CN112488759A

CN112488759A - Advertisement sorting system and method

Info

Publication number: CN112488759A
Application number: CN202011370462.4A
Authority: CN
Inventors: 刘利洁
Original assignee: Shanghai Kuliang Information Technology Co Ltd
Current assignee: Shanghai Kuliang Information Technology Co Ltd
Priority date: 2020-11-30
Filing date: 2020-11-30
Publication date: 2021-03-12

Abstract

An advertisement sequencing system comprises a directional module, a sequencing module, an adjusting module and a sending module; the targeting module is used for screening out advertisements meeting the conditions according to the information provided by the channel provider and the targeting conditions set by the advertiser; the ordering module is used for ordering the advertisements screened by the orientation module according to a first ordering condition; wherein the first ordering condition is effectiveness data of advertisement placement; the adjusting module is used for adjusting the sorting of the sorting module according to a second sorting condition; wherein, the second sorting condition may be one condition or a plurality of conditions; and the sending module is used for sending the advertisements to the channel trader according to the sequence of the adjusting module. The invention sequences the advertisements through different conditions, so that the high-quality advertisements obtain more display opportunities.

Description

Advertisement sorting system and method

Technical Field

The invention relates to the technical field of internet advertisements, in particular to an advertisement sequencing system and method.

Background

With the wide application of the internet, the internet advertisement is increasingly favored by advertisers. In an advertisement promotion method in which a settlement is performed with an effect such as CPC, CPS, CPA, ROI, etc., an advertisement promoter enhances collection of data such as number of clicks, download, sales, etc., and reduces collection of display-related data in order to save server resources. For the advertisement service provider, a plurality of advertisement contents need to be sent to the channel provider in a certain sequence, and then the channel provider carries out corresponding bidding on the advertisement, so that the advertisement is finally displayed. In the APX advertisement, the advertisement is not requested by the channel provider in real time, but is requested at intervals, so the advertisement ranking of the advertisement service provider has a very direct influence on the advertisement effect and the final profit.

In the current advertisement sorting technology, more is the sorting aiming at advertisement display. For example, an advertisement intelligent ranking algorithm includes the following steps.

Step 1: the method comprises the steps of order sorting, obtaining date and time interval information of all advertisement orders, sorting through screening software, and inserting all advertisement orders into a serial list according to date and time intervals, wherein in the serial list, when a next order advertisement is inserted, all order advertisements behind the inserted advertisement order in the serial list automatically move backwards by one position, so that a blank position is reserved at the insertion position of the inserted order advertisement, and the blank position is automatically filled with the inserted order advertisement.

Step 2: processing the advertisement at the designated position, placing the advertisement at the designated position according to the designated position, sorting and ordering all advertisement orders, setting the order as a preliminary playing order, then extracting all the advertisements at the designated position, and determining the playing order of the advertisements at the designated position by corresponding all the advertisements at the designated position to the designated position one by one.

And step 3: the method comprises the steps of processing advertisements at unspecified positions, inserting the advertisements at the unspecified positions one by one through efficient traversal, enabling the similar advertisements to automatically go down and follow, separating the similar advertisements, after determining the playing positions of the advertisements at the specified positions, arranging and playing the advertisements at the unspecified positions through an efficient traversal program, if the advertisements at the two adjacent unspecified positions belong to the same type of advertisements, determining the front one of the advertisements at the two adjacent unspecified positions to be at the playing position of the original sequence, and continuing the rear one of the advertisements to go down until the advertisements at the different types appear in the subsequent advertisements and the advertisements are assigned to the playing positions in sequence, and then assigning the rear one of the advertisements at the two adjacent unspecified positions to the subsequent playing positions, so that the similar advertisements can be effectively separated, and the dissatisfaction of audiences is avoided.

And 4, step 4: the method comprises the steps of determining sequencing parameters, learning the advertisement arrangement rules of the previous five days of the algorithm by a machine, converting the advertisement arrangement rules into sequencing parameters of the series list of the current day, opening the advertisement series list of the previous five days on the machine, automatically identifying the arrangement rules of the advertisement series list of the previous five days by the machine, converting the arrangement rules of the advertisement series list of the previous five days into corresponding sequencing parameters, and taking the sequencing parameters as the sequencing parameters of the series list of the sixth day, so that a large number of advertisement orders of the current day can be rapidly sequenced, and the working efficiency is greatly improved.

And 5: and finally, adjusting the sequence, namely, automatically converting the sequencing parameters of the serial list by the machine, carrying out final sequence adjustment according to the sequencing parameters, substituting the sequencing parameters into the serial list on the current day, carrying out sequence adjustment according to the sequencing parameters by the serial list on the current day, and finally playing the sequence to determine the final sequence.

Step 6: and inserting advertisements automatically according to two factors of the time interval free time and the sequencing parameters, and inserting the advertisements in sequence according to the time intervals and the position sequences of all the advertisements in an advertisement serial list, so that the time linkage can be completed, the complete playing of the advertisements is ensured, and the playing progress of the drama is not influenced.

Disclosure of Invention

The technical problem to be solved by the invention is to provide an advertisement sequencing system and method, which are used for sequencing advertisements under different conditions, so that high-quality advertisements can obtain more display opportunities.

In order to solve the technical problem, the invention provides an advertisement sequencing system, which comprises a directional module, a sequencing module and an adjusting module;

the targeting module is used for screening out advertisements meeting the conditions according to the information provided by the channel provider and the targeting conditions set by the advertiser;

the ordering module is used for ordering the advertisements screened by the orientation module according to a first ordering condition; wherein the first ordering condition is effectiveness data of advertisement placement;

the adjusting module is used for adjusting the sorting of the sorting module according to a second sorting condition; the second sorting condition may be one condition or a plurality of conditions.

And the sending module is used for sending the advertisements to the channel trader according to the sequence of the adjusting module.

In the technical scheme, the channel provider is an internet advertisement service provider which sends an advertisement request to the system, receives an advertisement sequence sent by the system and then carries out advertisement bidding and displaying.

As an improvement of the scheme, the advertisements screened by the targeting module comprise old advertisements and new advertisements; the old advertisements are advertisements with historical delivery data, and the new advertisements are advertisements without historical delivery data.

In the technical scheme, the old advertisements can be sorted according to the first sorting condition by using the historical delivery data to obtain a reasonable ranking, and the new advertisements cannot be ranked at the tail end of the sequence or even cannot obtain a display opportunity because the new advertisements do not have the historical delivery data and cannot be ranked corresponding to the new advertisements. By screening the new advertisements and the old advertisements together, all advertisements meeting the targeting requirements can be screened out, and the new advertisements and the old advertisements are guaranteed to have the same chance.

As an improvement of the above scheme, the first ordering condition is one of a conversion number, a conversion rate, a profit, a price, and a budget.

In the above technical solution, the conversion number refers to the number of successful conversions of the advertisement, which may be a viewing time, or may refer to downloading, installing, and the like. Each target emphasis point is different, but all targets can be the core targets of the advertisement.

As an improvement of the above, the first ordering condition may be the second ordering condition, but the first ordering condition is different from the second ordering condition in ordering the same advertisement.

In the technical scheme, the same group of advertisements are sequenced according to at least two conditions, so that the defect of a single parameter can be avoided, and the method is particularly friendly to new advertisements.

As a modification of the above, the second sorting condition at least comprises a piece sorting condition; the segmentation ordering condition is an index for dividing the advertisements into different echelons according to the condition and ordering the echelons.

In the technical scheme, the arrangement of the segmentation sorting condition can ensure that some single indexes are not high, but the advertisement with excellent quality can obtain more advanced ranking. Meanwhile, the ranks of the advertisements with prominent single index and inferior other indexes are limited. In addition, the technical scheme can also limit the ranking of the new advertisements, thereby realizing the smoothness of the overall ranking change.

As an improvement of the above scheme, the second sorting condition may also be advertisement stability and fund settlement rate.

In the technical scheme, although the advertisement stability has no direct influence on the income, the method has important significance on the advertisement display quality and the user experience, and is beneficial to the long-term development of the Internet advertisement industry. The funding rate is an index used to evaluate advertisers. Incorporating this index into the ranking can improve the advertisement ranking for a premium advertiser.

As an improvement of the above, the second ordering condition applies a different evaluation condition to the new advertisement than the old advertisement to advance the order of the new advertisement.

In the technical scheme, the new advertisements are evaluated according to the characteristics of the new advertisements, and certain ordering preference is given to the new advertisements, so that the new advertisements have display opportunities.

As an improvement of the above scheme, the second sorting condition has an attribute condition and a random condition for the sorting condition of the new advertisement; wherein the attribute condition refers to the targeting condition of the new advertisement or the attribute of the advertiser; the random condition is that the new advertisements are randomly ordered according to an equal probability method.

In the above technical solution, the attribute condition includes fixed information such as advertiser information and targeting condition. The advertisement is thus ranked better when the advertiser's other advertisements are of higher quality. When the advertisement is more targeted and matched with the advertisement request, the advertisement can obtain better ranking. And the random condition is used for randomly ordering the advertisements of the same type.

As an improvement of the scheme, the equal probability method comprises a Thompson algorithm, a random algorithm and a machine learning algorithm.

In the technical scheme, the three algorithms are effective sorting methods which can be used in advertisement sorting, and a proper sorting method can be selected according to different advertisement requirements.

Correspondingly, the invention also provides an advertisement sorting method, which comprises the following steps.

A. And screening out the advertisements meeting the conditions by using the targeting module according to the information provided by the channel provider and the targeting conditions set by the advertiser.

In the step, the server screens out an advertisement set meeting the requirement by receiving advertisement request information from the channel provider, obtains a preliminary advertisement set and defines a range for the subsequent steps.

B. Ordering the advertisements screened by the targeting module by using the ordering module according to a first ordering condition; wherein the first ordering condition is effectiveness data of advertisement placement.

In the step, all the advertisements in the advertisement set screened in the last step are sorted according to the first sorting condition, so that preliminary sorting is realized, and the advertisements with good performance are well ranked.

C. Adjusting the sorting of the sorting module according to a second sorting condition by using the adjusting module; the second sorting condition may be one condition or a plurality of conditions.

In the step, the ranking in the previous step is adjusted through one or more second ranking conditions, and unreasonable advertisement ranking in a single index is adjusted, so that high-quality advertisements can all obtain better ranking.

D. And sending the advertisements to the channel trader according to the sequence of the adjusting module by using the sending module.

The invention has the following beneficial effects.

The invention screens the advertisement request and the targeting condition of the advertiser, so that the sent advertisements all meet the requirement of the advertiser, and a better display effect can be obtained. The invention also enables the high-quality advertisements to obtain more display opportunities by using the advertisement putting data as a main sequencing method, thereby obtaining better display effect. The invention also adjusts the advertisement sequencing through the second sequencing condition, realizes the correction of the single index sequencing, and ensures that the advertisement income, the efficiency and the advertisement screening can obtain better performance.

Drawings

Fig. 1 is a schematic structural diagram of a first embodiment of an advertisement ranking system according to the present invention.

Fig. 2 is a flowchart of a first embodiment of an advertisement ranking method of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings.

In a first embodiment of the present invention, as shown in fig. 1, an advertisement ranking system is provided, which includes a targeting module 100, a ranking module 200, an adjusting module 300, and a sending module 400;

and the targeting module 100 is used for screening out advertisements meeting the conditions according to the information provided by the channel provider and the targeting conditions set by the advertiser.

Specifically, when placing an advertisement, the advertiser sets a series of targeting conditions for the advertisement, including country/region, terminal type, advertisement location, user gender, and the like. When the advertisement position is displayed for different advertisement target customers, the information of the advertisement target customers is obtained to be matched with the orientation conditions of the user. When a channel provider requests an advertisement from the system, the targeting module 100 screens out an advertisement that meets the advertisement targeting condition according to the channel provider information. Since the channel provider does not have user information when requesting advertisements for the same advertisement space, the advertisements screened by the embodiment are all the information related to the advertisement space. Therefore, when the advertisement ranking is sent to the channel provider, the targeting conditions set for the advertisement by the advertiser need to be sent to the channel together, so that when the channel receives the advertisement request of the user, the advertisement ranked at the top can be selected from the advertisement ranking according to the targeting conditions set by the advertiser to bid and display. The advertisements screened by the targeting module 100 include old advertisements and new advertisements. The old advertisement is an advertisement with historical placement data, and the new advertisement is an advertisement without historical placement data. Old advertisements are easier to sort because of historical placement data, thereby enabling a rating of the effectiveness of the advertisements, while new advertisements are less accurate to sort because of no support by relevant data.

