Marketing strategy generation method and system based on big data

Abstract

The invention relates to the technical field of big data analysis and marketing, and discloses a marketing strategy generation method and system based on big data, and the method comprises the steps: building a behavior reference through obtaining the historical interaction data of a consumer group, and obtaining the interaction information of an individual consumer for an innovative product in real time; and judging whether the interaction information exceeds the conventional performance, generating a potential interest indication, and adjusting the marketing content or the delivery mode according to the potential interest indication. The method effectively solves the problems that in the prior art, a marketing system is difficult to comprehensively capture the evolution trend of consumer behavior characteristics, and particularly, the identification of consumer groups which still belong to a traditional type on the surface but actually have potential interests in innovative products is very slow.

Description

Technical Field

[0001] This invention relates to the fields of big data analysis and marketing technology, and in particular to a method and system for generating marketing strategies based on big data. Background Technology

[0002] In today's rapidly changing market environment, companies often rely on marketing strategy generation systems when providing brand management services. These systems typically develop marketing plans by analyzing large amounts of consumer data.

[0003] However, existing systems face significant challenges in processing the evolution of consumer interests and behavioral patterns. They primarily focus on processing structured data such as past transaction records and browsing preferences to build user profiles and generate strategy rules. This approach works when the market is relatively stable, but it often falls short when dealing with the unstructured, real-time data from emerging channels like social media, and the subtle shifts in consumer interests reflected within it. This makes it difficult for systems to comprehensively capture the evolving trends in consumer behavior, particularly when identifying consumer groups who appear "traditional" but actually have a latent interest in innovative products.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] This invention provides a marketing strategy generation method and system based on big data, aiming to solve the problem that existing marketing strategy generation systems are unable to fully capture the evolution trend of consumer behavior characteristics when dealing with the evolution of consumer interests and behavior patterns, especially when identifying consumer groups that appear to be "traditional" but have actually developed a potential interest in innovative products.

[0006] The technical solution of this application is as follows: Firstly, this application discloses a method for generating marketing strategies based on big data, including: Acquire historical interaction data of consumer groups to establish behavioral references that characterize the regular interaction performance of consumer groups with marketing content; Real-time acquisition of individual consumer interaction information regarding innovative products; Based on behavioral references, determine whether the interactive information exceeds the norm for interactive behavior, and obtain the judgment result; Based on the degree of unusual interaction behavior in the judgment results, indicators of potential interest in innovative products for individual consumers are generated. Adjust the delivery of marketing content or messages to individual consumers based on potential interest indicators.

[0007] Through this technical solution, this application can effectively identify the potential interest of individual consumers in innovative products, even if their surface behavior is still routine, thereby overcoming the shortcomings of existing systems in capturing subtle changes in consumer interest and achieving more precise adjustments to marketing strategies.

[0008] Furthermore, in the above method, based on behavioral references, it is determined whether the interactive information exceeds the norm for interactive behavior, and the determination result includes: Define a set of interest states, which contains multiple interest states. Each interest state is associated with a set of behavioral patterns, which is used to distinguish between high-intent interactions and low-intent interactions. Construct an interest evolution trajectory graph, which contains multiple nodes and multiple edges. Each node represents an interest state, and each edge represents a migration path between interest states. Each migration path includes the behavioral conditions required for migration. Acquire behavioral sequences of individual consumers and innovative products, including user identifiers, event types, timestamps, interaction objects, and quantified behavioral metrics; Based on behavioral sequences and interest evolution trajectory maps, we analyze the migration paths of individual consumers in the interest evolution trajectory maps, and determine the current interest state of individual consumers based on the migration paths. When a behavioral sequence meets the condition of migrating from a low-intent state to a high-intent state, the individual consumer's interaction with the innovative product is judged as exceeding the normal interaction behavior of the consumer group, and a judgment result is obtained.

[0009] Through this technical solution, this application can accurately identify changes in consumer interest states, especially the migration from low intent to high intent, by constructing an interest evolution trajectory map and analyzing the behavioral sequences of individual consumers. This allows for a more detailed capture of the budding potential interests and improves the accuracy of the judgment.

[0010] Based on this, in the above method, according to the degree of unusual interaction performance in the judgment results, an indication of individual consumers' potential interest in innovative products is generated, including: Define the various interactive behaviors of individual consumers related to innovative products, as well as the causal chain between these interactive behaviors and the final high-value conversion. This causal chain includes the position and role of the interactive behaviors in the conversion path. Identify the sources of low-intention, high-noise interactions in marketing campaigns and quantify the contribution of each source to the overall noise level. Each interactive behavior is assigned an intention gain factor and a noise suppression factor; Adjust the intention gain factor and noise suppression factor to reflect the strength of the causal relationship between interactive behavior and final conversion; Based on the intention gain factor and noise suppression factor, the historical interaction data of the consumer group are weighted to establish a reference reflecting the consumer group's behavior towards innovative products; Based on the interaction information between individual consumers and innovative products, we can obtain individual consumers' cognitive deviations from innovative products; Based on the magnitude and frequency of cognitive deviations and their matching degree with the causal chain of high-value conversion, indicators of potential interest in innovative products for individual consumers are generated.

[0011] Through this technical solution, this application introduces an intention gain factor and a noise suppression factor, and combines them with the analysis of cognitive deviation, to more accurately quantify the potential interests of individual consumers, effectively distinguish between true intentions and noise, and thus generate more instructive indicators of potential interests.

[0012] In some preferred embodiments, the method described above involves adjusting the delivery of marketing content or information to individual consumers based on potential interest indicators, including: Establish adaptation rules between content formats and channels. These rules define the matching relationship between marketing content formats and marketing channels, and assess the expected effect of this matching relationship on the consumer group and the risk of content fatigue. Based on potential interest indicators, adaptation rules, and content fatigue risks, personalized marketing content is dynamically reorganized and generated. Combined with real-time feedback data, cross-channel marketing information delivery strategies are dynamically optimized to adjust the delivery method of marketing content or marketing information for individual consumers.

[0013] Through this technical solution, this application can dynamically generate personalized marketing content and optimize delivery strategies based on potential interest indicators and adaptation rules, effectively reducing the risk of content fatigue and thus improving the accuracy and effectiveness of marketing.

[0014] Furthermore, in the above method, personalized marketing content is dynamically reorganized and generated based on potential interest indicators, adaptation rules, and content fatigue risks, including: Maintain a content element library containing text snippets, visual materials, audio snippets, and interactive components, and associate each content element with one or more semantic tags and brand value tags; Maintain a content template library containing templates for various content formats. Each template has a pre-defined logic for combining content elements and a brand narrative structure. When a potential interest indication is received, one or more content templates are selected from the content template library based on the potential interest indication and the adaptation rules. Based on the pre-set combination logic and brand narrative structure in the content template, the content elements are arranged and reorganized to ensure the overall narrative coherence of the generated content and the consistency with the brand's core values.

[0015] Through this technical solution, this application achieves automated and personalized generation of marketing content by leveraging the synergistic effect of the content element library and the content template library, ensuring that the content satisfies individual interests while maintaining the consistency and coherence of the brand narrative.

[0016] Building upon the above, this application further proposes that, in the aforementioned method, personalized marketing content is dynamically reorganized and generated based on potential interest indicators, adaptation rules, and content fatigue risks. Combined with real-time feedback data, cross-channel marketing information delivery strategies are dynamically optimized to adjust the delivery method of marketing content or information targeting individual consumers, including: Establish a unified data access layer to receive and standardize real-time feedback data from different marketing channels. This real-time feedback data includes interactive behavior information, conversion event information, and user behavior path information on each channel. Configure data mapping rules to map standardized data to a unified data structure; Build a cross-channel data association engine to integrate standardized data from different channels in a unified data structure based on user identity and behavior timestamps, forming a complete view of individual consumer behavior across all channels; Generate attribution reports on cross-channel marketing performance based on a complete behavioral view; Based on attribution reports, dynamically adjust the way marketing content or marketing messages are delivered.

[0017] Through this technical solution, this application achieves the integration and analysis of real-time feedback data from multiple channels by constructing a unified data access layer and a cross-channel data association engine, thereby generating comprehensive attribution reports and providing a solid data foundation for dynamically optimizing marketing campaign strategies.

[0018] As a technological improvement, the above method generates an attribution report on cross-channel marketing effectiveness based on a complete behavioral view, including: Define a set of touchpoint types in a conversion path. This set of touchpoint types includes direct conversion touchpoints and auxiliary conversion touchpoints. The auxiliary conversion touchpoints do not directly generate conversions but promote conversions. Configure an initial contribution weight for each set of touchpoint types, which represents the expected role of the touchpoint type in the conversion path; Real-time monitoring of individual consumer behavior sequences across channels, and identification of touchpoint types within those sequences; Based on the touchpoint type and initial contribution weight in the behavior sequence, calculate the initial contribution value of each touchpoint to the final conversion; Establish a feedback mechanism to adjust the contribution weight of auxiliary conversion touchpoints based on actual conversion results and initial contribution values; Based on the adjusted contribution weights, the contribution of each touchpoint to the final conversion is recalculated, and an attribution report on cross-channel marketing performance is generated.

[0019] This technical solution introduces a feedback mechanism to dynamically adjust the contribution weight of auxiliary conversion touchpoints, enabling the attribution report to more accurately reflect the actual contribution of each touchpoint to the final conversion and improving the accuracy of attribution.

[0020] To improve the solution, in the above method, the contribution value of each touchpoint to the final conversion is recalculated based on the adjusted contribution weights, including: Identify the sequence relationship of all touchpoints in the conversion path, including the order in which the touchpoints appear and the interval between adjacent touchpoints; Evaluate the strength of the influence of each contact on subsequent contacts, which characterizes the promoting or reinforcing effect of a contact on subsequent contacts; Based on the sequence relationship and the influence intensity, a behavioral path dependency is constructed, which represents the nonlinear interaction between each touchpoint in the conversion path; Based on the adjusted contribution weights and behavioral path dependencies, the contribution value of each touchpoint to the final conversion is recalculated.

[0021] Through this technical solution, this application constructs behavioral path dependency by identifying the sequence relationship and influence intensity of the touch points, which enables a deeper understanding of the nonlinear interaction between each touch point, thereby making the calculation of contribution value more refined and accurate.

