiments of the invention are described below in conjunction with the accompanying drawings, which form part of this application and, together with the embodiments of the invention, serve to illustrate the innovative concept of the invention.

[0030] like Figure 1 As shown, this invention provides a multi-channel adaptive optimization method for intelligent sending of bank messages, comprising the following steps: S1. Multi-channel user data collection across the entire domain: The system establishes multi-source data acquisition interfaces to collect user data from all channels in real time and in batches, providing data support for subsequent channel status determination.

[0031] In the current app market, "omnichannel" refers to online channels such as users' bank apps, WeChat mini programs, and WeChat official accounts. User data for each channel includes behavioral data, permission data, and status data, covering multiple dimensions such as channel usage, permission activation, activity frequency, and online status.

[0032] Bank APP channel data: user registration status, login frequency in the past 30 days, last login time, APP push notification status, marketing message status, notification message status, device binding status, and APP online status.

[0033] WeChat Mini Program channel data: user authorization status, access frequency in the past 30 days, last access time, Mini Program message subscription switch status, whether it has been silent for a long time without access, and authorization expiration status.

[0034] WeChat Official Account Channel Data: User follow / unfollow status, message receiving on / off status, private message push permissions, interaction frequency in the past 30 days, last interaction time, and validity of official account authorization.

[0035] After data collection is completed, standardized data processing is performed to form the original user channel dataset using unique user identifiers. This includes cleaning, deduplication, and normalization of the collected raw data, and unifying unique user identifiers such as mobile phone numbers, ID card numbers, and customer numbers to form the original user channel dataset. ;in, The original dataset representing the channels of user u. For APP channel data, For mini-program channel data, Data from WeChat Official Accounts.

[0036] S2 and T+1 Channel Status Offline Statistics and Tag Modeling: Based on the original user channel dataset, a T+1 offline batch statistical model is used to process all user channel data collected the previous day on the same day, generating standardized channel status tags and constructing a global user channel status tag library. The standardized channel status tags include: validity tags, activity tags, and permission status tags for each user across each channel.

[0037] Considering the computational pressure of massive user data in banks, this invention adopts a T+1 offline batch statistical mechanism, which processes all user channel data collected the previous day offline at midnight every day, generates standardized channel status tags, avoids real-time big data calculation from occupying the core system's computing power, and ensures the stable operation of the business system.

[0038] In S2, the application process of the T+1 offline batch statistical model is as follows: S21. Quantification of Status Judgment Indicators: Set activity thresholds, validity thresholds, and permission compliance thresholds for each type of channel; for example: users who have logged in / accessed within the past 30 days are considered active users, users with the message switch turned on are considered to have valid permissions, and users with no operation for more than 90 days are considered to have the channel inactive.

[0039] S22. Multi-dimensional Status Scoring: For each user and each channel, calculate the channel effectiveness score separately. ; in: The channel activity score is the effectiveness score. Score the effectiveness of channel authorization. Score for message switch permissions. The score is based on the timeliness of recent use. , , , Preset weight parameters; S23. Generate standardized labels: Based on the channel effectiveness score and preset label classification rules, assign labels to each channel status. In S23, the labels include: Valid and Available, Conditionally Available, Invalid and Unavailable, and Fully Time-Limited. Specifically, based on the channel effectiveness score, each channel status is divided into four categories of labels: a. Effective and usable: Score ≥ high threshold, active channels, full permissions, can be pushed directly; b. Conditions available: The score is between the high and low thresholds, the channel is slightly active, and selective push can be performed; c. Invalid and unavailable: Score < low threshold, channel is silent, permissions are closed, authorization is invalid; d. Completely invalid: The user has not activated / has cancelled the corresponding channel.

[0040] S24. Build a global tag library: Update the standardized channel status tags of the previous day on the current day to form a global user channel status tag library that can be called in real time, so that the message push system can call it in real time.

[0041] S3. Task parsing and scenario matching: The system receives various message push tasks issued by the banking business terminal and completes task parsing and scenario rule matching.

[0042] include: S31. Analyze the core parameters of the message push task, including: target user list, message type, push priority, whether to force delivery, and message validity period; message types include: marketing, transaction reminder, risk control warning, and billing notification.

[0043] S32. Matching Differentiated Optimization Rules: Based on the core parameters, match the preset differentiated channel optimization rules; the channel optimization rules include the channel priority weights configured for different message types.

[0044] Specifically, channel priority weights are configured for different message types. Regular marketing messages strictly follow the priority order of "APP > Mini Program > Official Account > SMS". Emergency risk control and transaction warning messages can have relaxed channel restrictions. SMS is used as a backup after being pushed through multiple online channels in parallel to ensure timeliness.

[0045] S4. Online Channel Layering and Validation: Retrieve the global user channel status tag library and, according to the priority set by the channel selection rules, generally: according to the preset priority of APP, mini program, and official account, perform validity verification on each channel one by one to select the best available push channel; Specifically: The system uses a user T+1 channel tag library to perform channel verification and optimization for each target user, adhering to the principles of prioritizing low cost, high delivery speed, and active channels. First priority verification of bank APP: Determine whether the user's APP channel tag is valid / conditionally available and whether the push switch for the corresponding message type is turned on. If so, select the APP as the push channel. The second priority is to verify the WeChat Mini Program: If the APP channel is invalid, the permission is closed or not enabled, the validity of the Mini Program channel and the subscription switch status are verified. If it is valid, the Mini Program is selected as the push channel. The third priority is to verify the WeChat Official Account: If the APP and Mini Program are both invalid, verify the Official Account's follow status and message receiving permissions. If they are valid, the Official Account is selected as the push channel. Invalid channel filtering: Automatically filters all silent, disabled, and long-unused online channels to prevent invalid push notifications.

[0046] S5, Differentiated Message Intelligent Distribution: Push messages using the best available channels and record push logs and delivery status; For APP channels: Push corresponding types of APP pop-up notifications and message center notifications, adapting to user device types; For mini-program channels: push WeChat service notifications and mini-program subscription messages, in line with the WeChat ecosystem's reach logic; For WeChat Official Accounts: Push official account template messages and send private message notifications to ensure that messages are delivered in compliance with regulations.

[0047] After the push is completed, the push status, delivery result, and read status are recorded in real time to form a single-user push log.

[0048] SMS backup verification and resending: If all online channels are invalid or all channel message switches are turned off, an SMS backup push will be automatically triggered to ensure that the message is delivered.

[0049] Specifically, the precise triggering mechanism system has strict SMS fallback conditions. SMS messages are only sent when all of the following conditions are met, thus minimizing SMS costs: 1. The user has not activated any online channels such as the APP, mini-program, or official account; 2. Or all of the user's online channels are invalid or silent; 3. Or all push notifications for the user's online channels are turned off, making any online contact impossible.

[0050] The fallback SMS is only used to ensure message delivery. It can be appropriately exempted for non-urgent marketing messages, but is mandatory only for important messages such as risk control, transactions, and billing, in order to further optimize the cost structure.

[0051] S6. Push Data Feedback and Model Iteration: Summarize daily push results, channel delivery data, and user channel behavior change data, iterate and optimize channel selection rules and selection weights, and continuously improve push accuracy.

[0052] Specifically, the system aggregates all push data daily, constructs a data closed-loop iteration mechanism, and continuously optimizes push accuracy: Data feedback: aggregates push success rate, delivery rate, user click-through rate, SMS trigger ratio, and invalid push data from various channels; Rule iteration: dynamically adjusts channel effectiveness scoring weights, priority rules, and threshold parameters based on push performance data; Status update: revises T+1 tag modeling rules by combining the latest user channel behavior data for the day, reducing judgment errors; Anomaly optimization: optimizes verification logic for abnormal scenarios such as push failures, channel mismatches, and invalid SMS pushes, preventing the recurrence of similar problems.

[0053] Corresponding to the above method, the present invention provides a multi-channel adaptive optimization intelligent bank message sending system, comprising: Multi-source data acquisition module: Collects user data from all channels and performs standardized data processing to form the original user channel dataset through unique user identifiers; T+1 Offline Statistical Modeling Module: Based on the user's original channel dataset, the T+1 offline batch statistical model processes all user channel data collected the previous day on the same day, generates standardized channel status labels, and builds a global user channel status label library.

[0054] Push task parsing module: Receives push message tasks, parses the core parameters of the task, and matches them with the corresponding channel selection rules.

[0055] Multi-channel optimization and verification module: Retrieves the global user channel status tag library, performs validity verification on each channel one by one according to the priority set by the channel optimization rules, and selects the best available push channel; Online message distribution module: Push messages using the best available channels and record push logs and delivery status; SMS backup trigger module: used to automatically trigger SMS backup push when all online channels are in an invalid state or all channel message switches are turned off.

