CN117635191A - Full-link marketing method based on private domain flow pool data - Google Patents
Full-link marketing method based on private domain flow pool data Download PDFInfo
- Publication number
- CN117635191A CN117635191A CN202311593790.4A CN202311593790A CN117635191A CN 117635191 A CN117635191 A CN 117635191A CN 202311593790 A CN202311593790 A CN 202311593790A CN 117635191 A CN117635191 A CN 117635191A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- users
- marketing
- private
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000004458 analytical method Methods 0.000 claims abstract description 25
- 230000000694 effects Effects 0.000 claims abstract description 15
- 230000006870 function Effects 0.000 claims abstract description 8
- 238000012300 Sequence Analysis Methods 0.000 claims abstract description 6
- 238000007405 data analysis Methods 0.000 claims abstract description 4
- 238000013499 data model Methods 0.000 claims abstract description 4
- 230000006399 behavior Effects 0.000 claims description 29
- 230000008451 emotion Effects 0.000 claims description 22
- 230000003993 interaction Effects 0.000 claims description 19
- 238000012545 processing Methods 0.000 claims description 18
- 238000012360 testing method Methods 0.000 claims description 11
- 238000012731 temporal analysis Methods 0.000 claims description 9
- 238000000700 time series analysis Methods 0.000 claims description 9
- 238000004140 cleaning Methods 0.000 claims description 8
- 230000007774 longterm Effects 0.000 claims description 8
- 238000003058 natural language processing Methods 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 8
- 230000001932 seasonal effect Effects 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000005065 mining Methods 0.000 claims description 3
- 230000002860 competitive effect Effects 0.000 abstract description 5
- 230000008901 benefit Effects 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 description 4
- 230000006978 adaptation Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a full-link marketing method based on private flow pool data, which comprises the steps of obtaining private flow pool user data from various channels, carrying out univariate time sequence analysis on training set data, selecting proper components as the input of a model, referring to an SSA-SVM model to construct a classifier, determining optimal SVM kernel functions and super parameters by utilizing characteristic engineering and model training skills, classifying and observing users in the private flow pool through the trained model, predicting the second purchasing time of the users by utilizing a support vector regression method, quantifying and analyzing the life cycle value of each user, and finally optimizing all-link marketing strategies such as advertisement putting, product pricing, user experience and the like through data analysis and model prediction results. The full-link marketing method has the advantages that the private domain flow pool data is utilized to conduct accurate analysis and prediction, the marketing strategy can be optimized, the marketing effect is improved, and finally the increase of sales and the acquisition of competitive advantage are achieved.
Description
Technical Field
The invention relates to the technical field of full-link marketing, in particular to a full-link marketing method based on private domain flow pool data.
Background
The private domain flow pool is constructed to be in contact with clients for a long time, interact, transfer the product value, brand value and education concept of enterprises and merchants, shorten the distance between the private domain flow pool and the clients, strengthen trust, thereby being willing to experience your product, finally purchase your product and service, the flow pool refers to a channel with huge flow, continuously acquire new users, the full-link marketing emphasizes the marketing mode of consumer demand guidance, maintains a smooth communication marketing system, enables sales information to be smoothly presented to the consumers in the most appropriate form in the most appropriate time and the most appropriate place, is suitable for various enterprises and various types of marketing strategies, has a very good promotion effect on-line and off-line integration, can improve brand exposure and awareness, meets the demands of the clients, and finally, the full-link marketing can realize the increase of sales, and enables the enterprises to acquire competitive advantage, the core of the full-link marketing is a large amount of data, and the related operation information can be accurately positioned and analyzed by searching, accessing and accessing the related website to the consumers. Full link marketing relies on data technology solution, based on real-time analysis and customized analysis, the marketing data operation is made to realize enough diversification, and the marketing data operation is made more excellent in the aspect of refinement, so that an accurate marketing popularization group is found in the marketing link.
