Information flow advertisement module putting method

Abstract

The invention relates to the technical field of advertisement putting methods, in particular to an information flow advertisement module putting method, which specifically comprises the following steps of: 1, constructing a user portrait, collecting and analyzing behavior data of a user on various platforms, and constructing the user portrait; 2, analyzing and classifying advertisement materials; step 3, selecting and evaluating a delivery platform, namely selecting a corresponding delivery platform and evaluating the delivery platform; step 4, making a putting strategy, and making a specific putting strategy according to a user portrait, advertisement material classification and a putting platform evaluation result; 5, real-time monitoring and data collection: in the advertisement putting process, monitoring the advertisement putting effect in real time and collecting related data; according to the method and the device, the advertisement putting accuracy can be improved, an advertiser can accurately reach a target customer group, and the click rate and the conversion rate are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of advertising placement methods, and in particular to a method for placing information flow advertising modules. Background Art

[0002] In the context of digital marketing, information flow advertising has become one of the important means for enterprises to promote products and services. With the continuous growth of the number of Internet users and the increasing richness of user behavior data on various platforms, how to accurately push information flow advertising to target users and improve the advertising placement effect and return on investment has become a key issue faced by the advertising industry.

[0003] Traditional advertising placement methods often have many limitations. On the one hand, the placement accuracy is relatively low, and advertisers are difficult to accurately reach the target customer group, resulting in a large amount of advertising resources being wasted on non-target users. For example, on some social platforms, advertisements may only be simply targeted based on basic user attributes (such as age, gender, etc.), and it is impossible to deeply understand more valuable information such as users' interests and consumption habits, making the click-through rate and conversion rate of advertisements unsatisfactory. Summary of the Invention

[0004] In view of the technical problems raised in the background art, the present invention provides a method for placing information flow advertising modules.

[0005] The technical solution adopted by the present invention is: a method for placing information flow advertising modules, specifically including the following steps:

[0006] Step 1: User portrait construction, collecting and analyzing user behavior data on various platforms to construct a user portrait;

[0007] Step 2: Advertising material analysis and classification;

[0008] Step 3: Selection and evaluation of the placement platform, selecting the corresponding placement platform and evaluating the placement platform;

[0009] Step 4: Formulation of the placement strategy, formulating a specific placement strategy according to the results of user portrait construction, advertising material classification, and placement platform evaluation;

[0010] Step 5: Real-time monitoring and data collection, during the advertising placement process, real-time monitoring the advertising placement effect and collecting relevant data;

[0011] Step 6: Evaluation and analysis of the placement effect, evaluating and analyzing the advertising placement effect according to the data collected by real-time monitoring.

[0012] In one embodiment, in Step 1: User portrait construction, collecting and analyzing user behavior data on various platforms to construct a user portrait, the specific method is as follows;

[0013] Let the user set be U = {u1, u2, …, u n}, where n is the number of users. For each user u i , construct a feature vector P i = (p i1 , p i2 , …, p im ), where m represents the number of features; p i1 can represent the age of the user, p i2 represents the gender of the user; p i3 represents the geographical information of the user; p i4 is the interest tag of the user;

[0014] Let the number of times user u i views interest tag j be b ij , the number of searches be s ij , and the number of purchases be p ij . Then the interest preference degree I i of user u ij for interest tag j is calculated by the following formula:

[0015]

[0016] where α, β, γ are weight coefficients, and α + β + γ = 1.

[0017] In one embodiment, in step two: The method for advertising material analysis and classification is as follows:

[0018] Let the advertising material set be A = {a1, a2, …, a l}, where l is the number of advertising materials. For each advertising material a k , extract its feature vector F k = (f k1 , f k2 , …, f kn ); f k1 represents the theme of the advertising material, f k2 represents the style of the advertising material; f k3 represents the characteristics of the target audience group targeted by the advertising material;

[0019] Divide the advertising materials into K categories C1, C2, …, C K ;

[0020] Randomly select K initial cluster centers O1, O2, …, O K ;

[0021] For each advertising material a k , calculate its distance d(a to each cluster center.k , O j ), distance, the formula uses Euclidean distance:

[0022]

[0023] Among them, f ki is the i-th eigenvalue of the advertising material a k , and o j i is the i-th eigenvalue of the clustering center O j .

[0024] In one of the embodiments, in step three: the method for selecting and evaluating the placement platform is as follows:

[0025] Let the set of placement platforms be P = {p1, p2,..., p q}, where q is the number of platforms. For each platform p j , evaluate its multiple indicators;

[0026] The number of platform users N j represents the potential audience scale of the platform, and the matching degree M of the platform with the advertising target audience j ;

[0027] Let the user portrait feature vector be P pj = (p pj1 , p pj2 ,..., p pjn ), and the advertising target audience portrait feature vector be P a = (p a1 , p a2 ,..., p an ), then the matching degree M j is calculated by the following formula

[0028]

[0029] Introduce a comprehensive evaluation index E j , which is a weighted combination of the number of platform users, user activity, and matching degree, and is calculated as follows:

[0030]

[0031] Among them, ω1, ω2, ω3 are weight coefficients, and ω1 + ω2 + ω3 = 1; the weight coefficients are determined according to the advertiser's emphasis on different indicators.

[0032] In one of the embodiments, in step four: formulating the placement strategy, according to the results of user portrait, advertising material classification, and placement platform evaluation, the specific method for formulating the specific placement strategy is as follows:

[0033] Let the advertising delivery time interval be T = [t1, t2], which is divided into multiple time slices t s1 , t s2 , …, t sk ;

[0034] For each time slice t si , determine the advertising delivery volume Q si ;

[0035] Let the maximum allowed number of times user u i receives advertisements within a period of time be F maπi , and the actual number of times of receiving advertisements be F i ; In each time slice t s i, for user u i , if F i < F maπi , then advertisements can be delivered to him;

[0036] Let the total advertising budget be B, which is allocated to different delivery platforms and time slices;

[0037] For platform p j , the budget allocation in time slice t si is B jsi , and it satisfies ;

[0038] Let the revenue R j brought by the advertising delivery of platform p si in time slice t jsi , and the cost be C jsi , then the return on investment

[0039] Under the condition of satisfying the budget constraint, calculate the maximization of the total return on investment:

[0040]

[0041] The constraint conditions include:

[0042]

[0043] In one of the embodiments, in step five: real-time monitoring and data collection, during the advertising delivery process, the method for real-time monitoring of the advertising delivery effect and collecting relevant data is as follows:

[0044] Let the exposure volume of advertisement a k on platform p j be E kj , the click-through rate be the conversion rate be CR kj , and the user stay time be Tkj ; Click-through rate The calculation formula is as follows:

[0045]

[0046] Among them, Clicks skj is the number of clicks of advertisement a k on platform p j ;

[0047] The conversion rate CR kj The calculation formula is as follows:

[0048]

[0049] Among them, Conversions skj is the number of conversions of advertisement a k on platform p j such as purchase behavior, registration behavior, etc.);

[0050] By monitoring these data in real time, the performance of the advertisement on different platforms can be understood in a timely manner, providing data support for subsequent dynamic optimization.

[0051] In one of the embodiments, step six: evaluation and analysis of the delivery effect. The specific method for evaluating and analyzing the advertisement delivery effect according to the data collected by real-time monitoring is as follows:

[0052] Let the total return on investment of the advertisement delivery be ROI total , which is the weighted sum of the return on investment of each platform and time slice:

[0053]

[0054] By evaluating and analyzing the delivery effect, the problems existing in the advertisement delivery process can be found.

[0055] The beneficial effects of the present invention are as follows: Compared with the prior art, in the present invention, problems such as low accuracy in traditional advertising placement can be effectively solved, bringing significant beneficial effects. In the construction of user portraits, by collecting diverse behavioral data, quantifying interest preferences, analyzing and classifying advertising materials, and clustering according to characteristics such as theme and style, the materials are accurately matched with users. When selecting and evaluating the placement platform, the number of platform users, activity level, and matching degree with the target audience are comprehensively considered to select the optimal platform and avoid waste of resources. When formulating the placement strategy, the placement time, frequency, and budget are reasonably planned, and an optimization algorithm is used to maximize the return on investment. Key data such as exposure and click-through rate are obtained through real-time monitoring and data collection. Based on the evaluation and analysis of the placement effect, problems are identified to improve the accuracy of advertising placement, enabling advertisers to accurately reach the target customer group, increasing the click-through rate and conversion rate, and reducing the waste of advertising resources on non-target users. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a schematic flowchart of the present invention; DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "front", "upper", "lower", "left", "right", "vertical", "horizontal", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0058] In order to solve the problems existing in the background art, the present application proposes the following technical solution: A method for placing an information flow advertising module, specifically including the following steps:

[0059] Step 1: Construction of user portrait, collecting and analyzing the behavioral data of users on various platforms to construct a user portrait;

[0060] Step 2: Analysis and classification of advertising materials;

[0061] Step 3: Selection and evaluation of the placement platform, selecting the corresponding placement platform and evaluating the placement platform;

[0062] Step 4: Formulation of the placement strategy, formulating a specific placement strategy according to the results of the user portrait, advertising material classification, and placement platform evaluation;

[0063] Step 5: Real-time monitoring and data collection, during the advertising placement process, real-time monitoring the placement effect of the advertisement and collecting relevant data;

[0064] Step 6: Evaluation and analysis of the placement effect, evaluating and analyzing the advertising placement effect according to the data collected by real-time monitoring;

[0065] Among them, in Step 1: User portrait construction, collect and analyze the behavioral data of users on various platforms, and the specific method of constructing the user portrait is as follows;

[0066] The user behavioral data includes but is not limited to browsing records, search history, purchase behavior, social interaction, etc.

[0067] Let the user set be U = {u1, u2, …, u n}, where n is the number of users. For each user u i , construct a feature vector P i = (p i1 , p i2 , …, p im ), where m represents the number of features; p i1 can represent the age of the user. Users of different age groups have different acceptance levels and consumption tendencies towards advertisements.

[0068] For example, young people may be more concerned about advertisements for fashion and entertainment products, while middle-aged and elderly people may be more interested in advertisements for health and life services.

[0069] p i2 represents the gender of the user. Gender differences will affect the user's preference for products. For example, cosmetic advertisements are usually more inclined to be targeted at female users.

[0070] p i3 represents the geographical information of the user. There are differences in consumption habits and market demands among users in different regions. For example, the demand for warm-keeping supplies is relatively large in the northern region in winter, while the demand for heatstroke prevention and cooling products is more prominent in the southern region in summer;

[0071] p i4 is the interest tag of the user. By analyzing the user's browsing and search records, the fields that the user is interested in are extracted, such as sports, technology, food, etc. Suppose user u i frequently browses sports news websites and purchases sports goods within a certain period of time. Then the weight of sports-related in its interest tags will be relatively high. In order to quantify the user's interest preference, an interest preference degree formula is introduced.

[0072] Let the number of times user u i browses interest tag j be b ij , the number of searches be s ij , and the number of purchases be p ij . Then the interest preference degree I i of user u ij for interest tag j is calculated by the following formula:

[0073]

[0074] Among them, α, β, and γ are weight coefficients, and α + β + γ = 1. These weight coefficients are determined according to the importance of different behaviors in reflecting the user's interests;

[0075] For example, if the purchase behavior is considered the behavior that best reflects the user's true interests, then the value of γ can be relatively large. By constructing a user portrait, the advertising delivery system can deeply understand the characteristics and interest preferences of each user.

[0076] Among them, in step two: the method for analyzing and classifying advertising materials is as follows:

[0077] Advertising materials include various forms such as pictures, videos, and texts. Analyzing and classifying advertising materials helps to push appropriate advertising materials to target users according to the user portrait.

[0078] Let the advertising material set be A = {a1, a2, …, a l}, where l is the number of advertising materials. For each advertising material a k , extract its feature vector F k = (f k1 , f k2 , …, f kn ); f k1 can represent the theme of the advertising material, such as electronic products, clothing, food, etc. Through text analysis technology, extract keywords from the advertising copy to determine the theme category of the advertisement;

[0079] f k2 represents the style of the advertising material, such as simple style, gorgeous style, cartoon style, etc. For picture and video materials, their colors, compositions, etc. can be analyzed through image recognition technology to judge the style.

[0080] f k3 represents the characteristics of the target audience group targeted by the advertising material, such as age range, gender, interest hobbies, etc. When advertisers produce advertising materials, they usually clarify their target audience, and this information can be used as one of the characteristics of the advertising material. In order to classify the advertising materials, a clustering algorithm is introduced. In this embodiment, the K-Means clustering algorithm is adopted.

[0081] Classify the advertising materials into K categories C1, C2, …, C K ; the goal of clustering is to make the advertising materials in the same category as similar as possible in terms of characteristics, and the advertising materials in different categories are quite different in terms of characteristics. In the K-Means clustering algorithm, first

[0082] randomly select K initial clustering centers O1, O2, …, O K ;

[0083] For each advertising element structure a k , calculate its distance d(a k , O j ) from each cluster center. The distance formula uses the Euclidean distance:

[0084]

[0085] where f ki is the i-th eigenvalue of the advertising material a k , and o j i is the i-th eigenvalue of the cluster center O j . The advertising material a k is assigned to the category where the nearest cluster center is located. Then, recalculate the cluster center of each category, that is, the average value of the feature vectors of all advertising materials in that category, and repeat this process continuously until the cluster center no longer changes significantly to complete the classification of advertising materials. Through the analysis and classification of advertising materials, the user portrait and advertising materials can be better matched, improving the accuracy of advertising placement.

[0086] Among them, in step three: the selection and evaluation method of the placement platform is as follows:

[0087] Currently, there are many information flow advertising placement platforms in the market, such as social media platforms, news and information platforms, video platforms, etc. Different platforms have different user group characteristics, traffic scales, and advertising placement mechanisms. Therefore, choosing a suitable placement platform is crucial for the advertising placement effect.

[0088] Let the set of placement platforms be P = {p1, p2,..., p q}, where q is the number of platforms. For each platform p j , evaluate multiple of its indicators;

[0089] The number of platform users N j represents the potential audience scale of the platform. The more the number of platform users, the greater the possible exposure opportunity of the advertisement. The user activity A j can be measured by indicators such as the average daily login times and stay time of users on the platform. For platforms with high user activity, the possibility of user interaction with the advertisement is greater. The matching degree M j of the platform and the advertising target audience is calculated by analyzing the similarity between the portrait characteristics of platform users and the portrait characteristics of the advertising target audience. Let the portrait feature vector of users be P pj = (p pj1 , pp j2 , …, p pjn ), and the portrait feature vector of the advertising target audience be P a = (p a1 , pa2 ,…,p an ), then the matching degree M j can be calculated by the following formula:

[0090]

[0091] This formula is based on the cosine similarity of vectors. The closer the value is to 1, the higher the matching degree between the platform users and the target audience of the advertisement. In order to comprehensively evaluate the placement platform,

[0092] introduce a comprehensive evaluation index E j , which is a weighted synthesis of the number of platform users, user activity, and matching degree, and is calculated as follows:

[0093]

[0094] where ω1, ω2, and ω3 are weight coefficients, and ω1 + ω2 + ω3 = 1. The weight coefficients are determined according to the importance that the advertiser attaches to different indicators. For example, if the advertiser pays more attention to the precise placement of the advertisement, then the value of ω3 can be relatively large. By selecting and evaluating the placement platform, the most suitable platform for advertisement placement can be determined, improving the efficiency and effect of advertisement placement.

[0095] Among them, in step four: When formulating the placement strategy, according to the user portrait, advertisement material classification, and the results of the placement platform evaluation, the specific method for formulating the specific placement strategy is as follows:

[0096] The placement strategy includes aspects such as the time, frequency, and budget allocation of advertisement placement.

[0097] Let the advertisement placement time interval be T = [t1, t2], and divide it into multiple time slices t s1 , t s2 ,…, t sk ;

[0098] For each time slice t si , determine the advertisement placement quantity Q si , and the control of advertisement placement frequency is to avoid users getting bored with the advertisement.

[0099] Let user u i The maximum allowed number of times to receive advertisements within a period of time be F maπi , and the actual number of times to receive advertisements be F i ; In each time slice t s i, for user u i , if F i < F maπi , then advertisements can be placed to him, and budget allocation is an important part of the placement strategy.

[0100] Let the total advertising budget be B, and allocate it to different advertising platforms and time slices;

[0101] For platform p j , in time slice t s the budget allocation for i is B jsi , and it satisfies To optimize the budget allocation, the concept of return on investment (ROI) is introduced.

[0102] Let the revenue R j brought by the advertisement on platform p si in time slice t jsi , and the cost is C jsi , then the return on investment

[0103] Through an optimization algorithm (such as a linear programming algorithm). Under the condition of satisfying the budget constraint, calculate the maximum total return on investment:

[0104]

[0105] The constraint conditions include:

[0106]

[0107] Among them, in step five: real-time monitoring and data collection. During the advertisement placement process, the method of real-time monitoring of the advertisement placement effect and data collection is as follows:

[0108] The monitored data includes the exposure volume, click-through rate, conversion rate, user stay time, etc. of the advertisement.

[0109] Let advertisement a k have an exposure volume of E j on platform p kj , a click-through rate of a conversion rate of CR kj , and a user stay time of T kj .

[0110] The click-through rate is calculated as:

[0111]

[0112] Among them, Clicks skj is the number of clicks of advertisement a k on platform p j .

[0113] The conversion rate CR kj is calculated as:

[0114]

[0115] Among them, Conversions skj is the advertisement α k on the platform p j The number of conversions (such as purchase behavior, registration behavior, etc.).

[0116] By monitoring these data in real time, it is possible to timely understand the performance of the advertisement on different platforms and provide data support for subsequent dynamic optimization.`

[0117] Signature, in step six: Evaluation and analysis of the delivery effect. According to the data collected by real-time monitoring, the specific methods for evaluating and analyzing the advertisement delivery effect are as follows:

[0118] The evaluation indicators include click-through rate, conversion rate, return on investment, etc.

[0119] Let the overall return on investment of the advertisement delivery be ROI total , which is the weighted combination of the return on investment of each platform and time slice:

[0120]

[0121] By evaluating and analyzing the delivery effect, find out the problems existing in the advertisement delivery process. For example, if the click-through rate of a certain platform is low, it may be that the advertisement material does not match the users of this platform, or the delivery time is inappropriate; if the conversion rate is low, it may be necessary to optimize the advertisement copy or product page.

[0122] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for placing an information flow advertisement module, characterized in that, Specifically, it includes the following steps: Step 1: User portrait construction. Collect and analyze the behavioral data of users on various platforms to construct a user portrait; Step 2: Advertising material analysis and classification; Step 3: Selection and evaluation of the placement platform. Select the corresponding placement platform and evaluate the placement platform; Step 4: Formulation of the placement strategy. According to the results of user portrait construction, advertising material classification, and placement platform evaluation, formulate a specific placement strategy; Step 5: Real-time monitoring and data collection. During the advertising placement process, monitor the placement effect of the advertisement in real time and collect relevant data; Step 6: Evaluation and analysis of the placement effect. According to the data collected through real-time monitoring, evaluate and analyze the placement effect of the advertisement.

2. The method for placing an information flow advertisement module according to claim 1, characterized in that, In Step 1: For user portrait construction, the specific method of collecting and analyzing the behavioral data of users on various platforms to construct a user portrait is as follows; Let the user set be U = {u1, u2, …, u n}, where n is the number of users. For each user u i , construct a feature vector P i = (p i1 , p i2 , …, p im ), where m represents the number of features; p i1 can represent the user's age, p i2 represent the user's gender; p i3 represent the user's geographical information; p i4 is the user's interest tag; Let user u i has the browsing times b for interest tag j ij , the search times s ij , and the purchase times p ij , then the interest preference degree I of user u i for interest tag j ij is calculated by the following formula: Among them, α, β, and γ are weight coefficients, and α + β + γ = 1.

3. A method for placing an information flow advertisement module according to claim 1, characterized in that, In Step 2: The method of advertising material analysis and classification is as follows: Let the set of advertising materials be \(A = \{a_1, a_2, \ldots, a\) l \}, where \(l\) is the number of advertising materials. For each advertising material \(a\) k , extract its feature vector \(F\) k = (f k1 , f k2 , \ldots, f kn ); \(f\) k1 represents the theme of the advertising material, and \(f\) k2 represents the style of the advertising material; \(f\) k3 represents the characteristics of the target audience group targeted by the advertising material; Divide the advertising materials into K categories C1, C2, …, C K ; Randomly select K initial cluster centers O1, O2, …, O K ; For each advertising element structure a k , calculate its distance d(a k , O j ) from each cluster center. The distance is calculated using the Euclidean distance, and the calculation formula is as follows: Among them, f ki is the i-th eigenvalue of the advertisement material a k , and o j i is the i-th eigenvalue of the clustering center O j .

4. A method for placing an information flow advertisement module according to claim 1, characterized in that, In Step 3: The method of selection and evaluation of the placement platform is as follows: Let the set of placement platforms be \(P = \{p_1, p_2, \ldots, p_q\}\), where \(q\) is the number of platforms. For each platform \(p_i\), q evaluate multiple metrics for it; j ​ The number of platform users N j , representing the potential audience size of the platform, and the matching degree M between the platform and the advertising target audience j ; Let the user portrait feature vector be P pj =(p pj1 , p pj2 , …, p pjn ), and the advertising target audience portrait feature vector be P a =(p a1 , p a2 , …, p an ). Then the matching degree M j is calculated by the following formula: Introduce the comprehensive evaluation index E j , which is a weighted synthesis of the number of platform users, user activity, and matching degree, and is calculated as follows: Among them, ω1, ω2, and ω3 are weight coefficients, and ω1 + ω2 + ω3 = 1.

5. A method for placing an information flow advertisement module according to claim 1, characterized in that, In Step 4: For the formulation of the placement strategy, the specific method of formulating a specific placement strategy according to the results of user portrait construction, advertising material classification, and placement platform evaluation is as follows: Let the advertising time interval be T = [t1, t2], which is divided into multiple time slices t s1 , t s2 , …, t sk ; For each time slice t si , determine the ad delivery volume Q si ; Let user u i The maximum allowed number of receiving advertisements within a period of time is F maπi , and the actual number of receiving advertisements is F i ; at each time slice t si , for user u i , if F i <F maπi , then advertisements can be delivered to it; Let the total advertising budget be B, and allocate it to different placement platforms and time slices; For platform p j , at time slice t si the budget allocation is B jsi , and it satisfies Let platform p j At time slice t si The revenue R brought by the advertisement placement jsi , and the cost is C jsi , then the return on investment Under the condition of satisfying the budget constraint, calculate the maximized total return on investment: The constraint conditions include:

6. The method for placing an information flow advertisement module according to claim 1, wherein In Step 5: For real-time monitoring and data collection, the method of monitoring the placement effect of the advertisement in real time and collecting relevant data during the advertising placement process is as follows: Set advertisement a k On platform p j The exposure volume is E kj , and the click-through rate is The conversion rate is CR kj , and the user stay time is T kj ; Click-through rate The calculation formula is as follows: Among them, Clicks skj is the number of clicks of advertisement a k on platform p j ; Conversion rate CR kj The calculation formula is as follows: Among them, Conversions skj is the number of conversions of advertisement a k on platform p j above.

7. A method for placing an information flow advertisement module according to claim 1, characterized in that, In Step 6: For the evaluation and analysis of the placement effect, the specific method of evaluating and analyzing the placement effect of the advertisement according to the data collected through real-time monitoring is as follows: Let the overall return on investment (ROI) of the advertising placement be ROI total , which is a weighted combination of the return on investment for each platform and time slice: Through the evaluation and analysis of the placement effect, find out the problems existing in the advertising placement process.

Images