[0056] Data Iteration and Optimization Module: Used to summarize push effect data, iteratively optimize channel judgment rules, scoring weights and push strategies, and realize system adaptive upgrades.

[0057] Data storage module: Used to store core data such as user channel raw data, status tags, push logs, and iteration rule parameters.

[0058] Corresponding to the above methods and systems, the present invention also provides a multi-channel adaptive optimization intelligent bank message sending device, comprising: a processor and a memory, wherein the memory stores a computer program executable by the processor, and the processor implements the above methods when executing the computer program.

[0059] The innovation of this invention lies in: Innovation Point 1: A multi-dimensional, refined user channel identification mechanism based on T+1 offline statistics: Existing technologies simply determine the activation status of user channels, resulting in a single identification dimension, lagging data, and an inability to adapt to dynamic changes in user behavior. This invention breaks through the traditional static determination model, constructing a four-dimensional determination system of behavior, permissions, activity, and timeliness. By conducting daily T+1 offline batch statistics on key indicators such as the frequency of use of user apps, mini-programs, and official accounts, login status, message on / off permissions, and recent activity, it quantifies channel effectiveness and generates standardized status labels.

[0060] The core of this innovation lies in abandoning the crude binary judgment of "either open or closed" and realizing a refined hierarchical identification of users' online channels as "effectively available, conditionally available, ineffective and unavailable, and completely ineffective". At the same time, it adopts an offline computing mode to take into account the timeliness of data updates and the stability of system computing power, and completely solves the pain points of inaccurate identification of channel status and data lag in traditional technology.

[0061] Technical effects: Significantly improves the accuracy of determining the availability of user channels, fundamentally avoids invalid online pushes and channel mismatch issues, and provides a reliable data foundation for subsequent intelligent optimization and push.

[0062] Innovation Point Two: Hierarchical Adaptive Multi-Channel Intelligent Optimization Push Mechanism Existing push technologies lack a clear channel optimization logic, often employing parallel push across multiple channels or a fixed single channel, leading to severe SMS abuse and high costs. This invention proposes a unique layered adaptive push strategy that prioritizes apps, followed by mini-programs, supplemented by official accounts, and with SMS as a fallback. It dynamically selects the optimal zero-cost online channel based on the user's real-time channel status, triggering SMS as a fallback only when no online reach is possible.

[0063] The core of this innovation lies in: taking "cost reduction and efficiency improvement" as the core, maximizing the reuse of free online digital channels, strictly limiting the use scenarios of SMS channels, and combining message type-differentiated push rules to balance cost control of ordinary marketing messages with the reliability of emergency risk control messages.

[0064] Technical benefits: Significantly reduces redundant SMS messages, thereby reducing the operational costs of bank message push services; avoids the industry pain point of marketing SMS messages being blocked, and effectively improves the overall message delivery rate and reach quality.

[0065] Innovation Point 3: Closed-Loop Data Iteration and Precise Last-Hand Adaptation Mechanism Existing technologies lack a data optimization loop, resulting in rigid push strategies that cannot adapt to changes in user behavior and iterative business scenarios. This invention constructs a complete data loop encompassing "data collection - tag modeling - intelligent push - effect feedback - rule iteration," which can dynamically optimize channel judgment thresholds, selection weights, and push rules based on historical push data. Simultaneously, it sets up strict SMS fallback verification logic to accurately differentiate between different scenarios such as "channel not activated, permission closed, channel invalid," and prevent blind SMS resending.

[0066] The core of this innovation lies in: enabling adaptive and dynamic upgrades of push strategies, rather than rigidly executing fixed rules; and precisely balancing the two core requirements of "cost control" and "message delivery," thus avoiding both wasted SMS messages and missing important messages.

[0067] Technical effects: Continuously improves the accuracy and adaptability of push notifications, reduces invalid push notifications and harassment to users, optimizes customer experience, and ensures the integrity and reliability of message delivery across all scenarios.

[0068] The following is an embodiment of the present invention: intelligent preferred push scenario for bank marketing messages.

[0069] (1) The application background is as follows: Banks routinely send bulk marketing messages to existing users, including those promoting wealth management products, loan activities, and benefits. These messages are non-urgent and are traditionally sent via SMS, which suffers from high costs, high interception rates, extremely low delivery rates, and numerous user complaints. Meanwhile, some users actively use bank apps and WeChat mini-programs, which can be reached through free online channels, eliminating the need for wasted SMS resources. This embodiment employs the method of the present invention to achieve intelligent channel optimization for bulk user marketing message delivery.