However, the existing full-link marketing is inconvenient to combine private flow pool data to carry out marketing processing, is inconvenient to match preference and consumption level of each user during full-link marketing, is inconvenient to acquire brand loyalty of the private flow pool user during brand marketing, has poor brand matching efficiency during marketing, is easy to exist in the private flow pool for 'inactive' users, causes low quality of the private flow pool marketing, and increases cost required for marketing.
Disclosure of Invention
The invention aims to provide a full-link marketing method based on private flow pool data, which aims to solve the problems that the prior full-link marketing provided in the background art is inconvenient to combine with the private flow pool data to carry out marketing processing, is inconvenient to match favorites and consumption levels of all users during full-link marketing, is inconvenient to acquire brand loyalty of the private flow pool users during brand marketing, has poor brand matching efficiency during marketing, is easy to exist in a private flow pool, causes low quality of the private flow pool marketing and increases the cost required by marketing.
In order to achieve the above purpose, the present invention provides the following technical solutions: a full link marketing method based on private domain traffic pool data, the method comprising:
s01: acquiring relevant data of users of the private domain flow pool from each channel, cleaning and preprocessing the data, dividing the data into a 50% training set and a 50% testing set, and providing available data for subsequent steps;
s02: performing univariate time series analysis on the training set data acquired in the step S01, and selecting proper components as input of a subsequent model to provide effective characteristics for model input;
s03: referring to the SSA-SVM model, taking the time sequence component selected in the step S02 as an input characteristic of the SSA-SVM model, constructing a classifier, and determining an optimal SVM kernel function and super parameters by utilizing characteristic engineering and model training skills;
s04: testing the SSA-SVM model trained in the step S03 by using a test set, and performing model tuning and evaluation;
s05: classifying and insight users in the private domain flow pool by using a trained SSA-SVM model;
s06: predicting the second purchase time of the users by using a progressive Support Vector Regression (SVR) method, quantifying and analyzing the life cycle value (LTV) of each user, and further layering and formulating a targeted marketing strategy;
s07: and optimizing full-link marketing strategies such as advertisement putting, product pricing, user experience and the like through data analysis and model prediction results.
Further, the step S01 of acquiring user related data includes user behavior data, user portrait data, social media interaction data and related sales data, and cleaning and preprocessing the data, specifically includes the following steps:
s01.1: acquiring a user access page, the number of times the user clicks a link and the time the user stays on the page from a website analysis tool;
s01.2: acquiring the age, sex and geographic position of the user from the CRM system;
s01.3: obtaining praise, comment and shared interaction data of a user from a social media platform;
s01.4: sales, and purchase times are obtained from store sales records.
Further, processing the repeated data, filling the missing values and processing abnormal values of the data obtained in the step S01, performing standardization and normalization processing, performing NLP processing on interaction data such as praise, comment and share obtained in the step S01.3, and performing emotion analysis and theme extraction on the comment and interaction data by applying a natural language processing technology so as to know emotion attitudes and attention points of users;
the S01 step is used for acquiring user behavior data, user portrait data, social media interaction data and related sales data, cleaning, preprocessing and standardizing the data, providing accurate and reliable data basis for subsequent analysis and modeling, performing deduplication, missing value filling, outlier processing and normalization processing on the acquired data, and performing natural language processing on the social media interaction data, and performing emotion analysis and theme extraction to know emotion attitudes and attention points of users.
Further, in the step S02, univariate time series analysis is performed on the private domain traffic pool data, and the time series data is decomposed into a plurality of components, which specifically includes the following steps:
s02.1: performing time sequence analysis on the page access and the page stay time of the user, and analyzing long-term trend, periodicity and seasonal components;
s02.2: carrying out time sequence analysis on the sales data, and analyzing long-term trend, seasonality and periodicity components to know the periodicity and trend of sales conditions;
the step S02 is to decompose the private domain flow pool data into a plurality of components including long-term trend, periodicity and seasonal components through univariate time series analysis so as to know the trend, periodicity and seasonal of the user behavior and sales situation.
Further, training the SSA-SVM model in the step S03 specifically includes the following steps:
s03.1: training an SVM model to classify user behavior, such as high active users, low active users, etc., based on the time series components obtained in the step S02.1;
s03.2: training an SVM model to classify a population of users, such as young users, middle-aged users, male users, female users, with the user portrayal data as input features;
s03.3: training an SVM model by taking the extracted emotion characteristics as input characteristics to judge the attitudes and favorites of users, such as active users, passive users, users focusing on specific functions of products and the like;
s03.4: training an SVM model based on the time series components in the step S02.2 to predict the next purchase time of the user;
the step S03 is used for training the SSA-SVM model by using the selected time sequence components, the user portrait data and the emotion characteristics so as to classify the user behavior, the user group, the user attitude and predict the next purchase time.
Further, selecting an optimal SSA-SVM model through the step S04;
in the step S04, the trained SSA-SVM model is tested, optimized and evaluated through a test set, and optimal model parameters and kernel functions are selected, so that the accuracy and precision of the model are improved.
Further, in the step S05, the trained SSA-SVM model is used for analysis, which specifically includes the following steps:
s05.1: applying the trained SSA-SVM model to user data in a private domain flow pool;
s05.2: classifying users, which may be considered to be classified into different groups, into loyalty customers, potential customers, and churn customers;
s05.3: analyzing the characteristics of the user groups, and excavating insight information such as demand pain points, purchase intention and the like of the users by analyzing key characteristics such as interests, demands, values and the like of each user group;
in the step S05, the trained SSA-SVM model is utilized to classify the users in the private domain flow pool, the characteristics of the user group are analyzed, key information such as the demand pain points and the purchase intention of the users is mined, and a basis is provided for formulating a targeted marketing strategy;
firstly, classifying users in a private domain flow pool by using a trained SSA-SVM model, wherein the SSA-SVM model is a model based on time sequence components and user portrait characteristics, can use user historical behavior data and emotion characteristics to classify and predict user behaviors, and can divide the users in the private domain flow pool into different behavior groups by using the trained model, so that the behavior characteristics and purchasing habits of the users are deeply known;
second, the characteristics of the user population are analyzed to better understand the behavior and preferences of the different user populations. By comparing and analyzing the characteristics of different user groups, the difference and commonality among the user groups can be found, the characteristics include but are not limited to age, gender, region, purchasing preference, interaction behavior and the like, and by deeply analyzing the characteristics, key information about user requirements, pain points, purchasing intention and the like can be obtained, so that basis is provided for formulating targeted marketing strategies;
in the aspect of mining the demand pain points and the purchase intention of the users, the attention to the behavior characteristics and the emotion characteristics of the users is very important, the interest points, the doubtful points and the satisfaction degree of the users on products or services can be known by analyzing the behavior characteristics and the emotion expressions of the users in different time periods and different situations, and on the basis, the purchase intention, the purchase decision factors, the product preference, the service requirement and the like of the users can be further presumed, so that powerful support is provided for formulating targeted marketing strategies.
The trained SSA-SVM model is utilized to classify users in the private domain flow pool, analyze the characteristics of the user groups, mine key information such as demand pain points and purchase intention of the users, and the like, so that enterprises can be helped to accurately grasp the demands and interests of different user groups, and provide basis for formulating targeted marketing strategies, which is beneficial to improving marketing effects, enhancing user satisfaction and improving the competitiveness of the enterprises in competitive markets.
Further, the specific steps included in the step S06 are as follows:
s06.1: predicting the second purchase time of the user in the private domain flow pool by using a Support Vector Regression (SVR) method;
s06.2: quantifying and analyzing the life cycle value (LTV) of each user based on the prediction results, helping to further stratify and formulate targeted marketing strategies; users are classified into high-value users, ordinary users, low-value users, etc., so that limited resources and investment are used on the high-value users more pertinently;
s06.3: according to layering and analysis results, a corresponding marketing strategy is formulated; personalized customized services are provided for high value users, promotional campaigns are provided for general users, low value users are reactivated, and the like.
