Product attribution putting method and device, electronic equipment and storage medium

Abstract

The invention discloses a product attribution putting method and device, electronic equipment and a storage medium, and relates to the technical field of cloud computing, and the method comprises the steps: obtaining an effect evaluation index of a two-tuple composed of a user and a product and a three-tuple composed of the user, the product and a marketing scene, covering budget data, quantity starting capability data, attribution efficiency and an estimated conversion rate; the multi-dimensional evaluation can reflect the advertisement putting condition more comprehensively; according to a comparison result of effect evaluation indexes of different dimensions, an attribution threshold value is calibrated instead of adopting a fixed threshold value, and an estimated conversion rate threshold value, a budget threshold value and an attribution efficiency threshold value can be flexibly adjusted; based on the comparison of the calibrated attribution threshold value and the actual index, the pertinence of resource release can be ensured, and the input-output ratio of the whole marketing activity is improved, so that the market demand and the user behavior are matched more accurately.

Description

Technical Field

[0001] This application relates to the field of cloud computing technology, and in particular to a product attribution delivery method, apparatus, electronic device, and storage medium. Background Technology

[0002] In the advertising ecosystem, conversion attribution plays a crucial role, and its attribution capabilities and strategies profoundly impact advertising revenue. Specifically, the accuracy of advertising performance attribution not only directly determines the actual effectiveness of current campaigns but also plays a key role in guiding brands' long-term optimization strategies and business growth.

[0003] In real-world applications, advertisers typically run ads for a large number of products, and these different products vary significantly in terms of budget allocation and ad performance. Furthermore, even for the same product, attribution efficiency and final conversion rates differ depending on the ad placement.

[0004] Currently, most existing intelligent attribution delivery strategies rely on simple rules such as fixed thresholds for decision-making. This directly leads to uneven spending of advertisers' budgets across different products and ad placements, with some high-value traffic not being fully utilized, thus affecting the overall effectiveness of ad campaigns. Therefore, refining and optimizing existing attribution delivery strategies has become an urgent problem to be solved in the advertising industry. Summary of the Invention

[0005] In view of the above problems, this application provides a product attribution delivery method, apparatus, electronic device, and storage medium to overcome or at least partially solve the above problems. The technical solution is as follows: A product attribution-based ad placement method includes: obtaining a first performance evaluation metric consisting of a binary group of users and products, and a second performance evaluation metric consisting of a tripartite group of users, products, and marketing scenarios; wherein the performance evaluation metric includes budget data, scaling capability data, attribution efficiency, and estimated conversion rate; calibrating an attribution threshold based on a comparison between the first and second performance evaluation metrics; wherein the attribution threshold includes at least one of an estimated conversion rate threshold, a budget threshold, and an attribution efficiency threshold, wherein the estimated conversion rate threshold is used to adjust the product placement quota for different marketing scenarios, the budget threshold is used to characterize the available budget for different marketing scenarios, and the attribution efficiency threshold is used to adjust the budget allocation for different marketing scenarios; if the attribution threshold includes the estimated conversion rate threshold, and the actual estimated conversion rate of the target marketing scenario is higher than the estimated conversion rate threshold, then the product placement quota for the target marketing scenario is increased; if the attribution threshold includes the budget threshold, then the available budget corresponding to the target marketing scenario is determined; if the attribution threshold includes the attribution efficiency threshold, and the actual attribution efficiency of the target marketing scenario is higher than the attribution efficiency threshold, then the budget for the target marketing scenario is increased.

[0006] As an optional implementation method provided in this application, the attribution threshold is calibrated based on the comparison results of the first effect evaluation index and the second effect evaluation index, including: classifying and weighting the comparison results of the first effect evaluation index and the second effect evaluation index; wherein, the comparison results include at least two of the following: budget deviation, scaling capability deviation, attribution efficiency deviation, and estimated conversion rate deviation; multiplying the scale value corresponding to the comparison result by its weight, and then summing them to obtain a weighted total score; determining adjustment parameters to standardize the weighted total score into adjustment parameters; obtaining the adjustment sensitivity corresponding to the scaling capability data included in the first effect evaluation index; calculating the budget adjustment ratio based on the adjustment parameters and adjustment sensitivity, and adjusting the initial budget threshold according to the budget adjustment ratio to obtain the budget threshold.

[0007] As an optional implementation method provided in this application, determining adjustment parameters based on the comparison results of the first effect evaluation index and the second effect evaluation index includes: calculating the attribution efficiency deviation based on the attribution efficiency included in the first and second effect evaluation indices; calculating the budget deviation based on the budget data included in the first and second effect evaluation indices; multiplying the weights of the attribution efficiency deviation and the budget deviation to calculate a proportional coefficient; determining the cumulative deviation value of the adjustment parameters for the attribution efficiency deviation and the budget deviation within a preset time period, adding an integral limit, and calculating an integral coefficient; calculating the differential coefficient based on the rate of change of the deviations of the attribution efficiency deviation and the budget deviation within the preset time period; weighting the proportional coefficient, integral coefficient, and differential coefficient to obtain a combination of adjustment parameters; obtaining the adjustment sensitivity corresponding to the volume-generating capacity data included in the first effect evaluation index; calculating the budget adjustment ratio based on the combination of adjustment parameters and the adjustment sensitivity; and adjusting the initial budget threshold based on the budget adjustment ratio to obtain the budget threshold.

[0008] As an optional implementation method provided in this application, the attribution threshold is calibrated based on the comparison result of the first effect evaluation index and the second effect evaluation index, including: performing a ratio calculation on the first effect evaluation index and the second effect evaluation index to calculate a correction coefficient; and obtaining an initial attribution threshold. The initial attribution threshold is adjusted based on the correction factor to obtain the attribution threshold.

[0009] As an optional implementation provided in this application, the attribution threshold is calibrated based on the comparison result of the first effect evaluation index and the second effect evaluation index, including: calculating the attribution efficiency deviation based on the attribution efficiency included in the first effect evaluation index and the second effect evaluation index; determining the tilt coefficient corresponding to the attribution efficiency deviation; and adjusting the attribution threshold based on the tilt coefficient.

[0010] As an optional implementation provided in this application, the attribution threshold is calibrated based on the comparison result of the first effect evaluation index and the second effect evaluation index, including: calculating the long-term cumulative deviation and the short-term time period deviation based on the first effect evaluation index and the second effect evaluation index; calculating the adjustment amount of the attribution threshold based on the long-term cumulative deviation and the short-term time period deviation; and calibrating the initial attribution threshold based on the adjustment amount.

[0011] A product attribution delivery device, the device comprising: The acquisition module is used to acquire the first performance evaluation metric of the binary group consisting of users and products, and the second performance evaluation metric of the triple group consisting of users, products and marketing scenarios. The performance evaluation metrics include budget data, scaling capability data, attribution efficiency and estimated conversion rate. The threshold calibration module is used to calibrate the attribution threshold based on the comparison results of the first performance evaluation index and the second performance evaluation index. The attribution threshold includes at least one of the following: the estimated conversion rate threshold, the budget threshold, and the attribution efficiency threshold. The estimated conversion rate threshold is used to adjust the product placement quota for different marketing scenarios, the budget threshold is used to characterize the available budget for different marketing scenarios, and the attribution efficiency threshold is used to adjust the budget allocation for different marketing scenarios. The adjustment module is used to increase the product placement quota for the target marketing scenario if the attribution threshold includes the estimated conversion rate threshold and the actual estimated conversion rate of the target marketing scenario is higher than the estimated conversion rate threshold; if the attribution threshold includes the budget threshold, it determines the available budget for the target marketing scenario; if the attribution threshold includes the attribution efficiency threshold and the actual attribution efficiency of the target marketing scenario is higher than the attribution efficiency threshold, it increases the budget for the target marketing scenario.

[0012] A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the product attribution delivery method described in any of the preceding claims.

[0013] A computer program product includes a computer program that, when executed by a processor, implements the steps of any of the above-described product attribution and delivery methods.

[0014] This application obtains performance evaluation metrics for binary pairs consisting of users and products, as well as triplets consisting of users, products, and marketing scenarios. These metrics cover budget data, scaling capability data, attribution efficiency, and estimated conversion rate. This multi-dimensional evaluation can more comprehensively reflect the advertising performance. Based on the comparison results of performance evaluation metrics across different dimensions, the attribution threshold is calibrated instead of using a fixed threshold, allowing for flexible adjustment of the estimated conversion rate threshold, budget threshold, and attribution efficiency threshold. By comparing the calibrated attribution threshold with actual metrics, the targeting of resource allocation can be ensured, improving the overall return on investment of marketing campaigns and thus more accurately matching market demand and user behavior.

[0015] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0016] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1A flowchart illustrating a product attribution and placement method provided in an embodiment of this application is shown. Figure 2 This illustration shows a schematic diagram of a product attribution delivery device according to an embodiment of this application; Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0017] To more clearly illustrate the embodiments of this application, the technical terms used in the embodiments will be briefly introduced below: Attributing ad performance refers to identifying which ad behaviors (such as clicks, views, and impressions) ultimately led to user conversions (such as purchases, registrations, and downloads) and quantifying the contribution of different ad behaviors. In the short term, accurate attribution allows brands to understand which channels, ad placements, and creatives are effective, thereby assessing the current campaign performance (e.g., whether the money spent is worthwhile). In the long term, attribution results guide brands to adjust budget allocations and optimize campaign strategies, such as increasing investment in high-performing channels and eliminating ineffective creatives, ultimately impacting whether the brand can achieve sustainable growth.

[0018] Brands often run campaigns for multiple products simultaneously, such as an e-commerce platform promoting apparel, home appliances, and cosmetics. The budgets and scaling capabilities of these different products vary significantly. Higher-margin products may receive larger budgets, while new products may receive smaller ones. Some products, like bestsellers, have built-in traffic and can generate substantial conversions even with a small budget; others require a large budget to attract users. This difference leads to varying attribution requirements for different products. Products with strong scaling capabilities may need more granular attribution to maintain their advantage, while products with lower budgets may need attribution analysis to determine if increased investment is worthwhile.

[0019] Regarding attribution efficiency, some ad placements can clearly track user behavior, making attribution simpler, such as search ads which can directly link to the search-click-purchase path; while other ad placements, such as brand exposure ads, are difficult to track conversions directly, making attribution more challenging. Regarding conversion rates, some ad placements, such as e-commerce detail page ads, have clear user intent and high conversion rates; while other ad placements, such as social media feeds, may have users browsing casually, resulting in low conversion rates.

[0020] Currently, many brands use rather crude attribution methods. For example, if a user converts within 7 days of clicking an ad, it's attributed to that ad. While 7 days is a fixed threshold, the conversion cycle varies depending on the product and ad placement. For instance, the decision-making cycle for high-priced products might be 15 days, while for low-priced fast-moving consumer goods it might be 1 day. A fixed threshold leads to attribution bias. Another example is the last-click attribution method, which only considers the last ad click, ignoring the contributions of earlier exposures and browsing, resulting in incomplete attribution.

[0021] Scaling capability, also known as volume scaling capability or volume expansion capability, refers to the comprehensive ability of an advertisement or product to acquire effective traffic (such as impressions, clicks, conversions, etc.) quickly and on a large scale under certain budget, time, and targeting conditions. It is used to measure whether an advertisement / product can quickly get up and running during a campaign, that is, whether it can efficiently consume the budget and bring in sufficient traffic or conversions.

[0022] Predicted Conversion Rate (PCVR) is the probability that a user will complete a conversion (such as placing an order or registering) after a specific action (such as clicking an ad or browsing a product) based on historical data, user characteristics, and contextual factors, using an algorithmic model. It is usually expressed as a percentage.

[0023] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0024] Accuracy of advertising attribution is crucial, but in actual campaigns, there are complex differences between multiple products and multiple ad placements. Existing simple strategies, such as fixed thresholds, cannot cope with this complexity and may lead to attribution bias, ultimately affecting the brand's short-term performance and long-term growth.

[0025] To address the technical problem of inaccurate attribution affecting budget allocation and strategy optimization, this application provides a product attribution delivery method, such as... Figure 1 As shown, Figure 1 This is a schematic flowchart of a product attribution and placement method provided in an embodiment of this application. The method includes the following steps S101~S103: S101. Obtain the first performance evaluation metric of the binary combination of users and products, and the second performance evaluation metric of the ternary combination of users, products and marketing scenarios.

[0026] In this context, "user" refers to the product provider, such as an advertiser. "Product" refers to the specific goods, services, or content that the customer encounters, such as an advertisement. The binary pair consisting of "user" and "product" can be represented as [user, product].

[0027] Performance evaluation metrics include budget data, scaling capability data, attribution efficiency, and estimated conversion rate. Budget data includes expected budget, cumulative spending, and anticipated spending. Scaling capability data is the growth rate of exposure per unit time, which can be quantified by the historical budget spending rate versus conversion growth curve. Attribution efficiency is the ratio of conversions to sent messages. The estimated conversion rate can be predicted by combining historical conversions and current sent messages; all three metrics together provide a quantitative basis for judging the value of the marketing scenario.

[0028] Performance evaluation metrics may also include billing revenue, conversions, and peak traffic. Billing revenue is the direct income generated from the intersection of different tuples and dimensions, such as revenue generated when a customer clicks on a product through a secondary ad placement and pays accordingly. Billing revenue is linked to conversions; fluctuations in conversions will affect revenue data. Peak traffic can be defined as daily highs and lows in traffic.

[0029] Marketing scenarios include multi-level ad placements, such as primary and secondary ad placements. Primary ad placements are the core display medium and key touchpoints connecting customers and products, such as homepage banners and splash screen ads, which have high priority. Sub-display areas under primary ad placements represent more refined traffic distribution units, such as sub-modules within banners and specific rankings on list pages. The tripartite combination of user, product, and marketing scenario can be represented as [user, product, marketing scenario], for example, [user, product, primary ad placement, secondary ad placement].

[0030] For the binary group [user, product], we collect the first budget data (e.g., the total budget set by the advertiser for the product), the first scaling capability data (e.g., the historical rate of increase in ad impressions per unit time), the first attribution efficiency, and the first estimated conversion rate, and integrate them into the first performance evaluation metric. For the ternary group [user, product, marketing scenario], we collect segmented data for each specific marketing scenario according to the same metric type, forming the second performance evaluation metric.

[0031] For example, the system obtains the first billing revenue, first number of conversions, first estimated conversion rate, first attribution efficiency, first budget data, and first scaling capability data for [Advertiser, Ad]. It also obtains the second billing revenue, second number of conversions, second budget data, second scaling capability data, second attribution efficiency, and second estimated conversion rate for [Advertiser, Ad, Primary Ad Slot, Secondary Ad Slot].

[0032] The above embodiments obtain performance evaluation metrics for binary pairs consisting of users and products, as well as triplets consisting of users, products, and marketing scenarios, covering budget data, scaling capability data, attribution efficiency, and estimated conversion rates. Compared to traditional fixed-threshold strategies that only consider one or a few metrics, this multi-dimensional evaluation can more comprehensively reflect the advertising situation, making the evaluation more consistent with actual campaign performance and providing a richer data foundation for accurate attribution and strategy adjustments.

[0033] In some embodiments, a dynamic performance evaluation matrix is ​​constructed based on a second performance evaluation metric consisting of a tripartite structure of users, products, and marketing scenarios. Specifically, the attribution efficiency of the marketing scenario is correlated with budget data, and the estimated conversion rate is used as a third dimension to determine whether the marketing scenario is in an abnormal state of high cost and low efficiency. This scenario-based evaluation logic allows attribution to no longer rely on general product or user data, but to be precise down to the specific scenario and how efficient the product is at converting for that user, thus improving the scenario adaptability and accuracy of attribution.

[0034] S102. Based on the comparison results between the first effect evaluation index and the second effect evaluation index, the attribution threshold is calibrated.

[0035] The attribution threshold includes at least one of the following: estimated conversion rate threshold, budget threshold, and attribution efficiency threshold. The estimated conversion rate threshold is used to adjust product allocation for different marketing scenarios. The budget threshold represents the available budget for different marketing scenarios. The attribution efficiency threshold is used to adjust budget allocation for different marketing scenarios.

[0036] In this application, the first performance evaluation metric for the binary group consisting of users and their products is used as the baseline data, while the second performance evaluation metric for the tripartite group consisting of users, products, and marketing scenarios is used as the refined adjustment data. The deviation between the refined adjustment data and the baseline data is compared to calibrate the attribution threshold. The degree to which the refined adjustment data deviates from the baseline data is used to calibrate the attribution threshold. If the refined adjustment data is greater than the baseline data, the threshold is lowered; if the refined adjustment data is less than the baseline data, the threshold is increased. Calibrating the attribution threshold can prevent resource misallocation due to attribution bias.

[0037] We can first determine the initial attribution threshold benchmark, using the first performance evaluation metric of the binary [user, product] as the overall benchmark data. For example, we can use the estimated conversion rate of this binary as the initial estimated conversion rate threshold benchmark, and the total budget as the initial budget threshold benchmark. Then, we compare the second performance evaluation metrics of each triple [user, product, marketing scenario] with the overall benchmark data: For the estimated conversion rate threshold, if the estimated conversion rate of a certain triplet is higher than the overall benchmark, the estimated conversion rate threshold for that scenario will be lowered (e.g., if the overall benchmark is 20% and the scenario data is 30%, the threshold for that scenario will be lowered to 15%) to increase the probability of obtaining product placement quota for that scenario; if it is lower than the overall benchmark, the threshold will be appropriately increased.

[0038] Regarding the budget threshold, combining the scaling capability data of the triplet with the total budget of the binary, if the scaling capability data of a certain scenario is higher than the average level of the market (e.g., the average scaling capability of the binary is 500 impressions / hour, and the scenario data is 800 impressions / hour), then the budget threshold for that scenario will be increased within the total budget range (e.g., from 10% to 15% of the total budget); otherwise, it will be decreased.

[0039] Regarding the attribution efficiency threshold, if the attribution efficiency of a certain scenario is higher than the benchmark (e.g., the benchmark is 30% and the scenario data is 45%), the attribution efficiency threshold for that scenario will be lowered, and more budget will be allocated to that scenario first; if it is lower than the benchmark and lower than the standard deviation of the threshold (e.g., the standard deviation is set at 5% and the scenario data is 23%), a temporary threshold increase will be triggered to limit budget inflow.

[0040] In some embodiments, when performing step S102, the attribution efficiency deviation is first calculated based on the attribution efficiency included in the first effect evaluation index and the second effect evaluation index, then the tilt coefficient corresponding to the attribution efficiency deviation is determined, and then the attribution threshold is adjusted based on the tilt coefficient.

[0041] The attribution efficiency (labeled as primary attribution efficiency for clarity) included in the first performance evaluation metric reflects basic conversion capability. The attribution efficiency (i.e., secondary attribution efficiency) included in the second performance evaluation metric reflects the impact of the marketing scenario on conversion.

[0042] Specifically, the difference between the second attribution efficiency and the first attribution efficiency is first calculated. Then, the ratio of this difference to the first attribution efficiency is calculated to obtain the attribution efficiency deviation. The attribution efficiency deviation reflects the performance gap of the current ad placement.

[0043] In this application, different attribution efficiency biases correspond to different skew coefficients. The skew coefficient is used to adjust the attribution threshold. For example, the attribution efficiency bias is divided into 10 levels, and each level corresponds to a skew coefficient. An attribution efficiency bias of 10% corresponds to a skew coefficient of 10%, and an attribution efficiency bias of 20% corresponds to a skew coefficient of 20%.

[0044] In the above embodiments, the attribution efficiency deviation can capture the fluctuations in attribution efficiency caused by differences in marketing scenarios. Then, through a preset correspondence between the attribution efficiency deviation and the tilt coefficient, the attribution efficiency deviation is quantified into an operable tilt coefficient to adjust the attribution threshold. This mechanism makes the threshold no longer a fixed value, but a dynamic parameter that can be calibrated in real time according to changes in scenario efficiency, solving the problem that fixed thresholds cannot adapt to scenario differences.

[0045] In some embodiments, during step S102, the comparison results between the first and second performance evaluation indicators are categorized and weighted. The comparison results include at least two of the following: budget deviation, scaling-up capability deviation, attribution efficiency deviation, and estimated conversion rate deviation. Then, the corresponding tier value is multiplied by its weight, and the results are summed to obtain a weighted total score. This weighted total score is then standardized into adjustment parameters. Simultaneously, the adjustment sensitivity corresponding to the scaling-up capability data included in the first performance evaluation indicator is obtained. The budget adjustment ratio is then calculated based on the adjustment parameters and adjustment sensitivity, and the initial budget threshold is adjusted according to the budget adjustment ratio to obtain the budget threshold. The adjustment sensitivity reflects the product's response speed to budget adjustments.

[0046] In the specific implementation of the above embodiments, firstly, for each triplet (specific marketing scenario), the difference between its second performance evaluation indicator and the first performance evaluation indicator in each dimension is calculated. Assuming the total budget in the first performance evaluation indicator is B1, the scaling capability data is Q1, the attribution efficiency is E1, and the estimated conversion rate is P1, and the scenario budget consumption ratio in the second performance evaluation indicator is B2, the scenario scaling capability data is Q2, the scenario attribution efficiency is E2, and the scenario estimated conversion rate is P2, then the budget deviation ΔB = B2 - (1 ÷ total number of marketing scenarios) is calculated sequentially to determine whether the scenario budget consumption deviates from the average allocation ratio; the scaling capability deviation ΔQ = Q2 - Q1; the attribution efficiency deviation ΔE = E2 - E1; and the estimated conversion rate deviation ΔP = P2 - P1.

[0047] Then, the comparison results of the performance evaluation indicators (ΔB, ΔQ, ΔE, ΔP) are categorized and weighted. Each indicator deviation can be divided into 10 levels, with level 1 representing the worst performance and level 10 representing the best performance. For example, ΔE ≥ 15% indicates that the scenario attribution efficiency is 15% higher than the overall binary group, corresponding to level 10; ΔE ≤ -15% corresponds to level 1, and the intermediate range is divided proportionally. Similarly, ΔP ≥ 10% corresponds to level 10, ΔP ≤ -10% corresponds to level 1, and so on, ensuring that the deviation of different indicators can be quantified through levels to reflect their performance. Weights are assigned to the four indicator deviations, for example, attribution efficiency deviation (ΔE) weighted at 30%, estimated conversion rate deviation (ΔP) weighted at 25%, scaling capability deviation (ΔQ) weighted at 25%, and budget deviation (ΔB) weighted at 20%, prioritizing the impact of conversion-related indicators on the adjustment parameters.

[0048] First, multiply the deviation value of each indicator by its weight, then sum them to obtain the weighted total score. The standardized value of the weighted total score is then set to a preset range (e.g., 0.6-1.4). The standardization rules are as follows: the lowest weighted total score of 5 points corresponds to an adjustment parameter K=0.6, the highest score of 45 points corresponds to K=1.4, and scores in between are calculated as "K=0.6+(weighted total score-5)×0.02". Where K>1 indicates that the marketing scenario performs better than the overall level of the binary group [user, product], requiring an upward adjustment of the budget; K<1 indicates that the performance is worse than the overall level, requiring a downward adjustment of the budget; K=1 indicates that the performance is on par with the overall level, maintaining the basic adjustment direction.

[0049] In the above embodiments, when obtaining the adjustment sensitivity corresponding to the volume capacity data included in the first effect evaluation index, the volume capacity data affects the adjustment sensitivity. Products with high volume capacity have high sensitivity, and the adjustment parameters respond more violently to deviations. Budget adjustments can be quickly consumed and bring about conversion, supporting large-scale adjustments. Products with low volume capacity have low sensitivity, and the adjustment parameters are more gradual. Budget adjustments are consumed slowly, and the adjustment range of the budget threshold needs to be controlled to avoid resource waste.

[0050] Optionally, the initial performance evaluation index, Q1, can be divided into three levels: low running volume, medium running volume, and high running volume. A corresponding adjustment sensitivity S can be assigned to each level. The stronger the running volume capability, the more it can withstand larger budget adjustments, hence the higher the sensitivity.

[0051] For example, Q1≤5% (exposure growth rate per unit time≤5%) is considered low-volume performance, 5%<Q1≤15% is considered medium-volume performance, and Q1>15% is considered high-volume performance. Low-volume performance corresponds to S=0.8, indicating a smoother response to budget adjustments, avoiding difficulties in scaling up due to excessive adjustments; medium-volume performance corresponds to S=1.0, representing normal sensitivity; and high-volume performance corresponds to S=1.2, indicating a more responsive response to budget adjustments, allowing for rapid allocation of budget to high-quality scenarios or reduction of budget for inefficient scenarios.

[0052] Furthermore, the budget adjustment ratio is calculated based on the adjustment parameter K and the adjustment sensitivity S, and the initial budget threshold is adjusted according to the budget adjustment ratio to obtain the budget threshold. The initial budget threshold is a benchmark value set based on historical data or business objectives. It can be the amount of money evenly distributed across the total budget of the binary tuple [user, product] according to the number of marketing scenarios. Assuming the total budget is B and the number of marketing scenarios is N, then the initial budget threshold B0 = B / N.

[0053] In this step, the budget adjustment ratio R is first calculated based on the adjustment parameter K and the adjustment sensitivity S. This can be calculated using the formula R = K × S - constant term, where the constant term can be 0.5. The constant term ensures that the budget adjustment ratio is within a preset range (e.g., -0.1 to 0.3), avoiding excessively extreme adjustment. The initial budget threshold B0 is then adjusted based on the budget adjustment ratio R to obtain the budget threshold B_final = B0 × (1 + R).

[0054] In some embodiments, when performing step S102, firstly, the attribution efficiency deviation is calculated based on the attribution efficiency included in the first and second effect evaluation indicators; the budget deviation is calculated based on the budget data included in the first and second effect evaluation indicators; then, a proportional coefficient is calculated based on the weights of the attribution efficiency deviation and the budget deviation; an integral coefficient is calculated by adding an integral limit to the cumulative value of the deviations of the attribution efficiency deviation and the budget deviation within a preset time period; a differential coefficient is calculated based on the rate of change of the deviations of the attribution efficiency deviation and the budget deviation within the preset time period; the proportional coefficient, integral coefficient, and differential coefficient are weighted and calculated to obtain a combination of adjustment parameters; then, the adjustment sensitivity corresponding to the volume-generating capacity data included in the first effect evaluation indicator is obtained; the budget adjustment ratio is calculated based on the combination of adjustment parameters and the adjustment sensitivity; and the initial budget threshold is adjusted based on the budget adjustment ratio to obtain the budget threshold.

[0055] Specifically, firstly, for the first attribution efficiency E1 included in the first performance evaluation indicator and the second attribution efficiency E2 included in the second performance evaluation indicator, calculate the attribution efficiency deviation ΔE = E2 - E1. Then, for the first budget data B1 included in the first performance evaluation indicator and the second budget data B2 included in the second performance evaluation indicator, calculate the budget deviation. When calculating the budget deviation, first calculate the scenario budget consumption progress deviation ΔB1 = B2 - B1; then determine the scenario attribution efficiency level coefficient C based on the level of ΔE; budget deviation ΔB = ΔB1 × C. For example, if ΔB1 = 20% and C = 1.2, then ΔB = 24%, indicating that the consumption progress of this high-quality scenario is faster than the overall average, and the deviation weight should be increased to prioritize budget protection. If ΔB is positive, it indicates that the scenario budget consumption is faster than the overall average or that the scenario is high-quality and requires budget allocation; if it is negative, it indicates that the consumption is slow or the scenario is inefficient and requires budget control.

[0056] Then, based on the weights of attribution efficiency deviation and budget deviation, the proportional coefficient Kp is calculated. Kp responds to the deviation magnitude and determines the strength of threshold adjustment; the higher the attribution efficiency and the more sufficient the budget, the larger Kp. The basic weight Kp1 can be assigned according to the level of ΔE, for example, level 1 corresponds to 0.3, level 10 corresponds to 1.5, and each level increases by 0.13; the higher the level, the better the scenario, the larger Kp1, and the more sensitive the response to deviation; then, the correction weight Kp2 is assigned according to the direction of budget deviation ΔB, for example, Kp2=1.1 when ΔB>0, Kp2=1.0 when ΔB=0, and Kp2=0.9 when ΔB<0. Kp is increased for fast consumption or high-quality scenarios, and decreased for slow consumption or inefficient scenarios; finally, Kp=Kp1×Kp2. For example, if ΔE is 8 levels (Kp1=1.2) and ΔB=24% (Kp2=1.1), then Kp=1.2×1.1=1.32, indicating that this scenario requires a rapid response to deviations and priority should be given to adjusting the budget.

[0057] For the cumulative values ​​of attribution efficiency deviation and budget deviation within a preset time period, an integral limit is added to calculate the integral coefficient Ki. Ki responds to the trend of the cumulative deviation and determines the persistence of threshold adjustment. First, ΔE and ΔB for each time segment are summed and accumulated to calculate the cumulative deviation value ΣΔ=Σ(ΔE+ΔB) for the preset time period (e.g., 12 5-minute time segments). Then, the base value of Ki is determined based on the sign and magnitude of ΣΔ. When ΣΔ>0, the cumulative deviation is positive, and budget growth needs to be suppressed, Ki=0.2-0.001×ΣΔ; when ΣΔ≤0, the cumulative deviation is negative, and budget consumption needs to be promoted, Ki=0.2+0.001×|ΣΔ|. Finally, an integral limit is added to control Ki within a preset range (e.g., 0.05-0.5) to avoid excessive Ki due to long-term deviation, which could lead to budget overshoot or undershoot. For example, if ΣΔ=30% in the past hour, then Ki=0.2-0.001×30=0.17, and within the limit range, the cumulative positive deviation can be slowly eliminated.

[0058] The differential coefficient Kd is calculated based on the rate of change of attribution efficiency deviation and budget deviation within a preset time period. Kd responds to the rate of change of deviation and determines the timeliness of threshold adjustment. First, the rate of change of attribution efficiency deviation ΔE_rate = current ΔE - previous period ΔE, and the rate of change of budget deviation ΔB_rate = current ΔB - previous period ΔB are calculated; then the comprehensive deviation rate of change is calculated, such as Δ_rate = (ΔE_rate × 0.6 + ΔB_rate × 0.4); finally, Kd = 0.1 × |Δ_rate|. For example, assuming current ΔE = 5%, previous period ΔE = 3%, ΔE_rate = 2%; current ΔB = 24%, previous period ΔB = 20%, ΔB_rate = 4%; then Δ_rate = 2% × 0.6 + 4% × 0.4 = 2.8%, Kd = 0.1 × 2.8 = 0.28, indicating that the deviation is on the rise, and Kd needs to be used to suppress the excessively rapid growth of budget adjustments and maintain stability.

[0059] Referring to the foregoing embodiments, the adjustment sensitivity S corresponding to the volume-generating capacity data included in the first performance evaluation index is obtained. Then, based on the combination of adjustment parameters and the adjustment sensitivity, the budget adjustment ratio is calculated and applied to the initial budget threshold to obtain the budget threshold.

[0060] For example, the budget threshold Final_Budget is calculated according to the following formula (1): (1) In formula (1), Init_Budget is the initial budget threshold, and the adjustment parameter combination is a weighted calculation result of the proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd. If both the attribution efficiency deviation and the budget deviation are positive, the adjustment parameter combination is positive, increasing the budget threshold; if both the attribution efficiency deviation and the budget deviation are negative, the adjustment parameter combination is negative, decreasing the budget threshold. It should be noted that this application recalculates the attribution efficiency deviation, budget deviation, and adjustment parameters at preset intervals (e.g., 5 minutes) to dynamically update the final budget threshold, ensuring that it always adapts to the latest campaign performance.

[0061] The above embodiments simultaneously calculate attribution efficiency bias and budget bias, using both as core inputs to the adjustment parameters. This allows the adjustment of the budget threshold to no longer be limited by a single rule, but rather to balance the constraints of efficiency and budget. The adjustment parameters are dynamically determined by the attribution efficiency bias and budget bias, and combined with the adjustment sensitivity corresponding to the scaling capability data, ensuring that the adjustment parameters are adapted to product characteristics. This dynamic generation mechanism of adjustment parameters avoids the universality defects of empirical parameters, making adjustments more closely aligned with actual needs.

[0062] In some embodiments, when performing step S102, the long-term cumulative bias and short-term time bias are first calculated based on the first effect evaluation index and the second effect evaluation index; then, the adjustment amount of the attribution threshold is calculated based on the long-term cumulative bias and the short-term time bias; and then the initial attribution threshold is calibrated based on the adjustment amount.

[0063] Optionally, the adjustment amount pid_delta of the attribution threshold can be calculated according to the following formula (2): (2) In formula (2), pid_delta is used to correct the attribution threshold and offset bias. sum_bias is the long-term cumulative bias, which can be the cumulative budget consumption bias, the cumulative estimated conversion rate bias, or the cumulative attribution efficiency bias; timeslot_bias is the short-term time-slot bias, which can be the budget bias between adjacent time periods, the estimated conversion rate bias between adjacent time periods, or the attribution efficiency bias between adjacent time periods. The proportional coefficient Kp and the differential coefficient Kd are used to adjust the weight of the influence of the long-term cumulative bias sum_bias and the short-term time-slot bias timeslot_bias on the adjustment amount pid_delta. The values ​​of Kp and Kd need to be adjusted according to the business scenario (such as product type, deployment stage).

[0064] Formula (2) above calculates a reasonable adjustment amount through long-term cumulative deviation and short-term time-period deviation, and dynamically corrects the attribution threshold over time. Specifically, the proportional term (Kp*sum_bias) corrects the long-term trend deviation and adjusts for the long-term cumulative deviation to ensure that the overall deployment does not deviate from the long-term goal; the differential term (Kd * timeslot_bias) corrects the short-term fluctuation deviation and adjusts for the implementation deviation in the most recent period to quickly respond to sudden changes and improve the timeliness of adjustment.

[0065] The above embodiments achieve more accurate, timely, and business-adaptive dynamic calibration of the attribution threshold by dynamically weighting the attribution threshold through dual time-sensitivity dimensions.

[0066] Optionally, budget deviation includes cumulative consumption deviation and budget consumption deviation between adjacent time periods. Cumulative consumption deviation is the sum of budget consumption deviations from the start time of product launch to the current time. Adjacent time periods are two adjacent preset time periods, which can be 5 minutes apart. For example, the budget consumption deviation between the current time period and the adjacent previous time period. The budget consumption for any time period is calculated based on peak traffic and the expected budget.

[0067] Based on the budget data included in the first and second performance evaluation indicators, the cumulative consumption deviation and the budget consumption deviation between adjacent time periods are calculated. Then, based on the cumulative consumption deviation and the budget consumption deviation between adjacent time periods, the adjustment amount of the budget threshold is calculated, and the budget threshold is further calibrated based on this adjustment amount.

[0068] In some embodiments, during step S102, a ratio is calculated between the first effect evaluation index and the second effect evaluation index to determine a correction coefficient. Then, an initial attribution threshold is obtained, and the initial attribution threshold is adjusted based on the correction coefficient to obtain the final attribution threshold. This achieves calibration of the attribution threshold.

[0069] Among them, the correction coefficient reflects the impact of the marketing scenario on conversion and determines the adjustment range of the attribution threshold.

[0070] For example, firstly, the ratio of the first estimated conversion rate (included in the first performance evaluation indicator) to the second estimated conversion rate (included in the second performance evaluation indicator) is calculated: first estimated conversion rate / second estimated conversion rate. Then, an initial attribution threshold is obtained, and the attribution threshold is calculated using a Bayesian dynamic linear model according to the formula: Attribution threshold = Initial attribution threshold * (First estimated conversion rate / Second estimated conversion rate). The Bayesian dynamic linear model optimizes the calculation of the correction coefficient, making the threshold adjustment more closely reflect long-term trends and reducing bias caused by short-term random fluctuations.

[0071] In some embodiments, a budget threshold is set based on the remaining budget included in the budget data and the cumulative overspending risk data. Limiting the scope of the integral term prevents a single feedback data point from excessively influencing overall control.

[0072] Since different products have different budgets, traffic capacity and conversion rates in different ad slots, the estimated conversion rate threshold, budget threshold and attribution efficiency threshold can be flexibly adjusted based on real-time or periodic evaluation data.

[0073] In some embodiments, the calibration speed of the attribution threshold is adjusted based on the scaling capability data included in the first performance evaluation metric. Specifically, a correlation curve between budget consumption rate and conversion growth is plotted based on the scaling capability data, the slope of the correlation curve is determined, and the calibration speed is adjusted according to the slope. The steeper the slope of the correlation curve, the higher the conversion growth per unit budget, and the greater the adjustment in calibration speed, thus achieving a dynamic match between budget consumption and conversion efficiency.

[0074] S103a. If the attribution threshold includes the estimated conversion rate threshold, and the actual estimated conversion rate of the target marketing scenario is higher than the estimated conversion rate threshold, then increase the product placement quota for the target marketing scenario.

[0075] If the attribution threshold includes the estimated conversion rate threshold, and the actual estimated conversion rate of the target marketing scenario is higher than the estimated conversion rate threshold, then the ad allocation for the target marketing scenario is increased, such as from 10% to 15% of the total allocation, to give high-conversion-potential scenarios more exposure opportunities. Conversely, the allocation for inefficient scenarios is reduced to decrease ineffective ad placements.

[0076] In some embodiments, a standard deviation is calculated based on the estimated conversion rate and the estimated conversion rate threshold included in the second performance evaluation metric. When the estimated conversion rate is less than the standard deviation, a budget freeze is triggered.

[0077] The standard deviation is used to define a reasonable fluctuation range for the estimated conversion rate. When the estimated conversion rate in the second performance evaluation metric falls below its corresponding standard deviation, it indicates that the metric has significantly deviated from the reasonable fluctuation range. At this point, a budget freeze is triggered, immediately suspending budget allocation for that marketing scenario and preventing further resource consumption due to persistently low conversion rates. The frozen budget can then be quickly transferred to marketing scenarios with better estimated conversion rates, achieving dynamic reallocation of resources.

[0078] S103b. If the attribution threshold includes the budget threshold, then determine the available budget corresponding to the target marketing scenario.

[0079] If the attribution threshold includes the budget threshold, then the total budget for the tuple [user, product] is split according to the budget threshold ratio for each scenario to determine the available budget amount for the target marketing scenario. For example, if the total budget is 1 million yuan and the budget threshold for scenario A is 15%, then the available budget for scenario A is 150,000 yuan.

[0080] S103c. If the attribution threshold includes the attribution efficiency threshold, and the actual attribution efficiency of the target marketing scenario is higher than the attribution efficiency threshold, then increase the budget for the target marketing scenario.

[0081] If the attribution threshold includes the attribution efficiency threshold, and the actual attribution efficiency of the target marketing scenario is higher than the attribution efficiency threshold (e.g., the threshold is 25%, and scenario C actually reaches 35%), then additional budget will be allocated to scenario C. If the attribution efficiency of scenario D is lower than the threshold, then the remaining budget for scenario D will be frozen and transferred to high-efficiency scenario C to avoid budget waste.

[0082] In some embodiments, the attribution efficiency threshold and real-time budget consumption data are combined to monitor the attribution efficiency and budget consumption of each scenario according to a preset period: if the attribution efficiency of a certain scenario is consistently higher than the threshold and the budget is not exhausted, the budget is further tilted, such as adding 5%-10% of the available budget; if the attribution efficiency is lower than the threshold and the budget is consumed too quickly, the budget allocation for that scenario is suspended and the remaining budget is transferred to high-efficiency scenarios.

[0083] This approach of dynamically adjusting strategies based on actual conditions overcomes the rigidity of fixed threshold strategies, facilitating long-term optimization and growth in advertising campaigns. It makes attribution-based advertising more flexible, adaptable to different products and marketing scenarios, and enables dynamic and precise control over product attribution-based advertising.

[0084] In summary, this application provides an advertising attribution delivery method. This method obtains performance evaluation indicators for binary pairs consisting of users and products, as well as triplets consisting of users, products, and marketing scenarios. These indicators cover budget data, scaling capability data, attribution efficiency, and estimated conversion rate. This multi-dimensional evaluation can more comprehensively reflect the advertising delivery situation. Based on the comparison results of different performance evaluation indicators, the attribution threshold is calibrated instead of using a fixed threshold, allowing for flexible adjustment of the estimated conversion rate threshold, budget threshold, and attribution efficiency threshold. Adjusting the delivery quota through the estimated conversion rate threshold prioritizes resources towards scenarios with high conversion potential, improving the accuracy of product exposure. The budget threshold clarifies the upper limit of available budget for each scenario, achieving initial reasonable allocation of resources. The attribution efficiency threshold further optimizes budget allocation, tilting towards high-efficiency scenarios and improving the overall efficiency of marketing funds. Through a dynamic adjustment mechanism that compares actual metrics with thresholds, such as increasing the allocation for scenarios with high estimated conversion rates and adding budgets for scenarios with high attribution efficiency, we can respond to market changes in real time and maximize marketing effectiveness. This ensures the targeting of resource allocation and improves the return on investment of the overall marketing campaign, realizing the transformation of product placement from extensive to intensive.

[0085] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0086] like Figure 2 As shown, embodiments of this application also provide a product attribution delivery device, the device comprising: The acquisition module 201 is used to acquire the first performance evaluation index of the binary group consisting of users and products, and the second performance evaluation index of the triple group consisting of users, products and marketing scenarios; wherein, the performance evaluation index includes budget data, scaling capability data, attribution efficiency and estimated conversion rate. The threshold calibration module 202 is used to calibrate the attribution threshold based on the comparison results of the first performance evaluation index and the second performance evaluation index; wherein, the attribution threshold includes at least one of the estimated conversion rate threshold, the budget threshold, and the attribution efficiency threshold. The estimated conversion rate threshold is used to adjust the product placement quota for different marketing scenarios, the budget threshold is used to characterize the available budget for different marketing scenarios, and the attribution efficiency threshold is used to adjust the budget allocation for different marketing scenarios. The adjustment module 203 is used to increase the product placement quota for the target marketing scenario if the attribution threshold includes the estimated conversion rate threshold and the actual estimated conversion rate of the target marketing scenario is higher than the estimated conversion rate threshold; if the attribution threshold includes the budget threshold, the available budget for the target marketing scenario is determined; if the attribution threshold includes the attribution efficiency threshold and the actual attribution efficiency of the target marketing scenario is higher than the attribution efficiency threshold, the budget for the target marketing scenario is increased.

[0087] As an optional implementation provided in this application, the threshold calibration module 202 is specifically used for: classifying and weighting the comparison results of the first effect evaluation index and the second effect evaluation index; wherein, the comparison results include at least two of the following: budget deviation, volume-building capability deviation, attribution efficiency deviation, and estimated conversion rate deviation; multiplying the corresponding value of the comparison result by its weight, and then summing them to obtain a weighted total score; standardizing the weighted total score into an adjustment parameter; obtaining the adjustment sensitivity corresponding to the volume-building capability data included in the first effect evaluation index; calculating the budget adjustment ratio according to the adjustment parameter and the adjustment sensitivity, and adjusting the initial budget threshold according to the budget adjustment ratio to obtain the budget threshold.

[0088] As an optional implementation provided in this application, the threshold calibration module 202 is specifically used for: calculating the attribution efficiency deviation based on the attribution efficiency included in the first effect evaluation index and the second effect evaluation index; calculating the budget deviation based on the budget data included in the first effect evaluation index and the second effect evaluation index; multiplying the weight of the attribution efficiency deviation and the weight of the budget deviation to calculate the proportional coefficient; adding an integral limit to the cumulative value of the deviations of the attribution efficiency deviation and the budget deviation within a preset time period to calculate the integral coefficient; calculating the differential coefficient based on the rate of change of the deviations of the attribution efficiency deviation and the budget deviation within the preset time period; performing a weighted calculation on the proportional coefficient, the integral coefficient, and the differential coefficient to obtain the adjustment parameter combination; obtaining the adjustment sensitivity corresponding to the volume-generating capacity data included in the first effect evaluation index; calculating the budget adjustment ratio based on the adjustment parameter combination and the adjustment sensitivity; and adjusting the initial budget threshold according to the budget adjustment ratio to obtain the budget threshold.

[0089] As an optional implementation provided in this application, the threshold calibration module 202 is specifically used for: performing a ratio calculation on the first effect evaluation index and the second effect evaluation index to calculate a correction coefficient; obtaining an initial attribution threshold; and adjusting the initial attribution threshold according to the correction coefficient to obtain an attribution threshold.

[0090] As an optional implementation provided in this application, the threshold calibration module 202 is specifically used for: calculating the attribution efficiency deviation based on the first effect evaluation index and the attribution efficiency included in the second effect evaluation index; determining the tilt coefficient corresponding to the attribution efficiency deviation; and adjusting the attribution threshold based on the tilt coefficient.

[0091] As an optional implementation provided in this application, the threshold calibration module 202 is specifically used for: calculating the long-term cumulative deviation and the short-term time period deviation based on the first effect evaluation index and the second effect evaluation index; calculating the adjustment amount of the attribution threshold based on the long-term cumulative deviation and the short-term time period deviation; and calibrating the initial attribution threshold based on the adjustment amount.

[0092] Regarding the apparatus in the above embodiments, the specific manner in which each unit performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0093] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0094] For example, such as Figure 3 As shown, the electronic device includes a memory 301 and a processor 302. The memory 301 stores executable program code, and the processor 302 is used to call and execute the executable program code to perform a product attribution delivery method.

[0095] This embodiment can divide the electronic device into functional modules based on the above method example. For example, each function can be assigned to a separate module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods. When dividing each functional module according to its corresponding function, the electronic device may include: an acquisition module, a threshold calibration module, and a strategy adjustment module, etc. It should be noted that all relevant content of each step involved in the above method embodiment can be referenced from the functional description of the corresponding functional module, and will not be repeated here.

[0096] The electronic device provided in this embodiment is used to execute the above-described product attribution delivery method, and therefore can achieve the same effect as the above-described implementation method.

[0097] When using integrated units, the electronic device may include a processing module and a storage module. The processing module is used to control and manage the operation of the electronic device. The storage module is used to support the execution of program code and data by the electronic device.

[0098] The processing module may be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits as disclosed in this application. The processor may also be a combination of computing functions, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc., and the storage module may be a memory.

[0099] Embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above-described product attribution delivery method embodiments.

[0100] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above-described product attribution delivery method embodiments when it is run.

[0101] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0102] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above-described product attribution delivery method embodiments.

[0103] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in any of the above-described product attribution delivery method embodiments.

[0104] The beneficial effects of the above embodiments can be referred to the beneficial effects of the corresponding methods provided above, and will not be repeated here.

[0105] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0106] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or 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 device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0107] In the description of this application, it should be understood that if the terms "upper", "lower", "front", "rear", "left" and "right" are used to indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the position or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0108] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0109] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A product attribution placement method, characterized in that, include: The first performance evaluation metric is obtained from the binary combination of users and products, and the second performance evaluation metric is obtained from the tripartite combination of users, products and marketing scenarios. The performance evaluation metrics include budget data, scaling capability data, attribution efficiency and estimated conversion rate. Based on the comparison results between the first performance evaluation index and the second performance evaluation index, the attribution threshold is calibrated; wherein, the attribution threshold includes at least one of the estimated conversion rate threshold, the budget threshold, and the attribution efficiency threshold, the estimated conversion rate threshold is used to adjust the product placement quota for different marketing scenarios, the budget threshold is used to characterize the available budget for different marketing scenarios, and the attribution efficiency threshold is used to adjust the budget allocation for different marketing scenarios; If the attribution threshold includes the estimated conversion rate threshold, and the actual estimated conversion rate of the target marketing scenario is higher than the estimated conversion rate threshold, then the product placement quota for the target marketing scenario is increased. If the attribution threshold includes the budget threshold, then the available budget corresponding to the target marketing scenario is determined; If the attribution threshold includes the attribution efficiency threshold, and the actual attribution efficiency of the target marketing scenario is higher than the attribution efficiency threshold, then the budget for the target marketing scenario is increased.

2. The method according to claim 1, characterized in that, The step of calibrating the attribution threshold based on the comparison result between the first effect evaluation index and the second effect evaluation index includes: The comparison results between the first performance evaluation indicator and the second performance evaluation indicator are divided into levels and weighted accordingly; wherein, the comparison results include at least two of the following: budget deviation, scaling-up capability deviation, attribution efficiency deviation, and estimated conversion rate deviation. Multiply the gear value corresponding to the comparison result by its weight, and then sum them to obtain the weighted total score; The weighted total score is standardized into an adjustment parameter; Obtain the adjustment sensitivity corresponding to the volume-generating capacity data included in the first performance evaluation index; The budget adjustment ratio is calculated based on the adjustment parameters and the adjustment sensitivity, and the initial budget threshold is adjusted according to the budget adjustment ratio to obtain the budget threshold.

3. The method according to claim 1, characterized in that, The step of calibrating the attribution threshold based on the comparison result between the first effect evaluation index and the second effect evaluation index includes: Calculate the attribution efficiency bias based on the first performance evaluation index and the attribution efficiency included in the second performance evaluation index; Calculate the budget deviation based on the first performance evaluation index and the budget data included in the second performance evaluation index; The proportionality coefficient is calculated by multiplying the weight of the attribution efficiency bias and the weight of the budget bias. The cumulative values ​​of the attribution efficiency deviation and the budget deviation within a preset time period are added to the integral limit to calculate the integral coefficient. Calculate the differential coefficients based on the rate of change of the attribution efficiency deviation and the budget deviation within a preset time period; The proportional coefficient, the integral coefficient, and the derivative coefficient are weighted and calculated to obtain the combination of adjustment parameters; Obtain the adjustment sensitivity corresponding to the volume-generating capacity data included in the first performance evaluation index; Calculate the budget adjustment ratio based on the combination of adjustment parameters and the adjustment sensitivity; The initial budget threshold is adjusted according to the budget adjustment ratio to obtain the budget threshold.

4. The method according to claim 1, characterized in that, The step of calibrating the attribution threshold based on the comparison result between the first effect evaluation index and the second effect evaluation index includes: Calculate the correction coefficient by performing a ratio calculation between the first performance evaluation index and the second performance evaluation index; Obtain the initial attribution threshold; The initial attribution threshold is adjusted according to the correction coefficient to obtain the attribution threshold.

5. The method according to claim 1, characterized in that, The step of calibrating the attribution threshold based on the comparison result between the first effect evaluation index and the second effect evaluation index includes: Calculate the attribution efficiency bias based on the first performance evaluation index and the attribution efficiency included in the second performance evaluation index; Determine the tilt coefficient corresponding to the attribution efficiency bias; The attribution threshold is adjusted based on the tilt coefficient.

6. The method according to claim 1, characterized in that, The step of calibrating the attribution threshold based on the comparison result between the first effect evaluation index and the second effect evaluation index includes: Based on the first performance evaluation index and the second performance evaluation index, calculate the long-term cumulative deviation and the short-term period deviation; The adjustment amount of the attribution threshold is calculated based on the long-term cumulative deviation and the short-term time period deviation; The initial attribution threshold is calibrated based on the adjustment amount.

7. A product attribution delivery device, characterized in that, include: The acquisition module is used to acquire the first performance evaluation metric of the binary group consisting of users and products, and the second performance evaluation metric of the triple group consisting of users, products and marketing scenarios. The performance evaluation metrics include budget data, scaling capability data, attribution efficiency and estimated conversion rate. A threshold calibration module is used to calibrate the attribution threshold based on the comparison result between the first performance evaluation index and the second performance evaluation index; wherein, the attribution threshold includes at least one of an estimated conversion rate threshold, a budget threshold, and an attribution efficiency threshold, the estimated conversion rate threshold is used to adjust the product placement quota for different marketing scenarios, the budget threshold is used to characterize the available budget for different marketing scenarios, and the attribution efficiency threshold is used to adjust the budget allocation for different marketing scenarios; The adjustment module is configured to: increase the product placement quota for the target marketing scenario if the attribution threshold includes the estimated conversion rate threshold and the actual estimated conversion rate of the target marketing scenario is higher than the estimated conversion rate threshold; determine the available budget corresponding to the target marketing scenario if the attribution threshold includes the budget threshold; and increase the budget for the target marketing scenario if the attribution threshold includes the attribution efficiency threshold and the actual attribution efficiency of the target marketing scenario is higher than the attribution efficiency threshold.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the product attribution delivery method as described in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, it implements the steps of the product attribution delivery method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the product attribution delivery method as described in any one of claims 1 to 6.

Images