[0022] To enhance functionality, the above method evaluates the strength of each contact's influence on subsequent contacts. This strength characterizes the promoting or reinforcing effect of a contact on subsequent contacts, including: Real-time acquisition and analysis of individual consumers’ behavioral pattern changes before and after touchpoints in the conversion path, including interaction frequency, interaction depth and interaction content category; Based on changes in behavioral patterns, identify whether individual consumers' interest in innovative products has changed, including a shift from a low interest state to a high interest state. When interest states change, the intensity of the influence of the touchpoint on subsequent touchpoints is dynamically adjusted according to the magnitude and speed of the change, so as to reflect the modulation of the touchpoint effect by the evolution of consumer interest.

[0023] Through this technical solution, this application dynamically adjusts the intensity of touchpoint influence by analyzing changes in behavioral patterns and shifts in interest states in real time, enabling the attribution model to adapt to the dynamic evolution of consumer interests and further improving the accuracy and real-time nature of attribution.

[0024] Secondly, this application also discloses a marketing strategy generation system based on big data, including: Establish a platform to acquire historical interaction data of consumer groups in order to build behavioral references that characterize the regular interaction performance of consumer groups with marketing content; and acquire real-time interaction information of individual consumers with innovative products. The judgment end is used to determine whether the interactive information exceeds the normal interactive performance based on behavioral references, and obtain the judgment result; based on the degree of exceeding the normal interactive performance in the judgment result, an indication of the individual consumer's potential interest in the innovative product is generated. The adjustment end is used to adjust the delivery method of marketing content or marketing messages to individual consumers based on potential interest indicators.

[0025] This application provides a system-level solution through this technical solution. Through modular design, it enables accurate identification of potential consumer interests and dynamic adjustment of marketing strategies, providing enterprises with efficient marketing tools. Beneficial effects

[0026] This application discloses a marketing strategy generation method and system based on big data. It establishes behavioral references by acquiring historical interaction data of consumer groups and obtains real-time interaction information of individual consumers with innovative products. This allows for the determination of whether the interaction information exceeds conventional behavior, generating potential interest indicators, and adjusting marketing content or delivery methods accordingly. This method effectively solves the problem in existing technologies where marketing systems struggle to comprehensively capture the evolution of consumer behavior characteristics, particularly in identifying consumers who appear "traditional" but actually have a potential interest in innovative products. By introducing behavioral references and potential interest indicators, this application overcomes the limitations of traditional transaction data and demographic information, deeply mining unstructured, real-time data from emerging channels such as social media to identify subtle shifts in consumer interests. This enables businesses to identify consumers with potential interest in innovative products earlier and more accurately, avoiding missed opportunities due to misjudgment of market demand, and generating more targeted and attractive marketing content, significantly improving the conversion rate and ROI of marketing activities. Attached Figure Description

[0027] Figure 1 This is a flowchart of a marketing strategy generation method based on big data provided in an embodiment of the present invention; Figure 2This is a flowchart of a method for determining whether interactive information exceeds conventional interactive behavior based on behavioral references, provided by an embodiment of the present invention. Figure 3 This is a flowchart of a method for generating indicators of an individual consumer's potential interest in innovative products, provided by an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of a marketing strategy generation system based on big data provided in an embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Reference Figure 1 , Figure 1 This is a flowchart of a marketing strategy generation method based on big data provided in an embodiment of the present invention, including: S11, Obtain historical interaction data of the consumer group to establish a behavioral reference characterizing the consumer group's regular interaction performance with marketing content; S12, real-time acquisition of individual consumer interaction information on innovative products; S13, Based on the behavior reference, determine whether the interaction information exceeds the normal interaction performance, and obtain the judgment result; S14, Based on the degree to which the judgment result exceeds the normal interactive performance, generate an indication of the individual consumer's potential interest in the innovative product; S15, Based on the potential interest indication, adjust the delivery method of marketing content or marketing information targeting the individual consumer.

[0030] This application aims to address the shortcomings of existing marketing systems in capturing the evolution of consumer interests by introducing behavioral references to the regular interactive performance of consumer groups, acquiring individual consumer interaction information in real time, judging whether the interaction information exceeds the regular performance, generating potential interest indicators, and dynamically adjusting marketing content or delivery methods. This will enable more accurate identification of potential customers and optimization of marketing effectiveness.

[0031] To better understand the technical solution proposed in this application, some key terms involved will be explained first.

[0032] "Historical interaction data of a consumer group" refers to the data collection generated over a period of time by a large number of consumers interacting with various marketing content (such as advertisements, product pages, social media posts, etc.). This data can include records of behaviors such as clicks, browsing time, sharing, comments, and purchases, used to depict the overall behavioral patterns and preferences of the consumer group.

[0033] "Behavioral reference" is a benchmark model built on historical interaction data of a consumer group. It quantifies the level and pattern of interaction that a consumer group typically exhibits when faced with specific marketing content. This reference is used to subsequently determine whether an individual consumer's behavior is abnormal.

[0034] "Individual consumer interaction information with innovative products" refers to real-time behavioral data generated by individual consumers when they encounter new products or services, such as clicking on new product advertisements, browsing related content, discussing, liking, or sharing on social media. This information is key to determining the potential interests of individual consumers.

[0035] "Regular interaction performance" refers to the average or typical level of interaction that a consumer group exhibits with marketing content under normal circumstances.

[0036] A "potential interest indicator" is a quantitative metric generated based on the degree of deviation between an individual consumer's interaction information and their usual interaction performance. It reflects the individual consumer's potential, yet-to-be-fully-manifested, interest in innovative products. This indicator can be a score, a label, or an interest level.

[0037] "Marketing content" refers to various forms of information carriers used by businesses to promote their products or services, such as text advertisements, images, videos, and interactive pages.

[0038] "Marketing message delivery methods" refers to the channels, frequency, timing, and form used to deliver marketing content to target consumers, such as email, social media advertising, SMS, in-app push notifications, etc.

[0039] This application provides a big data-based marketing strategy generation method, the core of which lies in capturing consumers' potential interest in innovative products through refined data analysis, and dynamically adjusting marketing strategies accordingly.

[0040] There are several ways to acquire historical interaction data from consumer groups to establish behavioral benchmarks representing their typical interactions with marketing content. For example, one could collect browsing history, click behavior, search keywords, and purchase records from all consumers across the company's official website, mobile applications, and partner e-commerce platforms over the past year. After cleaning and aggregation, this data can be used to build a statistical model that can predict the average click-through rate, conversion rate, or engagement duration for a specific marketing campaign. Another approach is to leverage publicly available data from social media platforms, such as the volume of discussions on specific topics, the number of times related content is shared, and the frequency of user mentions of the brand or product, to establish behavioral benchmarks. This unstructured data can be analyzed using natural language processing techniques to extract the overall sentiment and level of attention of consumer groups towards a particular type of product or marketing content.

[0041] To acquire real-time information on individual consumer interactions with innovative products, a real-time data acquisition system can be deployed. For example, when a company releases an advertisement for a new smart home product, the system immediately tracks each individual consumer's actions, including clicking on the ad, browsing time, dwell time on the product page, adding the product to their cart, and sharing it with friends. This data is captured and transmitted to the processing module in real-time as streaming data. Alternatively, third-party data platforms or social media monitoring tools can be used to monitor individual consumers' mentions, comments, likes, and shares of innovative products on social media in real time. This interactive information, even if not direct purchasing behavior, can reflect consumers' level of interest in innovative products.

[0042] When determining whether interactive information exceeds typical interactive performance based on behavioral references, the real-time interactive information of an individual consumer can be compared with a pre-established behavioral reference. For example, if the behavioral reference shows that the average click-through rate (CTR) of a consumer group for similar new product advertisements is 5% in the past, and an individual consumer clicks on a new product advertisement multiple times within a short period and browses the product details page in depth, exhibiting a behavior pattern significantly higher than the 5% average, then their interactive information can be judged to exceed typical interactive performance. Alternatively, statistical methods can be used, such as setting a confidence interval. If an individual consumer's interactive behavior indicators (such as interaction frequency and interaction depth) exceed the confidence interval defined by the behavioral reference, then they are considered to have exceeded typical interactive performance.

[0043] When generating indicators of potential consumer interest in innovative products based on the degree to which interactive behavior deviates from the norm, potential interest can be quantified by the extent to which the interaction exceeds the norm. For example, the more frequent and in-depth an individual consumer's interaction with an innovative product, the higher the degree of deviation from the norm, and thus the stronger the generated indicator of potential interest. This indicator can be a score from 0 to 100, with higher scores indicating stronger potential interest. Alternatively, potential interest indicators can be generated based on the type and quality of the interactive behavior. For example, simply clicking on an advertisement might indicate low interest, while actively searching for related information, participating in product discussions, or adding the product to favorites indicates higher interest. Different types of behavior can be assigned different weights to comprehensively assess potential interest.

[0044] When adjusting the delivery of marketing content or messages to individual consumers based on potential interest indicators, subsequent marketing strategies can be dynamically adjusted according to the strength and type of the potential interest indicators. For example, if an individual consumer is identified as having high potential interest, the system can immediately push more attractive personalized marketing content to them, such as offering exclusive discounts, invitations to participate in product beta testing, or videos showcasing more product details. As another approach, the channels and frequency of marketing message delivery can be adjusted. For example, for consumers with high potential interest, the frequency of advertising on their frequently used social media platforms can be increased, or more detailed product information can be sent via email. For consumers with low potential interest, delivery may be temporarily reduced to avoid content fatigue.

[0045] The big data-based marketing strategy generation method proposed in this application works by constructing a dynamic and adaptive marketing closed loop. First, through in-depth analysis of historical interaction data of consumer groups, the system establishes a baseline behavioral reference, representing the "normal" response pattern of consumers to marketing content. This step is fundamental to understanding consumer behavior, providing context for subsequent anomaly behavior identification.

[0046] Subsequently, the system acquires real-time interaction information from individual consumers regarding innovative products. This includes various online behaviors such as clicks, browsing, sharing, and comments. This real-time data is crucial for capturing the evolution of consumer interests, as it reflects consumers' most immediate reactions and preferences.

[0047] The next key step is to determine, based on behavioral benchmarks, whether an individual consumer's interactions deviate from typical behavior. This judgment isn't simply a comparison of absolute values, but rather an identification of "abnormal" behaviors that might indicate emerging new interests by comparing individual behavior to a group benchmark. For example, a consumer who is normally uninterested in technology products might suddenly start frequently watching review videos of a particular smart home product; this deviation in behavioral pattern could be detected by the system.