In the step S06, the second purchase time of the user is predicted based on the advanced support vector regression method, and the life cycle value of each user is quantized and analyzed, so that targeted marketing strategies are further layered and formulated, and all-link marketing strategies such as advertisement delivery, product pricing, user experience and the like are optimized.
Further, the specific steps included in the step S07 are as follows:
s07.1: according to analysis of user group characteristics, personalized marketing schemes such as recommendation, preferential and service are designed;
s07.2: by subdividing user groups, different marketing strategies and activities are formulated for different groups; for male users and female users, products or services suitable for the male users and the female users can be pushed respectively;
in the step S07, personalized marketing schemes such as recommendation, preferential and service are designed according to the characteristic analysis of the user group, different marketing strategies and activities are formulated for different user groups, and the marketing effect and the user satisfaction are improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Compared with the prior art, the invention has the beneficial effects that: the invention provides a full-link marketing method based on private domain traffic pool data. Compared with the prior art, the invention has the following beneficial effects: the private domain flow pool data is cleaned, preprocessed and analyzed, the SSA-SVM model is utilized to conduct user classification, insight and prediction, the behavior, preference and demand of the user can be known more accurately, a basis is provided for formulating a targeted marketing strategy, thereby the marketing effect is improved, through analysis of user portrait data and social media interaction data, the hobbies and emotion attitudes of the user can be known, thereby advertisement delivery and product pricing are optimized, advertisements are delivered to target users more accurately, the product pricing is more in line with the consumption habit of the users, the advanced support vector regression method is utilized to predict the second purchase time of the users, and the life cycle value of each user is quantized and analyzed, the user can be layered and personalized marketing strategy is formulated, limited resources and investment are used for high-value users more targeted, the marketing efficiency is improved, and the cost is reduced: through optimizing full-link marketing strategies such as advertisement putting, product pricing, user experience and the like, the brand exposure rate and the popularity can be improved, the customer loyalty is increased, the marketing cost is reduced to the minimum, the marketing efficiency is improved, in a word, the full-link marketing method of the invention utilizes private domain flow pool data to carry out accurate analysis and prediction, the marketing strategy can be optimized, the marketing effect is improved, and finally the increase of sales and the acquisition of competitive advantage are realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those skilled in the art from this disclosure that the drawings described below are merely exemplary and that other embodiments may be derived from the drawings provided without undue effort.
FIG. 1 is a workflow diagram of a full link marketing method based on private traffic pool data according to the present invention.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples are not representative of all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus consistent with some aspects of the disclosure as detailed in the accompanying claims.
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, a full link marketing method based on private domain traffic pool data according to an embodiment of the present invention includes:
s01: acquiring relevant data of users of the private domain flow pool from each channel, cleaning and preprocessing the data, dividing the data into a 50% training set and a 50% testing set, and providing available data for subsequent steps;
s02: performing univariate time series analysis on the training set data acquired in the step S01, and selecting proper components as input of a subsequent model to provide effective characteristics for model input;
s03: referring to the SSA-SVM model, taking the time sequence component selected in the step S02 as an input characteristic of the SSA-SVM model, constructing a classifier, and determining an optimal SVM kernel function and super parameters by utilizing characteristic engineering and model training skills;
s04: testing the SSA-SVM model trained in the step S03 by using a test set, and performing model tuning and evaluation;
s05: classifying and insight users in the private domain flow pool by using a trained SSA-SVM model;
s06: predicting the second purchase time of the users by using a progressive Support Vector Regression (SVR) method, quantifying and analyzing the life cycle value (LTV) of each user, and further layering and formulating a targeted marketing strategy;
s07: and optimizing full-link marketing strategies such as advertisement putting, product pricing, user experience and the like through data analysis and model prediction results.
Further, the step S01 of acquiring user related data includes user behavior data, user portrait data, social media interaction data and related sales data, and cleaning and preprocessing the data, specifically includes the following steps:
s01.1: acquiring a user access page, the number of times the user clicks a link and the time the user stays on the page from a website analysis tool;
s01.2: acquiring the age, sex and geographic position of the user from the CRM system;
s01.3: obtaining praise, comment and shared interaction data of a user from a social media platform;
s01.4: sales, and purchase times are obtained from store sales records.