[0070] (2) The specific implementation steps are as follows: Step S101: Collect user data across multiple channels.

[0071] The system collects data from three major online channels from 10,000 existing users to be pushed to this time, including each user's APP login frequency, message on / off status, mini-program access records, official account follow status, permission status of each channel, and active behavior data in the past 30 days. The system then cleans the data and maps it to a unified user identifier to form a complete raw dataset of all user channels.

[0072] Step S102: T+1 Offline channel statistics and tag generation.

[0073] The system retrieves the offline statistical model from the previous day and performs quantitative scoring and tag modeling on channel data for 10,000 users: 1. Set the weight parameters: Activity weight 0.3, Permission weight 0.4, Timeliness weight 0.2, Authorization validity weight 0.1; 2. Quantitatively calculate the channel effectiveness scores for each user's APP, mini-program, and official account; 3. Generate channel status labels: valid and available, conditionally available, invalid and unavailable, completely invalid; Final statistics: 7,200 users have valid online channels, while 2,800 users have no available online channels or all permissions are closed.

[0074] Step S103: Push task parsing and rule matching.

[0075] The system receives marketing message push tasks for financial products, parses the message type as general marketing and non-urgent push, and matches the core rule: strictly implement online channel priority selection, and do not trigger SMS as a backup unless necessary.

[0076] Step S104: Optimization and verification of tiered online channels.

[0077] The system verified the validity of each of the 7,200 valid online channel users: 1. 4,500 users are active on the APP channel, marketing messages are enabled, and the APP channel is the preferred push channel; 2.1800 users were not active on the app, but frequently accessed the WeChat mini-program and had their subscriptions enabled, indicating that the mini-program channel was the preferred push channel; 3,900 users only follow the official account and have normal message receiving permissions, so the official account channel is the preferred push channel.

[0078] The system automatically filters out all invalid channels with closed permissions or that have been silent for a long time, and eliminates online channel mismatch push notifications.

[0079] Step S105: Intelligent distribution through online channels.

[0080] The system pushed 7,200 marketing messages through corresponding preferred channels, including 4,500 from the APP, 1,800 from the mini-program, and 900 from the official account. There was no SMS consumption throughout the process, and the delivery and retrieval status of each channel were recorded after the push was completed.

[0081] Step S106: Targeted SMS backup.

[0082] The push is targeted at the remaining 2,800 users who have no available online channels or whose online channels have all their messaging switches turned off. The system determines that it cannot reach them through online channels, so it triggers a backup SMS push to ensure full coverage of marketing messages while avoiding the waste of invalid SMS messages.

[0083] Step S107: Data feedback and rule iteration.

[0084] After the push notification task was completed, the system summarized all push data: the online channel delivery rate was 96.2%, far exceeding the 75% delivery rate of traditional SMS; the number of SMS messages sent decreased by 72%, significantly reducing operating costs. Based on this push data, the system fine-tuned the channel activity judgment threshold and optimized the preferred weight of marketing message channels.

[0085] (3) The effect achieved in this embodiment is: Cost optimization: Compared with traditional full-volume SMS push, the number of SMS messages sent in this task was reduced by 72%, significantly reducing the cost of bank marketing SMS messages; Improved delivery: Leveraging the advantages of online channels with no blocking, the overall message delivery rate has increased from the traditional 75% to 96.2%, solving the problem of marketing SMS blocking failure; Experience optimization: Eliminate sending redundant text messages to active online users, reduce user harassment complaints, and improve customer service experience; Precise Adaptation: Based on users' actual usage habits, the system intelligently matches push channels, eliminating invalid pushes and channel mismatch issues, thus significantly improving push accuracy.

[0086] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this invention should be included within the protection scope of this invention.

[0087] It should be understood that, in order to simplify the present invention and help those skilled in the art understand its various aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes described in a single embodiment or with reference to a single figure. However, the present invention should not be construed as implying that all features included in the exemplary embodiments are essential technical features of the claims of the present invention.

[0088] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0089] It should be understood that the modules, units, components, etc., included in the device of one embodiment of the present invention can be adaptively changed to be placed in a device different from that embodiment. Different modules, units, or components included in the device of the embodiment can be combined into a single module, unit, or component, or they can be divided into multiple sub-modules, sub-units, or sub-components.