[0048] Once interactions that deviate from the norm are identified, the system generates an indicator of a consumer's potential interest in the innovative product, based on the degree of deviation. This indicator is a quantitative assessment of the consumer's underlying intention, transforming vague interactive behavior into actionable marketing signals. The higher the degree of deviation, the stronger the potential interest indicator, indicating a higher level of consumer interest in the innovative product.

[0049] Finally, based on this potential interest indicator, the system dynamically adjusts the delivery of marketing content or information to that individual consumer. This means that marketing strategies are no longer static and one-size-fits-all, but are personalized based on each consumer's real-time changes in interest. For example, for consumers showing high potential interest, the system may push more in-depth product information, offer exclusive discounts, or reach them more frequently through their preferred channels. Conversely, for consumers with low potential interest, the system may adjust to a gentler approach to avoid over-marketing.

[0050] Through the close coordination of the above steps, the method of this application can effectively solve the pain points existing in the prior art. Traditional systems often struggle to capture the unstructured, real-time data exhibited by consumers on emerging channels such as social media, as well as the subtle shifts in consumer interests reflected therein. This makes it difficult for the system to comprehensively capture the evolutionary trends of consumer behavior characteristics, especially when identifying consumer groups that appear to be "traditional" but actually have a potential interest in innovative products.

[0051] The core innovation of this application lies in its ability to dynamically identify and quantify consumers' potential interest in innovative products, thereby enabling precise adjustments to marketing strategies. Unlike existing technologies that primarily rely on structured data such as past transaction records and browsing preferences to build user profiles, this application's method introduces behavioral references to the regular interactive performance of consumer groups and acquires individual consumer interaction information on innovative products in real time. This comparative analysis allows the system to capture consumer groups that, while superficially appearing "traditional," actually possess a potential interest in innovative products.

[0052] For example, in existing technologies, a consumer who previously only purchased traditional home appliances may not have their interest shifted even if they frequently like and share review videos of smart home products such as robot vacuums and air purifiers on social media, due to a lack of direct purchasing behavior. This results in marketing messages failing to reach these potential customers when companies promote innovative products, or even if they do reach them, the messages are pushed based on incorrect customer profiles, leading to poor marketing results.

[0053] This application identifies clicks that exceed typical consumer interaction patterns by determining whether they deviate from normal behavior, thus generating potential interest indicators based on the degree of deviation. For example, when a "traditional" consumer's interaction frequency and depth with smart home products significantly exceed their historical average or the typical behavior of their peer group, this application's method can immediately identify their potential interests and adjust marketing content and delivery methods accordingly. This allows businesses to push more attractive, personalized marketing content to these consumers, such as highlighting innovative features of smart kitchen devices like "AI intelligent ingredient recognition and personalized recipe recommendations" or "remote control for meal preparation on the way home from get off work," effectively stimulating their purchase desire and improving conversion rates.

[0054] Furthermore, the method presented in this application provides a clearer attribution chain, helping companies identify the root cause of problems. When the system identifies a potential shift in interests within a group, companies can adjust their product positioning and marketing strategies accordingly, avoiding missed opportunities due to misjudgments of actual market demand. This dynamic and adaptive marketing strategy generation method significantly improves the accuracy and efficiency of marketing, providing strong support for companies to gain an advantage in a highly competitive market.

[0055] In some of the embodiments described above in this application, when judging whether interactive information exceeds normal interactive performance based on behavioral references, there may be a problem of insufficient understanding of the evolution of individual consumer interests, resulting in the inability to accurately identify potential high-intent interactions. Traditional methods may only rely on simple thresholds or rules for judgment, making it difficult to capture the subtle migration process of consumer interests from low intent to high intent, thus affecting the accuracy of subsequent potential interest indications. To address this, this application further proposes a specific method for judging whether interactive information exceeds normal interactive performance based on behavioral references. This method defines interest states, constructs an interest evolution trajectory map, and analyzes the behavioral sequences of individual consumers to more precisely determine whether their interactive performance exceeds the norm.

[0056] Specifically, refer to Figure 2 , Figure 2 This is a flowchart of a method for determining whether interactive information exceeds normal interactive behavior based on behavioral references, provided by an embodiment of the present invention, including the following steps: S131, Define an interest state set, which contains multiple interest states, each interest state being associated with a set of behavioral patterns, which is used to distinguish between high-intent interactions and low-intent interactions. S132, Construct an interest evolution trajectory graph, which contains multiple nodes and multiple edges. The nodes represent the interest states, the edges represent the migration paths between the interest states, and the migration paths include the behavioral conditions required for migration. S133, Obtain the behavioral sequence of the individual consumer and the innovative product, the behavioral sequence including user identifier, event type, timestamp, interaction object and quantified behavioral indicators; S134, Based on the behavioral sequence and the interest evolution trajectory map, analyze the migration path of the individual consumer in the interest evolution trajectory map, and determine the current interest state of the individual consumer based on the migration path; S135, when the behavioral sequence meets the condition of migrating from a low-intent state to a high-intent state, the individual consumer's interaction information with the innovative product is judged to exceed the normal interaction performance of the consumer group, and the judgment result is obtained.

[0057] The set of interest states can be understood as a set of pre-defined, discrete stages that reflect a consumer's level of interest in an innovative product. For example, these interest states may include "initial browsing," "information gathering," "intent assessment," and "ready to buy." Each interest state is associated with a set of behavioral patterns designed to distinguish whether a consumer's interaction in the current state is high-intent or low-intent. Specifically, high-intent interactions might include actively searching, reading product details in depth, downloading a white paper, or adding the product to the shopping cart, while low-intent interactions might be limited to staying on a page, quickly browsing, or clicking on advertisements.

[0058] Furthermore, an interest evolution trajectory graph was constructed, consisting of multiple nodes and edges. The nodes represent the interest states defined above, while the edges represent possible migration paths between these interest states. Each migration path includes specific behavioral conditions required to complete the migration. For example, migrating from the "initial browsing" state to the "information gathering" state might require satisfying the behavioral condition of "staying on the product details page for more than X seconds and clicking to view user reviews." This graph visually demonstrates the dynamic process of consumer interest evolving from one stage to another.

[0059] Based on this, the system acquires the behavioral sequences of individual consumers and innovative products. These behavioral sequences are all records of a consumer's interactions with the innovative product over a period of time, including detailed information such as user identifiers, event types (e.g., clicks, views, downloads), timestamps of interactions, specific interaction objects (e.g., specific product pages, advertising creatives), and quantified behavioral metrics (e.g., dwell time, number of clicks). This data forms the basis for analyzing the evolution of consumer interests.

[0060] Subsequently, based on the acquired behavioral sequences and the constructed interest evolution trajectory map, the migration path of individual consumers within the map is analyzed. By matching the consumer's actual behavioral sequence with the migration path and behavioral conditions defined in the map, the current interest state of the individual consumer can be determined. For example, if the consumer's behavioral sequence satisfies the migration condition from "initial browsing" to "information gathering," then their current interest state is determined to be "information gathering."

[0061] Ultimately, when the behavioral sequence satisfies the condition of migrating from a low-intent state to a high-intent state, the individual consumer's interaction with the innovative product is judged to exceed the typical interaction behavior of the consumer group, thus leading to the judgment result. This means that when consumers exhibit clear signs of shifting from a lower or higher interest or intent state, their interaction behavior is considered to have significant potential value and deserves further attention and response.

[0062] This application's solution elevates the understanding of consumer interest in innovative products from a single interactive behavior judgment to a multi-stage, dynamically evolving state analysis by introducing an interest state set and an interest evolution trajectory map. Through this technical solution, the application can significantly improve the accuracy and depth of identifying the potential interests of individual consumers. Compared to methods based solely on simple rules or thresholds, this solution, by introducing an interest state set and an interest evolution trajectory map, can more precisely capture the dynamic evolution of consumer interest from low intent to high intent. This allows the marketing system to identify consumers whose individual behaviors may not be prominent, but whose overall behavioral sequence shows a clear trend of increasing interest, thus avoiding overlooking potential high-value users. Furthermore, by analyzing migration paths, the direction of consumer interest can be predicted earlier and more accurately, providing a more precise basis for subsequent adjustments to marketing content and optimization of delivery methods, effectively improving the efficiency and conversion rate of marketing activities and reducing the waste of marketing resources.

[0063] In some embodiments described above, potential interest indicators are generated by determining whether interactive information exceeds normal interactive behavior. However, in practical applications, judging potential interest solely based on the degree to which interactive behavior exceeds normal behavior may have certain limitations. For example, some interactive behaviors may not truly reflect a consumer's deep interests, but are merely accidental, low-intention, or noise interactions influenced by external factors. Failure to distinguish between these may lead to a decrease in the accuracy of potential interest indicators, thereby affecting the effectiveness of subsequent marketing strategies. Therefore, this application further proposes a method for generating potential interest indicators for individual consumers regarding innovative products, aiming to more accurately identify and quantify consumers' true interests, thereby improving the targeting and conversion efficiency of marketing strategies.

[0064] Specifically, refer to Figure 3 , Figure 3 This is a flowchart of a method for generating indicators of an individual consumer's potential interest in innovative products, provided by an embodiment of the present invention, including: S141 defines various interactive behaviors of individual consumers related to innovative products, as well as the causal chain between interactive behaviors and the final high-value conversion. The causal chain includes the position and role of interactive behaviors in the conversion path. S142, Identify the sources of low-intention, high-noise interactions in marketing campaigns and quantify the contribution of each source of noise to the overall noise level; S143 assigns an intention gain factor and a noise suppression factor to each interactive behavior; S144, adjust the intention gain factor and noise suppression factor to reflect the strength of the causal relationship between interactive behavior and final conversion; S145, based on the intention gain factor and noise suppression factor, the historical interaction data of the consumer group is weighted to establish a reference reflecting the consumer group’s behavior towards innovative products; S146, Obtain individual consumers’ cognitive deviations of innovative products based on interaction information between individual consumers and innovative products; S147 generates indicators of potential consumer interest in innovative products based on the magnitude and frequency of cognitive deviations and their matching degree with the causal chain of high-value conversion.

[0065] The definition of various interactive behaviors between individual consumers and innovative products refers to the classification and description of various behaviors that consumers may exhibit during their interaction with innovative products. These interactive behaviors may include, but are not limited to, browsing product pages, clicking on advertisements, watching product videos, sharing product information, adding items to the shopping cart, and saving items to favorites. Simultaneously, it is necessary to clarify the causal chain between these interactive behaviors and the final high-value conversion. High-value conversion typically refers to behaviors that provide significant value to the merchant, such as completing a purchase, subscribing to a service, or registering as a member. The construction of the causal chain aims to reveal the sequence, mutual influence, and respective roles and functions of different interactive behaviors in the consumer conversion path. For example, watching a product video may enhance a consumer's awareness of the product, thereby guiding them to add it to their shopping cart and ultimately complete the purchase.

[0066] Furthermore, identifying the sources of low-intent, high-noise interactions in marketing campaigns refers to distinguishing from massive amounts of consumer interaction data those interactions that do not truly reflect consumer interests or purchase intentions. Low-intent interactions may include accidental clicks and brief browsing, while high-noise interactions may stem from inappropriate ad placement, insufficient content appeal, or bot traffic. Identifying noise sources helps understand the reasons behind these ineffective or inefficient interactions. Quantifying the contribution of each noise source to the overall noise level can be achieved through statistical analysis, machine learning models, and other methods to assess the degree to which different noise sources interfere with data accuracy and the judgment of potential interests.

[0067] In a preferred implementation, each interactive behavior is assigned an intent gain factor and a noise suppression factor. The intent gain factor measures the extent to which the interactive behavior reflects the consumer's true interest or purchase intent; for example, adding a product to a shopping cart typically has a high intent gain factor. The noise suppression factor is used to reduce or eliminate the negative impact of noisy interactions on the judgment of potential interest; for example, a high noise suppression factor can be assigned to short-term browsing or accidental clicks. These factors can be initialized and adjusted based on historical data, expert experience, or through methods such as A / B testing.

[0068] In practical applications, adjusting the intent gain factor and noise suppression factor aims to more accurately reflect the strength of the causal relationship between interactive behavior and final conversion. For example, by analyzing a large number of successful conversion cases, it can be found that certain interactive behaviors (such as reading in-depth product reviews) have a stronger positive correlation with final purchase behavior; in this case, their intent gain factor can be increased. Conversely, if certain interactive behaviors (such as simply clicking on an ad without further browsing) have a weak correlation with conversion, their intent gain factor can be decreased or their noise suppression factor increased. This dynamic adjustment mechanism makes the judgment of potential interest more flexible and accurate.

[0069] Therefore, based on the adjusted intention gain factor and noise suppression factor, the historical interaction data of the consumer group is weighted to establish a behavioral reference reflecting the consumer group's attitude towards innovative products. This weighting process allows high-intention, low-noise interactions to have a greater weight in the behavioral reference, thus constructing a more representative and accurate benchmark for regular interaction performance. Once the interaction information between individual consumers and innovative products is obtained, the individual consumer's cognitive deviation from the innovative product can be determined based on this weighted behavioral reference. Cognitive deviation refers to the degree of difference between the individual consumer's actual interaction performance and the weighted group behavioral reference.

[0070] Ultimately, based on the magnitude and frequency of cognitive deviations and their alignment with the causal chain of high-value conversion, an indicator of individual consumers' potential interest in innovative products is generated. A greater magnitude of cognitive deviation generally indicates a stronger interest in the innovative product; a higher frequency of cognitive deviations may signify the persistence and stability of that interest. Furthermore, matching these deviations with predefined causal chains allows us to determine whether these deviations are on the critical path to high-value conversion. By comprehensively considering these three dimensions, a multi-dimensional and more insightful indicator of potential interest can be generated.

[0071] This application's solution, by introducing an intent gain factor and a noise suppression factor, and constructing a causal chain between interactive behavior and high-value conversion, can effectively distinguish the quality and intent of consumer interactive behavior. Traditional behavior-based judgments may only focus on the quantity of interactive behavior or whether it exceeds the norm, while this solution delves deeper into the essence of interactive behavior, identifying the true intent and potential noise behind it. By weighting historical interaction data and combining it with the magnitude and frequency of individual consumers' cognitive deviations and their matching degree with the conversion causal chain, this solution can more comprehensively and accurately assess individual consumers' potential interest in innovative products.

[0072] The aforementioned technical solutions significantly improve the accuracy and reliability of potential interest indicators, effectively avoiding misjudgments caused by low-intent or high-noise interactions. This allows subsequent adjustments to marketing content and optimization of delivery strategies to more accurately reach target consumers, improving the conversion efficiency and ROI of marketing campaigns. Furthermore, through refined analysis of interactive behavior, this solution helps businesses gain a deeper understanding of consumer behavior patterns and conversion paths, providing valuable insights for product optimization and market strategy development. This application further proposes adjusting the delivery method of marketing content or marketing information targeting the aforementioned individual consumers based on the aforementioned potential interest indications, including: Establish adaptation rules between content formats and channels. These adaptation rules define the matching relationship between the marketing content formats and marketing channels, and assess the expected effect of the matching relationship on the consumer group and the risk of content fatigue. Based on the aforementioned potential interest indicators, adaptation rules, and content fatigue risks, personalized marketing content is dynamically reorganized and generated. Combined with real-time feedback data, cross-channel marketing information delivery strategies are dynamically optimized to adjust the delivery methods of marketing content or information targeting the aforementioned individual consumers.

[0073] Specifically, "establishing adaptation rules between content formats and channels" refers to building a set of guiding principles or algorithmic models to determine which marketing channels (e.g., social media, search engines, email, SMS, offline advertising) should be used to deliver different types of marketing content (e.g., short videos, text and images, long articles, interactive H5 pages, etc.). These adaptation rules not only define the matching relationship between marketing content formats and marketing channels—for example, short video content is more suitable for short video platforms like Douyin and Kuaishou, while detailed product descriptions are more suitable for presentation on official websites or in emails—but also assess the expected effect of this matching relationship on specific consumer groups. For example, consumers of a certain age group may respond more positively to visually impactful short videos, while another group prefers in-depth text and image content. Furthermore, these adaptation rules also consider the risk of content fatigue, i.e., the potential boredom or aversion consumers may experience after receiving too much or repetitive marketing information. This risk can be mitigated by limiting the frequency of the same type of content being pushed out within a short period.

[0074] The phrase "dynamically reorganizing and generating personalized marketing content" can be understood as the system customizing marketing content in real-time or near real-time, based on individual consumers' potential interests, the aforementioned adaptation rules, and the risk of content fatigue. This may involve selecting text snippets, images, videos, audio, and other materials from a content element library and intelligently combining them according to preset templates or algorithmic logic to generate marketing content that highly matches individual consumers' interests and conforms to channel characteristics. For example, if a consumer shows strong interest in a specific feature of a product, the generated personalized content will highlight the advantages and usage scenarios of that feature.

[0075] In practical applications, "dynamically optimizing cross-channel marketing message delivery strategies by combining real-time feedback data" means that after marketing content is delivered, the system continuously collects and analyzes real-time feedback data from different marketing channels, such as click-through rate, interaction rate, conversion rate, dwell time, and sharing behavior. Based on this real-time feedback, the delivery strategy is dynamically adjusted and optimized. For example, if marketing content performs poorly on a certain channel, the system may automatically reduce the amount of content delivered on that channel or adjust the content format; if content performs well on a particular channel, the system may increase the delivery efforts on that channel. This optimization is cross-channel, meaning that the system comprehensively considers consumer behavior across all touchpoints to maximize overall marketing effectiveness.

[0076] This application's solution introduces adaptation rules between content format and channels, ensuring that the generation and delivery of marketing content are no longer arbitrary but based on a deep understanding of content characteristics, channel features, and consumer behavior patterns. These adaptation rules effectively guide the generation of personalized content, ensuring a high degree of consistency between content format and delivery channels, thereby improving content reach and appeal. Simultaneously, by assessing the risk of content fatigue, it avoids consumer aversion caused by excessive or inappropriate marketing, maintaining a positive user experience. Furthermore, by dynamically optimizing cross-channel delivery strategies using real-time feedback data, marketing campaigns can be rapidly iterated and adjusted based on actual results, ensuring efficient use of marketing resources and timely response to changes in consumer interests and market dynamics.

[0077] Through the aforementioned technical solution, this application significantly improves the accuracy of marketing content and the flexibility of delivery strategies. Compared to traditional methods that simply adjust based on potential interest indicators, this application effectively avoids content-channel mismatch and consumer fatigue by establishing adaptation rules and considering the risk of content fatigue, thereby improving the effective reach and conversion efficiency of marketing messages. Furthermore, dynamic optimization using real-time feedback data allows marketing strategies to continuously adapt to consumer behavior and market changes, ensuring that marketing activities remain at their optimal state, thus achieving a better return on marketing investment and a more positive consumer interaction experience.

[0078] The above-mentioned dynamic reorganization and generation of personalized marketing content based on potential interest indicators, adaptation rules, and content fatigue risk can be achieved in the following ways.

[0079] Maintain a content element library, which includes text snippets, visual materials, audio snippets, and interactive components, and associate each content element with one or more semantic tags and brand value tags; Maintain a content template library, which contains templates of various content formats. Each template has a preset combination logic of content elements and a brand narrative structure. When the potential interest indication is received, one or more content templates are selected from the content template library according to the potential interest indication and the adaptation rules; Based on the pre-set combination logic and brand narrative structure in the content template, the content elements are arranged and reorganized to ensure the overall narrative coherence of the generated content and the consistency with the brand's core values.

[0080] The content element library can be understood as a collection storing all the smallest reusable units that can be used to build marketing content. These units can be text descriptions, images, videos, audio clips, or even interactive components. Semantic tags and brand value tags are associated with each content element to provide a structured description of the content. For example, semantic tags can describe the content's theme, emotional tone, product characteristics, etc., while brand value tags can associate it with the brand's core philosophy, tone, or the goals of a specific marketing campaign.

[0081] The content template library contains a variety of preset content structures and layouts, such as templates for different channels (e.g., social media, email, app push notifications) and different marketing objectives (e.g., brand promotion, product promotion, user education). Each template predefines the combination logic of content elements and the brand narrative structure. This means that the template not only defines the layout of the content but also specifies the placement and quantity limits of different types of content elements (e.g., headlines, body text, images, calls to action), as well as the narrative order and brand expression methods that they should follow.

[0082] When the system receives indications of potential interests from individual consumers, it intelligently selects one or more content templates from its content template library that best match the current consumer's interests and the characteristics of the distribution channel, based on pre-established adaptation rules between content formats and channels. Subsequently, according to the preset combination logic and brand narrative structure within the selected templates, the system automatically arranges and reorganizes content elements with corresponding semantic and brand value tags from its content element library. This process aims to ensure that the final personalized marketing content not only accurately matches the interests of individual consumers but also maintains the coherence of the overall narrative and a high degree of consistency with the brand's core values.

[0083] This application's solution achieves personalized marketing content generation by constructing a content element library and a content template library, and dynamically selecting and recombining them based on potential interest indicators and adaptation rules. Specifically, the content element library provides rich atomic content with semantic and brand tags, while the content template library provides a structured content framework and combination logic. When the system identifies the potential interests of an individual consumer, it can intelligently select the most suitable template based on their interest preferences and the adaptation rules of the content delivery channel. Subsequently, through the template's preset combination logic and brand narrative structure, relevant content elements are extracted and arranged from the content element library, thereby efficiently and automatically generating highly personalized marketing content that aligns with the brand's tone. This mechanism avoids the inefficiency of manually creating large amounts of personalized content and ensures a balance between content diversity and consistency.

[0084] The aforementioned technical solutions significantly improve the efficiency and personalization of marketing content generation. Leveraging the synergy of content element and template libraries, the system can quickly and flexibly combine diverse marketing content based on real-time consumer interests and behavioral data, effectively avoiding the problems of content homogenization and user fatigue inherent in traditional marketing. Furthermore, pre-defined brand narrative structures and semantic tags ensure that even with high personalization, marketing content maintains brand image consistency and information delivery coherence, thereby enhancing the overall effectiveness of marketing campaigns and user experience.

[0085] In some embodiments of this application, the above-described method of dynamically reorganizing and generating personalized marketing content based on potential interest indicators, adaptation rules, and content fatigue risk, while capable of dynamically optimizing cross-channel marketing information delivery strategies by combining real-time feedback data, does not detail how to effectively integrate real-time feedback data from different marketing channels, or how to accurately attribute cross-channel performance based on this data, thereby achieving truly refined marketing strategy adjustments. In practical applications, differences in data formats, collection methods, and user identifiers across different marketing channels can lead to difficulties in data integration, making it challenging to form a complete view of individual consumer behavior across all channels, thus affecting the accuracy of marketing effectiveness evaluation and strategy optimization.

[0086] In response, this application further proposes a method for dynamically optimizing cross-channel marketing information delivery strategies by combining real-time feedback data, aiming to solve the aforementioned data integration and attribution challenges and achieve more precise marketing adjustments.

[0087] In some embodiments of this application, the step of dynamically reorganizing and generating personalized marketing content based on the potential interest indication, the adaptation rules, and the content fatigue risk, and dynamically optimizing cross-channel marketing information delivery strategies in conjunction with real-time feedback data to adjust the delivery method of marketing content or marketing information targeting the individual consumer, includes: Establish a unified data access layer to receive and standardize real-time feedback data from different marketing channels. The real-time feedback data includes interactive behavior information, conversion event information, and user behavior path information on each channel. Configure data mapping rules to map standardized data to a unified data structure; A cross-channel data association engine is built to integrate standardized data from different channels in the unified data structure based on user identity and behavior timestamps, forming a complete view of an individual consumer's behavior across all channels; Based on the complete behavioral view, generate an attribution report on cross-channel marketing effectiveness; Based on the attribution report, the delivery method of the marketing content or marketing information is dynamically adjusted.

[0088] Specifically, establishing a unified data access layer refers to building an interface or platform capable of collecting data from various marketing channels (such as social media platforms, email marketing systems, search engine advertising platforms, and push services within mobile applications). This access layer is responsible for the initial processing of the received raw data, including data cleaning, format conversion, and deduplication, to ensure data quality, consistency, and usability. Real-time feedback data can be understood as data generated immediately after consumers interact with marketing content or products during the marketing campaign. This includes interactive behavior information such as clicks, views, shares, and comments, as well as conversion event information such as changes in user status resulting from these behaviors (e.g., from potential customers to paying customers), and behavioral path information of users jumping and interacting between different channels.

[0089] Furthermore, configuring data mapping rules refers to defining a set of rules to unify standardized data from different sources and formats into a pre-defined data model. For example, user IDs from different channels can be mapped to a unified user identity identifier, and event names from different channels can be mapped to a unified event type, ensuring that all data remains logically consistent and facilitating subsequent integration and analysis. These rules can be predefined or dynamically adjusted according to actual needs.

[0090] Furthermore, building a cross-channel data association engine involves developing a core processing module that integrates standardized data from different channels, mapped to a unified data structure, using unified user identification (e.g., through device ID, login account, cookies, phone number, etc.) and behavioral timestamps. This integration allows tracking of all interactions by individual consumers across different touchpoints and channels throughout the marketing journey, creating a comprehensive and coherent view of their behavior across all channels. This view clearly shows the entire consumer journey from initial contact to final conversion, such as clicking on social media ads, browsing the website, consulting at a physical store, and ultimately making a purchase.

[0091] Therefore, generating an attribution report on cross-channel marketing effectiveness based on the complete behavioral view refers to evaluating the contribution of different marketing touchpoints or channels to the final conversion using integrated consumer behavior data and specific attribution models (e.g., first-interaction attribution, last-interaction attribution, linear attribution, time-decay attribution, or U-shaped attribution). The attribution report quantifies the role of each channel or touchpoint in driving conversion, helping marketers understand which channel combinations or interaction paths are most effective, thus providing data support for marketing budget allocation and strategy optimization.

[0092] Ultimately, dynamically adjusting the delivery method of marketing content or information based on the attribution report means optimizing marketing strategies in real-time or near real-time based on the channel performance and consumer behavior patterns revealed in the attribution report. For example, this could involve increasing budget allocations to high-contribution channels, adjusting marketing content or delivery frequency for inefficient channels, or selecting marketing content formats and timings that better align with the cross-channel behavior paths of specific consumer groups.

[0093] This application's solution effectively addresses the challenges of inconsistent data formats and integration across different marketing channels by establishing a unified data access layer and configuring data mapping rules, ensuring the standardization and consistency of real-time feedback data. Building upon this foundation, a cross-channel data association engine is constructed, enabling deep integration of standardized data from various channels to create a complete behavioral view of individual consumers across all channels. This comprehensive behavioral view is crucial for precise marketing strategy adjustments, as it provides a complete consumer interaction trajectory, allowing marketers to fully understand consumer behavior across different touchpoints and channels. Analyzing this complete behavioral view generates accurate cross-channel marketing performance attribution reports, quantifying the contribution of each marketing touchpoint or channel to the final conversion. The availability of accurate attribution reports allows marketers to dynamically adjust the delivery of marketing content or information based on accurate and comprehensive data insights, avoiding resource waste and strategy deviations caused by data silos and inaccurate attribution.

[0094] Through the aforementioned technical solution, this application effectively addresses the problems of inaccurate marketing effectiveness evaluation and delayed strategy adjustments caused by fragmented and difficult-to-integrate data in traditional marketing. Specifically, by establishing a unified data access layer and data mapping rules, seamless integration and standardization of multi-channel data are achieved, significantly improving data processing efficiency and accuracy. The construction of a cross-channel data correlation engine allows for a complete presentation of individual consumer behavior across all channels, providing a comprehensive perspective for a deeper understanding of consumer behavior. The resulting cross-channel marketing effectiveness attribution report accurately quantifies the contribution of each marketing touchpoint, making marketing decisions data-driven and avoiding subjective judgment and empiricism. Ultimately, dynamically adjusting marketing content and delivery methods based on accurate attribution reports significantly improves the ROI of marketing campaigns, reduces the risk of content fatigue, and provides individual consumers with a more personalized and attractive marketing experience, thereby enhancing user stickiness and promoting high-value conversions.

[0095] The above-mentioned attribution report on cross-channel marketing performance, generated based on a complete behavioral view, specifically includes: Define a set of touchpoint types in a conversion path, which includes direct conversion touchpoints and auxiliary conversion touchpoints. The auxiliary conversion touchpoints do not directly generate conversions but promote conversions. An initial contribution weight is configured for each set of touchpoint types, the initial contribution weight representing the expected role of the touchpoint type in the conversion path; Real-time monitoring of the individual consumer's behavioral sequence on the channel, and identification of the touchpoint types in the behavioral sequence; Based on the touchpoint type in the behavior sequence and the initial contribution weight, calculate the preliminary contribution value of each touchpoint to the final conversion; Establish a feedback mechanism to adjust the contribution weight of auxiliary conversion touchpoints based on the actual conversion results and the initial contribution value; Based on the adjusted contribution weights, the contribution value of each touchpoint to the final conversion is recalculated, and an attribution report of the cross-channel marketing effectiveness is generated.

[0096] Specifically, the set of touchpoint types in the conversion path can be understood as all the interactions a consumer has with a marketing campaign before completing the final conversion (e.g., purchase, registration, download). Direct conversion touchpoints are those interactions that directly lead to the final conversion, such as clicking a purchase button and completing payment. Secondary conversion touchpoints, on the other hand, are those interactions that, while not directly causing a conversion, guide, educate, or enhance the consumer's decision-making process, such as browsing product detail pages, watching brand promotional videos, and reading relevant blog posts. These touchpoints play different roles in the conversion path, and their contribution needs to be differentiated.

[0097] The purpose of assigning initial contribution weights to each set of touchpoint types is to provide a benchmark for subsequent attribution calculations. For example, touchpoints that directly convert can be assigned higher initial weights, while touchpoints that assist in conversion can be assigned relatively lower initial weights. These initial weights can be set based on historical data analysis, industry experience, or expert knowledge.

[0098] In practical applications, real-time monitoring of individual consumer behavior sequences across channels refers to continuously tracking every interaction a consumer makes on different marketing channels (such as social media, search engines, email, websites, mobile applications, etc.), including clicks, browsing, dwell time, and search keywords. By analyzing these behavior sequences, it is possible to identify each touchpoint and its type experienced by the consumer in the conversion path.

[0099] Furthermore, based on the touchpoint type and initial contribution weight in the behavioral sequence, the preliminary contribution value of each touchpoint to the final conversion is calculated. The purpose is to quantify the role of each touchpoint before the feedback mechanism intervenes. For example, simple models such as linear attribution models or first / last interaction attribution models can be used for preliminary calculations.

[0100] This application's solution establishes a feedback mechanism to adjust the contribution weights of auxiliary conversion touchpoints based on actual conversion results and initial contribution values. This feedback mechanism is the core of the solution, aiming to correct biases in the initial weights and make the attribution model more realistic. For example, when an auxiliary touchpoint demonstrates a significant promoting effect in multiple actual conversions, its contribution weight will be dynamically increased; conversely, if an auxiliary touchpoint, although assigned a high initial weight, has an insignificant effect in actual conversions, its weight may be decreased. Actual conversion results can include specific metrics such as purchase amount, user retention rate, and registration success rate.

[0101] Therefore, based on the adjusted contribution weights, the contribution value of each touchpoint to the final conversion is recalculated, and an attribution report of the cross-channel marketing performance is generated. This recalculation process ensures the dynamism and accuracy of the attribution results, enabling the final attribution report to more accurately reflect the actual role of each marketing touchpoint in the complex conversion path.

[0102] The solution presented in this application optimizes the generation of attribution reports for cross-channel marketing effectiveness because it introduces a dynamic feedback mechanism. Traditional attribution methods often rely on pre-defined, static weighting rules, which are ill-suited to the complexity and variability of consumer behavior. This application provides a basic framework for attribution analysis by first defining different types of touchpoints and configuring initial weights. More importantly, by monitoring consumer behavior sequences in real time and establishing a feedback mechanism based on actual conversion results, the contribution weights of auxiliary conversion touchpoints can be dynamically adjusted according to their performance in the actual conversion path. It is precisely this weight adjustment based on actual results that allows the attribution model to continuously learn and optimize, thereby overcoming the limitations of static attribution models and ensuring the accuracy and practicality of the attribution report.

[0103] Through the aforementioned technical solution, this application significantly improves the accuracy of cross-channel marketing performance attribution. Compared to attribution methods that rely solely on static rules or preliminary calculations, this application, by introducing a feedback mechanism based on actual conversion results, can more realistically and precisely assess the actual contribution of direct and auxiliary conversion touchpoints in the consumer conversion path. This enables marketers to obtain more reliable data insights, thereby more effectively identifying high-value marketing touchpoints and channels, optimizing marketing budget allocation, and dynamically adjusting marketing content and delivery strategies, ultimately improving the overall marketing ROI.

[0104] In some preferred embodiments, a specific example is given below. Suppose a consumer goes through the following sequence of actions before purchasing an innovative product: first, seeing an advertisement for the product on social media (auxiliary conversion touchpoint); then clicking on the advertisement to visit the brand's official website and browse the product details page (auxiliary conversion touchpoint); next, receiving an email containing product promotional information (auxiliary conversion touchpoint); and finally, searching for the product name through a search engine and clicking to enter the official website to complete the purchase (direct conversion touchpoint).

[0105] In the initial phase, the system assigned lower initial contribution weights to auxiliary conversion touchpoints such as social media ads, product detail page views, and emails, while assigning higher initial contribution weights to the final search engine click-through purchase. Once the consumer completed a purchase, the feedback mechanism was triggered. The system analyzed the consumer's complete behavioral sequence and actual purchase outcome, finding that while social media ads and emails did not directly drive the purchase, they played a significant guiding and facilitating role in the consumer's decision-making process. For example, by analyzing the consumer's dwell time on the product detail page, click depth, and subsequent email open and click-through rates, the system identified that these auxiliary touchpoints had a significant positive impact on the final conversion.

[0106] Based on this, the feedback mechanism dynamically adjusts the contribution weights of these auxiliary conversion touchpoints according to actual conversion results, for example, appropriately increasing the contribution weights of social media ads and emails. Subsequently, the system uses the adjusted contribution weights to recalculate the contribution value of each touchpoint to this purchase conversion. In this way, the final attribution report will more accurately reflect the actual value of social media ads and emails throughout the conversion path, avoiding the underestimation of their role due to static weight allocation. This allows marketing teams to gain a clearer understanding of the true utility of different marketing channels and content in the consumer journey, thereby making more informed marketing decisions.

[0107] In some of the embodiments described above in this application, although a feedback mechanism is established and the contribution weights of auxiliary conversion touchpoints are adjusted based on actual conversion results to recalculate the contribution value of each touchpoint to the final conversion, thereby generating an attribution report for cross-channel marketing effectiveness, this method may not fully consider the complex sequential relationships, time dependencies, and non-linear interactions between touchpoints in the conversion path during its implementation. Simply adjusting the weights may not accurately capture the mutually promoting or reinforcing effects between touchpoints, resulting in room for improvement in the accuracy of the attribution results and their reflection of actual user behavior. If the above problems are not addressed, the generated attribution report may not provide sufficiently accurate guidance for the refined adjustment of marketing strategies, thereby affecting the optimal allocation of marketing resources and overall marketing effectiveness.

[0108] In response, this application proposes a more refined method for recalculating contribution values. By identifying the sequential relationships between touchpoints, assessing the influence strength between touchpoints, and constructing behavioral path dependencies, this method can more comprehensively reflect the true contribution of each touchpoint in the conversion path.

[0109] Based on the adjusted contribution weights, the contribution value of each touchpoint to the final conversion is recalculated, including: Identify the sequence relationship of all touchpoints in the conversion path, the sequence relationship including the order in which the touchpoints appear and the interval between adjacent touchpoints; Evaluate the strength of the influence of each contact on subsequent contacts, whereby the strength of the influence characterizes the promoting or reinforcing effect of the contact on the subsequent contacts; Based on the sequence relationship and the influence intensity, a behavioral path dependency relationship is constructed, which characterizes the nonlinear interaction between each touchpoint in the conversion path. Based on the adjusted contribution weights and the behavioral path dependencies, the contribution value of each touchpoint to the final conversion is recalculated.

[0110] Specifically, the sequence relationship refers to the order in which various marketing touchpoints (such as ad impressions, clicks, content views, registrations, adding items to carts, etc.) occur during an individual consumer's final conversion process, as well as the time intervals between these touchpoints. For example, a consumer might first see a social media ad, then click to enter a product page, receive an email notification a few days later, and finally complete the purchase. The order and duration of these touchpoints are crucial for understanding the conversion path.

[0111] Assessing the impact of each touchpoint on subsequent touchpoints can be understood as quantifying the promoting or reinforcing effect of a specific touchpoint on the conversion efficiency of immediately following touchpoints or a series of subsequent touchpoints. This impact can be positive, such as a high-quality content touchpoint significantly increasing the depth of browsing and conversion intention on subsequent product detail pages; it can also be negative, such as excessively frequent ad exposure leading to user fatigue and reducing the effectiveness of subsequent touchpoints. The aim is to more precisely characterize the dynamic relationships between touchpoints.

[0112] In practical applications, the behavioral path dependency relationship is specifically constructed based on the identified sequence relationships and assessed influence strength, creating a model or graph that characterizes the nonlinear interactions between various touchpoints in the conversion path. For example, Markov chains, deep learning models (such as recurrent neural networks), or graph neural networks can be used to capture this complex dependency relationship. This dependency relationship not only considers the independent contribution of individual touchpoints but, more importantly, reveals the synergistic effect of touchpoint combinations and sequences on the final conversion. The aim is to provide a more comprehensive and realistic conversion path model.

[0113] This application addresses the problem that traditional attribution models may overlook complex interactions between touchpoints by incorporating considerations of the sequence relationships, the strength of influence between touchpoints, and behavioral path dependencies within the conversion path. Through this technical solution, the application significantly improves the accuracy and reliability of cross-channel marketing performance attribution reports. By deeply analyzing the sequence relationships and time intervals of touchpoints, the evolution of user behavior paths can be understood more accurately, avoiding the bias of simply normalizing conversions across different paths. By assessing the strength of the influence of each touchpoint on subsequent touchpoints, the synergistic or inhibitory effects between touchpoints can be quantified, enabling the attribution model to capture deeper-level conversion drivers. By constructing behavioral path dependencies, this application comprehensively reflects the non-linear interactions between touchpoints in the conversion path, thereby generating more refined contribution values ​​that more closely reflect the actual user conversion process. This improved attribution method provides marketers with more insightful decision-making support, helping to more effectively optimize marketing budget allocation, adjust content strategies and channel selection, and ultimately maximize marketing ROI.

[0114] In some of the embodiments described above in this application, the assessment of the impact of each touchpoint on subsequent touchpoints in the conversion path may be based solely on preset rules or historical statistical data. However, the evolution of an individual consumer's interest in an innovative product is a dynamic and complex psychological process, and their behavioral patterns and interest states may change significantly in a short period of time. Failure to capture and reflect this dynamic change in real time may result in an inaccurate assessment of the touchpoint's impact, thereby affecting the accuracy of cross-channel marketing performance attribution reports and making it impossible for marketing strategy adjustments to adequately adapt to the real-time interests of individual consumers.

[0115] In this regard, this application further proposes the above-mentioned assessment of the influence intensity of each contact on subsequent contacts, wherein the influence intensity characterizes the promoting or reinforcing effect of the contact on the subsequent contacts, including: Real-time acquisition and analysis of the individual consumer’s behavioral pattern changes before and after the touchpoint in the conversion path, including interaction frequency, interaction depth and interaction content category; Based on the changes in the behavioral patterns, identify whether the individual consumer's interest in the innovative product has changed, including a shift from a low interest state to a high interest state. When the state of interest changes, the influence intensity of the touch point on subsequent touch points is dynamically adjusted according to the magnitude and speed of the change, so as to reflect the modulation of the touch point effect by the evolution of consumer interest.

[0116] Specifically, real-time acquisition and analysis of individual consumer behavior patterns before and after touchpoints in the conversion path refers to the system continuously monitoring and collecting a series of behavioral data on consumers before and after encountering a specific marketing touchpoint (such as clicking an ad, browsing a product page, or watching a video). This behavioral data can include the frequency of consumer interaction with marketing content, such as the number of clicks or views within a specific time period; the depth of interaction, such as the time spent on the page, the percentage of scrolling, the duration of video viewing, and comments or sharing behavior; and the type of interactive content, such as the product features, content themes, or types of marketing campaigns that consumers primarily focus on. These changes in behavioral patterns can provide initial clues to consumer interests and intentions.

[0117] Identifying whether an individual consumer's interest in innovative products has shifted based on changes in behavioral patterns involves in-depth analysis of the aforementioned real-time acquired behavioral pattern changes to determine whether the consumer's interest in the innovative product has migrated from one state to another. For example, if a consumer's interaction frequency significantly increases and interaction depth deepens after a touchpoint appears, and they begin to focus on content categories more relevant to their purchase decision, this can be identified as a migration from a low-interest state to a high-interest state. A low-interest state might manifest as occasional browsing and superficial interaction, while a high-interest state might manifest as proactive searching, in-depth research, or even adding items to the shopping cart. This identification can be performed using a pre-defined interest state model or machine learning algorithms.

[0118] In practical applications, when a consumer's interest state changes, the influence of a touchpoint on subsequent touchpoints is dynamically adjusted based on the magnitude and speed of the change. This means that once a change in a consumer's interest state is detected, the system adjusts the influence weight of that touchpoint on other touchpoints in the subsequent conversion path in real time, according to the degree and speed of this change. For example, if a consumer's interest rapidly and significantly shifts from a low-interest state to a high-interest state, the facilitating or reinforcing effect of that touchpoint on subsequent touchpoints will be significantly increased; conversely, if the change is small or slow, the adjustment will be correspondingly smaller. This dynamic adjustment aims to more accurately reflect the modulation of the actual effect of touchpoints on the evolution of consumer interest, ensuring that the attribution model can capture subtle changes in the consumer's psychological state.

[0119] This application's solution captures the dynamic evolution of consumer interests by monitoring changes in individual consumer behavior patterns before and after touchpoints in the conversion path in real time. It is precisely through detailed analysis of changes in behavioral patterns such as interaction frequency, interaction depth, and interaction content categories that the system can identify shifts in consumer interest states, such as migration from low to high interest. By directly linking these shifts in interest states, especially the magnitude and speed of the shift, to adjusting the intensity of the touchpoint's influence on subsequent touchpoints, this application's solution overcomes the limitations of traditional static or historical data-based evaluation methods. Therefore, the intensity of a touchpoint's influence is no longer fixed but can be dynamically modulated according to the real-time evolution of individual consumer interests, thus more accurately reflecting the actual role and value of the touchpoint throughout the entire conversion path.

[0120] Through the aforementioned technical solution, this application enables a more refined and dynamic assessment of the influence intensity of touchpoints in the conversion path. Compared to assessments based solely on preset rules or historical data, this application can capture the real-time evolution of individual consumers' interest in innovative products and dynamically adjust the influence intensity of touchpoints based on the magnitude and speed of changes in interest status. Consequently, the marketing effectiveness attribution report will more accurately reflect the true contribution of each touchpoint in the consumer conversion process, especially when consumer interests change rapidly or their understanding of innovative products deepens, effectively avoiding misallocation of marketing resources due to inaccurate assessments. This dynamic adjustment mechanism significantly enhances the flexibility and responsiveness of marketing strategies, allowing adjustments to marketing content or delivery methods targeting individual consumers to more precisely match their current interests and intentions, thereby improving overall marketing efficiency and conversion rates.

[0121] In today's rapidly changing market environment, businesses often rely on marketing strategy generation systems when providing brand management services. These systems typically develop marketing plans by analyzing large amounts of consumer data. However, existing systems face significant challenges in handling the evolution of consumer interests and behavioral patterns. They primarily focus on processing structured data such as past transaction records and browsing preferences to build user profiles and strategy generation rules. This approach works when the market is relatively stable, but it often falls short when dealing with the unstructured, real-time data contained in emerging channels such as social media, and the subtle shifts in consumer interests reflected within it. This makes it difficult for these systems to comprehensively capture the evolving trends in consumer behavior, especially when identifying consumer groups who appear "traditional" but actually have a latent interest in innovative products.

[0122] In this regard, refer to Figure 4 , Figure 4This is a schematic diagram of the structure of a marketing strategy generation system based on big data provided in an embodiment of the present invention, including: Establish a platform to acquire historical interaction data of consumer groups in order to establish behavioral references that characterize the regular interaction performance of the consumer groups with marketing content; and acquire real-time interaction information of individual consumers with innovative products. The judgment end is used to determine whether the interaction information exceeds the normal interaction performance based on the behavior reference, and obtain a judgment result; based on the degree of exceeding the normal interaction performance in the judgment result, it generates the potential interest indication of the individual consumer in the innovative product. The adjustment terminal is used to adjust the delivery method of marketing content or marketing information targeting the individual consumer based on the potential interest indication.

[0123] This application proposes a big data-based marketing strategy generation system aimed at addressing the shortcomings of existing marketing systems in capturing the evolution of consumer interests, thereby more accurately identifying potential customers and optimizing marketing effectiveness. The system forms a dynamic, adaptive marketing closed loop through the collaborative work of three components: the data acquisition end and the behavioral reference construction end; the judgment end is responsible for analyzing individual consumer behavior and identifying potential interests; and the adjustment end optimizes marketing strategies in real time based on the identification results. This systematic design enables enterprises to effectively respond to market changes and improve the accuracy and efficiency of marketing activities.

[0124] Specifically, the system proposed in this application includes an establishment end, a judgment end, and an adjustment end.

[0125] The establishment module is configured to acquire historical interaction data of a consumer group to establish behavioral references characterizing the group's regular interaction with marketing content, and to acquire real-time interaction information of individual consumers with innovative products. The specific methods for acquiring historical interaction data of the consumer group and acquiring real-time interaction information of individual consumers with innovative products have already been described in the above embodiments and will not be repeated here. It is important to emphasize that the establishment module can be a data acquisition and preprocessing module that integrates multiple data interfaces, such as API interfaces, web crawler modules, or data stream processing engines, to collect structured and unstructured data from different data sources (such as CRM systems, e-commerce platforms, social media, etc.). As one implementation method, the establishment module can employ batch processing or stream processing techniques to clean, standardize, and aggregate the raw data, thereby providing a high-quality data foundation for subsequent behavioral reference establishment and real-time interaction information acquisition.

[0126] The judgment terminal is configured to perform functions such as determining whether the interaction information exceeds the normal interaction behavior based on the behavioral reference, obtaining a judgment result, and generating an indication of the individual consumer's potential interest in the innovative product based on the degree to which the interaction information exceeds the normal interaction behavior in the judgment result. The specific methods for determining whether the interaction information exceeds the normal interaction behavior and generating the indication of potential interest have already been described in the above embodiments and will not be repeated here. It is important to emphasize that the judgment terminal can be a data analysis and decision-making module, which can have built-in statistical models, machine learning algorithms, or rule engines to compare and analyze the individual consumer's real-time interaction information with the behavioral reference provided by the judgment terminal. For example, the judgment terminal can use anomaly detection algorithms to identify interactive behaviors that exceed the normal behavior, or use classification models to quantify the degree of potential interest.

[0127] The adjustment module is configured to perform the function of adjusting the delivery method of marketing content or marketing information targeting the individual consumer based on the potential interest indication. The specific methods for adjusting the delivery method of marketing content or marketing information have been described in the above embodiments and will not be repeated here. It is important to emphasize that the adjustment module can be a strategy generation and execution module that can be integrated with a content management system, advertising platform, or marketing automation tool. For example, the adjustment module can dynamically select personalized marketing content templates based on the potential interest indication generated by the judgment module, and combine them with preset delivery rules to push customized marketing information to individual consumers through different channels (such as email, social media advertising, in-app push, etc.). As a preferred embodiment, the adjustment module can also include a feedback loop for receiving marketing performance data and optimizing subsequent adjustment strategies accordingly.

[0128] The big data-based marketing strategy generation system proposed in this application works by constructing a dynamic and adaptive marketing closed loop. First, the system establishes a baseline behavioral reference by deeply mining historical interaction data of consumer groups. This reference represents the "normal" response pattern of consumers to marketing content. This step is fundamental to understanding consumer behavior and provides context for subsequent identification of abnormal behavior.

[0129] Subsequently, the system acquires real-time information on individual consumer interactions with innovative products. This includes various online behaviors such as clicks, browsing, sharing, and comments. This real-time data is crucial for capturing the evolution of consumer interests, as it reflects consumers' most immediate reactions and preferences.

[0130] Next, the judgment unit, based on the behavioral references provided by the establishment unit, determines whether the individual consumer's interaction information exceeds the normal interaction performance. This judgment is not a simple comparison of absolute values, but rather an identification of "abnormal" behaviors that may indicate the emergence of new interests by comparing individual behavior with group benchmarks. For example, a consumer who is not usually interested in technology products might suddenly start frequently watching review videos of a smart home product; this deviation in behavioral pattern may be detected by the judgment unit.

[0131] Once the system identifies interactions that deviate from the norm, it generates an indicator of the individual consumer's potential interest in the innovative product, based on the degree of deviation. This indicator is a quantitative assessment of the consumer's underlying intention, transforming vague interactive behavior into actionable marketing signals. The higher the degree of deviation, the stronger the potential interest indicator, indicating a higher level of consumer interest in the innovative product.

[0132] Finally, the adjustment mechanism dynamically adjusts the delivery of marketing content or information to individual consumers based on the potential interest indicators generated by the judgment mechanism. This means that marketing strategies are no longer static and one-size-fits-all, but are personalized according to each consumer's real-time interest changes. For example, for consumers showing high potential interest, the adjustment mechanism may push more in-depth product information, offer exclusive discounts, or reach them more frequently through their preferred channels. Conversely, for consumers with low potential interest, the approach may be adjusted to a gentler one to avoid over-marketing.

[0133] Through the systematic design and collaborative work described above, the system of this application can effectively solve the pain points existing in the prior art. Traditional systems often struggle to capture the unstructured, real-time data exhibited by consumers on emerging channels such as social media, as well as the subtle shifts in consumer interests reflected therein. This makes it difficult for the system to comprehensively capture the evolutionary trends of consumer behavior characteristics, especially when identifying consumer groups that appear to be "traditional" but have actually developed a potential interest in innovative products.

[0134] The core innovation of this system lies in its ability to dynamically identify and quantify consumers' potential interest in innovative products, thereby enabling precise adjustments to marketing strategies. Unlike existing technologies that primarily rely on structured data such as past transaction records and browsing preferences to build user profiles, this system establishes profiles by referencing consumers' regular interactive behaviors and acquiring real-time information on individual consumers' interactions with innovative products. Through comparative analysis, the system can identify consumers who, while superficially appearing "traditional," have actually developed a potential interest in innovative products.

[0135] For example, in existing technologies, a consumer who previously only purchased traditional home appliances may not have their interest shifted even if they frequently like and share review videos of smart home products such as robot vacuums and air purifiers on social media, due to a lack of direct purchasing behavior. This results in marketing messages failing to reach these potential customers when companies promote innovative products, or even if they do reach them, the messages are pushed based on incorrect customer profiles, leading to poor marketing results.

[0136] The system in this application identifies whether an individual consumer's interaction information exceeds normal interaction behavior through a judgment terminal. It can promptly identify such "above-background noise" click behavior and generate potential interest indicators based on the degree of exceedance. For example, when a "traditional" consumer's interaction frequency and depth with smart home products are significantly higher than their historical average or the normal performance of similar groups, the system can immediately identify their potential interest and adjust marketing content and delivery methods accordingly. This allows businesses to push more attractive personalized marketing content to these consumers, such as highlighting innovative features of smart kitchen equipment like "AI intelligent ingredient recognition and personalized recipe recommendations" or "remote control, allowing you to prepare meals on your way home from get off work," thereby effectively stimulating their purchase desire and improving conversion rates.

[0137] Furthermore, the system in this application provides a clearer attribution chain, helping companies identify the root cause of problems. When the system identifies a potential shift in interests within a group, companies can adjust their product positioning and marketing strategies accordingly, avoiding missed opportunities due to misjudgments of actual market demand. This dynamic and adaptive marketing strategy generation system significantly improves the accuracy and efficiency of marketing, providing strong support for companies to gain an advantage in a highly competitive market.

[0138] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for generating marketing strategies based on big data, characterized in that, include: Acquire historical interaction data of the consumer group to establish behavioral references that characterize the consumer group's regular interaction performance with marketing content; Real-time acquisition of individual consumer interaction information regarding innovative products; Based on the behavioral reference, determine whether the interactive information exceeds the normal interactive behavior, and obtain a judgment result; Based on the degree to which the judgment results exceed the normal interactive performance, an indication of the individual consumer's potential interest in the innovative product is generated; Based on the indicated potential interests, adjust the delivery method of marketing content or information to the individual consumers.

2. The marketing strategy generation method based on big data according to claim 1, characterized in that, The step of determining whether the interactive information exceeds the normal interactive behavior based on the behavioral reference, and obtaining a determination result, includes: Define an interest state set, which contains multiple interest states. Each interest state is associated with a set of behavioral patterns, which is used to distinguish between high-intent interactions and low-intent interactions. Construct an interest evolution trajectory graph, which contains multiple nodes and multiple edges. The nodes represent the interest states, the edges represent the migration paths between the interest states, and the migration paths include the behavioral conditions required for migration. Obtain the behavioral sequence of the individual consumer and the innovative product, the behavioral sequence including user identifier, event type, timestamp, interaction object and quantified behavioral indicators; Based on the behavioral sequence and the interest evolution trajectory map, analyze the migration path of the individual consumer in the interest evolution trajectory map, and determine the current interest state of the individual consumer based on the migration path; When the behavioral sequence meets the condition of migrating from a low-intent state to a high-intent state, the individual consumer's interaction information with the innovative product is judged to exceed the normal interaction performance of the consumer group, and the judgment result is obtained.

3. The marketing strategy generation method based on big data according to claim 1, characterized in that, The step of generating an indication of the individual consumer's potential interest in the innovative product based on the degree to which the judgment result exceeds the normal interactive performance includes: Define the various interactive behaviors of the individual consumer in relation to the innovative product, and the causal chain between the interactive behaviors and the final high-value conversion, wherein the causal chain includes the position and role of the interactive behaviors in the conversion path; Identify the sources of low-intention, high-noise interactions in marketing campaigns and quantify the contribution of each of these sources to the overall noise level. Each of the aforementioned interactive behaviors is assigned an intent gain factor and a noise suppression factor; Adjust the intent gain factor and the noise suppression factor to reflect the strength of the causal relationship between the interactive behavior and the final conversion; Based on the intent gain factor and the noise suppression factor, the historical interaction data of the consumer group are weighted to establish a reference reflecting the consumer group's behavior towards the innovative product; Based on the interaction information between the individual consumer and the innovative product, obtain the individual consumer's cognitive deviation from the innovative product; Based on the magnitude and frequency of the cognitive deviation and its degree of matching with the causal chain of high-value conversion, an indication of the individual consumer's potential interest in the innovative product is generated.

4. The method for generating marketing strategies based on big data according to claim 1, characterized in that, The step of adjusting the delivery method of marketing content or marketing information targeting individual consumers based on the potential interest indication includes: Establish adaptation rules between content formats and channels, wherein the adaptation rules define the matching relationship between the marketing content formats and marketing channels, and evaluate the expected effect of the matching relationship on the consumer group and the risk of content fatigue; Based on the potential interest indications, the adaptation rules, and the content fatigue risk, personalized marketing content is dynamically reorganized and generated. Combined with real-time feedback data, cross-channel marketing information delivery strategies are dynamically optimized to adjust the delivery method of marketing content or marketing information for the individual consumers.

5. The marketing strategy generation method based on big data according to claim 4, characterized in that, The step of dynamically reorganizing and generating personalized marketing content based on the potential interest indications, the adaptation rules, and the content fatigue risk includes: Maintain a content element library, which includes text snippets, visual materials, audio snippets, and interactive components, and associate each content element with one or more semantic tags and brand value tags; Maintain a content template library, which contains templates of various content formats. Each template has a preset combination logic of content elements and a brand narrative structure. When the potential interest indication is received, one or more content templates are selected from the content template library according to the potential interest indication and the adaptation rules; Based on the pre-set combination logic and brand narrative structure in the content template, the content elements are arranged and reorganized to ensure the overall narrative coherence of the generated content and the consistency with the brand's core values.

6. The marketing strategy generation method based on big data according to claim 4, characterized in that, The process of dynamically reorganizing and generating personalized marketing content based on the potential interest indications, the adaptation rules, and the content fatigue risk, and dynamically optimizing cross-channel marketing information delivery strategies in conjunction with real-time feedback data to adjust the delivery method of marketing content or marketing information targeting individual consumers, includes: Establish a unified data access layer to receive and standardize real-time feedback data from different marketing channels. The real-time feedback data includes interactive behavior information, conversion event information, and user behavior path information on each channel. Configure data mapping rules to map standardized data to a unified data structure; A cross-channel data association engine is built to integrate standardized data from different channels in the unified data structure based on user identity and behavior timestamps, forming a complete view of an individual consumer's behavior across all channels; Based on the complete behavioral view, generate an attribution report on cross-channel marketing effectiveness; Based on the attribution report, the delivery method of the marketing content or marketing information is dynamically adjusted.

7. The marketing strategy generation method based on big data according to claim 6, characterized in that, The process of generating an attribution report on cross-channel marketing effectiveness based on the complete behavioral view includes: Define a set of touchpoint types in a conversion path, which includes direct conversion touchpoints and auxiliary conversion touchpoints. The auxiliary conversion touchpoints do not directly generate conversions but promote conversions. An initial contribution weight is configured for each set of touchpoint types, the initial contribution weight representing the expected role of the touchpoint type in the conversion path; Real-time monitoring of the individual consumer's behavioral sequence on the channel, and identification of the touchpoint types in the behavioral sequence; Based on the touchpoint type in the behavior sequence and the initial contribution weight, calculate the preliminary contribution value of each touchpoint to the final conversion; Establish a feedback mechanism to adjust the contribution weight of auxiliary conversion touchpoints based on the actual conversion results and the initial contribution value; Based on the adjusted contribution weights, the contribution value of each touchpoint to the final conversion is recalculated, and an attribution report of the cross-channel marketing effectiveness is generated.

8. The marketing strategy generation method based on big data according to claim 7, characterized in that, The step of recalculating the contribution value of each touchpoint to the final conversion based on the adjusted contribution weight includes: Identify the sequence relationship of all touchpoints in the conversion path, the sequence relationship including the order in which the touchpoints appear and the interval between adjacent touchpoints; Evaluate the strength of the influence of each contact on subsequent contacts, whereby the strength of the influence characterizes the promoting or reinforcing effect of the contact on the subsequent contacts; Based on the sequence relationship and the influence intensity, a behavioral path dependency relationship is constructed, which characterizes the nonlinear interaction between each touchpoint in the conversion path. Based on the adjusted contribution weights and the behavioral path dependencies, the contribution value of each touchpoint to the final conversion is recalculated.

9. The method for generating marketing strategies based on big data according to claim 8, characterized in that, The evaluation of the influence intensity of each contact on subsequent contacts, wherein the influence intensity characterizes the promoting or reinforcing effect of the contact on the subsequent contacts, includes: Real-time acquisition and analysis of the individual consumer’s behavioral pattern changes before and after the touchpoint in the conversion path, including interaction frequency, interaction depth and interaction content category; Based on the changes in the behavioral patterns, identify whether the individual consumer's interest in the innovative product has changed, including a shift from a low interest state to a high interest state. When the state of interest changes, the influence intensity of the touch point on subsequent touch points is dynamically adjusted according to the magnitude and speed of the change, so as to reflect the modulation of the touch point effect by the evolution of consumer interest.

10. A marketing strategy generation system based on big data, characterized in that, include: Establish a endpoint to acquire historical interaction data of the consumer group in order to establish a behavioral reference characterizing the consumer group's regular interaction performance with marketing content; Real-time acquisition of individual consumer interaction information regarding innovative products; The judgment end is used to determine whether the interaction information exceeds the normal interaction performance based on the behavior reference, and obtain a judgment result; based on the degree of exceeding the normal interaction performance in the judgment result, it generates the potential interest indication of the individual consumer in the innovative product. The adjustment terminal is used to adjust the delivery method of marketing content or marketing information targeting the individual consumer based on the potential interest indication.

Images