Further, processing the repeated data, filling the missing values and processing abnormal values of the data obtained in the step S01, performing standardization and normalization processing, performing NLP processing on interaction data such as praise, comment and share obtained in the step S01.3, and performing emotion analysis and theme extraction on the comment and interaction data by applying a natural language processing technology so as to know emotion attitudes and attention points of users;
the S01 step is used for acquiring user behavior data, user portrait data, social media interaction data and related sales data, cleaning, preprocessing and standardizing the data, providing accurate and reliable data basis for subsequent analysis and modeling, performing deduplication, missing value filling, outlier processing and normalization processing on the acquired data, and performing natural language processing on the social media interaction data, and performing emotion analysis and theme extraction to know emotion attitudes and attention points of users.
Further, in the step S02, univariate time series analysis is performed on the private domain traffic pool data, and the time series data is decomposed into a plurality of components, which specifically includes the following steps:
s02.1: performing time sequence analysis on the page access and the page stay time of the user, and analyzing long-term trend, periodicity and seasonal components;
s02.2: carrying out time sequence analysis on the sales data, and analyzing long-term trend, seasonality and periodicity components to know the periodicity and trend of sales conditions;
the step S02 is to decompose the private domain flow pool data into a plurality of components including long-term trend, periodicity and seasonal components through univariate time series analysis so as to know the trend, periodicity and seasonal of the user behavior and sales situation.
Further, training the SSA-SVM model in the step S03 specifically includes the following steps:
s03.1: training an SVM model to classify user behavior, such as high active users, low active users, etc., based on the time series components obtained in the step S02.1;
s03.2: training an SVM model to classify a population of users, such as young users, middle-aged users, male users, female users, with the user portrayal data as input features;
s03.3: training an SVM model by taking the extracted emotion characteristics as input characteristics to judge the attitudes and favorites of users, such as active users, passive users, users focusing on specific functions of products and the like;
s03.4: training an SVM model based on the time series components in the step S02.2 to predict the next purchase time of the user;
the step S03 is used for training the SSA-SVM model by using the selected time sequence components, the user portrait data and the emotion characteristics so as to classify the user behavior, the user group, the user attitude and predict the next purchase time.
Further, selecting an optimal SSA-SVM model through the step S04;
in the step S04, the trained SSA-SVM model is tested, optimized and evaluated through a test set, and optimal model parameters and kernel functions are selected, so that the accuracy and precision of the model are improved.
Further, in the step S05, the trained SSA-SVM model is used for analysis, which specifically includes the following steps:
s05.1: applying the trained SSA-SVM model to user data in a private domain flow pool;
s05.2: classifying users, which may be considered to be classified into different groups, into loyalty customers, potential customers, and churn customers;
s05.3: analyzing the characteristics of the user groups, and excavating insight information such as demand pain points, purchase intention and the like of the users by analyzing key characteristics such as interests, demands, values and the like of each user group;
in the step S05, the trained SSA-SVM model is utilized to classify the users in the private domain flow pool, the characteristics of the user group are analyzed, key information such as the demand pain points and the purchase intention of the users is mined, and a basis is provided for formulating a targeted marketing strategy;
firstly, classifying users in a private domain flow pool by using a trained SSA-SVM model, wherein the SSA-SVM model is a model based on time sequence components and user portrait characteristics, can use user historical behavior data and emotion characteristics to classify and predict user behaviors, and can divide the users in the private domain flow pool into different behavior groups by using the trained model, so that the behavior characteristics and purchasing habits of the users are deeply known;
second, the characteristics of the user population are analyzed to better understand the behavior and preferences of the different user populations. By comparing and analyzing the characteristics of different user groups, the difference and commonality among the user groups can be found, the characteristics include but are not limited to age, gender, region, purchasing preference, interaction behavior and the like, and by deeply analyzing the characteristics, key information about user requirements, pain points, purchasing intention and the like can be obtained, so that basis is provided for formulating targeted marketing strategies;
in the aspect of mining the demand pain points and the purchase intention of the users, the attention to the behavior characteristics and the emotion characteristics of the users is very important, the interest points, the doubtful points and the satisfaction degree of the users on products or services can be known by analyzing the behavior characteristics and the emotion expressions of the users in different time periods and different situations, and on the basis, the purchase intention, the purchase decision factors, the product preference, the service requirement and the like of the users can be further presumed, so that powerful support is provided for formulating targeted marketing strategies.