[0090] The modules, units, or components in the embodiments of the present invention can be implemented in hardware, in software running on one or more processors, or in a combination thereof. Those skilled in the art should understand that... In practice, microprocessors or digital signal processors (DSPs) can be used to implement embodiments of the invention. The invention can also be implemented on computer program products or computer-readable media for performing some or all of the methods described herein.

Claims

1. A multi-channel adaptive optimization method for intelligent bank message sending, characterized by: Includes the following steps: S1. Collect user data from all channels and perform standardized data processing to form the original user channel dataset through unique user identifiers; S2. Based on the user's original channel dataset, the T+1 offline batch statistical model is used to process the full amount of user channel data collected the previous day on the same day, generate standardized channel status labels, and build a global user channel status label library. S3: Receive push notification tasks, parse the core parameters of the tasks, and match the corresponding channel selection rules; S4. Retrieve the global user channel status tag library, and perform validity checks on each channel one by one according to the priority set by the channel selection rules, and select the best available push channel; S5. Use the best available channel to push messages and record push logs and delivery status; if all online channels are invalid or all channel message switches are turned off, automatically trigger SMS backup push.

2. The intelligent bank message sending method with multi-channel adaptive optimization according to claim 1, characterized in that: User data for each channel includes behavioral data, permission data, and status data.

3. The intelligent bank message sending method with multi-channel adaptive optimization according to claim 1, characterized in that: In S2, standardized channel status labels include: validity labels, activity labels, and permission status labels for each user and each channel.

4. The intelligent bank message sending method with multi-channel adaptive optimization according to claim 1, characterized in that: in S2, the application process of the T+1 offline batch statistical model is as follows: S21. Set activity thresholds, validity thresholds, and permission compliance thresholds for each type of channel; S22. For each user and each channel, calculate the channel effectiveness score separately: ； in: The channel activity score is the effectiveness score. Score the effectiveness of channel authorization. Score for message switch permissions. The score is based on the timeliness of recent use. , , , Preset weight parameters; S23. Based on the channel effectiveness score and the preset label classification rules, assign labels to the status of each channel; S24. Update the standardized channel status tags of the previous day on the same day to form a global user channel status tag library.

5. The intelligent bank message sending method with multi-channel adaptive optimization according to claim 4, characterized in that: In S23, the tags include: valid and available, conditionally available, invalid and unavailable, and fully expired.

6. The intelligent bank message sending method with multi-channel adaptive optimization according to claim 1, Its characteristics are: S3 includes: S31. Analyze the core parameters of the message push task; S32. Match the preset differentiated channel selection rules according to the core parameters; the channel selection rules include the channel priority weights configured for different message types.

7. The intelligent bank message sending method with multi-channel adaptive optimization according to claim 6, characterized in that: S31. The core parameters of the task include: target user list, message type, push priority, whether to force delivery, and message validity period; message types include: marketing, transaction reminder, risk control warning, and billing notification.

8. The intelligent bank message sending method with multi-channel adaptive optimization according to claim 1, characterized in that: After S5, S6 is executed: summarizing daily push results, channel delivery data, and user channel behavior change data, and iteratively optimizing channel selection rules and selection weights.

9. A multi-channel adaptive optimization intelligent bank message sending system, characterized by: include: Multi-source data acquisition module: Collects user data from all channels and performs standardized data processing to form the original user channel dataset through unique user identifiers; T+1 Offline Statistical Modeling Module: Based on the user's original channel dataset, the T+1 offline batch statistical model processes all user channel data collected the previous day on the same day, generates standardized channel status labels, and builds a global user channel status label library. Push task parsing module: Receives push message tasks, parses the core parameters of the tasks, and matches them with the corresponding channel selection rules; Multi-channel optimization and verification module: Retrieves the global user channel status tag library, performs validity verification on each channel one by one according to the priority set by the channel optimization rules, and selects the best available push channel; Online message distribution module: Pushes messages using the best available channels and records push logs and delivery status; SMS backup trigger module: used to automatically trigger SMS backup push when all online channels are in an invalid state or all channel message switches are turned off.

10. A multi-channel adaptive optimization intelligent bank message sending device, characterized in that: include: A processor and a memory, the memory storing a computer program executable by the processor, wherein the processor, when executing the computer program, implements the method of any one of claims 1-8.

Images