The ordering module 200 is used for ordering the advertisements screened by the targeting module according to a first ordering condition; wherein the first ordering condition is effectiveness data of advertisement placement.

Specifically, the ranking module 200 ranks the advertisements screened by the targeting module 100, where there are both new advertisements and old advertisements. The ranking module 200 ranks according to one of a number of conversions, a conversion rate, an income, a price, a budget. For example, the ranking module 200 ranks high conversion advertisements first and low conversion advertisements last, if the ranking module 200 ranks high conversion. And the new ad is ranked last because it has no conversion data. The rank of ads with the same conversion rate is randomly generated.

An adjusting module 300, configured to adjust the sorting of the sorting module according to a second sorting condition; the second sorting condition may be one condition or a plurality of conditions.

Specifically, the index not used in the first sorting condition may also be the second sorting condition, for example, if the first sorting condition is conversion rate, the sorting module 200 may sort using conversion number, profit, price, budget, etc., but may not use conversion rate to perform repeated sorting. The second sorting condition comprises at least one segment sorting condition; the segmentation ordering condition is an index for dividing the advertisements into different echelons according to the condition and ordering the echelons. The different echelons are spaced at unequal intervals. For example, the sorting module 200 uses the conversion number as the segment sorting condition, and the conversion number is 0 to 100,000, the sorting module 200 can divide the conversion number into 7 different echelons, and the corresponding conversion numbers are: 0 to 30,000, 30,001 to 45,500, 45,501 to 70,000, 70,001 to 90,000, 90001 to 97000, 97001 to 99000, and 99,001 to 100,000. In the above 7 banks, the data interval of each bank is not the same. The higher the flight, the fewer the number of advertisements in the flight that perform better. For example, the first echelon with the conversion number of 99,001-100,000 has 5 advertisements, and the second echelon with the conversion number of 97001-99000 has 100 advertisements. If the adjusted advertisement rank corresponding to the first platoon is within 20 and the advertisement ranks discharged in the sorting module 200 are 1, 4, 10, 18, and 24, respectively, the adjusting module 300 will adjust the advertisement ranks other than 20 to the rank 20, and the advertisement ranks in the range will not change, so the rank corresponding to the advertisement of the first platoon is adjusted to 1, 4, 10, 18, and 20. While the original advertisement rank 20 is adjusted to 21, and the subsequent advertisements are adjusted back in sequence. And then, the advertisements of the second echelon are adjusted until all the advertisements obtain the ranking meeting the requirements.

In the case where the second sorting condition employs the index in the first sorting condition, it is also necessary to employ an index other than the first sorting condition, that is, two second sorting conditions are required at minimum. The second ordering condition applies a different evaluation condition to the new advertisement than the old advertisement to advance the order of the new advertisement. The second sorting condition has an attribute condition and a random condition for the sorting condition of the new advertisement; wherein the attribute condition refers to the targeting condition of the new advertisement or the attribute of the advertiser; the random condition is that the new advertisements are randomly ordered according to an equal probability method. When the targeting conditions for new advertisements are more numerous, they will get better ranking than other new advertisements. When the advertisement is sent by the channel provider, the advertisement with more targeting conditions is ranked at the top, so the advertisement with more targeting conditions is matched preferentially. If the matching is successful, the advertisement is recommended, and if the matching is unsuccessful, the advertisement is recommended. If a ad with a high targeting condition is ranked lower than an ad with a low targeting condition, there will be little bidding opportunity and will perform very poorly. On the other hand, the advertisement with more orientation conditions has more requirements, and the targeted advertisement target client is more accurate, so the promotion effect is theoretically better, and the advertisement promotion effect can be improved by giving better ranking.

The equal probability method for the new advertisement comprises a Thompson algorithm, a random algorithm and a machine learning algorithm. The principle behind thompson sampling is the Beta distribution, which has two parameters a and b. If the parameter a is the number of clicks of the recommended user and the parameter b is the number of clicks of the recommended user, the Thompson algorithm process is as follows:

1. and extracting parameters a and b corresponding to each candidate. 2. For each candidate a and b are used as parameters, a random number is generated using a beta distribution. 3. And sorting according to random numbers, and outputting the candidate corresponding to the maximum value. 4. And observing user feedback, and adding 1 to a of the corresponding candidate if the user clicks, or adding 1 to b.

In fact, in the recommendation system, to store a set of parameters for each user, for example, m candidates and n users, 2 m n parameters are stored.

The thompson algorithm is effective for the following reasons.

1) If a candidate is selected a large number of times, i.e. a + b is large, its distribution is narrow, in other words the benefit of the candidate is well defined, i.e. almost well defined regardless of whether the distribution center is close to 0 or 1. It is used to generate random numbers, substantially near the center position, approaching average profit.

2) If a candidate is not only a + b very large, i.e. the distribution is very narrow, but also a/(a + b) is very large, close to 1, it is determined that this is a good candidate, the average gain is very good, and each time the selection is dominant, the utilization phase is entered. Otherwise, the average distribution is close to 0, and almost no more days exist.

3) If a candidate with a + b small is widely distributed, i.e. not selected too many times, indicating that the candidate is good or bad or not very certain, the distribution is skipped, which may be good this time, which may be bad the next time, i.e. there is still a chance to exist, which is not completely discarded. It is possible to obtain a larger random number by generating the random number using it, and the larger random number is preferentially output in sorting, which serves as a screening function.

The card shuffling algorithm adopts a Fisher-Yates Shuffle algorithm, and the basic idea is to randomly take a number which is not taken before from an original array into a new array, and the method specifically comprises the following steps:

1. initializing an original array and a new array, wherein the original array has the length of n (known); 2. randomly generating a number p between [0, k) (assuming the array starts from 0) from the array that has not been processed (if there are k remaining); 3. taking out the p-th number from the rest k numbers; 4. repeating the steps 2 and 3 until all the numbers are taken out; 5. the number sequence extracted from step 3 is a scrambled number sequence.

The machine learning algorithm adopts a Deep Belief Network (DBN) model. The limited Boltzmann machine adds a limit to the Boltzmann machine, and the limit changes the complete graph into a bipartite graph. The neuron is composed of a visible layer and a hidden layer, and the neurons of the visible layer and the neurons of the hidden layer are in full bidirectional connection.

h represents a hidden layer, v represents a visible layer in the RBM, a weight w between any two connected neurons represents the connection strength of the neurons, and each neuron has a bias coefficient b (for the visible layer neuron) and a bias coefficient c (for the hidden layer neuron) to represent the weight of the neuron itself. The DBN is a probabilistic generative model, as opposed to the neural network of the traditional discriminant model, which builds a joint distribution between observed data and labels, and evaluates both P (observer | Label) and P (Label | observer), while the discriminant model has only evaluated the latter, i.e., P (Label | observer). The DBN is composed of layers of constrained Boltzmann Machines (Restricted Boltzmann Machines), a typical type of neural network being shown. These networks are "constrained" to a visible layer and a hidden layer, with connections between layers, but no connections between cells within a layer. Hidden layer units are trained to capture the dependencies of the high-order data represented in the visible layer.

A gradient approach is still employed for ranking of new advertisements in the overall advertisement ranking. For example, consider the top 15 in the ordering of advertisements as a first platoon and 16-50 as a second platoon, where 5 new advertisements are allocated in the first platoon, 15 new advertisements are allocated in the second platoon, and the new advertisements are all located at the last position of the ranking in the first platoon. When the ranking of the new advertisement conflicts with other ranking rules for the same ranking, the advertisements specified by other rankings are ranked forward based on the ranking specified by the ranking rules of the new advertisement.

For the invention, the combination of different indexes can generate different sequencing effects, so the selection of the indexes is very critical. The number of conversions or the conversion rate is generally selected in a first ranking condition of an advertising promotion style of conversion settlement. And the other unselected one of the conversion number or the conversion rate in the second sequencing condition is selected, and other and new advertisement sequencing conditions are supplemented, so that various high-quality advertisements can obtain better ranking, and better advertisement promotion effect and income are realized.

A sending module 400, configured to send the advertisement to the distributor according to the ordering of the adjusting module.

In particular, the sending module 400 may send the latest advertisement ranking for each advertisement request. The system recalculates the ranking each time a distributor requests an advertisement to make the ranking of advertisements more accurate. The sorted index data is updated once a day. When the new advertisement putting time is within 24 hours, all the new advertisements are regarded as new advertisements, and the data of the new advertisements are not used as the sequencing indexes.

Correspondingly, as shown in fig. 2, the present invention further provides an advertisement ranking method, which includes the following steps.

And S001, screening out advertisements meeting the conditions by using the orientation module according to the information provided by the channel provider and the orientation conditions set by the advertiser.

In this step, the server forms an advertisement set by screening advertisements from targeting conditions preset by the advertiser, and defines a range for subsequent advertisement sequencing.

S002, sorting the advertisements screened out by the orientation module according to a first sorting condition by using the sorting module; wherein the first ordering condition is effectiveness data of advertisement placement.

In this step, the advertisements screened in step S001 are sorted according to the most critical effect indicators, so that an advertisement sorting with better overall quality can be formed, but the problems that other indicators cannot be taken care of and new advertisements are not friendly in a single indicator sorting still exist, and therefore the sorting needs to be further adjusted.

S003, adjusting the sorting of the sorting module by using the adjusting module according to a second sorting condition; the second sorting condition may be one condition or a plurality of conditions.

In this step, one or more second sorting conditions are set, and advertisements with excellent other indexes are sorted in a former position, and meanwhile, ranking guarantee can be performed on new advertisements, so that overall income is guaranteed, and advertisements with better quality can obtain more showing opportunities. In addition, the method is more friendly to the new advertisements, so that the new advertisements can obtain more display opportunities, and therefore when the historical advertisement data is owned and the old advertisements become, the advertisement ranking matched with the advertisement quality can be obtained.

And S004, sending the advertisements to the channel traders according to the sequence of the adjusting module by using the sending module.

In this step, the advertisement sets are sent to the channel provider together according to the final ranking of step S003 for the channel provider to bid, show, etc. for advertisements. Each new ad request triggers a re-execution from steps S001-S004 to obtain the latest ad ranking.

While the invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not to be limited to the disclosed embodiment, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. An advertisement ranking system, characterized by: the device comprises an orientation module, a sequencing module, an adjustment module and a sending module;

the adjusting module is used for adjusting the sorting of the sorting module according to a second sorting condition; wherein, the second sorting condition may be one condition or a plurality of conditions;

2. The advertisement ranking system of claim 1 wherein the advertisements screened by the targeting module include old advertisements and new advertisements; the old advertisements are advertisements with historical delivery data, and the new advertisements are advertisements without historical delivery data.

3. The advertisement ranking system of claim 1 wherein the first ranking condition is one of a number of conversions, a conversion rate, an avail, a price, a budget.

4. An advertisement ranking system according to claim 1 wherein the first ranking condition is also the second ranking condition, but the first ranking condition is not the same as the second ranking condition in ranking the same advertisement.

5. The advertisement ranking system of claim 1 wherein said second ranking condition comprises at least one segment ranking condition; the segmentation ordering condition is an index for dividing the advertisements into different echelons according to the condition and ordering the echelons.

6. The system of claim 1, wherein the second ranking condition is advertisement stability and fund settlement rate.

7. An advertisement ranking system according to claim 1 wherein the second ranking condition applies different rating conditions to the new advertisements than to the old advertisements to advance the ranking of the new advertisements.

8. The advertisement ranking system of claim 7 wherein the second ranking condition has an attribute condition and a random condition for the ranking condition of the new advertisement; wherein the attribute condition refers to the targeting condition of the new advertisement or the attribute of the advertiser; the random condition is that the new advertisements are randomly ordered according to an equal probability method.

9. The advertisement ranking system of claim 7 wherein the equal probability method comprises Thompson's algorithm, stochastic algorithm, machine learning algorithm.

10. A method of ordering advertisements, comprising:

A. screening out eligible advertisements by using the targeting module according to the information provided by the channel provider and the targeting conditions set by the advertiser according to any one of claims 1 to 9;

B. ranking the advertisements screened by the targeting module according to a first ranking condition using a ranking module according to any of claims 1-9; wherein the first ordering condition is effectiveness data of advertisement placement;

C. adjusting the sorting of the sorting module according to a second sorting condition using an adjustment module according to any one of claims 1-9; wherein, the second sorting condition may be one condition or a plurality of conditions;

D. transmitting advertisements to the distributor in the order of the adjustment module using the transmission module of any of claims 1-9.