A budget allocation and rhythm control method based on delivery value feedback

Abstract

This invention relates to the field of programmatic advertising technology. More specifically, it relates to a budget allocation and pacing control method based on campaign value feedback. The method first divides the advertising campaign period into multiple time intervals according to a preset timeframe; it obtains the initial target spending for each time interval and confirms the actual spending and campaign result value of the ads already placed within each time interval, constructing a campaign time-series dataset; it calculates the unit budget value efficiency based on the campaign time-series dataset, constructs a relative score for each time interval, and constructs a target spending curve based on the relative score; it obtains the actual cumulative spending for each time interval to calculate the cumulative spending deviation, and determines the update direction of the campaign adjustment factor based on the sign direction of the cumulative spending deviation; based on the collected and calculated change in the ratio between actual spending and target spending and the change in the campaign adjustment factor, it calculates effectiveness evaluation data.

Description

Technical Field

[0001] This invention relates to the field of programmatic advertising technology. More specifically, this invention relates to a method for budget allocation and pacing control based on campaign value feedback. Background Technology

[0002] Programmatic advertising systems utilize big data and other technologies to integrate information on digital media advertising placement. Through the collaborative operation of core components such as demand-side platforms, supply-side platforms, and ad exchanges, they automate the purchase and sale of ad space. Essentially, this system acts as an automated transaction hub connecting advertisers and media resources, completing the entire process—from sending ad requests and making bidding decisions among multiple parties to determining the winner—within milliseconds. For advertisers, its value lies in improving campaign efficiency and cost control, as well as providing closed-loop optimization support for personalized marketing. The core of a programmatic advertising system is its budget allocation and campaign timing control mechanism.

[0003] In existing technologies, the core consumption curve of the aforementioned systems is typically constructed using methods such as uniform allocation, static allocation based on traffic prediction, or static allocation based on historical performance. After the target consumption curve is determined, the actual consumption is then adjusted to match the target curve through mechanisms such as adjusting the bid multiplier and exposure probability. However, the aforementioned technologies have the following shortcomings: First, budget allocation relies on external prediction models and lacks adaptive capabilities based on user profiles. Taking the Chinese patent application CN121526712A as an example, this technology primarily focuses on audience targeting, and its budget allocation strategy still relies on preset rules or traffic prediction models. However, when user behavior patterns change or the market competition environment fluctuates, budget allocation struggles to respond in real time, leading to missed opportunities to reach high-value users and failing to achieve accurate matching between budget and user value. Second, the campaign pace control lacks a mechanism for judging the effectiveness of execution; parameters rely on manual settings, easily leading to ineffective adjustments. Third, budget allocation and campaign control are disconnected; evaluation results are not effectively fed back to the budget allocation module, preventing the budget allocation strategy from being continuously optimized based on actual conversion value. Summary of the Invention

[0004] To address the aforementioned technical problems, such as the disconnect between budget allocation and delivery control, and the lack of a closed-loop feedback mechanism based on the value of delivery results, this invention discloses a budget allocation and pace control method based on delivery value feedback.

[0005] In a first aspect, this invention discloses a budget allocation and pace control method based on deployment value feedback, comprising: The advertising campaign is divided into multiple time intervals according to a preset advertising cycle; Obtain the initial target spending for multiple time intervals, and confirm the actual spending and value of the ads delivered within the time intervals to construct a time-series dataset for ad delivery. The unit budget value efficiency is calculated based on the aforementioned time series dataset, a relative score is constructed for each time interval, and a target consumption curve is constructed based on the relative score. Obtain the actual cumulative consumption for each time interval to calculate the cumulative consumption deviation, and determine the update direction of the delivery adjustment factor based on the sign direction of the cumulative consumption deviation; Based on the change in the ratio between actual consumption and target consumption and the change in the delivery adjustment factor obtained from the data collection and calculation, the effectiveness evaluation data is calculated. The sign direction of the cumulative consumption deviation is extracted, and the update magnitude of the delivery adjustment factor is determined in combination with the preset delivery accessibility index. The delivery adjustment factor is updated based on the update direction and the update magnitude to generate an advertising target consumption curve, thereby constructing a budget delivery control strategy and sending it to the advertising delivery system for execution.

[0006] The beneficial effects of this invention are as follows: First, it collects actual consumption and campaign result value data (such as conversions, order amounts, etc.) for each time interval to construct a time-series dataset, providing a fine-grained data foundation. Second, it calculates the unit and full-cycle average value efficiency to generate a normalized relative score, eliminating interference from budget scale differences and achieving value comparability across intervals. Third, it uses the relative score to adaptively correct the initial target consumption curve, and after normalization, ensures total budget constraints, feeding back historical effects to budget decisions in real time, forming a positive data flow. Furthermore, it calculates the cumulative consumption deviation and jointly analyzes the changes in the campaign adjustment factor and consumption ratio to generate campaign accessibility and effectiveness indicators, reflecting execution efficiency and identifying the impact of the market environment, providing support for the adjustment factor update range. Finally, it determines the update range based on the direction of the cumulative consumption deviation and the campaign accessibility indicators, adaptively updates the adjustment factor, and generates and distributes control strategies to the campaign system. The continuous data flow throughout the entire process drives the online update of the campaign adjustment factor, breaking down data silos in budget allocation and execution control, and significantly improving the system's dynamic environmental adaptability.

[0007] Preferably, constructing the deployment time-series dataset includes: Obtain the advertising campaign period and total advertising budget; The advertising campaign period is divided into intervals according to a preset time granularity to obtain multiple time intervals. Each time interval is a preset time granularity, resulting in a set of time intervals. The total advertising budget is evenly distributed across the time interval set, and the initial target expenditure for the time interval set is calculated to obtain the initial target expenditure curve data. Collect actual consumption data of the advertising system within the time interval, record the amount of budget consumed at the end of each time interval, and obtain actual consumption time series data; Collect campaign result value data within each time interval to generate campaign result value time series data; wherein, the campaign result value data includes at least one or a weighted combination of conversion number, conversion value, and order amount; The actual consumption time-series data and the delivery result value time-series data are timestamped and integrated according to time intervals to obtain the delivery time-series dataset.

[0008] Preferably, constructing a target consumption curve based on the relative score includes: The ratio of the value of the delivery results to the actual consumption is calculated for each time interval to obtain the unit budget value efficiency sequence; the unit budget value efficiency is used to quantify the average business value that each unit budget can generate within the time interval. The total actual consumption is obtained by summing the actual consumption over the time interval, and the total value of the campaign results is obtained by summing the value of the campaign results over the time interval. The ratio of the total value of the campaign results to the total actual consumption is calculated to obtain the average value efficiency over the entire cycle. The average value efficiency over the entire cycle is used to quantify the average business value generated by each unit of budget during the entire campaign cycle. The relative score sequence is constructed by calculating the ratio of the unit budget value efficiency of each time interval to the average value efficiency over the entire period. Adjust the target consumption based on the initial target consumption and relative score for each time interval to obtain the updated target consumption; The total update target consumption is obtained by summing the updated target consumption for each time interval. The ratio of the total budget to the total update target consumption is calculated to generate a normalization coefficient and construct the target consumption curve.

[0009] Preferably, determining the update direction of the delivery adjustment factor based on the sign direction of the cumulative consumption deviation includes: Initialize the delivery adjustment factor to its default value; Obtain the actual cumulative consumption and target cumulative consumption for each time interval; The actual consumption in each time interval is summed to obtain the first consumption sequence data; The cumulative consumption of the target in each time interval is summed to obtain the second consumption sequence data; The cumulative difference between the first consumption sequence data and the second consumption sequence data is calculated to obtain the cumulative consumption deviation for each time interval; The sign direction of the cumulative consumption deviation in each time interval is extracted to obtain the sign direction of the cumulative consumption deviation, which is used as the update direction of the delivery adjustment factor.

[0010] Preferably, the sign direction of the cumulative consumption deviation for each time interval is extracted, including: If the cumulative consumption deviation is greater than zero, the sign direction is positive; if the cumulative consumption deviation is less than zero, the sign direction is negative; if the cumulative consumption deviation is equal to zero, the sign direction is zero.

[0011] Preferably, effectiveness evaluation data is calculated based on the change in the ratio between actual consumption and target consumption and the change in the dosage adjustment factor obtained from the data collection and calculation, including: Obtain the delivery adjustment factor for adjacent time intervals, calculate the difference between the delivery adjustment factor for the next time interval and the delivery adjustment factor for the previous time interval, and obtain the change in delivery adjustment factor. The ratio of actual consumption to target consumption is calculated by adding a very small positive number to each time interval, and the difference in consumption ratio between adjacent time intervals is also calculated to obtain the target consumption ratio. Based on the changes in the adjustment factors and the ratios of the deployment, a deployment accessibility index is constructed to obtain data for evaluating the effectiveness of deployment adjustment.

[0012] Preferably, the deployment accessibility index constructed based on the change in the deployment adjustment factor and the change in the ratio includes: The absolute value of the change in the delivery adjustment factor is calculated to the maximum value and compared with the preset minimum delivery adjustment range. The larger value is taken as the denominator, and the change in the ratio of actual consumption to target consumption is taken as the numerator. The ratio is calculated, and the calculation result is cropped so that its value range is between 0 and 1 to obtain the original delivery accessibility index data for each time interval. The original delivery accessibility index sequence is subjected to exponential smoothing. A preset smoothing coefficient is introduced, and the original delivery accessibility index of the current time interval is weighted and summed with the smoothed delivery accessibility index of the previous time interval to obtain the delivery accessibility index. When the accessibility index is lower than the preset limit threshold, the update range of the delivery adjustment factor is limited or the adjustment is temporarily suspended.

[0013] Preferably, the original delivery accessibility index data and the calculation formula for the delivery accessibility index include: The formula for calculating the original accessibility index data is as follows:

[0014] In the formula, This represents the change in the ratio of actual consumption to target consumption. To adjust the change in the adjustment factor; This is the preset minimum adjustment range for the delivery; This means cropping the calculation results to the range of 0 to 1; This is the original data on the accessibility of the delivery system.

[0015] Preferably, the generated ad target consumption curve includes: Obtain the total budget data; calculate the ratio of the cumulative consumption deviation for each time interval to the total budget data to obtain the normalized cumulative consumption deviation; The update magnitude of the delivery adjustment factor is determined by using the cumulative consumption deviation and delivery accessibility indicators, and the increase or decrease trend of the delivery adjustment factor is adjusted in combination with the update direction. The difference between the current time interval's delivery adjustment factor and the update magnitude and cumulative consumption deviation sign direction is calculated, and the calculation result is pruned to obtain the updated delivery adjustment factor. The intensity of ad delivery is adjusted using the updated delivery adjustment factor, and a final ad target consumption curve is generated, which is used to drive the execution of the ad delivery system.

[0016] Secondly, the present invention also discloses a budget allocation and pacing control system based on delivery value feedback, characterized in that it includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the budget allocation and pacing control method based on delivery value feedback described in the first aspect is implemented.

[0017] The beneficial effects of this invention are as follows: (1) Compared with the prior art, the method of the present invention solves the problems of budget allocation and delivery control being separated and lacking a closed-loop feedback mechanism based on the value of delivery results in the prior art.

[0018] (2) Compared with the prior art, the method of the present invention collects the actual consumption and deployment result value data of each time interval, constructs the unit budget value efficiency and relative scoring index, and significantly improves the accuracy and robustness of matching budget allocation with deployment benefits.

[0019] (3) Compared with the prior art, the method of the present invention introduces the delivery accessibility index. This index can determine in real time whether the adjustment measures are effective in the current market environment by analyzing the quantitative relationship between the change of the delivery adjustment factor and the degree of consumption response, and adaptively adjust the update range accordingly, effectively avoiding ineffective adjustment and improving the stability and adaptability of delivery rhythm control. Attached Figure Description

[0020] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein: Figure 1 This is a flowchart of the budget allocation and pace control method based on the value feedback of the deployment in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram illustrating the principle of target consumption curve generation and delivery rhythm control based on delivery result value feedback in Embodiment 1 of the present invention. Figure 3 This is a schematic diagram of the budget allocation and rhythm control system based on the feedback of the delivery value in Embodiment 2 of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Example 1 like Figure 1 As shown, this embodiment discloses a budget allocation and pace control method based on deployment value feedback, including: S10: Divide the advertising campaign into multiple time intervals according to the preset advertising cycle; In this embodiment, the total duration of the advertising campaign period and the preset time granularity (e.g., per hour) are obtained. The continuous time axis is divided into multiple mutually exclusive and continuous time intervals with a fixed step size. A unique sequence number and data structure are assigned to each interval to store fields such as initial target consumption, actual consumption, and campaign result value.

[0023] S20: Obtain the initial target consumption for multiple time intervals, and confirm the actual consumption and value of the advertising results within the time intervals to construct a time-series dataset for advertising. In this embodiment, the initial target consumption value for each time interval is read from the initial target consumption curve array; the actual consumption data for each interval is collected through the advertising delivery system log interface, and conversion behavior data is collected through the tracking callback. The data is attributed to the corresponding interval based on the timestamp to form the delivery result value data; after validating the collected data, the initial target consumption, actual consumption, and delivery result value for each interval are horizontally concatenated by sequence number to generate structured records and arrange them in order to construct the delivery time series dataset.

[0024] S30: Calculate the unit budget value efficiency based on the time series dataset, construct the relative score for each time interval, and construct the target consumption curve based on the relative score; In this embodiment, the actual consumption and campaign result value for each time interval are extracted from the campaign time-series dataset. The unit budget value efficiency (i.e., the ratio of campaign result value to actual consumption) for each interval is calculated, along with the overall average value efficiency (i.e., the ratio of the sum of campaign result value across all intervals to the sum of actual consumption). The unit budget value efficiency for each interval is divided by the overall average value efficiency to obtain a relative score. A score greater than 1 indicates that the value efficiency is above average, while a score less than 1 indicates that it is below average. Preset maximum and minimum adjustment multipliers are obtained. After cropping the relative score, it is multiplied by the initial target consumption for the corresponding interval to obtain the updated target consumption value. Finally, the updated target consumption values ​​for all intervals are summed, and the total advertising budget is divided by this sum to obtain a normalization coefficient. This coefficient is then multiplied by the updated target consumption value for each interval, ensuring that the sum of the target consumption for each interval is exactly equal to the total budget, thereby generating a target consumption curve that satisfies the budget constraint.

[0025] S40: Obtain the actual cumulative consumption for each time interval, calculate the cumulative consumption deviation, and determine the update direction of the adjustment factor based on the sign direction of the cumulative consumption deviation. In this embodiment, the actual consumption values ​​for each time interval are sequentially accumulated to obtain the actual cumulative consumption sequence; the target consumption values ​​in the target consumption curve are also accumulated to obtain the target cumulative consumption sequence; the actual cumulative consumption for each time interval is subtracted from the target cumulative consumption to obtain the cumulative consumption deviation. The sign of this deviation is determined: a positive sign indicates that the actual consumption is ahead, a negative sign indicates that it is behind, and zero indicates that it is consistent. This sign is directly used as the update direction of the delivery adjustment factor; a positive sign decreases the adjustment factor to reduce the delivery intensity, and a negative sign increases the adjustment factor to increase the delivery intensity.

[0026] S50: Based on the change in the ratio between actual consumption and target consumption and the change in the delivery adjustment factor obtained from the data collection and calculation, the effectiveness evaluation data is calculated. In this embodiment, actual consumption and target consumption are obtained from two adjacent time intervals. The ratios of these two values ​​are calculated, and the difference is taken to obtain the change in ratio. Simultaneously, the difference in the adjustment factor applied between adjacent intervals is calculated to obtain the change in factor. The change in ratio is divided by the larger of the absolute value of the factor change and the preset minimum adjustment range, and the result is cropped to the 0-1 interval to obtain the original effectiveness evaluation data. This data is then subjected to exponential smoothing, i.e., weighted summation with the evaluation data from the previous interval, to obtain the effectiveness evaluation data.

[0027] S60: Extract the sign direction of the cumulative consumption deviation and determine the update range of the delivery adjustment factor in combination with the preset delivery accessibility index; In this embodiment, the sign direction (positive, negative, or zero) is extracted from the cumulative consumption deviation. A preset basic update magnitude and delivery accessibility compensation coefficient are obtained, and the delivery accessibility index (between 0 and 1) for the current time interval is read. The compensation term is calculated: subtract the index from 1, multiply by the compensation coefficient, add 1, and multiply by the basic update magnitude to obtain the update magnitude. This mechanism ensures that the lower the delivery accessibility index (the less effective the adjustment), the larger the update magnitude; conversely, it reverts to the basic value.

[0028] S70: Update the delivery adjustment factor based on the update direction and update magnitude, generate the advertising target consumption curve, and construct a budget delivery control strategy to be sent to the advertising delivery system for execution.

[0029] In this embodiment, the current ad delivery adjustment factor is read, and addition or subtraction operations are performed according to the update direction (positive signs add, negative signs subtract, zero remains unchanged) and update magnitude. The result is then cropped to between a preset minimum and maximum value to obtain the updated ad delivery adjustment factor. The target consumption curve and the updated ad delivery adjustment factor sequence are integrated into a structured control strategy, which is then sent to the ad delivery execution module via API or message queue. After parsing the strategy, the execution module allocates the budget according to the target consumption value for each interval and adjusts the bid multiplier, bidding probability, or request approval ratio based on the ad delivery adjustment factor to ensure that the actual consumption matches the target curve.

[0030] like Figure 2 As shown, the adaptive allocation of advertising budget and the control of the delivery rhythm in this embodiment are composed of a target consumption curve generation layer and a delivery rhythm control layer. The target consumption curve generation layer adaptively corrects the initial target consumption for each time interval based on the delivery value feedback to generate a target consumption curve that meets the total budget constraint. The delivery rhythm control layer dynamically adjusts the advertising delivery intensity based on the target consumption curve and the delivery adjustment factor to make the actual consumption as close as possible to the target consumption curve, thereby forming a closed-loop linkage between budget allocation and delivery execution.

[0031] Preferably, constructing the deployment time-series dataset includes: Obtain the advertising campaign period and total advertising budget; The advertising campaign period is divided into intervals according to a preset time granularity to obtain multiple time intervals. Each time interval is a preset time granularity, resulting in a set of time intervals. The total advertising budget is evenly distributed across the time interval set, and the initial target expenditure for the time interval set is calculated to obtain the initial target expenditure curve data. Collect actual consumption data of the advertising system within the time interval, record the amount of budget consumed at the end of each time interval, and obtain actual consumption time series data; Collect campaign result value data within each time interval to generate campaign result value time series data; wherein, the campaign result value data includes at least one or a weighted combination of conversion number, conversion value, and order amount; The actual consumption time-series data and the delivery result value time-series data are timestamped and integrated according to time intervals to obtain the delivery time-series dataset.

[0032] In this embodiment, the total duration of the campaign period (in days or hours) set by the advertiser and the total budget amount for this campaign are obtained. The campaign period is discretized according to a preset time granularity (e.g., hourly), dividing the continuous time axis into multiple mutually exclusive and continuous time intervals. Each interval has a unique start and end timestamp, forming a set of time intervals. Based on this, the system evenly distributes the total advertising budget according to the total number of time intervals, that is, by dividing the total budget by the number of intervals, calculating the initial target consumption value for each time interval under ideal conditions. These values ​​are arranged in interval order to obtain the initial target consumption curve data. During the campaign execution phase, the system collects the actual budget amount consumed in each time interval with millisecond-level precision through the real-time log interface of the advertising engine, and aggregates and accumulates the consumption in each interval at the end of each interval to form actual consumption time-series data.

[0033] Simultaneously, the system collects user conversion behavior data through event tracking callbacks or API interfaces. Based on the advertiser's preset value definition rules (such as order amount corresponding to a single conversion, number of conversions, etc.), it calculates the cumulative value of the campaign results within each time interval, forming time-series data of the campaign result value. After collection, the system performs timestamp alignment processing on the actual consumption time-series data and the campaign result value time-series data, matching the actual consumption value and campaign result value belonging to the same time interval into the records of that interval, and removing misaligned data caused by data delays or loss. Following ascending order of time interval numbers, the system integrates the initial target consumption, actual consumption, and campaign result value of each interval into a structured record, and all records constitute the campaign time-series dataset.

[0034] Preferably, constructing a target consumption curve based on the relative score includes: Calculate the ratio of the value of the campaign results to the actual consumption for each time interval to obtain the unit budget value efficiency sequence; the unit budget value efficiency is used to quantify the average business value that each unit budget can generate within a time interval; its calculation formula is:

[0035] In the formula, To prevent the division by zero assumption of extremely small positive numbers; For the first The value of investment results collected within a specific time interval; For the first The actual budget amount consumed within each time interval; Efficiency per unit of budget value; The total actual consumption is obtained by summing the actual consumption over the time intervals, and the total campaign result value is obtained by summing the campaign result value over the time intervals. The ratio of the total campaign result value to the total actual consumption is calculated to obtain the average value efficiency over the entire campaign period. The average value efficiency over the entire campaign period is used to quantify the average business value generated by each unit of budget during the entire campaign period; its calculation formula is as follows:

[0036] In the formula, This represents the sum of the delivery results across all time intervals. This represents the total actual consumption across all time intervals. This represents the average efficiency per unit of budget value. To prevent division by zero errors for extremely small positive numbers.

[0037] The relative score sequence is constructed by calculating the ratio of the unit budget value efficiency of each time interval to the average value efficiency over the entire period. The initial target consumption and relative score for each time interval are adjusted to obtain the updated target consumption; the update formula for the relative score adjustment is as follows:

[0038] In the formula, This indicates a numerical cropping operation, where the relative score sequence is less than 1. When, the value is When the relative rating sequence is greater than When, the value is Otherwise, the value is a relative score sequence; To achieve the maximum adjustment ratio, Minimum adjustment ratio; This represents the initial target consumption for each time interval; Represented as a relative rating sequence; The total update target consumption is obtained by summing the updated target consumption for each time interval. The ratio of the total budget to the total update target consumption is calculated to generate a normalization coefficient and construct the target consumption curve.

[0039] In the embodiments of the present invention, time intervals are used as the basic index units. Interval-by-interval ratio calculations are performed on the value data of the delivery results and the corresponding actual consumption data within each time slice. The results are encoded into a unit budget value efficiency sequence. The sequence is structurally represented as a one-dimensional numerical vector arranged over time, where each element reflects the value output intensity corresponding to the unit budget within the corresponding time interval. Based on this, aggregation operations are performed on the actual consumption data and the value data of the deployment results across the entire cycle. The total actual consumption scalar and the total deployment result value scalar are formed by summing these data, and their ratio is further constructed to generate the average value efficiency parameter for the entire cycle. This parameter is then embedded into the subsequent calculation process as a global reference benchmark. Each element in the unit budget value efficiency sequence is mapped to this global benchmark on an element-by-element basis to obtain a relative score sequence. This sequence, in terms of numerical distribution, represents a standardized characterization of the degree of efficiency deviation in each time interval, and its value range reflects the degree of strengthening or weakening of each interval relative to the overall average level. During the target consumption adjustment phase, the budget values ​​for each time interval in the initial target consumption curve are coupled with the corresponding relative scores. The original budget distribution is proportionally reconstructed through multiplication or weighting, thereby generating an updated target consumption sequence. This sequence exhibits a redistribution characteristic consistent with the efficiency distribution in the time dimension. The updated target consumption sequence is summed to obtain the total updated target consumption value. The ratio of this value to the predetermined total budget is calculated to generate a uniform normalization coefficient. This coefficient is then used to scale the updated target consumption sequence so that it meets the total amount constraint while maintaining the relative distribution shape. Finally, a target consumption curve data structure that is continuously unfolded on the time axis is formed.

[0040] Preferably, determining the update direction of the delivery adjustment factor based on the sign direction of the cumulative consumption deviation includes: Initialize the delivery adjustment factor to its default value; Obtain the actual cumulative consumption and target cumulative consumption for each time interval; The actual consumption for each time interval is summed to obtain the first consumption sequence data; the calculation formula for the first consumption sequence data is as follows:

[0041] In the formula, This is represented as the actual consumption for each time interval; Indicated as a time period; This is represented as the first consumed sequence data; The cumulative consumption of the target over each time interval is summed to obtain the second consumption sequence data; the calculation formula for the second consumption sequence data is as follows:

[0042] In the formula, Accumulate the consumption for the target in each time interval; Indicated as a time period; This is represented as the second consumed sequence data; The cumulative difference between the first consumption sequence data and the second consumption sequence data is calculated to obtain the cumulative consumption deviation for each time interval; the formula for calculating the cumulative consumption deviation is as follows:

[0043] In the formula, This indicates the direction of the update of the adjustment factor. The sign direction of the cumulative consumption deviation in each time interval is extracted to obtain the sign direction of the cumulative consumption deviation, which is used as the update direction of the delivery adjustment factor.

[0044] Preferably, the sign direction of the cumulative consumption deviation for each time interval is extracted, including: If the cumulative consumption deviation is greater than zero, the sign direction is positive; if the cumulative consumption deviation is less than zero, the sign direction is negative; if the cumulative consumption deviation is equal to zero, the sign direction is zero. Here, the symbol sign(x) represents a sign direction as follows:

[0045] In the formula, This is expressed as the cumulative consumption deviation.

[0046] In an embodiment of the invention, the cumulative consumption deviation value for each time interval is read sequentially from the cumulative consumption deviation sequence according to the time interval number. This value is a floating-point number representing the difference between the actual cumulative consumption and the target cumulative consumption. The system performs a sign judgment logic on this value: calling a sign function, taking the cumulative consumption deviation as input, and outputting a discrete sign value. Specifically, if the deviation value is greater than zero, a positive sign (usually represented as +1) is output; if the deviation value is less than zero, a negative sign (represented as -1) is output; if the deviation value is equal to zero (or the absolute value is less than a preset minimum threshold), a zero sign (represented as 0) is output. This judgment process traverses all time intervals, generating a sign direction sequence of the same length as the number of time intervals, with each element taking the value of +1, -1, or 0. The sign direction sequence is stored in memory as an array, serving as the input parameter for the subsequent adjustment factor update step. When the sign is positive, it indicates that the current actual consumption progress is ahead of the target progress, and the delivery intensity needs to be reduced. Therefore, the system sets the subsequent update direction to negative (i.e., reduces the delivery adjustment factor). When the sign is negative, it indicates that the actual consumption is lagging behind, and the delivery intensity needs to be increased. The update direction is set to positive (increases the delivery adjustment factor). When the sign is zero, it indicates that the consumption progress is consistent with the target, and the update direction is set to zero (keeping the delivery adjustment factor unchanged).

[0047] Preferably, effectiveness evaluation data is calculated based on the change in the ratio between actual consumption and target consumption and the change in the dosage adjustment factor obtained from the data collection and calculation, including: Obtain the delivery adjustment factor for adjacent time intervals, calculate the difference between the delivery adjustment factor for the later time interval and the delivery adjustment factor for the earlier time interval, and obtain the change in delivery adjustment factor; the formula for calculating the change in delivery adjustment factor is as follows:

[0048] In the formula, Represented as the first The difference between each adjustment factor; Represented as the first One adjustment factor for the distribution; Represented as the first One adjustment factor for the distribution; The ratio of actual consumption to target consumption is calculated by adding a very small positive number to each time interval, and then calculating the difference in consumption ratios between adjacent time intervals.

[0049] In the formula, Indicates the first The time interval relative to the first The change in the ratio of actual consumption to target consumption over a given time interval; This represents the actual consumption in the (t+1)th time interval; This represents the actual consumption in the t-th time interval; This is represented as the target consumption for the (t+1)th time interval; This is represented as the target consumption for the t-th time interval; Minimal positive numbers to prevent division by zero errors; Based on the changes in the adjustment factors and the ratios of the deployment, a deployment accessibility index is constructed to obtain data for evaluating the effectiveness of deployment adjustment.

[0050] In an embodiment of the invention, the values ​​of the delivery adjustment factor for two adjacent time intervals (interval t and interval t+1) are read from the delivery adjustment factor sequence. The difference between the latter interval and the former period is calculated to obtain the change in the delivery adjustment factor. This change can be positive or negative, representing the magnitude and direction of the adjustment of delivery intensity. Simultaneously, the actual consumption value and the target consumption value for these two time intervals are extracted from the delivery time-series dataset. To prevent division by zero due to a zero target consumption value, a very small positive number (e.g., 1e-8) is superimposed on each target consumption value. The ratio of actual consumption to target consumption for each interval is calculated. Then, the ratio of the latter interval is subtracted from the ratio of the former interval to obtain the change in the ratio between actual consumption and target consumption. This change reflects the fluctuation of the consumption response as the delivery intensity is adjusted.

[0051] Next, the system takes the absolute value of the change in the adjustment factor and compares it with the preset minimum adjustment range (a positive number used to stabilize the denominator when the adjustment range is too small), taking the larger value as the denominator. The system then uses the change in the ratio of actual consumption to target consumption as the numerator and performs a division operation to calculate the consumption response level caused by a unit adjustment intensity. To prevent the calculation results from exceeding a reasonable range due to noise or outliers, the system prunes this ratio: if it is less than 0, it is set to 0; if it is greater than 1, it is set to 1, obtaining the original effectiveness assessment data.

[0052] Finally, an exponential smoothing method is used to weight and sum the original effectiveness assessment data for the current time interval with the smoothed effectiveness assessment data for the previous time interval. The weights are controlled by a preset smoothing coefficient (between 0 and 1) to generate the final effectiveness assessment data (i.e., the delivery accessibility index). This index ranges from 0 to 1 and is used to quantify the effectiveness of the delivery adjustment factor in controlling actual consumption.

[0053] Preferably, the deployment accessibility index constructed based on the change in the deployment adjustment factor and the change in the ratio includes: The absolute value of the change in the delivery adjustment factor is calculated to the maximum value and compared with the preset minimum delivery adjustment range. The larger value is taken as the denominator, and the change in the ratio of actual consumption to target consumption is taken as the numerator. The ratio is calculated, and the calculation result is cropped so that its value range is between 0 and 1 to obtain the original delivery accessibility index data for each time interval. The original accessibility index sequence is exponentially smoothed by introducing a preset smoothing coefficient. The original accessibility index for the current time interval is then weighted and summed with the smoothed accessibility index for the previous time interval to obtain the accessibility index. The calculation formula for the accessibility index is as follows:

[0054] In the formula, For the first Deploy accessibility metrics for one time interval. For the first Accessibility metrics for delivery within a specific time interval; This refers to the original accessibility index data for delivery. The preset smoothing coefficient satisfies 0 < <1; When the accessibility index is lower than the preset limit threshold, the update range of the delivery adjustment factor is limited or the adjustment is temporarily suspended.

[0055] In embodiments of the present invention, the changes in the delivery adjustment factor and the changes in the ratio of actual consumption to target consumption are obtained for adjacent time intervals. The absolute value of the change in the delivery adjustment factor is taken and compared with a preset minimum delivery adjustment amplitude (a positive number, such as 0.01), and the larger value is used as the denominator; the change in the ratio of actual consumption to target consumption is used as the numerator, and a division operation is performed. This ratio reflects the degree of consumption response caused by a unit adjustment intensity. To prevent the calculation results from exceeding a reasonable range due to noise or outliers, the system prunes the calculation results: if it is less than 0, it is set to 0; if it is greater than 1, it is set to 1, thereby obtaining the original delivery accessibility index data for each time interval, with values ​​ranging from 0 to 1.

[0056] To eliminate potential random fluctuations from single sampling, the system introduces exponential smoothing: It retrieves the previously calculated delivery accessibility index (initial value can be set to 0.5 or a default value), and then weights and sums the original delivery accessibility index for the current time interval with the smoothed index from the previous time interval. A preset smoothing coefficient (typically between 0.2 and 0.3) controls the weighting of historical information and current observations, resulting in the final delivery accessibility index for the current time interval. This index also remains within the range of 0 to 1; a higher value indicates more effective control of actual consumption through delivery adjustments, while a lower value indicates that the adjustment measures are becoming ineffective.

[0057] The system further sets a preset threshold (e.g., 0.2 or 0.3). After calculating the accessibility index for each time interval, it compares it with this threshold. If the current index is lower than the threshold, it is determined that the adjustment measures are basically ineffective or have a weak effect in the current market environment. Based on this, the system limits the update range of the adjustment factor (e.g., multiplies the update range by a decay coefficient less than 1) or directly suspends the current adjustment (i.e., skips the update of the adjustment factor for this interval), thereby avoiding system oscillation or resource waste caused by ineffective adjustment.

[0058] Preferably, the original delivery accessibility index data and the calculation formula for the delivery accessibility index include: The formula for calculating the original accessibility index data is:

[0059] In the formula, This represents the change in the ratio of actual consumption to target consumption. To adjust the change in the adjustment factor; This is the preset minimum adjustment range for the delivery; This means cropping the calculation results to the range of 0 to 1; This is the original data on the accessibility of the delivery system.

[0060] Preferably, the generated ad target consumption curve includes: Obtain the total budget data; calculate the ratio of the cumulative consumption deviation for each time interval to the total budget data to obtain the normalized cumulative consumption deviation; the formula for calculating the cumulative consumption deviation is as follows:

[0061] In the formula, To account for cumulative consumption deviation, For the total advertising budget, To prevent extremely small positive numbers from being divided by zero; The update magnitude of the delivery adjustment factor is determined using the cumulative consumption deviation and delivery accessibility indicators, while the increase or decrease trend of the delivery adjustment factor is adjusted in conjunction with the update direction; the calculation formula for the update magnitude of the delivery adjustment factor is as follows:

[0062] In the formula, The preset basic update range and satisfying , The preset accessibility compensation coefficient is used and satisfies the following conditions. , For the first Accessibility metrics for delivery within a specific time interval; The updated delivery adjustment factor is obtained by performing a difference calculation between the current time interval's delivery adjustment factor and the update magnitude and cumulative consumption deviation sign direction, and then pruning the calculation results. The update formula is as follows:

[0063] In the formula, For the first The adjustment factor for the delivery of each time interval, For the first The adjustment factor for the delivery of each time interval, To adjust the adjustment factor update magnitude, The sign direction of the cumulative consumption deviation. The preset minimum delivery adjustment factor, This is the preset maximum delivery adjustment factor; The intensity of ad delivery is adjusted using the updated delivery adjustment factor, and a final ad target consumption curve is generated, which is used to drive the execution of the ad delivery system.

[0064] In an embodiment of the present invention, the total budget data set by the advertiser is obtained, and the cumulative consumption deviation value of the current time interval is read from the cumulative consumption deviation sequence. The deviation value is divided by the total budget data (while superimposing a very small positive number to prevent division by zero) to obtain the normalized cumulative consumption deviation, which is used to eliminate the dimensional influence of different budget sizes on the deviation magnitude.

[0065] The system reads the current time interval's reachability index (range 0 to 1) and preset basic update magnitude and compensation coefficient. It calculates the update magnitude of the ad adjustment factor according to the formula: when the reachability index is low, the update magnitude automatically increases to compensate for adjustment failures; when the index is high, the update magnitude returns to the basic value. Simultaneously, the system determines the adjustment factor's increase / decrease trend based on the sign of the cumulative consumption deviation: a positive sign requires decreasing the adjustment factor to reduce ad intensity, while a negative sign requires increasing the adjustment factor to increase ad intensity. Next, the system performs a difference calculation between the current time interval's ad adjustment factor value and the result of multiplying the update magnitude by the sign direction (i.e., subtracting the update magnitude when the sign is positive, and adding the update magnitude when the sign is negative), and then trims the result to ensure the updated ad adjustment factor falls between the preset minimum and maximum values ​​(e.g., 0 to 1). After the update, the system uses this ad adjustment factor to actually adjust the ad intensity, including adjusting the bid multiplier, bid participation probability, or exposure request release ratio.

[0066] The system integrates the target consumption value for each time interval (obtained through relative scoring normalization) with the updated delivery adjustment factor sequence to generate the final advertising target consumption curve. This curve is encapsulated in a structured data format and sent to the execution module of the advertising delivery system through an interface. This drives the system to allocate the budget according to the target consumption value for each time interval and adjust the delivery intensity in real time according to the delivery adjustment factor, so that the actual consumption fits the target curve.

[0067] Most importantly, the delivery adjustment factor affects the delivery process by adjusting at least one of the following methods: adjusting the ad bid multiplier, adjusting the probability of the ad participating in the bidding, and adjusting the release ratio of exposure requests: Adjusting ad delivery intensity using the updated delivery adjustment factors includes at least one method: adjusting the ad bid multiplier, adjusting the probability of the ad participating in the auction, and adjusting the approval rate for impression requests. By adjusting the ad bid multiplier, the campaign process is affected. Specifically, the original bid is multiplied by a campaign adjustment factor. The campaign adjustment factor and the adjustment intensity are monotonically positively correlated. The bid is adjusted proportionally. When the campaign adjustment factor is large, the ad bid is increased to make it easier for the ad to win in the bidding and increase actual consumption. When the campaign adjustment factor is small, the ad bid is decreased to reduce the probability of winning in the bidding and slow down the consumption rate. The probability of an ad participating in bidding is adjusted to influence the ad delivery process. Specifically, a delivery adjustment factor is used to adjust the probability of an ad participating in bidding. When the delivery adjustment factor is large, the probability of an ad participating in bidding is increased, allowing the ad to participate in bidding in more requests to increase the chances of being consumed. When the delivery adjustment factor is small, the probability of an ad participating in bidding is decreased, allowing the ad not to participate in bidding in some requests to slow down the consumption rate. The release ratio of exposure requests is adjusted to affect the delivery process. Specifically, during the process of ad requests going through budget checks and delivery strategy screening, the release ratio of ad requests is adjusted using a delivery adjustment factor. When the delivery adjustment factor is large, the release ratio of requests is increased, allowing more ad requests to enter the delivery process and increasing actual spending. When the delivery adjustment factor is small, the release ratio of requests is decreased, causing some requests to be intercepted in advance to slow down the spending rate.

[0068] It should be noted that the comparison table between this solution and existing technologies is shown below:

[0069] Example 2 like Figure 3As shown, this embodiment discloses a budget allocation and pacing control system based on delivery value feedback, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the budget allocation and pacing control method based on delivery value feedback described in Embodiment 1 is implemented.

[0070] The system also includes other components well known to those skilled in the art, such as communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0071] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions that can be stored or otherwise maintained by such a computer-readable medium.

[0072] In the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

[0073] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A budget allocation and pace control method based on investment value feedback, characterized in that, include: The advertising campaign is divided into multiple time intervals according to a preset advertising cycle; Obtain the initial target spending for multiple time intervals, and confirm the actual spending and value of the ads delivered within the time intervals to construct a time-series dataset for ad delivery. The unit budget value efficiency is calculated based on the aforementioned time series dataset, a relative score is constructed for each time interval, and a target consumption curve is constructed based on the relative score. Obtain the actual cumulative consumption for each time interval to calculate the cumulative consumption deviation, and determine the update direction of the delivery adjustment factor based on the sign direction of the cumulative consumption deviation; Based on the change in the ratio between actual consumption and target consumption and the change in the delivery adjustment factor obtained from the data collection and calculation, the effectiveness evaluation data is calculated. The sign direction of the cumulative consumption deviation is extracted, and the update magnitude of the delivery adjustment factor is determined in combination with the preset delivery accessibility index. The delivery adjustment factor is updated based on the update direction and the update magnitude to generate an advertising target consumption curve, thereby constructing a budget delivery control strategy and sending it to the advertising delivery system for execution.

2. The budget allocation and pace control method based on investment value feedback as described in claim 1, characterized in that, Constructing the time-series dataset for deployment includes: Obtain the advertising campaign period and total advertising budget; The advertising campaign period is divided into intervals according to a preset time granularity to obtain multiple time intervals. Each time interval is a preset time granularity, resulting in a set of time intervals. The total advertising budget is evenly distributed across the time interval set, and the initial target expenditure for the time interval set is calculated to obtain the initial target expenditure curve data. Collect actual consumption data of the advertising system within the time interval, record the amount of budget consumed at the end of each time interval, and obtain actual consumption time series data; Collect campaign result value data within each time interval to generate campaign result value time series data; wherein, the campaign result value data includes at least one or a weighted combination of conversion number, conversion value, and order amount; The actual consumption time-series data and the delivery result value time-series data are timestamped and integrated according to time intervals to obtain the delivery time-series dataset.

3. The budget allocation and pace control method based on investment value feedback as described in claim 1, characterized in that, Constructing a target consumption curve based on the relative score includes: The ratio of the value of the delivery results to the actual consumption is calculated for each time interval to obtain the unit budget value efficiency sequence; the unit budget value efficiency is used to quantify the average business value that each unit budget can generate within the time interval. The total actual consumption is obtained by summing the actual consumption over the time intervals, and the total value of the campaign results is obtained by summing the value of the campaign results over the time intervals. The ratio of the total value of the campaign results to the total actual consumption is calculated to obtain the average value efficiency over the entire campaign period. The average value efficiency over the entire campaign period is used to quantify the average business value generated by each unit of budget during the entire campaign period. The relative score sequence is constructed by calculating the ratio of the unit budget value efficiency of each time interval to the average value efficiency over the entire period. Adjust the target consumption based on the initial target consumption and relative score for each time interval to obtain the updated target consumption; The total update target consumption is obtained by summing the updated target consumption for each time interval. The ratio of the total budget to the total update target consumption is calculated to generate a normalization coefficient and construct the target consumption curve.

4. The budget allocation and pace control method based on deployment value feedback according to claim 1, characterized in that, The update direction of the delivery adjustment factor is determined based on the sign direction of the cumulative consumption deviation, including: Initialize the delivery adjustment factor to its default value; Obtain the actual cumulative consumption and target cumulative consumption for each time interval; The actual consumption in each time interval is summed to obtain the first consumption sequence data; The cumulative consumption of the target in each time interval is summed to obtain the second consumption sequence data; The cumulative difference between the first consumption sequence data and the second consumption sequence data is calculated to obtain the cumulative consumption deviation for each time interval; The sign direction of the cumulative consumption deviation in each time interval is extracted to obtain the sign direction of the cumulative consumption deviation, which is used as the update direction of the delivery adjustment factor.

5. The budget allocation and pace control method based on deployment value feedback according to claim 4, characterized in that, The sign and direction of the cumulative consumption deviation for each time interval are extracted, including: If the cumulative consumption deviation is greater than zero, the sign direction is positive; if the cumulative consumption deviation is less than zero, the sign direction is negative; if the cumulative consumption deviation is equal to zero, the sign direction is zero.

6. The budget allocation and pace control method based on delivery value feedback according to claim 1, characterized in that, Based on the change in the ratio between actual consumption and target consumption and the change in the adjustment factor obtained from the data collection and calculation, effectiveness evaluation data are calculated, including: Obtain the delivery adjustment factor for adjacent time intervals, calculate the difference between the delivery adjustment factor for the next time interval and the delivery adjustment factor for the previous time interval, and obtain the change in delivery adjustment factor. The ratio of actual consumption to target consumption is calculated by adding a very small positive number to each time interval, and the difference in consumption ratio between adjacent time intervals is also calculated to obtain the target consumption ratio. Based on the changes in the adjustment factors and the ratios of the deployment, a deployment accessibility index is constructed to obtain data for evaluating the effectiveness of deployment adjustment.

7. The budget allocation and pace control method based on deployment value feedback according to claim 6, characterized in that, The deployment accessibility index is constructed based on the changes in the deployment adjustment factor and the changes in the ratio, including: The absolute value of the change in the delivery adjustment factor is calculated to the maximum value and compared with the preset minimum delivery adjustment range. The larger value is taken as the denominator, and the change in the ratio of actual consumption to target consumption is taken as the numerator. The ratio is calculated to obtain the original delivery accessibility index data for each time interval. The original delivery accessibility index sequence is subjected to exponential smoothing. A preset smoothing coefficient is introduced, and the original delivery accessibility index of the current time interval is weighted and summed with the smoothed delivery accessibility index of the previous time interval to obtain the delivery accessibility index. When the accessibility index is lower than the preset limit threshold, the update range of the delivery adjustment factor is limited or the adjustment is temporarily suspended.

8. The budget allocation and pace control method based on deployment value feedback according to claim 7, characterized in that, The original accessibility index data and the calculation formula for the accessibility index include: The formula for calculating the original accessibility index data is as follows: In the formula, This represents the change in the ratio of actual consumption to target consumption. To adjust the change in the adjustment factor; This is the preset minimum adjustment range for the delivery; This means cropping the calculation results to the range of 0 to 1; This is the original data on the accessibility of the delivery system.

9. The budget allocation and pace control method based on deployment value feedback according to claim 1, characterized in that, The generated ad spend curve includes: Obtain total budget data; The ratio of the cumulative consumption deviation for each time interval to the total budget data is calculated to obtain the normalized cumulative consumption deviation. The update magnitude of the delivery adjustment factor is determined by using the cumulative consumption deviation and delivery accessibility indicators, and the increase or decrease trend of the delivery adjustment factor is adjusted in combination with the update direction. The difference between the current time interval's delivery adjustment factor and the update magnitude and cumulative consumption deviation sign direction is calculated, and the calculation result is pruned to obtain the updated delivery adjustment factor. The intensity of ad delivery is adjusted using the updated delivery adjustment factor, and a final ad target consumption curve is generated, which is used to drive the execution of the ad delivery system.

10. A budget allocation and pacing control system based on investment value feedback, characterized in that, It includes a processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement the budget allocation and pacing control method based on delivery value feedback according to any one of claims 1-9.

Images