CN115170167A - Advertisement diagnosis method, device, computer equipment and storage medium - Google Patents

Advertisement diagnosis method, device, computer equipment and storage medium Download PDF

Info

Publication number
CN115170167A
CN115170167A CN202110363406.6A CN202110363406A CN115170167A CN 115170167 A CN115170167 A CN 115170167A CN 202110363406 A CN202110363406 A CN 202110363406A CN 115170167 A CN115170167 A CN 115170167A
Authority
CN
China
Prior art keywords
deviation
cost
advertisement
estimated
conversion
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110363406.6A
Other languages
Chinese (zh)
Inventor
陶进
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to CN202110363406.6A priority Critical patent/CN115170167A/en
Publication of CN115170167A publication Critical patent/CN115170167A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0251—Targeted advertisements
    • G06Q30/0269—Targeted advertisements based on user profile or attribute
    • G06Q30/0271—Personalized advertisement
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0273—Determination of fees for advertising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0277—Online advertisement

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application relates to an advertisement diagnosis method, an advertisement diagnosis device, computer equipment and a storage medium. The method comprises the following steps: acquiring hour level data and day level data of indexes required by diagnosis according to advertisement related data of the advertisement to be diagnosed; acquiring index data of the advertisement to be diagnosed in a target time interval from the hour-level data and the day-level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate estimation deviation and conversion rate estimation deviation; obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption; and when the first cost deviation meets the over-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion. By adopting the method, the comprehensiveness of the analysis visual angle and the accuracy of the diagnosis result can be improved.

Description

Advertisement diagnosis method, device, computer equipment and storage medium
Technical Field
The present application relates to the field of internet advertisement technologies, and in particular, to an advertisement diagnosis method, apparatus, computer device, and storage medium.
Background
With the rapid development of the internet, the internet advertisement has a fast propagation speed and a good effect, and has become an important part of modern marketing strategies. The method comprises the steps that an advertiser puts advertisements through media, ocpx is an intelligent bidding putting mode, the advertiser gives an optimization target and expected conversion cost, the system predicts the conversion probability of putting opportunities each time through machine learning and automatically bids in combination with the expected conversion cost to help the advertiser to effectively control the conversion cost and improve the advertisement efficiency.
In the related technology, the system diagnoses according to the learning state and the actual performance of each advertisement, the diagnosis can mark a learning state label on the advertisement according to the putting duration, cost and exposure performance of a new advertisement, if the advertisement fails to learn, the exposure is not expected or the subsequent performance cost is poor, shutdown can be recommended, and once the shutdown recommendation is given, the subsequent change is not carried out. However, the above diagnosis suggestions depend on system internal parameters, such as initial model expression, filtering code filtering ratio and other factors, and are not very consistent with the actual needs of customers, so that the problem that the analysis view angle is not comprehensive exists, and more accurate diagnosis suggestions are lacked.
Disclosure of Invention
In view of the above, it is desirable to provide an advertisement diagnosis method, apparatus, computer device, and storage medium capable of improving comprehensiveness of analysis.
An advertisement diagnostic method, the method comprising:
acquiring hour level data and day level data of indexes required by diagnosis according to advertisement related data of the advertisement to be diagnosed;
acquiring index data of the advertisement to be diagnosed in a target time interval from the hour level data and the day level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate estimation deviation and conversion rate estimation deviation;
obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption;
and when the first cost deviation meets an over-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion.
An advertisement diagnostic device, the device comprising:
the data processing module is used for obtaining hour level data and day level data of indexes required by diagnosis according to advertisement related data of the advertisements to be diagnosed;
the data acquisition module is used for acquiring index data of the advertisement to be diagnosed in a target time interval from the hour level data and the day level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate prediction deviation and conversion rate prediction deviation;
the cost deviation determining module is used for obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption;
and the diagnosis module is used for obtaining a diagnosis result according to the index data when the first cost deviation meets an over-cost condition, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion.
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring hour level data and day level data of indexes required by diagnosis according to advertisement related data of the advertisement to be diagnosed;
acquiring index data of the advertisement to be diagnosed in a target time interval from the hour level data and the day level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate prediction deviation and conversion rate prediction deviation;
obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption;
and when the first cost deviation meets an over-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring hour level data and day level data of indexes required by diagnosis according to advertisement related data of the advertisement to be diagnosed;
acquiring index data of the advertisement to be diagnosed in a target time interval from the hour level data and the day level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate prediction deviation and conversion rate prediction deviation;
obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption;
and when the first cost deviation meets an over-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion.
According to the advertisement diagnosis method, the advertisement diagnosis device, the computer equipment and the storage medium, the hour level data and the day level data of the index required by diagnosis are obtained according to the advertisement related data of the advertisement to be diagnosed; acquiring index data of the advertisement to be diagnosed in a target time interval from the hour-level data and the day-level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate estimation deviation and conversion rate estimation deviation; obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption; and when the first cost deviation meets the over-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion. Therefore, the cost-exceeding reason can be analyzed in a full link, the comprehensiveness of an analysis view angle and the accuracy of a diagnosis result are improved, and the method is helpful for providing richer and more accurate operation suggestions.
Drawings
FIG. 1 is a diagram of an application environment of a method for advertisement diagnosis in one embodiment;
FIG. 2 is a flow diagram illustrating a method of advertisement diagnosis in one embodiment;
FIG. 3 is a schematic diagram of diagnostic logic in one embodiment;
FIG. 4 is a block diagram showing the structure of an advertisement diagnosis apparatus according to an embodiment;
FIG. 5 is a diagram illustrating an internal structure of a computer device according to an embodiment;
FIG. 6 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The advertisement diagnosis method provided by the application can be applied to the application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. A user may access a platform (e.g., a media-side internal diagnostic system, a delivery-side external diagnostic system, an advertisement operation tool, etc.) providing diagnostic services through a terminal 302, and a server 304 may be a server corresponding to the platform. The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as cloud service, a cloud database, cloud computing, cloud functions, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, and a big data and artificial intelligence platform. The terminal may be, but is not limited to, a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and the like. The terminal and the server may be directly or indirectly connected through wired or wireless communication, and the application is not limited herein.
In one embodiment, as shown in fig. 2, an advertisement diagnosis method is provided, which is described by taking the method as an example applied to the server in fig. 1, and includes the following steps S202 to S206.
S202, according to the advertisement related data of the advertisement to be diagnosed, the hour level data and the day level data of the index required by diagnosis are obtained.
The advertisement to be diagnosed is an advertisement which needs to be subjected to super-cost judgment and super-cost reason analysis, and the advertisement related data can include but is not limited to: price adjustment data, operation data, competition performance data and effect data.
The advertisement-related data may be understood as data required for diagnosis, wherein the pricing data is derived from a system pricing log, comprising data on changes in the system pricing factors. The operation data is derived from the operation log and comprises operation information of the advertiser on the advertisement, such as input target conversion cost and the like, and the target conversion cost represents the conversion cost expected by the advertiser, specifically the conversion unit price. The competitive performance data is derived from advertising competitive performance data stored in a tracking log (tracklog). The effectiveness data is derived from the file system (hdfs) and includes post-advertisement exposure performance data such as actual number of conversions, which represents the number of actual conversions formed by clicking on the advertisement, and the like.
Processing the advertisement-related data may generate data for an indicator needed for diagnosis for diagnostic analysis. In one embodiment, existing data computation tools may be utilized to perform hourly quasi-real-time summary computation on advertisement-related data for each exposed advertisement for up to 5 days, while obtaining hourly-level data and daily-level data for the indices required for each diagnosis. Wherein the hour-level data indicates data in hours as a time unit, the hour-level data of the index includes a data value of the index in each hour, the day-level data indicates data in days as a time unit, and the day-level data of the index includes a data value of the index in each day.
S204, acquiring index data of the advertisement to be diagnosed in the target time interval from the hour-level data and the day-level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate prediction deviation and conversion rate prediction deviation.
The target time period may be a day level, for example, a certain day or a certain number of days, and if the target time period is a certain day, the index data of the target time period includes data of each index within the day. The target time period may also be an hour level, such as an hour or several hours, and if the target time period is an hour, the index data of the target time period includes data of the index accumulated to the hour on the day.
Actual consumption represents the actual cost to the advertiser for all exposures, clicks, or conversions. The estimated deviation of click rate (denoted pctrbias) is determined from the deviation of the estimated click rate (denoted pctr) from the actual click rate (denoted ctr), pctrbias = pctr/ctr-1. The predicted deviation of conversion (denoted pcvrbias) was determined from the deviation of the predicted conversion (denoted pcvr) from the actual conversion (denoted cvr), pcvrbias = pcvr/cvr-1.
And S206, obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption.
The target consumption (denoted by gmv) was obtained from the product of the target conversion cost (denoted by targetcpa) and the actual number of conversions (denoted by N), gmv = targetcpa N. From the deviation of the actual consumption (denoted by cost) from the target consumption, a first cost deviation (denoted by cpabias 1) is obtained, cpabias1= cost/gmv-1.
And S208, when the first cost deviation meets the super-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises a super-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion.
The over-cost condition may be that the first cost deviation is greater than a preset first cost deviation threshold, and the first cost deviation threshold may be set in combination with an actual requirement, which is not limited herein. In one embodiment, the first cost variance threshold is set to 30% when the advertisement is targeted for preferential volume and 20% when the advertisement is targeted for steady volume. And when the first cost deviation meets the super-cost condition, analyzing the super-cost reason according to the index data to obtain the super-cost reason of the advertisement to be diagnosed and a corresponding operation suggestion.
And when the first cost deviation does not meet the over-cost condition, judging that the advertisement is not over-cost, and recommending to keep stable delivery. After the advertisement is judged not to be over-cost, whether the first cost deviation meets the underreceiving condition or not can be further judged, and when the first cost deviation meets the underreceiving condition, the advertisement is judged to be under-received. The underrun condition may be that the first cost deviation is greater than a preset second cost deviation threshold, which may be set in conjunction with the actual demand, for example to-20%.
According to the advertisement related data of the advertisement to be diagnosed, acquiring hour level data and day level data of indexes required by diagnosis; acquiring index data of the advertisement to be diagnosed in a target time interval from the hour-level data and the day-level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate estimation deviation and conversion rate estimation deviation; obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption; and when the first cost deviation meets the over-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion. Therefore, the cost-exceeding reason can be analyzed in a full link, the comprehensiveness of an analysis view angle and the accuracy of a diagnosis result are improved, and the method is helpful for providing richer and more accurate operation suggestions.
In one embodiment, the step of obtaining a diagnostic result from the metric data comprises: obtaining the starting number according to the ratio of the actual consumption to the target conversion cost; obtaining a first diagnostic result when the number of starts is less than a first number threshold; the first diagnosis result is used for indicating that the excess cost reason of the advertisement to be diagnosed in the target time interval is insufficient, and the capacity can be improved through competitive analysis; the diagnostic result includes a first diagnostic result.
When the reason of the advertisement over cost is analyzed, whether the starting amount is sufficient or not is judged at first. Taking the ratio of actual consumption to target conversion cost as the number of starts (denoted by n), n = cost/targetcpa. The first quantity threshold (denoted by n 1) may be understood as the minimum quantity of starting material required to consider the starting material sufficient, and when the starting material quantity is less than the first quantity threshold (i.e., n < n 1), it is determined that the starting material is insufficient and the cost is not trusted. In this case the following diagnostic recommendations are given: advertisements are not sufficient today and capacity can be increased by competitive analysis. Wherein n1 may be set according to actual requirements, and is not limited herein. In one embodiment, n1 is set to 3.
In one embodiment, the step of obtaining a diagnostic result based on the indicator data further comprises: when the number of starts is greater than or equal to a first number threshold, acquiring estimated backflow of a future preset time period; subtracting the estimated backflow from the actual consumption to obtain estimated actual consumption, determining a second cost deviation according to the deviation between the estimated actual consumption and the target consumption, and obtaining a second diagnosis result when the second cost deviation does not meet the over-cost condition; the second diagnosis result is used for indicating that the reason of the cost exceeding of the advertisement to be diagnosed in the target time interval is that the advertisement is not completely refluxed and can be kept to be delivered; the diagnostic result includes a second diagnostic result.
When the starting number is greater than or equal to the first number threshold (that is, n is greater than or equal to n 1), the starting number is considered to be sufficient, and whether the conversion is not completed due to slow backflow is judged next, specifically, whether the estimated backflow in the future preset time period does not exceed the cost can be judged, for example, the future preset time period can be 5 days in the future. The estimated backflow may be calculated using existing or future backflow estimation methods, which are not limited herein. The estimated actual consumption is obtained by subtracting the estimated reflux (indicated by back) from the actual consumption (cost), and a second cost deviation (indicated by cpabias), cpabias = (cost-back)/gmv-1, is determined from the deviation of the estimated actual consumption from the target consumption (gmv). And when the second cost deviation does not meet the over-cost condition, judging that the backflow is not complete. In this case the following diagnostic recommendations are given: the advertisement may not have been completely reflowed, a cost may be expected to be pulled back after reflow, a patience waiting is recommended, and the delivery may be maintained.
In one embodiment, the step of obtaining a diagnostic result based on the indicator data further comprises: if the second cost deviation meets the over-cost condition, obtaining a third diagnosis result when the estimated click rate deviation is larger than the first deviation threshold, the ratio of the estimated click rate deviation to the second cost deviation is larger than the first ratio threshold, and the estimated click rate deviation is larger than the estimated conversion rate deviation; the third diagnosis result is used for indicating that the excessive cost reason of the advertisement to be diagnosed in the target time interval is that the estimated deviation of the click rate is higher, and the estimated model can be corrected; the diagnostic result includes a third diagnostic result.
And judging whether the estimated click rate deviation (pctrbias) is high or not by combining the estimated backflow and considering that the backflow is complete. When pctrbias > e1, pctrbias/cpabias > g1 and pctrbias > pcvrbias are satisfied, it is determined that pctrbias is higher and pctrbias is the largest influence factor. In this case the following diagnostic recommendations are given: the advertisement pctrbias is high and it is recommended to correct the prediction model. Where e1 and g1 denote a first deviation threshold and a first ratio threshold, respectively. e1 and g1 can be set according to actual requirements, and are not limited herein. In one embodiment, e1 is set to 25% and g1 is set to 40%.
In one embodiment, further comprising: and acquiring the estimated deviation of the last click rate, the second cost deviation and the actual consumption of the advertisement to be diagnosed in the last time interval of the target time interval. In order to analyze the fluctuation condition of pctrbias, pctrbias in a target time interval and a previous time interval is compared and analyzed, and the deviation is estimated by the previous click rate (using pctrbias) t-1 Expressed), last second cost bias (in cpabias) t-1 Expressed) and last actual consumption (in cost) t-1 Expressed) respectively refer to the estimated deviation of click rate, the second cost deviation and the actual consumption of the previous period. For example, assuming that the target period is 12 o ' clock of the day and the previous period may be 11 o ' clock of the day, pctrbias, cpabias and cost respectively represent the estimated deviation of the click rate, the second cost deviation and the actual consumption accumulated to 12 o ' clock of the day, pctrbias t-1 、cpabias t-1 And cost t-1 Respectively representing estimated deviation of click rate accumulated to 11 points in the day, second cost deviation and actual consumption.
In one embodiment, the step of obtaining a diagnostic result based on the indicator data further comprises: if the second cost deviation meets the over-cost condition, obtaining a fourth diagnosis result when judging that the click rate estimation deviation fluctuation condition is met according to the click rate estimation deviation, the last click rate estimation deviation, the second cost deviation, the last second cost deviation, the actual consumption, the last actual consumption, the actual conversion number and the target conversion cost; the fourth diagnosis result is used for indicating that the super-cost reason of the advertisement to be diagnosed in the target time interval is that the click rate estimation deviation fluctuates, whether the flow has dimension migration or not can be observed firstly, and if not, the estimation model is checked; the diagnostic result includes a fourth diagnostic result.
Whether pctrbias fluctuates can be further judged, even if pctrbias does not exceed the standard, due to the existence of the price adjustment factor, the fluctuation (such as-50% to 0%) of pctrbias still causes excessive cost, and the influence caused by the fluctuation needs to be captured. Specifically, the cpabias change is in the same direction as the pctrbias change, and the influence factor of pctrbias on cpabias (denoted as cost _ effect _ factor) is first determined, and cost _ effect _ factor = alpha (pctrbias-pctrbias) t-1 )/(cpabias-cpabias t-1 ) Where alpha is a constant, if the calculated cost _ effect _ factor is greater than 1, let cost _ effect _ factor be 1. Then calculate the amount of excess (denoted by cost _ effect) brought by pctrbias fluctuation, cost _ effect = (cost-cost) t-1 -gmv) cost effect factor. And calculating the cost deviation (represented by cpa _ bias _ pc) after deducting the excess amount, wherein cpa _ bias _ pc = (cost _ sum-cost _ effect _ sum)/N/targetcpa-1. And when the cpa _ bias _ pc is smaller than a preset value, judging that the click rate estimated deviation fluctuation condition is met. In this case the following diagnostic recommendations are given: the advertisement pctrbias generates fluctuation, and the proposal is to firstly observe whether the flow has dimension migration or not, and if not, the estimation model is checked. The preset value may be set in combination with actual requirements, for example, to 0.2.
In the above embodiment, the same direction influence on cpabias is generated by utilizing the fluctuation of pctrbias, the influence factor cost _ effect _ factor of pctrbias on cpabias is firstly quantized, then the excess amount cost _ effect caused by the fluctuation of pctrbias is calculated according to the influence factor, and finally whether the cost is pulled back after the amount is deducted is judged, if so, the excess cost is caused by the fluctuation of pctrbias. In practice, the hours with the greatest influence can be located according to the influence degree of each hour, and the hour with the greatest influence is used as a mutation time point for later positioning.
In one embodiment, further comprises: and acquiring the estimated deviation of the last conversion rate of the advertisement to be diagnosed in the last time interval of the target time interval. In order to analyze the fluctuation condition of the pcvrbias, the pcvrbias of the target time interval and the previous time interval are compared and analyzed, and the deviation is estimated by the previous conversion rate (using the pcvrbias) t-1 Expressed) refers to the predicted deviation in conversion over the previous period.
In one embodiment, the step of obtaining a diagnostic result based on the indicator data further comprises: if the second cost deviation satisfies the over-cost condition, if at least one of the following two items is satisfied: the first item is that the estimated conversion rate deviation is greater than a second deviation threshold, the ratio of the estimated conversion rate deviation to the second cost deviation is greater than a second ratio threshold, and the estimated conversion rate deviation is greater than the estimated click rate deviation; the second item judges that the estimated conversion rate deviation fluctuation condition is met according to the estimated conversion rate deviation, the estimated last conversion rate deviation, the second cost deviation, the actual consumption, the actual conversion number and the target conversion cost; obtaining a fifth diagnostic result when the number of starts is less than the second number threshold; the fifth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the conversion is too sparse, the capacity can be improved through competitive analysis, or the originality can be optimized; the diagnostic result includes a fifth diagnostic result.
When pcvrbias > e2, pcvrbias/cpabias > g2 and pcvrbias > pctrbias are satisfied, it is determined that pcvrbias is higher and pcvrbias is the largest influence factor. Where e2 and g2 represent the second deviation threshold and the second ratio threshold, respectively. e2 and g2 can be set according to actual requirements, and are not limited herein. In one embodiment, e2 is set to 25% and g2 is set to 40%. In this case, the next decision is continued to analyze the more refined over-cost cause that leads to higher pcvrbias. Specifically, the next step is to determine whether bias is visible, and when the number of starts (n) is less than a second number threshold (denoted by n 2), it is determined that bias is not visible, giving the following diagnostic advice: the conversion is too sparse, and the yield is improved by analyzing and improving the yield capacity through the competitive power or optimizing the originality. Wherein the second number threshold (n 2) is greater than the first number threshold (n 1), and in one embodiment, n1 is set to 3 and n2 is set to 6.
Whether the pcvrbias fluctuates can be further judged, even if the pcvrbias does not exceed the standard, due to the existence of the price adjusting factor, the fluctuation (such as from-50% to 0%) of the pcvrbias still causes excessive cost, and the influence caused by the fluctuation needs to be captured. Specifically, the cpabias change and the pcvrbias change are in the same direction, and the influence factor of the pcvrbias on the cpabias (represented by cost _ effect _ factor) is determined first, and the cost _ effect _ factor = alpha (pcvrbias-pcvrbias) t-1 )/(cpabias-cpabias t-1 ) Where alpha is a constant, if the calculated cost _ effect _ factor is greater than 1, let cost _ effect _ factor be 1. Then calculate the amount of excess (denoted by cost _ effect) brought by the pcvrbias fluctuation, cost _ effect = (cost-cost) t-1 -gmv) cost effect factor. And calculating the cost deviation (represented by cpa _ bias _ pc) after deducting the excess amount, wherein cpa _ bias _ pc = (cost _ sum-cost _ effect _ sum)/N/targetcpa-1. And when the cpa _ bias _ pc is smaller than a preset value, judging that the conversion rate estimated deviation fluctuation condition is met. In this case, the next decision is continued to analyze the more refined over-cost cause that leads to higher pcvrbias. Specifically, the next step is to determine whether bias is visible, and when the number of starts (n) is less than a second number threshold (denoted by n 2), it is determined that bias is not visible, giving the following diagnostic advice: and the transformation is too sparse, and the transformation rate is improved by improving the capacity through competitive analysis or optimizing the originality. Wherein the second number threshold (n 2) is greater than the first number threshold (n 1), and in one embodiment, n1 is set to 3 and n2 is set to 6. The preset value may be set in combination with the actual demand, for example to 0.2.
In an embodiment, after it is determined that the conversion rate estimated deviation fluctuation condition is satisfied, the influence factors corresponding to each hour may be further obtained according to the conversion rate estimated deviation, the last conversion rate estimated deviation, the second cost deviation, and the last second cost deviation, and an hour corresponding to a largest influence factor among the influence factors corresponding to each hour is taken as a mutation time point. Wherein, the influence factor is used for measuring the influence degree of the estimated deviation fluctuation of the conversion rate on the cost deviation fluctuation.
For example, assume the target period is on the order of hours, e.g.At 12 th day, the current value is changed by the formula cost _ effect _ factor = alpha (pctrbias-pctrbias) t-1 )/(cpabias-cpabias t-1 ) The influence factors of each hour from 0 point to 12 points on the day can be calculated, and the 10 points are taken as mutation time points if the influence factors corresponding to the 10 points are maximum. In practice, the hours with the greatest impact can be located according to the degree of impact of each hour, and the hour with the greatest impact can be used as a mutation time point for later location.
In one embodiment, the metric data further includes an actual conversion rate (denoted by cvr) obtained for an advertisement to be diagnosed at a previous slot of the target slot (denoted by cvr) t-1 Representation). The step of obtaining a diagnostic result from the indicator data further comprises: if the number of starts is greater than or equal to the second number threshold, obtaining a sixth diagnostic result when the conversion rate abnormal fluctuation condition is met; the sixth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is abnormal fluctuation of the conversion rate, and can be used for checking holidays, natural attenuation, abnormal backflow and follow-up marketing follow-up conditions; the diagnostic result includes a sixth diagnostic result.
Determining that a conversion rate abnormal fluctuation condition is satisfied when at least one of the following two items is satisfied: a first term, wherein the rate of decrease of the actual conversion relative to the last actual conversion is less than a rate of decrease threshold; in the second term, the ratio of the difference between the actual conversion rate and the last actual conversion rate to the difference between the estimated deviation of the conversion rate and the last estimated deviation of the conversion rate is greater than a predetermined value.
In one embodiment, the threshold value of the descent rate is set to-20%, the preset value is set to 0.4, and when cvr/cvr is used, the threshold value and the preset value can be set according to actual requirements t-1 When-1 < -20%, judging that the abnormal fluctuation condition of the conversion rate is satisfied, and when the time point of the mutation is (cvr-cvr) t-1 )/(pcvrbias-pcvrbias t-1 )>At 0.4, it is also judged that the conversion rate abnormal fluctuation condition is satisfied. In this case the following diagnostic recommendations are given: advertisement conversion rate is reduced obviously today compared with yesterday, and analysis of reasons can be tried, and possible reasons include: holidays, natural decay, abnormal customer return or follow-up marketing follow-up, etc.
In one embodiment, the step of obtaining a diagnostic result based on the indicator data further comprises: if the conversion rate abnormal fluctuation condition is not met, obtaining a seventh diagnosis result when modified data aiming at the advertisement putting information is monitored; the seventh diagnosis result is used for indicating that the over-cost reason of the advertisement to be diagnosed in the target time interval is that the conversion rate estimation deviation is unstable due to the operation of the client, and the reduction of the operation can be recommended; the diagnostic result includes a seventh diagnostic result.
If the conversion rate abnormal fluctuation condition is not met, whether the client has modification operation aiming at the advertisement delivery information (such as orientation, bid price, material, optimization target, expansion state, landing page and the like) can be further judged, and specifically, whether the advertisement delivery information is modified in the first 6 hours of the mutation time point can be monitored. And when the modification data is monitored, judging that the client performs modification operation. In this case the following diagnostic recommendations are given: it was found that the client modified the data, resulting in pcvrbias instability, suggesting that operations be minimized.
In one embodiment, the step of obtaining a diagnostic result based on the indicator data further comprises: if the modification data aiming at the advertisement putting information is not monitored, obtaining an eighth diagnosis result when the estimated deviation of the overall conversion rate is greater than or equal to a third deviation threshold value; the eighth diagnosis result is used for indicating that the cost-exceeding reason of the advertisement to be diagnosed in the target time interval is that the conversion rate prediction deviation generates flow grade deviation in the whole dimension, and the prediction model can be checked; the diagnostic result includes an eighth diagnostic result. .
If the modification data aiming at the advertisement putting information is not monitored, whether the commodity type conversion target (ptbo) is abnormal or not can be further judged. The estimated deviation of the overall conversion may be understood as the estimated deviation of the overall conversion of the ptbo particle size, the third deviation threshold may be set to 20%, and when the estimated deviation of the overall conversion is greater than or equal to 20%, it is determined that the ptbo is abnormal. In this case the following diagnostic recommendations are given: and (4) carrying out flow magnitude deviation on the pcvrbias in the whole dimension, and suggesting to carry out investigation on the pre-estimated model.
In one embodiment, the step of obtaining a diagnostic result based on the indicator data further comprises: if the estimated deviation of the overall conversion rate is smaller than a third deviation threshold value, then: obtaining a ninth diagnostic result when the number of starts is greater than or equal to a third number threshold; the ninth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the estimated deviation of the conversion rate is higher, and the estimated model can be corrected; obtaining a tenth diagnostic result when the number of starts is less than the third number threshold; the tenth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the conversion is too sparse, the capacity can be improved through competitive analysis, or the creative idea can be optimized; the diagnostic result includes a ninth diagnostic result or a tenth diagnostic result.
If the estimated deviation of the overall conversion rate is smaller than the third deviation threshold value, whether the bias is credible or not can be further judged. When the number of starts (n) is greater than or equal to a third number threshold (denoted by n 3), bias confidence is determined, giving the following diagnostic recommendation: pcvrbias is high or fluctuates, and the correction of the prediction model is recommended. When the number of starts (n) is less than a third number threshold (n 3), bias is judged to be unreliable, giving the following diagnostic recommendation: the conversion is too sparse, and the yield is improved by analyzing and improving the yield capacity through the competitive power or optimizing the originality. Wherein the third number threshold (n 3) is greater than the second number threshold (n 2), and in one embodiment, n2 is set to 6 and n3 is set to 20.
In one embodiment, the metric data further comprises: bid data and estimated conversion; the step of obtaining a diagnostic result from the indicator data further comprises: if the condition of excessive cost is met by combining the estimated backflow judgment, obtaining an eleventh diagnosis result when the condition of excessive high price of the system is met by judging based on the bid data, the estimated conversion rate and the target conversion cost; the eleventh diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the system bid is too high, and abnormal price-adjusting factors can be checked; the diagnostic result includes an eleventh diagnostic result.
The system bid data may be a single click cost (denoted cpc) or a thousand exposure cost (denoted cpm), and the system bid is determined to be too high when cpc > targetcpa pcvr a or cpm > targetcpa pcvr pctr a, where a is a constant and may be set to 1.3. In this case the following diagnostic recommendations are given: and if the system bids excessively, the abnormal price adjustment factor is suggested to be checked. Specifically, the relationship between each factor of thousands of hourly display revenue (ecpm) and the cost deviation can be calculated by the following formula:
Figure BDA0003006447810000131
where X represents the set of some ecpm factor (some coefficient affecting the system bid) over the time series, X i Representing the value of the factor at time point i, Y representing the set of cumulative cost deviations of the day, Y i Representing the cost deviation from 0 to time point i. The factor has the effect of reversely regulating the cost, and if the cost is increased, the factor is reduced, and if the cost is reduced, the factor is increased. Therefore, if the factor is negatively correlated with the cost deviation, it indicates that it is effectively adjusting the cost, and if the factor is positively correlated with the cost deviation, it indicates that it is not normally adjusting, and it is regarded as an abnormal price-adjusting factor.
In the embodiment, the operation suggestions can be adjusted in a quasi-real-time manner according to the current actual performance through hour-level super-cost analysis and reason positioning, so that the customer experience is improved; the analysis of the full link is realized for the reason of the super cost, and the analysis not only comprises a pcvr/pctr model, but also comprises a plurality of modules of system bidding, advertisement competition and the like; providing richer and more accurate operation suggestions, comprising not only observation and shutdown, but also: starting amount optimization, model correction, advertisement modification and other operation suggestions; the complete diagnosis of the quantity of tens of millions of advertisements is realized and the advertisements are put in storage, so that the subsequent backtracking analysis is facilitated.
It can be understood that there is a certain correlation between the diagnosis schemes of the above embodiments, specifically, for the advertisement with excessive cost, it is determined whether the starting amount is sufficient, if the starting amount is sufficient, it is determined whether the conversion is not completed, if the conversion is completed, it is determined whether the click rate estimated deviation, the conversion rate estimated deviation and the system bid are abnormal, if so, the user goes down to find the reason of the abnormality, and the reason of the excessive cost is analyzed step by step.
In one embodiment, the t +5 day no-over-cost rate (expressed by rate) of all advertisements below the time may be counted respectively according to industry, optimization goal, daily cumulative consumption, daily cumulative cost bias, and several dimensions of the time. If rate <10%, the operation recommendation is: advertisements are today more difficult to pull back costs, suggesting immediate shutdown and rebuilding, e.g., more than 15. If the rate is more than or equal to 10% and less than 40%, the operation proposal is as follows: the remaining costs of eventual pullback are less likely to be consumed today, suggesting a close focus, shutting down immediately if the cost performance deteriorates and restarting on the second day. If the rate is more than or equal to 40% and less than 70%, the operation proposal is as follows: the possibility of the residual consumption final pull-back cost is high today, and there is a certain probability of exceeding the cost, and the customer continues to put the product but needs to pay certain attention. If 70% ≦ rate, the operating recommendations were: the advertisement probability does not exceed the cost, and can be released safely.
In one embodiment, the ocpx casting policy may not be continuously adjusted by a new policy or an old policy, so that the diagnosis logic also needs to be adjusted correspondingly frequently. As shown in fig. 3, a tree-like diagnostic logic diagram is provided. For the specific judgment logic, reference may be made to the foregoing embodiments, which are not described herein again.
Specifically, the checking logic is organized into a tree and stored in a dictionary, and each node comprises: the type, description and function list represent the specific subsequent functions which are required to be performed when the final ultra-cost reason type id, chinese description and diagnosis logic are performed to the node if the node is the leaf diagnosis logic terminal. And the function list is stored with three nodes of function name, left branch logic and right branch logic, wherein the function name is used for calling specific functions according to the name, and each function is finally obtained according to internal calculation: and (4) returning values of 'yes' and 'no', if yes is returned, continuing to check according to left branch logic, otherwise, continuing according to right branch logic until the logic tree is ended.
The specific function of each node is stored in a function library, and each function transmits and returns all data of the advertisement and returns 'yes' or 'no' according to specific logic. When the advertisement needs to be diagnosed, the system recursively calls the corresponding function from the root node according to the logic tree, and continues subsequent diagnosis according to the return value.
Therefore, when the diagnosis logic needs to be adjusted, the diagnosis logic tree is only needed to be modified, the position of the newly added logic in the logic tree is selected, the function name is set, and the function is added into the function library, so that the whole diagnosis logic has flexible and extensible attributes.
In one embodiment, the advertisement diagnosis method is applied to an internal diagnosis system for operation developers, and focuses on the positioning of reasons and provides basic dialogs for feedback clients. Specifically, an operation developer can input a date and an advertisement on an advertisement diagnosis page of the internal diagnosis system and trigger a diagnosis operation, if the advertisement is over-cost, the 'abnormity' is displayed beside the over-cost diagnosis selection column, and at the moment, the diagnosis result can be obtained by clicking the over-cost diagnosis.
In one embodiment, the advertisement diagnosis method is applied to an advertisement operation tool, and mainly analyzes the account health status from the view point of a customer account, checks the distribution of account cost excess reasons and gives diagnosis suggestions at the account level. In particular, the operations developer may enter one or more account ids and select a date in the operations tool page, and then select cost stability to view cost performance of those accounts over that time. Then, under the cost standard-reaching performance, the user can select to press an account number or an advertisement, and input corresponding id or other screening conditions to inquire corresponding details. In cost compliance analysis, specific categories of over-cost diagnostic results may be reviewed, and the diagnostic results may be counted by consumption or by number of advertisements. After clicking on a single category, all advertisement details of the corresponding category may also be retrieved and entered into the advertisement diagnosis page.
In one embodiment, the advertisement diagnosis method is applied to an external diagnosis system of a delivery end, is directly oriented to a client, and provides diagnosis analysis and operation suggestions of some non-system reasons.
It should be understood that, although the steps in the flowcharts related to the above embodiments are shown in sequence as indicated by the arrows, the steps are not necessarily executed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in each flowchart related to the above embodiment may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least a part of the steps or stages in other steps.
In one embodiment, as shown in fig. 4, there is provided an advertisement diagnosis apparatus 400, which may be a part of a computer device using a software module or a hardware module, or a combination of the two, the apparatus specifically includes: a data processing module 410, a data acquisition module 420, a cost bias determination module 430, and a diagnostic module 440, wherein:
and the data processing module 410 is configured to obtain hour-level data and day-level data of an index required for diagnosis according to advertisement-related data of an advertisement to be diagnosed.
And the data acquisition module 420 is configured to acquire index data of the advertisement to be diagnosed in the target time period from the hour-level data and the day-level data, where the index data includes a target conversion cost, an actual conversion number, actual consumption, a click rate prediction deviation, and a conversion rate prediction deviation.
And a cost deviation determining module 430, configured to obtain a target consumption according to a product of the target conversion cost and the actual conversion number, and determine a first cost deviation according to a deviation between the actual consumption and the target consumption.
And the diagnosis module 440 is configured to obtain a diagnosis result according to the index data when the first cost deviation satisfies the super-cost condition, where the diagnosis result includes a super-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion.
In one embodiment, the diagnostic module 440 is specifically configured to: obtaining the starting number according to the ratio of the actual consumption to the target conversion cost; obtaining a first diagnostic result when the number of starts is less than a first number threshold; the first diagnosis result is used for indicating that the cost-exceeding reason of the advertisement to be diagnosed in the target time interval is insufficient, and the capacity can be improved through competitive analysis.
In one embodiment, the diagnostic module 440 is further configured to: when the number of starts is larger than or equal to the first number threshold, obtaining estimated backflow in a future preset time period, subtracting the estimated backflow from actual consumption to obtain estimated actual consumption, determining a second cost deviation according to the deviation between the estimated actual consumption and target consumption, and obtaining a second diagnosis result when the second cost deviation does not meet an over-cost condition; the second diagnosis result is used for indicating that the cost-exceeding reason of the advertisement to be diagnosed in the target time interval is that the advertisement is not completely reflowed and can be kept to be delivered.
In one embodiment, the diagnostic module 440 is further configured to: if the second cost deviation meets the over-cost condition, obtaining a third diagnosis result when the estimated click rate deviation is larger than the first deviation threshold, the ratio of the estimated click rate deviation to the second cost deviation is larger than the first ratio threshold, and the estimated click rate deviation is larger than the estimated conversion rate deviation; the third diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the estimated deviation of the click rate is higher, and the estimated model can be corrected.
In one embodiment, the data obtaining module 420 is further configured to obtain a last click rate estimated deviation, a last second cost deviation and a last actual consumption of the advertisement to be diagnosed in a last period of the target period; the diagnostic module 440 is further configured to: if the second cost deviation meets the over-cost condition, obtaining a fourth diagnosis result when judging that the click rate estimated deviation fluctuation condition is met according to the click rate estimated deviation, the previous click rate estimated deviation, the second cost deviation, the previous second cost deviation, the actual consumption, the previous actual consumption, the actual conversion number and the target conversion cost; and the fourth diagnosis result is used for indicating that the over-cost reason of the advertisement to be diagnosed in the target time interval is that the click rate estimation deviation fluctuates, whether the flow has dimension migration or not can be observed firstly, and if not, the estimation model is checked.
In one embodiment, the data obtaining module 420 is further configured to obtain an estimated deviation of a last conversion rate of the advertisement to be diagnosed in a last time period of the target time period; the diagnostic module 440 is further configured to: if the second cost deviation satisfies the over-cost condition, at least one of the following two items is satisfied: the first item is that the estimated conversion rate deviation is greater than a second deviation threshold, the ratio of the estimated conversion rate deviation to the second cost deviation is greater than a second ratio threshold, and the estimated conversion rate deviation is greater than the estimated click rate deviation; the second item judges that the conversion rate estimated deviation fluctuation condition is met according to the conversion rate estimated deviation, the last conversion rate estimated deviation, the second cost deviation, the last second cost deviation, the actual consumption, the last actual consumption, the actual conversion number and the target conversion cost; obtaining a fifth diagnostic result when the number of starts is less than the second number threshold; the fifth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the conversion is too sparse, the capacity can be improved through competitive analysis, or the originality can be optimized.
In one embodiment, the diagnostic module 440 is further configured to: after judging that the conversion rate estimated deviation fluctuation condition is met, obtaining an influence factor corresponding to each hour according to the conversion rate estimated deviation, the previous conversion rate estimated deviation, the second cost deviation and the previous second cost deviation, wherein the influence factor is used for measuring the influence degree of the conversion rate estimated deviation fluctuation on the cost deviation fluctuation; the hour corresponding to the largest influencing factor among the influencing factors corresponding to the respective hours was set as the mutation time point.
In one embodiment, the indicator data further comprises an actual conversion; the data obtaining module 420 is further configured to obtain a last actual conversion rate of the advertisement to be diagnosed in a last time period of the target time period; the diagnostic module 440 is further configured to: if the number of starts is greater than or equal to the second number threshold, obtaining a sixth diagnostic result when the conversion rate abnormal fluctuation condition is met; the sixth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is abnormal fluctuation of the conversion rate, and can be used for checking holidays, natural attenuation, abnormal backflow and follow-up marketing follow-up conditions; determining that a conversion rate abnormal fluctuation condition is satisfied when at least one of the following two items is satisfied: a first term, wherein the rate of decrease of the actual conversion relative to the last actual conversion is less than a rate of decrease threshold; in the second term, the ratio of the difference between the actual conversion rate and the last actual conversion rate to the difference between the estimated deviation of the conversion rate and the last estimated deviation of the conversion rate is greater than a predetermined value.
In one embodiment, the diagnostic module 440 is further configured to: if the conversion rate abnormal fluctuation condition is not met, obtaining a seventh diagnosis result when modified data aiming at the advertisement putting information is monitored; the seventh diagnosis result is used for indicating that the over-cost reason of the advertisement to be diagnosed in the target time interval is that the conversion rate estimation deviation is unstable due to the operation of the client, and reduction operation can be recommended; if the modification data aiming at the advertisement putting information is not monitored, obtaining an eighth diagnosis result when the estimated deviation of the overall conversion rate is greater than or equal to a third deviation threshold value; and the eighth diagnosis result is used for indicating the over-cost reason of the advertisement to be diagnosed in the target time interval as the traffic level deviation of the conversion rate prediction deviation in the whole dimension, and the prediction model can be checked.
In one embodiment, the diagnostic module 440 is further configured to: if the estimated deviation of the overall conversion rate is smaller than a third deviation threshold value, then: obtaining a ninth diagnostic result when the number of starts is greater than or equal to a third number threshold; the ninth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the estimated deviation of the conversion rate is higher, and the estimated model can be corrected; obtaining a tenth diagnostic result when the number of starts is less than the third number threshold; the tenth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the conversion is too sparse, the capacity can be improved through competitive analysis, or the creative idea can be optimized.
In one embodiment, the indicator data further comprises bid data and estimated conversion; the diagnostic module 440 is further configured to: if the second cost deviation meets the over-cost condition, obtaining an eleventh diagnosis result when the over-cost condition of the system is judged to be met based on the bid data, the estimated conversion rate and the target conversion cost; and the eleventh diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the system bid too high, and abnormal price-adjusting factors can be checked.
For the specific definition of the advertisement diagnosis device, reference may be made to the definition of the advertisement diagnosis method above, and details are not repeated here. The modules in the advertisement diagnosis device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 5. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement an advertisement diagnosis method.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 6. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement an advertisement diagnosis method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on a shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the configurations shown in fig. 5 or fig. 6 are only block diagrams of some configurations relevant to the present application, and do not constitute a limitation on the computer apparatus to which the present application is applied, and a particular computer apparatus may include more or less components than those shown in the drawings, or may combine some components, or have a different arrangement of components.
In one embodiment, a computer device is further provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method embodiments when executing the computer program.
In an embodiment, a computer-readable storage medium is provided, in which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.
In one embodiment, a computer program product or computer program is provided that includes computer instructions stored in a computer-readable storage medium. The computer instructions are read by a processor of a computer device from a computer-readable storage medium, and the computer instructions are executed by the processor to cause the computer device to perform the steps in the above-mentioned method embodiments.
It should be understood that the terms "first", "second", etc. in the above-described embodiments are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implying any number of technical features indicated. For the description of numerical ranges, the term "plurality" is understood to be equal to or greater than two.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above may be implemented by hardware instructions of a computer program, which may be stored in a non-volatile computer-readable storage medium, and when executed, may include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
All possible combinations of the technical features in the above embodiments may not be described for the sake of brevity, but should be considered as being within the scope of the present disclosure as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent application shall be subject to the appended claims.

Claims (10)

1. An advertisement diagnosis method, characterized in that the method comprises:
acquiring hour level data and day level data of indexes required by diagnosis according to advertisement related data of the advertisement to be diagnosed;
acquiring index data of the advertisement to be diagnosed in a target time interval from the hour level data and the day level data, wherein the index data comprises target conversion cost, actual conversion number, actual consumption, click rate prediction deviation and conversion rate prediction deviation;
obtaining target consumption according to the product of the target conversion cost and the actual conversion number, and determining a first cost deviation according to the deviation between the actual consumption and the target consumption;
and when the first cost deviation meets an over-cost condition, obtaining a diagnosis result according to the index data, wherein the diagnosis result comprises an over-cost reason of the advertisement to be diagnosed in the target time period and a corresponding operation suggestion.
2. The method of claim 1, wherein obtaining a diagnostic result from the metric data comprises:
obtaining the starting number according to the ratio of the actual consumption to the target conversion cost;
obtaining a first diagnostic result when the number of starts is less than a first number threshold; the first diagnosis result is used for indicating that the cause of the super cost of the advertisement to be diagnosed in the target time interval is insufficient, and the capacity can be improved through competitive analysis;
when the number of starts is larger than or equal to the first number threshold, obtaining estimated backflow of a future preset time period, subtracting the estimated backflow from the actual consumption to obtain estimated actual consumption, determining a second cost deviation according to the deviation between the estimated actual consumption and the target consumption, and obtaining a second diagnosis result when the second cost deviation does not meet the over-cost condition; the second diagnosis result is used for indicating that the reason of the over-cost of the advertisement to be diagnosed in the target time interval is that the advertisement to be diagnosed is not completely reflowed and can be kept to be delivered; the diagnostic result includes the first diagnostic result or the second diagnostic result.
3. The method of claim 2, wherein obtaining a diagnostic result based on the metric data further comprises:
if the second cost deviation meets the super-cost condition, obtaining a third diagnosis result when the estimated click rate deviation is larger than a first deviation threshold, the ratio of the estimated click rate deviation to the second cost deviation is larger than a first ratio threshold, and the estimated click rate deviation is larger than the estimated conversion rate deviation;
the third diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the estimated deviation of the click rate is high, and the estimated model can be corrected; the diagnostic result includes the third diagnostic result.
4. The method of claim 2, further comprising: acquiring the estimated deviation of the last click rate, the second cost deviation and the actual consumption of the advertisement to be diagnosed in the last time period of the target time period; obtaining a diagnostic result from the indicator data, further comprising:
if the second cost deviation meets the super-cost condition, obtaining a fourth diagnosis result when judging that the click rate estimated deviation fluctuation condition is met according to the click rate estimated deviation, the last click rate estimated deviation, the second cost deviation, the last second cost deviation, the actual consumption, the last actual consumption, the actual conversion number and the target conversion cost;
the fourth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the click rate estimation deviation fluctuates, whether the flow has dimension migration or not can be observed firstly, and if not, the estimation model is checked; the diagnostic result comprises the fourth diagnostic result.
5. The method of claim 2, further comprising: acquiring the predicted deviation of the last conversion rate of the advertisement to be diagnosed in the last time interval of the target time interval; obtaining a diagnostic result from the indicator data, further comprising:
if the second cost deviation satisfies the over-cost condition, then at least one of:
the conversion rate estimated deviation is greater than a second deviation threshold, the ratio of the conversion estimated deviation to the second cost deviation is greater than a second ratio threshold, and the conversion rate estimated deviation is greater than the click rate estimated deviation;
a second term for determining that a conversion rate estimated deviation fluctuation condition is satisfied according to the conversion rate estimated deviation, the previous conversion rate estimated deviation, the second cost deviation, the previous second cost deviation, the actual consumption, the previous actual consumption, the actual conversion number and the target conversion cost;
obtaining a fifth diagnostic result when the number of starts is less than a second number threshold; the fifth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the conversion is too sparse, the capacity can be improved through competitive analysis, or the creative idea is optimized; the diagnostic result includes the fifth diagnostic result.
6. The method of claim 5, further comprising: after judging that the conversion rate estimated deviation fluctuation condition is met, the method further comprises the following steps:
obtaining an influence factor corresponding to each hour according to the conversion rate estimated deviation, the last conversion rate estimated deviation, the second cost deviation and the last second cost deviation, wherein the influence factor is used for measuring the influence degree of the conversion rate estimated deviation fluctuation on the cost deviation fluctuation;
the hour corresponding to the largest influencing factor among the influencing factors corresponding to the respective hours was set as the mutation time point.
7. The method of claim 5, wherein the indicator data further comprises an actual conversion rate, the method further comprising: acquiring the last actual conversion rate of the advertisement to be diagnosed in the last time interval of the target time interval; obtaining a diagnostic result from the indicator data, further comprising:
if the number of starts is greater than or equal to the second number threshold, obtaining a sixth diagnostic result when a conversion rate anomalous fluctuation condition is satisfied; the sixth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is abnormal fluctuation of conversion rate, and can be used for checking holidays, natural attenuation, backflow abnormity and subsequent marketing follow-up conditions; the diagnostic result comprises the sixth diagnostic result;
determining that the conversion rate abnormal fluctuation condition is satisfied when at least one of the following two items is satisfied:
a first term, a rate of decrease of the actual conversion relative to the last actual conversion being less than a rate of decrease threshold;
and in the second term, the ratio of the difference between the actual conversion rate and the last actual conversion rate to the difference between the estimated conversion rate deviation and the estimated last conversion rate deviation is larger than a preset value.
8. The method of claim 7, wherein obtaining a diagnostic result based on the metric data further comprises:
if the conversion rate abnormal fluctuation condition is not met, obtaining a seventh diagnosis result when modified data aiming at advertisement putting information are monitored; the seventh diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the estimated deviation of the conversion rate is unstable due to the operation of a client, and reduction operation can be recommended;
if the modification data aiming at the advertisement putting information is not monitored, obtaining an eighth diagnosis result when the estimated deviation of the overall conversion rate is greater than or equal to a third deviation threshold value; the eighth diagnosis result is used for indicating that the over-cost reason of the advertisement to be diagnosed in the target time interval is that the conversion rate prediction deviation generates flow grade deviation in the whole dimension, and a prediction model can be checked; the diagnostic result includes the seventh diagnostic result or the eighth diagnostic result.
9. The method of claim 8, further comprising: obtaining a diagnostic result from the indicator data, further comprising: if the estimated deviation of the overall conversion rate is smaller than the third deviation threshold value, then:
obtaining a ninth diagnostic result when the number of starts is greater than or equal to a third number threshold; the ninth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the estimated deviation of the conversion rate is higher, and the estimated model can be corrected;
obtaining a tenth diagnostic result when the number of starts is less than the third number threshold; the tenth diagnosis result is used for indicating that the reason of the super cost of the advertisement to be diagnosed in the target time interval is that the conversion is too sparse, the capacity can be improved through competitive analysis, or the creative idea is optimized; the diagnostic result includes the ninth diagnostic result or the tenth diagnostic result.
10. The method of claim 2, wherein the metric data further comprises: bid data and estimated conversion; obtaining a diagnostic result from the indicator data, further comprising:
if the excess cost condition is satisfied in combination with the second cost deviation, obtaining an eleventh diagnostic result when it is determined that a system over-bid condition is satisfied based on the bid data, the estimated conversion rate, and the target conversion cost; the eleventh diagnosis result is used for indicating that the over-cost reason of the advertisement to be diagnosed in the target time interval is over-high bid, and abnormal price adjustment factors can be checked; the diagnostic result includes the eleventh diagnostic result.
CN202110363406.6A 2021-04-02 2021-04-02 Advertisement diagnosis method, device, computer equipment and storage medium Pending CN115170167A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110363406.6A CN115170167A (en) 2021-04-02 2021-04-02 Advertisement diagnosis method, device, computer equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110363406.6A CN115170167A (en) 2021-04-02 2021-04-02 Advertisement diagnosis method, device, computer equipment and storage medium

Publications (1)

Publication Number Publication Date
CN115170167A true CN115170167A (en) 2022-10-11

Family

ID=83475630

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110363406.6A Pending CN115170167A (en) 2021-04-02 2021-04-02 Advertisement diagnosis method, device, computer equipment and storage medium

Country Status (1)

Country Link
CN (1) CN115170167A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115982455A (en) * 2022-12-20 2023-04-18 贝壳找房(北京)科技有限公司 Flow adjusting method and device based on fuzzy breakpoint regression model and electronic equipment

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115982455A (en) * 2022-12-20 2023-04-18 贝壳找房(北京)科技有限公司 Flow adjusting method and device based on fuzzy breakpoint regression model and electronic equipment
CN115982455B (en) * 2022-12-20 2023-09-15 贝壳找房(北京)科技有限公司 Flow adjustment method and device based on fuzzy breakpoint regression model and electronic equipment

Similar Documents

Publication Publication Date Title
JP5801425B2 (en) Ad progressive pricing method
US10546305B2 (en) Pay-per-action system for selling advertisements
US10453118B2 (en) Custom user definable keyword bidding system and method
US8473339B1 (en) Automatically switching between pricing models for services
US8676646B2 (en) Response attribution valuation
US8676647B2 (en) Response attribution valuation
US20110231239A1 (en) Method and system for attributing an online conversion to multiple influencers
US20080065479A1 (en) System and method for optimizing online advertisement auctions by applying linear programming using special ordered sets
US20120130798A1 (en) Model sequencing for managing advertising pricing
CN103942236A (en) System and method for serving electronic content
WO2008024997A2 (en) System and method for modeling value of an on-line advertisement campaign
US9875484B1 (en) Evaluating attribution models
US20190228422A1 (en) Method and system for determining and monitoring brand performance
US20110055004A1 (en) Method and system for selecting and optimizing bid recommendation algorithms
US20130197996A1 (en) System and method for allocating unspent advertising budget and generating keyword bids
US20120130828A1 (en) Source of decision considerations for managing advertising pricing
CN111899041A (en) Information delivery processing method, information delivery device, information delivery equipment and storage medium
US11144968B2 (en) Systems and methods for controlling online advertising campaigns
US10181130B2 (en) Real-time updates to digital marketing forecast models
Hummel et al. Loss functions for predicted click‐through rates in auctions for online advertising
CN115170167A (en) Advertisement diagnosis method, device, computer equipment and storage medium
US20200034859A1 (en) System and method for predicting stock on hand with predefined markdown plans
US20110270671A1 (en) Predicting number of selections of advertisement using hierarchical Bayesian model
WO2015179717A1 (en) Determination of initial value for automated delivery of news items
US20130282433A1 (en) Methods and apparatus to manage marketing forecasting activity

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination