Advertisement display control method and related device

Abstract

Embodiments of the present application disclose an advertisement display control method, an advertisement display control device and a computer readable storage medium, which are used for advertisement display control in the case of improving overall advertisement revenue increment. The method comprises: obtaining, for each first user, an advertisement revenue ECPM value obtainable per thousand times of display of the first user in a future period, determining a first user whose ECPM value is greater than or equal to a preset ECPM threshold as a second user, obtaining, for each second user, an increment of target advertisement exposure times, an advertisement tolerance value and a DAU probability value of the second user in the future period, establishing an integer programming model according to the ECPM value, the increment of target advertisement exposure times, the advertisement tolerance value and the DAU probability value corresponding to each second user, solving the integer programming model, determining a third user, and executing an aggressive expansion strategy on the third user.

Description

Technical Field

[0001] This application relates to the field of advertising display control, and more specifically, to advertising display control methods, advertising display control devices, and computer-readable storage media. Background Technology

[0002] Display control is a crucial aspect of advertising, determining the frequency and timing of ad presentations. Common ad display controls include the maximum number of ad displays, the minimum time interval between ad displays, and the minimum content consumption requirement for ad displays. The maximum number of ad displays refers to the maximum number of times an ad can be shown to a user in a given ad slot within a specific timeframe (e.g., one day). Once this maximum number of displays is exceeded, the ad will not be shown again. The minimum time interval between ad displays refers to the minimum time between ad displays in a given ad slot. If the time since the last ad display is less than this minimum time interval, the ad will not be shown. The minimum content consumption requirement for ad displays refers to the minimum amount of content a user must consume to display an ad in a given ad slot. If the number of views (or completions) of content since the last ad display is less than this minimum content consumption requirement, the ad will not be shown again.

[0003] Ad display control aims to balance ad revenue and user experience. On the one hand, overly conservative ad display control can severely impact ad impressions, thereby affecting ad revenue. On the other hand, overly aggressive ad display control can severely impact user experience, making users feel that ads are too numerous and frequent, affecting user retention, and ultimately impacting the business itself, as well as ad revenue in the long run. Therefore, appropriate and effective ad display control is crucial for both ad revenue and user experience.

[0004] Existing ad display control methods involve first estimating the user's future ECPM value, then segmenting users based on this estimated future ECPM value, and finally applying different ad display control strategies to different user segments. Specifically, user segmentation can be achieved by first setting a segmentation percentage (e.g., the top 15% of users by ECPM value) or a segmentation threshold (e.g., users with an ECPM value greater than 18.5), and then segmenting users based on this percentage or threshold. Applying different ad display control strategies to different user segments can involve using a more aggressive strategy for high ECPM users, such as setting a shorter minimum ad display interval (e.g., 5-10 minutes), and a more conservative strategy for users with medium to low ECPM values, such as setting a longer minimum ad display interval (e.g., 30 minutes).

[0005] However, for users with high ECPM values, implementing aggressive ad display control strategies can make users feel that ads are too frequent, which will seriously affect the user experience, user retention, and consequently the business itself. In the long run, it will also affect advertising revenue. Therefore, after ad display control, the user experience is poor and the overall increase in advertising revenue is low. Summary of the Invention

[0006] This application provides an advertising display control method, an advertising display control device, and a computer-readable storage medium, which can control advertising display while increasing overall advertising revenue.

[0007] In a first aspect, embodiments of this application provide an advertising display control method, including:

[0008] For each first user, obtain the ECPM value of the advertising revenue that the first user can obtain for every thousand impressions in future periods;

[0009] The first user whose ECPM value is greater than or equal to the preset ECPM threshold is identified as the second user;

[0010] For each second user, obtain the target ad exposure increment, ad tolerance value, and daily active user (DAU) probability value for the second user in the future period; wherein, the target ad exposure increment for the second user in the future period is the ad exposure increment after the implementation of the incentive control strategy.

[0011] An integer programming model is established based on the ECPM value, the incremental number of times the target ad can be exposed, the ad tolerance value, and the DAU probability value for each second user;

[0012] Solve the integer programming model to determine the third user, and then execute a trigger control strategy on the third user.

[0013] Optionally, the step of establishing an integer programming model based on the ECPM value corresponding to each second user, the incremental number of times the target ad can be exposed, the ad tolerance value, and the DAU probability includes:

[0014] An objective function is established based on the ECPM value corresponding to each second user, the incremental number of times the target ad can be exposed, and the DAU probability; wherein, the objective function is used to calculate the maximum value of the overall ad revenue increment;

[0015] Define decision variable x i Where i represents the second user i, and the decision variable x i The value represents whether the activation control strategy is executed for the second user i;

[0016] The target constraints are determined based on the advertising tolerance value and / or the DAU probability corresponding to each second user;

[0017] Based on the decision variable x i The integer programming model is established by defining the objective function and objective constraints.

[0018] Optionally, determining the target constraint based on the advertising tolerance value and / or the DAU probability corresponding to each second user includes:

[0019] Establish a first constraint condition, which is used to control the user coverage scale for implementing the aggressive control strategy in the future period to not exceed a first preset number; and / or

[0020] A second constraint is established to control the DAU (Daily Active Users) coverage size for daily active users in the future period from not exceeding a second preset number; wherein the DAU coverage size is determined based on the DAU probability corresponding to each second user; and / or

[0021] A third constraint is established to control the overall user advertising intolerance value from not exceeding a preset overall advertising intolerance threshold in the future; wherein the overall user advertising intolerance value is determined based on the advertising tolerance value corresponding to each second user;

[0022] The first constraint, the second constraint, and / or the third constraint are used as the target constraint.

[0023] Optionally, obtaining the incremental number of times the target advertisement can be exposed for each second user in future periods includes:

[0024] For each second user, obtain the incremental number of times the second user's first ad could be displayed in the past period;

[0025] The increment of the second user's second ad exposure in the future period is determined based on the increment of the first ad exposure in the past period.

[0026] The incremental number of times the second advertisement can be exposed for the second user is fine-tuned according to a preset fine-tuning rule to obtain the incremental number of times the target advertisement can be exposed.

[0027] Optionally, determining the increment of the second user's second ad exposures in the future based on the increment of the first ad exposures in the past period includes:

[0028] Obtain the expected value of the increase in the number of times the second user's first advertisement can be exposed in the past period;

[0029] The expected value is used as the increment of the second user's second advertisement exposure in future periods.

[0030] Optionally, the step of fine-tuning the incremental number of times the first advertisement can be exposed for the second user according to a preset fine-tuning rule to obtain the incremental number of times the target advertisement can be exposed includes:

[0031] For each second user, if the increase in the number of times the first advertisement can be exposed is greater than or equal to the preset threshold for the number of times the advertisement can be exposed, then the increase in the number of times the target advertisement can be exposed for the second user is determined as the preset threshold for the number of times the advertisement can be exposed.

[0032] If the increase in the number of times the first advertisement can be exposed for the second user is less than the preset threshold for the number of times the advertisement can be exposed, then the increase in the number of times the target advertisement can be exposed for the second user is determined as the increase in the number of times the first advertisement can be exposed.

[0033] Optionally, obtaining the advertising tolerance value of each second user in future periods includes:

[0034] For each second user, a pre-trained binary classification model is used to score the second user's advertising tolerance level to obtain the advertising tolerance value.

[0035] Optionally, before scoring the advertising tolerance level of each second user using a pre-trained binary classification model to obtain the advertising tolerance value, the method further includes:

[0036] Obtain positive and negative user samples; each positive and negative user sample is labeled with an ad tolerance value.

[0037] The positive and negative user samples are input into a binary classification model to obtain the predicted ad tolerance value output by the binary classification model.

[0038] The loss between the predicted ad tolerance value and the labeled ad tolerance value is calculated based on the regression loss function. When the loss satisfies the convergence condition, the trained binary classification model is obtained.

[0039] Optionally, obtaining the ECPM (expendable advertising revenue per thousand impressions) for each first user in future periods includes:

[0040] For each first user, obtain the ECPM value of the first user in the past period;

[0041] Obtain the expected value of the first user's ECPM value over a past period;

[0042] The expected value is used as the ECPM value for the first user in the future period.

[0043] Optionally, obtaining the probability value of daily active users (DAU) for each second user in a future period includes:

[0044] For each second user, obtain the DAU probability value of the second user in the past period;

[0045] Obtain the expected value of the DAU probability value of the second user in the past period;

[0046] The expected value is used as the probability value of the second user's DAU in the future period.

[0047] Secondly, embodiments of this application provide an advertising display control device, including:

[0048] The acquisition unit is used to acquire, for each first user, the ECPM value of advertising revenue that the first user can obtain for every thousand impressions in a future period;

[0049] The determining unit is used to identify the first user whose ECPM value is greater than or equal to a preset ECPM threshold as the second user;

[0050] The obtaining unit is further configured to obtain, for each second user, the target ad exposure increment, ad tolerance value, and daily active user (DAU) probability value for the second user in the future period; wherein, the target ad exposure increment for the second user in the future period is the ad exposure increment after the execution of the incentive control strategy.

[0051] A unit is established to build an integer programming model based on the ECPM value corresponding to each second user, the incremental number of times the target advertisement can be exposed, the advertisement tolerance value, and the DAU probability value.

[0052] The solving unit is used to solve the integer programming model, determine the third user, and execute the activation control strategy on the third user.

[0053] Thirdly, embodiments of this application provide an advertising display control device, including:

[0054] Central processing unit, memory, input / output interfaces, wired or wireless network interfaces, and power supply;

[0055] The memory is either a short-term storage memory or a persistent storage memory;

[0056] The central processing unit is configured to communicate with the memory and execute instructions in the memory to perform the aforementioned advertising display control method.

[0057] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the aforementioned advertising display control method.

[0058] Fifthly, embodiments of this application provide a computer program product containing instructions that, when run on a computer, cause the computer to execute the aforementioned advertising display control method.

[0059] As can be seen from the above technical solutions, the embodiments of this application have the following advantages: They can take into account the problems of low ad exposure increment and poor ad tolerance that may exist when implementing a stimulus control strategy for users with high ECPM values. An integer programming model can be established based on the ECPM value corresponding to each second user, the target ad exposure increment, and the ad tolerance value to determine the third user, and the stimulus control strategy can be implemented for the third user. This can improve the overall ad revenue increment after ad display control while balancing user experience. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the architecture of an advertising display control system disclosed in an embodiment of this application;

[0061] Figure 2 This is a flowchart illustrating an advertising display control method disclosed in an embodiment of this application;

[0062] Figure 3 This is a schematic diagram illustrating the calculation of the increase in the number of times an advertisement can be exposed due to the adjustment of the display control strategy disclosed in an embodiment of this application;

[0063] Figure 4 This is a schematic diagram of the structure of an advertising display control device disclosed in an embodiment of this application. Detailed Implementation

[0064] This application provides an advertising display control method, an advertising display control device, and a computer-readable storage medium for controlling advertising display while increasing overall advertising revenue.

[0065] Please see Figure 1 The architecture of the advertising display control system in this application embodiment includes:

[0066] The system includes an advertising display control device 101 and a client 102. When controlling advertising display, the advertising display control device 101 can connect to the client 102. The advertising display control device 101 can obtain target information sent by the client 102, determine a third user based on the target information, and execute a trigger control strategy on the third user.

[0067] based on Figure 1 Please refer to the advertising display control system shown. Figure 2 , Figure 2 This is a flowchart illustrating an advertising display control method disclosed in an embodiment of this application. The method includes:

[0068] 201. For each first user, obtain the ECPM value of the advertising revenue that the first user can obtain for every thousand impressions in the future period.

[0069] In this embodiment, when controlling ad display, the ECPM value of the ad revenue that the first user can obtain for each thousand displays in the future can be obtained for each first user.

[0070] 202. The first user whose ECPM value is greater than or equal to the preset ECPM threshold is identified as the second user.

[0071] For each first user, after obtaining the ECPM value of the advertising revenue that the first user can obtain per thousand impressions in the future period, the first user whose ECPM value is greater than or equal to the preset ECPM threshold can be identified as the second user.

[0072] 203. For each second user, obtain the second user's target ad exposure increment, ad tolerance value, and daily active user (DAU) probability value in the future period; wherein, the second user's target ad exposure increment in the future period is the ad exposure increment after the implementation of the incentive control strategy.

[0073] After identifying the first user whose ECPM value is greater than or equal to the preset ECPM threshold as the second user, for each second user, we can obtain the target ad exposure increment, ad tolerance value, and daily active user (DAU) probability value for the second user in the future period; wherein, the target ad exposure increment for the second user in the future period is the ad exposure increment after the implementation of the incentive control strategy.

[0074] 204. Based on the ECPM value, the incremental number of times the target ad can be exposed, the ad tolerance value, and the DAU probability value for each second user, establish an integer programming model.

[0075] For each second user, after obtaining the incremental target ad impressions, ad tolerance, and daily active user (DAU) probability for that user in future periods, an integer programming model can be established based on the corresponding ECPM, incremental target ad impressions, ad tolerance, and DAU probability for each second user. It can be understood that the method for establishing an integer programming model based on the corresponding ECPM, incremental target ad impressions, ad tolerance, and DAU probability for each second user can be as follows: first, establish an objective function based on the corresponding ECPM, incremental target ad impressions, and DAU probability for each second user; where the objective function is used to calculate the maximum increase in overall ad revenue, and then the decision variable x is set. i Where i represents the second user i, and x is the decision variable. i The value of represents whether to execute the incentive control strategy for the second user i. Then, the target constraints are determined based on the advertising tolerance value and / or DAU probability corresponding to each second user. Finally, the decision variable x is used as the basis for the determination of the target constraints. i The objective function and objective constraints can be used to establish an integer programming model, or other reasonable methods can be used, which are not limited here.

[0076] 205. Solve the integer programming model, determine the third user, and implement the activation control strategy for the third user.

[0077] After establishing an integer programming model based on the ECPM value, target ad exposure increment, ad tolerance value, and DAU probability value for each second user, the integer programming model can be solved to determine the third user, and an incentive control strategy can be implemented for the third user.

[0078] In this embodiment, the potential impact on user experience, such as low ad exposure increment and poor ad tolerance, when implementing an aggressive ad control strategy on users with high ECPM values, can be considered. An integer programming model can be established based on the ECPM value, the target ad exposure increment, and the ad tolerance value corresponding to each second user to determine the third user. An aggressive ad control strategy can then be implemented on the third user. This approach can improve the overall ad revenue increment after ad display control while balancing user experience.

[0079] In this embodiment of the application, there are various methods for establishing an integer programming model based on the ECPM value, the incremental number of target ad impressions, the ad tolerance value, and the DAU probability for each second user. Figure 2 The advertising display control methods shown are described below, and one of them is described in detail below.

[0080] In this embodiment, when controlling ad display, the ECPM value of the ad revenue that the first user can obtain for each thousand displays in the future can be obtained for each first user.

[0081] One method for obtaining the ECPM value of advertising revenue that a first user can obtain per thousand impressions in the future for each first user is to obtain the ECPM value of the first user in the past period, obtain the expected value of the ECPM value of the first user in the past period, and use the expected value as the ECPM value of the first user in the future period.

[0082] Specifically, ECPM represents the revenue generated per thousand impressions. The future period could be tomorrow, and the ECPM for all users in the future (e.g., tomorrow) can be estimated based on data statistics or machine learning. For example, using data statistics, a user's recent ECPM can be used to approximate their future ECPM; please refer to Formula 1:

[0083]

[0084] Specifically, Formula 1 It is the ECPM of user i during time period t, calculated by dividing the total ad spend of user i during time period t by the total number of ad impressions; The value of represents whether user i is a DAU during time period t, with a value of 1 indicating DAU and a value of 0 indicating no DAU; w t It refers to weights, which can be based on empirical presets or obtained through machine learning. Machine learning can supplement more features to learn and predict, such as user profiles and user behavior; TMR is an abbreviation for tomorrow.

[0085] It is understandable that, in addition to Formula 1, Formula 2 could also be used; please refer to Formula 2:

[0086]

[0087] Other reasonable methods can be used to approximate a user's future ECPM by utilizing the user's recent ECPM, but specific methods are not limited here.

[0088] For each first user, after obtaining the ECPM value of the advertising revenue that the first user can obtain per thousand impressions in the future period, the first user whose ECPM value is greater than or equal to the preset ECPM threshold can be identified as the second user.

[0089] Understandably, users with low ECPM values ​​can be identified as not being the target users for the aggressive progress control strategy and can be directly excluded. The preset ECPM threshold can be set to ECPM. LB If the user's ECPM value is lower than ECPMLB If it is not included in subsequent calculations, then it will not be included in subsequent calculations.

[0090] After identifying the first user whose ECPM value is greater than or equal to the preset ECPM threshold as the second user, for each second user, we can obtain the target ad exposure increment, ad tolerance value, and daily active user (DAU) probability value for the second user in the future period; wherein, the target ad exposure increment for the second user in the future period is the ad exposure increment after the implementation of the incentive control strategy.

[0091] One method for obtaining the incremental number of times the target ad can be displayed for each second user in the future is as follows: First, for each second user, obtain the incremental number of times the first ad can be displayed for the second user in the past period. Then, based on the incremental number of times the first ad can be displayed for the second user in the past period, determine the incremental number of times the second ad can be displayed for the second user in the future period. Finally, fine-tune the incremental number of times the second ad can be displayed for the second user according to the preset fine-tuning rules to obtain the incremental number of times the target ad can be displayed.

[0092] One method for determining the second user's second ad exposure increment in the future period based on the second user's first ad exposure increment in the past period is to first obtain the expected value of the second user's first ad exposure increment in the past period, and then use the expected value as the second user's second ad exposure increment in the future period.

[0093] Specifically, the increase in ad impressions resulting from future adjustments to display control strategies for all users can be estimated based on data statistics or machine learning. For example, using data statistics, the increase in ad impressions resulting from daily adjustments to display control strategies over a recent period can be used to approximate the increase in ad impressions resulting from future adjustments to display control strategies. Please refer to Formula 3:

[0094]

[0095] Among them, Formula 3 This refers to the increase in ad exposures resulting from adjustments to the display control strategy for user i during time period t. The calculation method can be found in [link to relevant documentation]. Figure 3 , Figure 3 This is a schematic diagram illustrating the calculation of the increase in ad exposure frequency resulting from an adjustment in the display control strategy disclosed in this application. For each user, ad placement exposure reports are collected, and the ad exposure frequency before and after the display control strategy adjustment is calculated to obtain the increase in ad exposure frequency resulting from the adjustment of the display control strategy. Figure 3It is known that before the display control strategy was adjusted, the minimum time interval for ad display was 30 minutes. For example, the times when an ad slot could be displayed but the ad could not be shown were 09:03, 09:21, 09:37, 09:45, 09:55, 10:10, 11:35, and 11:52. The times when an ad slot could be displayed and the ad could be shown were 09:03, 09:37, 10:10, and 11:35. That is, when a user opened the target application (such as Douyin) at eight times, the ad could be shown at four of those times. In other words, before the display control strategy was adjusted, the corresponding number of ad exposures was 4. After adjusting the display control strategy, for example, by adjusting the minimum time interval for ad display to 15 minutes, the times when the ad slot is visible but not yet exposed can still be 09:03, 09:21, 09:37, 09:45, 09:55, 10:10, 11:35, and 11:52. However, the times when the ad slot is visible and exposed can be 09:03, 09:21, 09:37, 09:55, 10:10, 11:35, and 11:52. This means that if a user opens the target application (such as TikTok) at any of the eight times, the ad will be visible (displayed) at seven of those times. Therefore, the ad exposure count after the display control strategy adjustment is 7 times. Thus, the increase in the number of ad exposures for the user due to the display control strategy adjustment is 7 - 4 = 3 times. In Formula 3, T represents the time range under consideration; v t These are weights, which can be based on empirical presets or obtained through machine learning. Machine learning can supplement more features to learn and predict, such as user profiles and user behavior, etc., but we will not limit them here.

[0096] The method for fine-tuning the incremental number of times the first advertisement can be exposed for the second user according to a preset fine-tuning rule to obtain the incremental number of times the target advertisement can be as follows: for each second user, if the incremental number of times the first advertisement can be exposed for the second user is greater than or equal to the preset threshold for the incremental number of times the advertisement can be exposed, then the incremental number of times the target advertisement can be exposed for the second user is determined as the preset threshold for the incremental number of times the advertisement can be exposed; if the incremental number of times the first advertisement can be exposed for the second user is less than the preset threshold for the incremental number of times the advertisement can be exposed, then the incremental number of times the target advertisement can be exposed for the second user is determined as the incremental number of times the first advertisement can be exposed.

[0097] Specifically, the increase in user ad impressions resulting from the adjustment of the display control strategy needs to be fine-tuned after calculation. The starting point for fine-tuning is to avoid an excessive increase in user ad impressions leading to the selection of users with low to medium ECPM. By setting an upper limit on the increase in ad impressions resulting from the adjustment of the display control strategy, the fine-tuning of the increase in user ad impressions resulting from the adjustment of the display control strategy can be referred to Formula 4:

[0098]

[0099] Specifically, in Formula 4 To fine-tune according to preset rules and obtain the incremental number of times the target ad can be exposed, ΔN UB This is the threshold for the number of times an ad can be displayed.

[0100] One method to obtain the advertising tolerance value of each second user in the future period is to use a pre-trained binary classification model to score the advertising tolerance level of each second user and obtain the advertising tolerance value.

[0101] Specifically, for each second user, a pre-trained binary classification model is used to score the user's ad tolerance level. Before obtaining the ad tolerance value, the binary classification model can be trained first. The training method can be as follows: first, obtain positive and negative user samples; each positive and negative user sample is labeled with an ad tolerance value. Then, input the positive and negative user samples into the binary classification model to obtain the predicted ad tolerance value output by the model. Finally, calculate the loss between the predicted ad tolerance value and the labeled ad tolerance value based on a regression loss function. When the loss meets the convergence condition, the trained binary classification model is obtained.

[0102] Specifically, ad tolerance for all users can be assessed based on statistical data or machine learning. Taking machine learning as an example, users who have received ads and experienced strong negative feedback (such as customer complaints) and whose activity has significantly decreased or even churned are collected as negative samples. Users who have received ads and are consistently active are collected as positive samples. These samples are then combined with user profiles, user behavior, and other features to train and validate a binary classification model. This model is then used to score all users, representing their ad tolerance. It's understood that for ad tolerance, the closer the value is to 1, the stronger the tolerance; the closer it is to 0, the weaker the tolerance. Similarly, ad intolerance can be defined. For ad intolerance, the closer the value is to 1, the weaker the tolerance; the closer it is to 0, the stronger the tolerance. It's also understood that users with high ad intolerance should not be subject to aggressive performance control strategies to avoid decreased activity or churn. Therefore, an upper limit for ad intolerance can be set for each user. If a user's ad intolerance exceeds this limit, they are directly excluded from subsequent calculations.

[0103] One method for obtaining the probability value of daily active users (DAU) for each second user in the future period is to first obtain the probability value of the second user's DAU in the past period, then obtain the expected value of the probability value of the second user's DAU in the past period, and finally use the expected value as the probability value of the second user's DAU in the future period.

[0104] Specifically, the probability of all users becoming DAU in the future (e.g., tomorrow) can be estimated based on data statistics or machine learning. For example, using data statistics, the probability of a user becoming a DAU in the future can be predicted by checking whether they have been a DAU each day recently. Please refer to Formula 5:

[0105]

[0106] Specifically, u in Formula 5 t These are weights, which can be based on empirical presets or obtained through machine learning. Machine learning can supplement these weights with more features to learn and predict, such as user profiles and user behavior.

[0107] For each second user, after obtaining the incremental number of target ad impressions, ad tolerance value, and daily active user (DAU) probability value for the second user in the future period, an integer programming model can be established based on the corresponding ECPM value, incremental number of target ad impressions, ad tolerance value, and DAU probability value for each second user.

[0108] One method for establishing an integer programming model based on the ECPM value, the incremental number of target ad impressions, the ad tolerance value, and the DAU probability for each second user is to first establish an objective function based on the ECPM value, the incremental number of target ad impressions, and the DAU probability for each second user. The objective function is used to calculate the maximum increase in overall ad revenue, and then the decision variable x is set. i Where i represents the second user i, and x is the decision variable. i The value of represents whether to execute the activation control strategy for the second user i. Then, the target constraints are determined based on the advertising tolerance value and / or DAU probability corresponding to each second user. Finally, based on the decision variable x... i Establish an integer programming model based on the objective function and objective constraints.

[0109] The method for determining the target constraint based on the advertising tolerance value and / or DAU probability corresponding to each second user can be as follows: First, establish a first constraint to control the user coverage scale for implementing the aggressive control strategy in the future period to not exceed a first preset number; and / or establish a second constraint to control the DAU daily active user coverage scale for daily active users in the future period to not exceed a second preset number; wherein the DAU daily active user coverage scale is determined based on the DAU probability corresponding to each second user; and / or establish a third constraint to control the overall user advertising intolerance value in the future period to not exceed a preset overall advertising intolerance threshold; wherein the overall user advertising intolerance value is determined based on the advertising tolerance value corresponding to each second user, and then use the first constraint, the second constraint, and / or the third constraint as the target constraint.

[0110] Specifically, the following is a concrete example of the established integer programming model.

[0111] The first preset quantity can be set to N. count The second preset quantity N dau and preset overall advertising intolerance threshold A 0-1 programming model can be established by comprehensively considering the user's future ECPM value, the increase in the number of ad impressions the user can have in the future due to the adjustment of the incentive control strategy, the user's ad tolerance value, and the user's future DAU probability value. This model includes the following decision variables, objective function, and objective constraints (first constraint, second constraint, and third constraint):

[0112] Wherein, the decision variable is x i x i =1 indicates that an incentive control policy is applied to user i, x i =0 indicates that no incentive control policy is applied to user i;

[0113] The objective function is used to calculate the maximum increase in overall advertising revenue; please refer to Formula Six:

[0114]

[0115] The first constraint is to ensure that the user coverage scale for implementing the aggressive control strategy in the future does not exceed a first preset number N. count At the same time, users with low DAU probability are restricted from being selected:

[0116]

[0117] The second constraint is to ensure that the daily active users (DAU) coverage of the targeted development control strategy in the future does not exceed a second preset number N.dau To more precisely control the actual impact of the stimulus control strategy:

[0118]

[0119] The third constraint is to ensure that the overall advertising intolerance value of users who are determined to implement the aggressive progress control strategy in the future does not exceed a preset overall advertising intolerance threshold.

[0120]

[0121] After establishing an integer programming model based on the ECPM value, target ad exposure increment, ad tolerance value, and DAU probability value for each second user, the integer programming model can be solved to determine the third user, and an incentive control strategy can be implemented for the third user.

[0122] For example, if there are 200 million users in total, and the ECPM value is less than 3 yuan, then some users are removed. If the ad tolerance value is less than 0.15, then another group is removed (corresponding to the third constraint), leaving 150 million users. Of these 150 million users, only 20% can execute the trigger-based progress control strategy under their own self-defined constraints, equivalent to N... count The constraint is 30 million users (corresponding to the first constraint); however, whether these 30 million users will come or not, another constraint can be set to allow 1 / 3 of the users to come, i.e., 1,000 users (corresponding to the second constraint).

[0123] Understandably, solving this integer programming model yields a user selection scheme for the stimulus control strategy, which can improve advertising revenue while balancing user experience.

[0124] It's worth noting that, besides the ad display control method described above that applies aggressive display control strategies to users with ECPM values ​​greater than or equal to a preset ECPM threshold, another approach is to apply a more conservative display control strategy to users with low ECPM values ​​and a more aggressive strategy to users with medium to high ECPM values. Specifically, for most users with low ECPM values, the ad revenue per unit impression is low. Applying a more aggressive display control strategy to these users would not significantly improve overall ad revenue but would instead negatively impact the user experience for most. Therefore, a more conservative display control strategy can be used for most users. Conversely, for a small group of users with high ECPM values, the ad revenue per unit impression is high. Applying a more aggressive display control strategy to these users would significantly improve overall ad revenue while having a limited impact on user experience. Therefore, a more aggressive display control strategy can be used for this smaller group of users.

[0125] It's also worth mentioning that, in this embodiment, in addition to considering the user's ECPM value, the important factor of the increase in ad exposure frequency resulting from adjustments to the display control strategy is also taken into account. Specifically, even if a user's ECPM value is high, if the increase in ad exposure frequency resulting from adjustments to the display control strategy is low, adopting an aggressive display control strategy will not bring a significant increase in advertising revenue. Therefore, by comprehensively considering the user's ECPM value and the increase in ad exposure frequency resulting from adjustments to the display control strategy, the overall increase in advertising revenue after ad display control can be improved while balancing user experience. Secondly, the important factor of user ad tolerance is also considered. Specifically, for some users with poor ad tolerance, applying an aggressive display control strategy to these users may lead to a decrease in user activity or even churn. Therefore, when faced with the choice between selecting one user with high ad value but low ad tolerance or selecting two users with medium ad value but high ad tolerance, selecting two users with medium ad value but high ad tolerance can improve the overall increase in advertising revenue after ad display control while balancing user experience. Furthermore, currently, the reach of aggressive advertising strategies is determined by the user ECPM stratification ratio or stratification threshold. However, the actual impact of aggressive advertising strategies can fluctuate significantly. For example, an increase or overestimation of a user's ECPM value may lead to more users being targeted by aggressive advertising strategies. Alternatively, even if the user ECPM stratification ratio remains constant, significant fluctuations in the DAU proportion of high-ECPM user strata can significantly impact the actual reach of aggressive advertising strategies. These factors can negatively affect the stability of the user experience. This application's embodiments also consider the user's DAU probability value, which can improve the accuracy of controlling the actual reach of aggressive advertising strategies. Therefore, while maintaining a stable and balanced user experience, it can maximize the overall increase in advertising revenue.

[0126] It is understandable that, in addition to the method described above for establishing an integer programming model based on the ECPM value, target ad impression increment, ad tolerance value, and DAU probability for each second user; in addition to the method described above for determining target constraints based on the ad tolerance value and / or DAU probability for each second user; in addition to the method described above for obtaining the target ad impression increment for each second user in future periods; in addition to the method described above for determining the second ad impression increment for the second user in future periods based on the first ad impression increment for the second user in past periods; in addition to the method described above for fine-tuning the first ad impression increment for the second user according to a preset fine-tuning rule to obtain the target ad impression increment; in addition to the method described above for obtaining the ad tolerance value for each second user in future periods; in addition to the method described above for training a binary classification model; in addition to the method described above for obtaining the ECPM value of ad revenue that a first user can obtain per thousand impressions in future periods; in addition to the method described above for obtaining the DAU probability value for each second user in future periods; other reasonable methods may also be used, and specific methods are not limited here.

[0127] In this embodiment, the potential impact on user experience when implementing a targeted ad display control strategy on users with high ECPM values, such as lower incremental ad exposures and poor ad tolerance, can be addressed by establishing an integer programming model based on the ECPM value, the incremental target ad exposures, and the ad tolerance value for each second user. This model determines the third user, and the targeted ad display control strategy is then implemented on this third user. This approach can improve the overall ad revenue increase after ad display control while maintaining a balanced user experience. Furthermore, introducing the user's DAU probability value can improve the accuracy of controlling the actual impact of the targeted ad display control strategy, thereby maximizing the overall ad revenue increase while maintaining a balanced user experience.

[0128] The advertising display control method in the embodiments of this application has been described above. The advertising display control device in the embodiments of this application is described below. Please refer to [link / reference]. Figure 2 One embodiment of the advertising display control device in this application includes:

[0129] The acquisition unit is used to acquire, for each first user, the ECPM value of advertising revenue that the first user can obtain for every thousand impressions in a future period;

[0130] The determining unit is used to identify the first user whose ECPM value is greater than or equal to a preset ECPM threshold as the second user;

[0131] The obtaining unit is further configured to obtain, for each second user, the target ad exposure increment, ad tolerance value, and daily active user (DAU) probability value for the second user in the future period; wherein, the target ad exposure increment for the second user in the future period is the ad exposure increment after the execution of the incentive control strategy.

[0132] A unit is established to build an integer programming model based on the ECPM value corresponding to each second user, the incremental number of times the target advertisement can be exposed, the advertisement tolerance value, and the DAU probability value.

[0133] The solving unit is used to solve the integer programming model, determine the third user, and execute the activation control strategy on the third user.

[0134] In this embodiment, the potential impact on user experience, such as low ad exposure increment and poor ad tolerance, when implementing an aggressive ad control strategy on users with high ECPM values, can be considered. An integer programming model can be established based on the ECPM value, the target ad exposure increment, and the ad tolerance value corresponding to each second user to determine the third user. An aggressive ad control strategy can then be implemented on the third user. This approach can improve the overall ad revenue increment after ad display control while balancing user experience.

[0135] The advertising display control device in the embodiments of this application is described in detail below. Please refer to [link / reference]. Figure 3 Another embodiment of the advertising display control device in this application includes:

[0136] The acquisition unit is used to acquire, for each first user, the ECPM value of advertising revenue that the first user can obtain for every thousand impressions in a future period;

[0137] The determining unit is used to identify the first user whose ECPM value is greater than or equal to a preset ECPM threshold as the second user;

[0138] The obtaining unit is further configured to obtain, for each second user, the target ad exposure increment, ad tolerance value, and daily active user (DAU) probability value for the second user in the future period; wherein, the target ad exposure increment for the second user in the future period is the ad exposure increment after the execution of the incentive control strategy.

[0139] A unit is established to build an integer programming model based on the ECPM value corresponding to each second user, the incremental number of times the target advertisement can be exposed, the advertisement tolerance value, and the DAU probability value.

[0140] The solving unit is used to solve the integer programming model, determine the third user, and execute the activation control strategy on the third user.

[0141] The establishment unit is specifically used to establish an objective function based on the ECPM value corresponding to each second user, the incremental number of times the target advertisement can be exposed, and the DAU probability; wherein, the objective function is used to calculate the maximum value of the overall advertising revenue increment, and the decision variable x is set. i Where i represents the second user i, and the decision variable x i The value of represents whether to execute the stimulus control strategy for the second user i. Target constraints are determined based on the advertising tolerance value and / or the DAU probability corresponding to each second user, and based on the decision variable x. i The integer programming model is established by defining the objective function and objective constraints.

[0142] The establishing unit is specifically used to establish a first constraint condition, which is used to control the user coverage scale for implementing the incentive control strategy in the future period to not exceed a first preset number; and / or establish a second constraint condition, which is used to control the DAU daily active user coverage scale for daily active users in the future period to not exceed a second preset number; wherein the DAU daily active user coverage scale is determined based on the DAU probability corresponding to each second user; and / or establish a third constraint condition, which is used to control the overall user advertising intolerance value in the future period to not exceed a preset overall advertising intolerance threshold; wherein the overall user advertising intolerance value is determined based on the advertising tolerance value corresponding to each second user, and the first constraint condition, the second constraint condition, and / or the third constraint condition are used as the target constraint condition.

[0143] The obtaining unit is specifically used to obtain, for each second user, the increment of the first ad exposure count of the second user in the past period, determine the increment of the second ad exposure count of the second user in the future period based on the increment of the first ad exposure count of the second user in the past period, and fine-tune the increment of the second ad exposure count of the second user according to a preset fine-tuning rule to obtain the increment of the target ad exposure count.

[0144] The obtaining unit is specifically used to obtain the expected value of the increment of the first ad exposure times of the second user in the past period, and use the expected value as the increment of the second ad exposure times of the second user in the future period.

[0145] The obtaining unit is specifically configured to, for each second user, if the increase in the number of times the first advertisement can be exposed for the second user is greater than or equal to a preset threshold for the number of times the first advertisement can be exposed, then determine the increase in the number of times the target advertisement can be exposed for the second user as the preset threshold for the number of times the target advertisement can be exposed; if the increase in the number of times the first advertisement can be exposed for the second user is less than the preset threshold for the number of times the target advertisement can be exposed, then determine the increase in the number of times the target advertisement can be exposed for the second user as the increase in the number of times the first advertisement can be exposed.

[0146] The obtaining unit is specifically used to score the advertising tolerance level of each second user using a pre-trained binary classification model to obtain the advertising tolerance value.

[0147] The advertising display control device also includes an input unit and a calculation unit;

[0148] The obtaining unit is also used to obtain positive user samples and negative user samples; the positive user samples and negative user samples are respectively labeled with advertising tolerance values;

[0149] The input unit is used to input the positive user samples and negative user samples into the binary classification model to obtain the predicted advertising tolerance value output by the binary classification model.

[0150] The calculation unit is used to calculate the loss between the predicted ad tolerance value and the labeled ad tolerance value according to the regression loss function. When the loss meets the convergence condition, the trained binary classification model is obtained.

[0151] The obtaining unit is specifically configured to, for each first user, obtain the ECPM value of the first user in a past period, obtain the expected value of the ECPM value of the first user in a past period, and use the expected value as the ECPM value of the first user in a future period.

[0152] The obtaining unit is specifically used to obtain, for each second user, the DAU probability value of the second user in the past period, obtain the expected value of the DAU probability value of the second user in the past period, and use the expected value as the DAU probability value of the second user in the future period.

[0153] In this embodiment, each unit in the advertising display control device performs as described above. Figure 2 The operation of the advertising display control device in the illustrated embodiment will not be described in detail here.

[0154] Please refer to the following: Figure 4 Another embodiment of the advertising display control device 400 in this application includes:

[0155] Central processing unit 401, memory 405, input / output interface 404, wired or wireless network interface 403, and power supply 402;

[0156] Memory 405 is either a short-term storage memory or a persistent storage memory;

[0157] The central processing unit 401 is configured to communicate with the memory 405 and execute instructions stored in the memory 405 to perform the aforementioned operations. Figure 2 The method in the illustrated embodiment.

[0158] This application also provides a computer-readable storage medium, which includes instructions that, when executed on a computer, cause the computer to perform the aforementioned actions. Figure 2 The method in the illustrated embodiment.

[0159] This application also provides a computer program product containing instructions, which, when run on a computer, causes the computer to perform the aforementioned... Figure 2 The method in the illustrated embodiment.

[0160] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0161] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0162] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0163] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0164] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0165] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. An advertisement display control method characterized by comprising: The method comprises the following steps: obtaining, for each first user, an ECPM value available for each thousand times of display of the first user in a future period; determining, as a second user, the first user whose ECPM value is greater than or equal to a preset ECPM threshold value; obtaining, for each second user, an increment of target ad exposure times, an ad tolerance value and a DAU probability value of the second user in a future period; wherein the increment of target ad exposure times of the second user in the future period is an increment of ad exposure times after performing an aggressive expansion control strategy adjustment; establishing an integer programming model according to the ECPM value, the increment of target ad exposure times, the ad tolerance value and the DAU probability value corresponding to each second user; solving the integer programming model to determine a third user and performing an aggressive expansion control strategy on the third user.

2. The method of claim 1, wherein, The step of establishing an integer programming model according to the ECPM value, the increment of target ad exposure times, the ad tolerance value and the DAU probability value corresponding to each second user comprises the following steps: establishing a target function according to the ECPM value, the increment of target ad exposure times and the DAU probability value corresponding to each second user; wherein the target function is used to calculate a maximum value of an overall ad revenue increment; Setting a decision variable ; wherein, for indicating a second user , the decision variable has a value representing whether to execute an aggressive prosecution strategy for the second user ; determining a target constraint condition according to the ad tolerance value and / or the DAU probability value corresponding to each second user; based on the decision variables , an objective function and objective constraints to establish the integer programming model.

3. The method of claim 2, wherein, The step of determining a target constraint condition according to the ad tolerance value and / or the DAU probability value corresponding to each second user comprises the following steps: establishing a first constraint condition for controlling that the user coverage scale determined to perform an aggressive expansion control strategy in a future period does not exceed a first preset number; and / or establishing a second constraint condition for controlling that a DAU coverage scale of daily active users in a future period does not exceed a second preset number; wherein the DAU coverage scale is determined according to the DAU probability value corresponding to each second user; and / or establishing a third constraint condition for controlling that a user overall ad intolerance value in a future period does not exceed a preset overall ad intolerance threshold value; wherein the user overall ad intolerance value is determined according to the ad tolerance value corresponding to each second user; taking the first constraint condition, the second constraint condition and / or the third constraint condition as the target constraint condition.

4. The method of claim 1, wherein, The step of obtaining, for each second user, an increment of target ad exposure times of the second user in a future period comprises the following steps: obtaining, for each second user, an increment of first ad exposure times of the second user in a past period; determining an increment of second ad exposure times of the second user in a future period based on the increment of first ad exposure times of the second user in a past period; adjusting the increment of second ad exposure times corresponding to the second user according to a preset fine-tuning rule to obtain the increment of target ad exposure times.

5. The method of claim 4, wherein, The determining the second advertisement exposure times increment of the second user in the future period based on the first advertisement exposure times increment of the second user in the past period comprises: obtaining an expected value of the first advertisement exposure times increment of the second user in the past period; using the expected value as the second advertisement exposure times increment of the second user in the future period.

6. The method of claim 4, wherein, The fine-tuning the second advertisement exposure times increment of the second user according to the preset fine-tuning rule to obtain the target advertisement exposure times increment comprises: for each second user, if the second advertisement exposure times increment of the second user is greater than or equal to a preset advertisement exposure times increment threshold, determining the target advertisement exposure times increment of the second user as the preset advertisement exposure times increment threshold; if the second advertisement exposure times increment of the second user is less than the preset advertisement exposure times increment threshold, determining the target advertisement exposure times increment of the second user as the second advertisement exposure times increment.

7. The method of claim 1, wherein, The obtaining, for each second user, the advertisement tolerance value of the second user in the future period comprises: for each second user, scoring the advertisement tolerance degree of the second user by using a pre-trained binary classification model to obtain the advertisement tolerance value.

8. The method of claim 7, wherein, Before the scoring, for each second user, the advertisement tolerance degree of the second user by using a pre-trained binary classification model to obtain the advertisement tolerance value, the method further comprises: obtaining user positive samples and user negative samples, wherein the user positive samples and the user negative samples are respectively labeled with an advertisement tolerance value; inputting the user positive samples and the user negative samples into a binary classification model to obtain a predicted advertisement tolerance value output by the binary classification model; calculating a loss between the predicted advertisement tolerance value and the labeled advertisement tolerance value according to a regression loss function, and obtaining a trained binary classification model when the loss meets a convergence condition.

9. The method of claim 1, wherein, The obtaining, for each first user, an ECPM value of advertisement revenue obtainable per thousand displays of the first user in the future period comprises: for each first user, obtaining an ECPM value of the first user in the past period; obtaining an expected value of the ECPM value of the first user in the past period; using the expected value as the ECPM value of the first user in the future period.

10. The method of claim 1, wherein, The obtaining, for each second user, a DAU probability value of the second user in the future period comprises: for each second user, obtaining a DAU probability value of the second user in the past period; obtaining an expected value of the DAU probability value of the second user in the past period; using the expected value as the DAU probability value of the second user in the future period.

11. An advertisement display control device characterized by comprising: comprises: a central processing unit, a memory, an input and output interface, a wired or wireless network interface, and a power supply; the memory is a volatile memory or a persistent storage memory; the central processing unit is configured to communicate with the memory and execute instruction operations in the memory to perform the method in any one of claims 1 to 10.

12. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises instructions which, when run on a computer, cause the computer to perform the method of any one of claims 1 to 10.

Images