The trained SSA-SVM model is utilized to classify users in the private domain flow pool, analyze the characteristics of the user groups, mine key information such as demand pain points and purchase intention of the users, and the like, so that enterprises can be helped to accurately grasp the demands and interests of different user groups, and provide basis for formulating targeted marketing strategies, which is beneficial to improving marketing effects, enhancing user satisfaction and improving the competitiveness of the enterprises in competitive markets.
Further, the specific steps included in the step S06 are as follows:
s06.1: predicting the second purchase time of the user in the private domain flow pool by using a Support Vector Regression (SVR) method;
s06.2: quantifying and analyzing the life cycle value (LTV) of each user based on the prediction results, helping to further stratify and formulate targeted marketing strategies; users are classified into high-value users, ordinary users, low-value users, etc., so that limited resources and investment are used on the high-value users more pertinently;
s06.3: according to layering and analysis results, a corresponding marketing strategy is formulated; personalized customized services are provided for high-value users, promotion activities are provided for common users, and low-value users are reactivated;
in the step S06, the second purchase time of the user is predicted based on the advanced support vector regression method, and the life cycle value of each user is quantized and analyzed, so that targeted marketing strategies are further layered and formulated, and all-link marketing strategies such as advertisement delivery, product pricing, user experience and the like are optimized.
Further, the specific steps included in the step S07 are as follows:
s07.1: according to analysis of user group characteristics, personalized marketing schemes such as recommendation, preferential and service are designed;
s07.2: by subdividing user groups, different marketing strategies and activities are formulated for different groups; for male users and female users, products or services suitable for the male users and the female users can be pushed respectively;
in the step S07, personalized marketing schemes such as recommendation, preferential and service are designed according to the characteristic analysis of the user group, different marketing strategies and activities are formulated for different user groups, and the marketing effect and the user satisfaction are improved.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope of the disclosure being indicated by the following claims.
Claims (9)
1. A full link marketing method based on private domain traffic pool data, the method comprising:
s01: acquiring relevant data of users of the private flow pool from each channel, cleaning and preprocessing the data, and dividing the data into a 50% training set and a 50% testing set;
s02: performing univariate time series analysis on the training set data acquired in the step S01, and selecting proper components as input of a subsequent model;
s03: referring to the SSA-SVM model, taking the time sequence component selected in the step S02 as an input characteristic of the SSA-SVM model, constructing a classifier, and determining an optimal SVM kernel function and super parameters by utilizing characteristic engineering and model training skills;
s04: testing the SSA-SVM model trained in the step S03 by using a test set, and performing model tuning and evaluation;
s05: classifying and insight users in the private domain flow pool by using a trained SSA-SVM model;
s06: predicting the second purchase time of the users by using an advanced support vector regression method, quantifying and analyzing the life cycle value of each user, and further layering and formulating a targeted marketing strategy;
s07: and optimizing full-link marketing strategies such as advertisement putting, product pricing, user experience and the like through data analysis and model prediction results.
2. The full-link marketing method based on private traffic pool data according to claim 1, wherein the step S01 of obtaining user related data includes user behavior data, user portrait data, social media interaction data and related sales data, and cleaning and preprocessing the data, specifically comprising the steps of:
s01.1: acquiring a user access page, the number of times the user clicks a link and the time the user stays on the page from a website analysis tool;
s01.2: acquiring the age, sex and geographic position of the user from the CRM system;
s01.3: obtaining praise, comment and shared interaction data of a user from a social media platform;
s01.4: sales, and purchase times are obtained from store sales records.
3. The full-link marketing method based on private domain flow pool data according to claim 2, wherein the processing of removing duplicate data, filling up missing values, processing abnormal values, and performing standardization and normalization processing is performed on the data obtained in the step S01, and NLP processing is performed on praise, comment and share interaction data obtained in the step S01.3, and natural language processing technology is applied to perform emotion analysis and topic extraction on comment and interaction data so as to know emotion attitudes and points of interest of users.
4. The full-link marketing method based on private flow pool data according to claim 1, wherein in the step S02, univariate time series analysis is performed on the private flow pool data, and the time series data is decomposed into a plurality of components, specifically comprising the following steps:
s02.1: performing time sequence analysis on the page access and the page stay time of the user, and analyzing long-term trend, periodicity and seasonal components;
s02.2: and (3) carrying out time series analysis on the sales data, and analyzing long-term trend, seasonality and periodicity components to know the periodicity and trend of sales.
5. The full-link marketing method based on private traffic pool data according to claim 1, wherein the training of the SSA-SVM model in the step S03 specifically comprises the following steps:
s03.1: training an SVM model based on the time series components obtained in the step S02.1 to classify user behaviors;
s03.2: training an SVM model by taking user portrait data as input characteristics so as to classify user groups;
s03.3: training an SVM model by taking the extracted emotion characteristics as input characteristics to judge the attitude and preference of a user;
s03.4: based on the time series components in the step S02.2, an SVM model is trained to predict the next purchase time of the user.
6. The full link marketing method based on private domain traffic pool data according to claim 1, wherein the optimal SSA-SVM model is selected by the step S04.
7. The full-link marketing method based on private traffic pool data according to claim 1, wherein the step S05 uses a trained SSA-SVM model for analysis, and specifically comprises the steps of:
s05.1: applying the trained SSA-SVM model to user data in a private domain flow pool;
s05.2: classifying users, which may be considered to be classified into different groups, into loyalty customers, potential customers, and churn customers;
s05.3: and analyzing the characteristics of the user groups, and mining the demand pain points and the purchase intention insight information of the users by analyzing the key characteristics such as interests, demands, values and the like of each user group.
8. The method of claim 1, wherein the step S06 comprises the following steps,
s06.1: predicting the second purchase time of the user in the private domain flow pool by using a further support vector regression method;
s06.2: quantifying and analyzing the life cycle value of each user according to the prediction result, and helping to further layer and formulate a targeted marketing strategy; users are classified into high-value users, normal users and low-value users, so that limited resources and investment are used on the high-value users more pertinently;
s06.3: according to layering and analysis results, a corresponding marketing strategy is formulated; personalized customized services are provided for high-value users, promotion activities are provided for common users, and low-value users are reactivated.
9. The full-link marketing method based on private domain traffic pool data according to claim 1, wherein the specific steps included in the step S07 are as follows:
s07.1: according to analysis of user group characteristics, personalized recommendation, preferential and service marketing schemes are designed; providing a loyalty customer with a dedicated gift or discount, providing a potential customer with customized product information, and providing an churn customer with a return offer;
s07.2: by subdividing user groups, different marketing strategies and activities are formulated for different groups; for male users and female users, products or services suitable for them can be pushed separately.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311593790.4A CN117635191A (en) | 2023-11-27 | 2023-11-27 | Full-link marketing method based on private domain flow pool data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311593790.4A CN117635191A (en) | 2023-11-27 | 2023-11-27 | Full-link marketing method based on private domain flow pool data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117635191A true CN117635191A (en) | 2024-03-01 |
Family
ID=90033216
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311593790.4A Pending CN117635191A (en) | 2023-11-27 | 2023-11-27 | Full-link marketing method based on private domain flow pool data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117635191A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111598648A (en) * | 2020-04-16 | 2020-08-28 | 上海源慧信息科技股份有限公司 | Full-link online marketing method based on fast-moving industrial commodities |
| CN112365285A (en) * | 2020-11-13 | 2021-02-12 | 上海源慧信息科技股份有限公司 | Full link marketing method based on private domain flow pool data |
| CN112700286A (en) * | 2020-12-23 | 2021-04-23 | 罗科仕管理顾问有限公司 | Deep learning model of customer lifecycle values for customer classification and multi-entity matching |
| CN116304374A (en) * | 2023-05-19 | 2023-06-23 | 云印技术(深圳)有限公司 | Customer matching method and system based on package data |
| CN116485424A (en) * | 2023-06-19 | 2023-07-25 | 江西倬慧信息科技有限公司 | Intelligent marketing method, system, equipment terminal and readable storage medium |
-
2023
- 2023-11-27 CN CN202311593790.4A patent/CN117635191A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111598648A (en) * | 2020-04-16 | 2020-08-28 | 上海源慧信息科技股份有限公司 | Full-link online marketing method based on fast-moving industrial commodities |
| CN112365285A (en) * | 2020-11-13 | 2021-02-12 | 上海源慧信息科技股份有限公司 | Full link marketing method based on private domain flow pool data |
| CN112700286A (en) * | 2020-12-23 | 2021-04-23 | 罗科仕管理顾问有限公司 | Deep learning model of customer lifecycle values for customer classification and multi-entity matching |
| CN116304374A (en) * | 2023-05-19 | 2023-06-23 | 云印技术(深圳)有限公司 | Customer matching method and system based on package data |
| CN116485424A (en) * | 2023-06-19 | 2023-07-25 | 江西倬慧信息科技有限公司 | Intelligent marketing method, system, equipment terminal and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hoban et al. | Effects of internet display advertising in the purchase funnel: Model-based insights from a randomized field experiment | |
| Van den Poel et al. | Predicting online-purchasing behaviour | |
| US10360568B2 (en) | Customer state-based targeting | |
| Neslin et al. | Defection detection: Measuring and understanding the predictive accuracy of customer churn models | |
| US10902443B2 (en) | Detecting differing categorical features when comparing segments | |
| US20110231246A1 (en) | Online and offline advertising campaign optimization | |
| US20110231245A1 (en) | Offline metrics in advertisement campaign tuning | |
| US11734711B2 (en) | Systems and methods for intelligent promotion design with promotion scoring | |
| US20110231244A1 (en) | Top customer targeting | |
| US20240005368A1 (en) | Systems and methods for an intelligent sourcing engine for study participants | |
| WO2018213019A1 (en) | Systems and methods for intelligent promotion design with promotion selection | |
| Neslin | Customer relationship management (CRM) | |
| US20230368226A1 (en) | Systems and methods for improved user experience participant selection | |
| Tahoun et al. | Artificial intelligence as the new realm for online advertising | |
| CN117635191A (en) | Full-link marketing method based on private domain flow pool data | |
| US20170316449A1 (en) | Systems and methods for intelligent promotion design with promotion selection | |
| US20030208494A1 (en) | System and method for multidimensional valuation of consumer technology customers | |
| Singh et al. | A/B Testing and Audience Creation for Effective Digital Marketing: Evidences from Facebook Analytics | |
| Tănase | Predictive Marketing: Anticipating Market Demand with Proactive Action | |
| US11941659B2 (en) | Systems and methods for intelligent promotion design with promotion scoring | |
| US20240177204A1 (en) | Systems and methods for attribute characterization of usability testing participants | |
| Febrianti et al. | Investigated The Role Of Celebrity Endorsements And Influencers On Marketing Performance With Social Media As A Intervening Variable | |
| Gajanova et al. | Digital Marketing in the Context of Consumer Behaviour in the ICT Industry: The Case Study of the Slovak Republic | |
| Arjmand et al. | The impact of customer clubs on lifetime value of banking customers | |
| Hng et al. | The forefront of mobile shopping: An emerging economy's perspective |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |