A Method and Application of Automobile Enterprise Activity Operation KPI Allocation Based on a Sliding Sales Funnel

Abstract

The present application discloses a method and application for allocating KPI of automobile enterprise activities operation based on a sliding sales funnel. The method includes: obtaining historical data and inputting the historical data into a funnel table in units of cycles; calculating the correlation between key indicators and multiple attributes based on the historical data; extracting key indicators with a window length of a set cycle and multiple attributes associated therewith from the funnel table as training data to construct an index prediction model for the current cycle; predicting the key indicator values of the current cycle according to the index prediction model; obtaining the real key indicator values of the current cycle and filling them into the funnel table, sliding the window of the funnel table forward by one cycle, extracting key indicator values with a set cycle therefrom to construct an index prediction model for the next cycle, and predicting the key indicator values of the next cycle. The present invention processes data in a sliding window manner, which can make the calculated indicators have better interpretability and accuracy, and improve the accuracy of index prediction and the rationality of index adjustment.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and more specifically, to a method and application for allocating KPIs for automobile enterprise activity operations based on a sliding sales funnel. Background Art

[0002] An activity is an instantiation of a strategy. After enterprise operators host an activity operation, they will evaluate the effects of various measures. Whether the measures are good or bad, they will serve as a reference for subsequent similar activities. The operation effect of a certain activity will be similar to the effect of the most similar activities in the time dimension, that is, a certain activity has a higher correlation with several adjacent activities of this activity. And during the execution of the activity, there are often risks that the operation effect cannot meet the expectations, causing negative impacts and increasing enterprise operation costs, etc. Therefore, if the historical data of activity operations can be used to adjust and optimize the activity operations in the next stage, it will help to improve the effect of activity operations.

[0003] Taking automobile enterprises as an example, since automobiles are low-frequency and high-value purchase goods, automobile enterprises generally use activity operations to maintain brand exposure and conduct in-depth interactions with car owners, but the quantitative evaluation of activity operation effects has always been a difficult point.

[0004] The invention patent application with the patent application number CN201911346348.5 discloses an operation system and method based on user behavior data and user portrait data. The basic steps of this method include: 1. Obtain user behavior data and user portrait data; 2. Configure the acquisition rules for the target user group based on the user behavior data and portrait data; 3. Configure the update period and push rules for the target user group; 4. Regularly push the target user group to the operation system; 5. After the operation is completed, evaluate the operation effect. This method uses user behavior data and user portrait data, and is more suitable for the user operation of high-frequency purchase goods. It is a static ex-post operation effect evaluation, without establishing an operation effect prediction model, and dynamically correcting the prediction model according to the obtained operation data during the operation process.

[0005] Currently, in the activity operations and digital marketing of automobile enterprises, the most commonly used user operation model is the AARRR pirate model and its variants, which is a typical sales funnel model. The invention patent application with the patent number CN202011319286.1 discloses a SFA management method and system based on a sales funnel. In the lead management module of this method, the sales funnel model is used to model and manage sales leads. Essentially, it is a refined operation of stratifying the leads flowing into the system, and there is no effective means to deal with the allocation and prediction of key performance indicators in different operation stages. Summary of the Invention

[0006] In view of at least one defect or improvement requirement of the prior art, the present invention provides a method and application for allocating KPIs for automotive enterprise activity operation based on a sliding sales funnel, which belongs to a data processing method. By mining the relationship between key indicators and other attributes from historical data, a historical data with a determined length unit is dynamically taken out by using a sliding window to construct an index prediction model for different statistical periods, and the prediction of key indicators for the corresponding period is executed to realize the pre-allocation of key indicators in different statistical periods. Based on the above method to guide the optimization of the activity operation plan, the operation effect can be effectively improved.

[0007] To achieve the above object, according to one aspect of the present invention, there is provided a method for allocating KPIs for automotive enterprise activity operation based on a sliding sales funnel, which includes:

[0008] Obtain historical data of multiple statistical periods in activity operation, and input the historical data into a funnel table in units of periods;

[0009] Calculate the correlation degree between key indicators and multiple attributes associated therewith based on the historical data, and obtain several attributes with the largest correlation degree with the key indicators;

[0010] Extract key indicators with a window length of a set period and multiple attributes associated therewith from the funnel table as training data to construct an index prediction model for the current statistical period; predict the key indicator value of the current statistical period according to the index prediction model;

[0011] Obtain the real key indicator value of the current statistical period and fill it into the funnel table, slide the window of the funnel table forward by one statistical period, extract the key indicator value with the set period and including the current period from it to construct an index prediction model for the next statistical period, and predict the key indicator value of the next statistical period until the activity deadline is reached.

[0012] Preferably, in the above method for allocating KPIs for automotive enterprise activity operation based on a sliding sales funnel, the calculation of the correlation degree between key indicators and multiple attributes associated therewith based on the historical data includes:

[0013] Construct a relationship matrix based on key indicators and multiple attributes associated therewith, and the relationship matrix is an [,n + 1]-dimensional matrix, where n represents the number of periods included in the historical operation data, and m represents the total number of key indicators and multiple attributes;

[0014] Calculate the variance between each key indicator and the associated attribute, and use the least squares method to calculate the correlation coefficient between each attribute and the key indicator;

[0015] Standardize the correlation coefficient to obtain the correlation degree between each attribute and the key indicator.

[0016] Preferably, in the above-mentioned method for allocating KPIs for automotive enterprise activity operation based on a sliding sales funnel, the historical data includes attribute data of users generated during the digital operation process, user behavior data with timestamps, and operation behavior data with timestamps.

[0017] Preferably, the above-mentioned method for allocating KPIs for automotive enterprise activity operation based on a sliding sales funnel further includes:

[0018] Calculating the deviation between the key indicator value predicted by the indicator prediction model for the current statistical period and the predetermined value and outputting it.

[0019] According to the second aspect of the present invention, there is also provided a device for allocating KPIs for automotive enterprise activity operation based on a sliding sales funnel, which is characterized by including:

[0020] An acquisition unit, configured to obtain historical data of multiple statistical periods in activity operation and input the historical data into a funnel table in units of periods;

[0021] An association unit, calculating the correlation between a key indicator and multiple attributes associated therewith based on the historical data, and obtaining several attributes with the highest correlation with the key indicator;

[0022] A model construction unit, configured to extract the key indicator with a window length of a set period and multiple attributes associated therewith from the funnel table as training data to construct an indicator prediction model for the current statistical period; predicting the key indicator value for the current statistical period according to the indicator prediction model;

[0023] A sliding unit, configured to obtain the actual key indicator value for the current statistical period and fill it into the funnel table, slide the window of the funnel table forward by one statistical period, extract the key indicator value with the set period and including the current period therefrom to construct an indicator prediction model for the next statistical period, and predict the key indicator value for the next statistical period until the activity end time is reached.

[0024] Preferably, in the above-mentioned device for allocating KPIs for automotive enterprise activity operation based on a sliding sales funnel, the association unit includes:

[0025] A matrix construction module, configured to construct a relationship matrix based on the key indicator and multiple attributes associated therewith, the relationship matrix being an [m, n + 1]-dimensional matrix, where n represents the number of periods included in the historical operation data, and m represents the total number of the key indicator and multiple attributes;

[0026] A correlation calculation module, configured to calculate the variance between the key indicator and each associated attribute, and calculate the correlation coefficient between each attribute and the key indicator using the least squares method;

[0027] A normalization module, configured to normalize the correlation coefficients to obtain the correlation degrees between each attribute and the key indicators.

[0028] Preferably, in the above-mentioned automotive enterprise activity operation KPI allocation device based on a sliding sales funnel, the historical data includes attribute data of users generated during the digital operation process, user behavior data with timestamps, and operation behavior data with timestamps.

[0029] Preferably, in the above-mentioned automotive enterprise activity operation KPI allocation device based on a sliding sales funnel, it further includes:

[0030] A statistics module, configured to calculate and output the deviation between the key indicator value predicted by the indicator prediction model in the current statistical period and the predetermined value.

[0031] According to one aspect of the present invention, there is also provided a computer device, which includes at least one processing unit and at least one storage unit. Among them, the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the steps of the method described in any one of the above.

[0032] According to one aspect of the present invention, there is also provided a computer-readable medium, characterized in that it stores a computer program executable by a computer device, and when the computer program runs on the computer device, the computer device executes the steps of the method described in any one of the above.

[0033] Generally speaking, compared with the prior art by the above technical solutions conceived by the present invention, the following beneficial effects can be achieved:

[0034] The data processing method provided by the present invention utilizes the correlation between the data of different cycles of the same activity, dynamically trains a corresponding indicator prediction model according to the historical data of the window length before the current cycle, and uses this model to calculate the key indicators of the current cycle, and evaluates the difference between the predicted value and the allocated value or the target value. In the next cycle, the sliding window position moves to the next cycle, and the prediction model is dynamically corrected according to the obtained operation data, and the indicator prediction model for the next cycle is retrained. Subsequently, it recurs to the deadline or the activity deadline cycle. In this way, the entire funnel table is filled dynamically. According to the data in the funnel table, it can be judged and analyzed whether the KPI target value can be completed according to the current model prediction value, and the data has better readability. Furthermore, the resource input and plan of the activity operation can be dynamically adjusted during the operation process according to the current actual completion situation, and the key indicators of the subsequent cycles can be dynamically readjusted and allocated in an iterative and refined manner, ultimately improving the activity operation efficiency and achieving the goal of completing the total activity operation indicators.

[0035] Data processing is carried out based on the sliding window method, which quantifies the capabilities of operation personnel, enables the calculated metrics to have better interpretability and accuracy, and improves the precision of metric prediction and the rationality of metric adjustment. Brief Description of the Drawings

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0037] Figure 1 It is a flowchart of a data processing method provided in this embodiment;

[0038] Figure 2 It is a flowchart of a data processing device provided in this embodiment;

[0039] Figure 3 It is a flowchart of a computer device provided in this embodiment. Detailed Embodiments

[0040] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0041] The terms "first", "second", "third", etc. in the specification, claims and drawings of the present application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally further include steps or units not listed, or may optionally further include other steps or units inherent to these processes, methods, products or devices.

[0042] Taking the enterprise activity operation process as an example, the technical idea of ​​the present invention is: within a certain period range, the historical activity data of the previous period or several periods is used to dig out the relationship between key indicators and several attributes, and this relationship is modeled to dynamically predict the key indicators or effects of the next period, so that the enterprise operation personnel can configure activity resources according to the prediction results and improve the operation effect. After the current cycle activity is completed, the actual KPI is filled into the funnel table (such as the sales funnel table), that is, the funnel table is slid by a window. According to the actual key indicators, the prediction model of the current cycle is corrected, and the key indicators in the next cycle are predicted using the modified prediction model, and the entire sales funnel table is filled in turn, so as to achieve the dynamic allocation of key indicators and resource configuration of activity operation by sliding the sales funnel method, continuously iterate and optimize the prediction model, and realize the purpose of dynamic allocation of key indicators of activity operation. In addition, the prediction value of the key indicator output by the prediction model of each cycle can be used to evaluate the effect of activity operation.

[0043] Figure 1 is a flow chart of a method for allocating KPIs for automobile enterprise activity operations based on a sliding sales funnel provided in this embodiment. Figure 1 As shown, the method comprises the following steps:

[0044] S1 obtains historical data of multiple statistical periods during the operation of the activity, inputs the historical data into the sales funnel table in units of periods, and completes the initialization of the sales funnel table;

[0045] In this embodiment, the historical data of activity operation refers to various data generated during the digital operation process of the enterprise, including trend data, source data, access data, page data, specifically user attribute data, user behavior data with timestamps, and operational behavior data with timestamps. Among them, activity data needs to be distinguished from general operational data, so specific activity identifiers or activity-provided filtering items are required to distinguish activity data from operational data. That is, historical data needs to have distinguishable identifiers or filtering conditions.

[0046] The collected historical data has a time tag, which is accurate to the finest known granularity, and the historical data can be entered into the sales funnel table in cycles based on the time tag.

[0047] Optionally, after collecting historical data, it is necessary to clean a large amount of historical data, and sort out the data related to key indicators and activity-related data from it; in a specific example, based on the input pattern of data characteristic values, refer to the rules set by humans to clean up invalid data. The purpose of data cleaning is mainly to detect outliers and perform integrity checks to clean up valid activity operation data. For example, activity operation data mainly includes activity types, activity goals (such as increasing 10,000 fans, increasing 1,000 leads, conversion rate of 8%, etc.), activity input resources (such as activity funds, activity props, activity gifts, number of copywriting videos, etc.), activity cycle, etc.

[0048] Initialize the sales funnel table, mainly by using the AARRR model to construct the funnel table and write historical data into the funnel table. Here, taking the fan-increasing activity of a car company as an example, a sales funnel model is constructed through the fan layer, intention layer, active layer, and lead layer. Generally speaking, the number decreases continuously from the fan layer to the intention layer, from the intention layer to the active layer, and finally from the active layer to the lead layer, thus forming a sales funnel model. During the process of user operation using this sales funnel model, some historical data will be generated. The most typical data among them are the corresponding number of fans in the fan layer, the number of intentions in the intention layer, the number of active users in the active layer, and the number of leads in the lead layer, etc. Construct a sales funnel table based on historical operation data, activity input resource data, actual historical KPI completion data, etc., and initialize the sales funnel table with the historical data that has occurred (from July to September). The initialized sales funnel table after importing historical data is shown in Table 1:

[0049] Table 1 Initialized sales funnel table

[0050]

[0051] Refer to Table 1, where increasing fans and leads are both key indicators (important KPIs) of activity operation. Taking the fan-increasing and lead-generation activities of a certain car company's APP as an example, the total goal is to increase 100,000 fans, and the total activity cycle is from July to December. Each evaluation and prediction cycle is in units of months. In the following table, July, August, and September are the real historical data that have occurred. Therefore, it is still necessary to increase 86,190 fans from October to December. Here, the activity funds (unit: yuan) considered to be most relevant to the fan-increasing activity are taken as the input resources. The following will take increasing fans and leads as examples, and in combination with Tables 1 to 4, further describe this method in detail.

[0052] S2 Calculate the correlation between key indicators and multiple attributes associated with them based on historical data, and obtain several attributes with the highest correlation with the key indicators;

[0053] In this embodiment, data analysis means are used to mine the relationships between key indicators and multiple attributes associated with them from historical data to form a relationship model; the relationship model refers to representing the relationships between key indicators and other attributes in a model manner. The other attributes in the operation data except for the key indicators are sorted out, the key indicators are used as dependent variables, and the other attributes are used as independent variables to calculate the importance of all independent variables, and 1 to 3 attributes with the highest importance are selected for regression analysis. Among them, f is the mined relationship model, and X i is the data of the i-th period, and Opdata is the collected historical operation data:

[0054]

[0055] Furthermore, the correlation degrees between the key indicators and multiple attributes associated with them are calculated, including:

[0056] A relationship matrix is constructed based on the key indicators and multiple attributes associated with them. The relationship matrix is an [m, n+1]-dimensional matrix, where n represents the number of periods included in the historical operation data, and m represents the total number of key indicators and multiple attributes;

[0057] The variances between the key indicators and each associated attribute are calculated, and the least squares method is used to calculate the correlation coefficients between each attribute and the key indicators;

[0058] The correlation coefficients are standardized to obtain the correlation degrees between each attribute and the key indicators.

[0059] Specifically, the key indicators and data of multiple attributes associated with the key indicators are extracted from the historical data. For example, for the i-th period, the data is filled in the corresponding positions in the form according to KPI, y (1) 、y (2) …y (n) A total of n periods can obtain an [m, n+1]-dimensional matrix. Among them, KPI, y (1) 、y (2) …y (n) represent the key indicator and n attributes respectively, and their variances are The coefficients α1, α1, …α n of the n attributes are obtained using the least squares method, and the standardized coefficients are the variable importance The variable importance is sorted, and the top 3 attributes with the largest values (such as activity type, funds, days) are taken, and the key indicator is linearly correlated with these 3 attributes to obtain a linear equation:

[0060] KPI 10 =α1y (type) +α2y (days) +α3y (cost)

[0061] S3 extracts key indicators with a window length of a set period and multiple attributes associated therewith from the sales funnel table as training data to construct an indicator prediction model for the current statistical period; predicts the key indicator values for the current statistical period according to the indicator prediction model;

[0062] In this embodiment, the indicator prediction model is a regression model that constructs a prediction of the key indicator for the next target period based on the relationship model between the key indicators formed in step S2 and multiple attributes associated therewith. When the target period changes, the indicator prediction model needs to be dynamically adjusted according to the key indicator values that actually occurred in the previous period. Select the indicator prediction model corresponding to the current period to predict the key indicators for the current period.

[0063] As shown in Table 2, the red data are the fan growth data predicted by the indicator prediction model for the next period (October).

[0064] Table 2 Updated sales funnel table

[0065]

[0066] If October is the end month of the current activity operation, the current data processing process ends; otherwise, input the actual fan growth data in October into the sales funnel table to replace the predicted value of the indicator prediction model, and enter the data processing process for the next period (January).

[0067] S4 obtains the actual key indicator values for the current statistical period and fills them into the sales funnel table, slides the window of the sales funnel table forward by one statistical period, extracts from it key indicators with the set period and including the key indicator values for the current period to construct an indicator prediction model for the next statistical period, and predicts the key indicator values for the next statistical period until the activity end time is reached.

[0068] In this embodiment, an indicator prediction model is periodically constructed to predict the values of key indicators for the corresponding statistical period: on the basis of the divided periods, for the periods to be calculated in chronological order, an indicator prediction model for each period is constructed in turn to predict the key indicator values for the current period. At the i-th period, according to the historical data X before the i-th period i-1 Construct an indicator prediction model f (i) , calculate the key performance indicator KPI for the current i-th period i . Among them, the data for each period can be regarded as the key performance indicator KPI, the activity historical data X i-1 , the activity attribute data y i-1 (such as the funds to be invested in the activity, etc.). Then at the i-th period, calculate the corresponding indicator prediction model f (i) and predict the key performance indicator KPIi :

[0069] x i-1 ={KPI i-1 ,y i-1}

[0070] X i-1 ={x j |j < i, j ∈ Z+}

[0071]

[0072] KPI i = f (i) (y i )

[0073] Construct a regression model for each period to be calculated based on the divided periods and the mined correlation attributes, predict the key indicator values for the periods that have not occurred in the sales funnel table, and fill the predicted values of the key indicators into the corresponding period data in the sales funnel table; then slide forward by one statistical period and repeat the modeling and prediction until the entire sales funnel table is filled.

[0074] In this embodiment, an indicator prediction model is constructed based on the sliding window method. Since historical data that is far from the current time has less impact on the current modeling, while historical data that is close to the current time has a greater impact on the current modeling, a certain length unit of historical data is discarded, and the remaining length unit is the window. Taking the time unit as an example, at the (+1)th period, set the window length to n, and then use the window data X i and the mined indicator prediction model f (i+1) , calculate the key indicator KPI i+1 :

[0075] X i ={x j |i - n ≤ j ≤ i, j ∈ N}

[0076] card(X i ) = n

[0077] KPI i+1 = f (i+1) (y i+1 )

[0078] In this embodiment, first, it is determined whether the current cycle is the cut-off cycle, and it is decided whether to slide the sales funnel table. The cut-off date is set to December 31st. After calculating the key indicators for October, it is determined that October does not include the cut-off date, and the training index prediction model is returned. The current cycle is incremented by one cycle or the funnel table is slid forward by one window. Then, based on the data for August, September, and October, the index prediction model for November is trained, and the key indicator values predicted by the model for November are written into the sales funnel table. An example is shown in Table 3, where the red data are the predicted data.

[0079] Table 3 Updated Sales Funnel Table

[0080]

[0081] In the same way, the data for increasing the number of fans in December is predicted. It is determined that December includes the activity cut-off date, so the model iteration terminates, and the sliding sales funnel table is filled. As shown in Table 4, where the green number 86190 represents the current target, that is, a total of 86,190 fans need to be added in October, November, and December; the red number 58619 represents the predicted number of fans added by the model, which represents the expectation of the activity execution. Obviously, the model-predicted increase in fans is less than the target number, so it is judged that it is highly probable that the target cannot be achieved according to this plan, and measures need to be taken, such as setting a new activity operation plan, such as increasing the activity operation funds or modifying the index allocation, etc. It can be seen that by using the data processing method provided in this solution, the key indicator prediction data for different cycles can be iteratively generated, the final effect of the activity operation can be evaluated and analyzed, and it can be used to guide the optimal allocation of the activity operation.

[0082] Table 4 Updated Sales Funnel Table

[0083]

[0084] It should be noted that although in the above embodiments, the operations of the method of the embodiments of this specification are described in a specific order, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the order of the steps depicted in the flowchart can be changed. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution.

[0085] This embodiment also provides a device for allocating KPIs for vehicle enterprise activities operation based on a sliding sales funnel. Refer to Figure 2 , this device includes:

[0086] An acquisition unit, configured to obtain historical data for multiple statistical cycles, and input the historical data into the sales funnel table in units of cycles;

[0087] The association unit calculates the correlation between key metrics and multiple attributes associated therewith based on historical data, and obtains several attributes with the highest correlation with the key metrics;

[0088] The model construction unit is configured to extract key metrics with a window length of a set period and multiple attributes associated therewith from the sales funnel table as training data to construct an index prediction model for the current statistical period; predict the key metric value for the current statistical period according to the index prediction model;

[0089] The sliding unit is configured to obtain the actual key metric value for the current statistical period and fill it into the sales funnel table, slide the window of the sales funnel table forward by one statistical period, extract key metric values with the set period and including the current period therefrom to construct an index prediction model for the next statistical period, and predict the key metric value for the next statistical period until the activity cut-off time is reached.

[0090] Further, the association unit includes:

[0091] The matrix construction module is configured to construct a relationship matrix based on the key metrics and multiple attributes associated therewith, and the relationship matrix is an [m, n + 1]-dimensional matrix, where n represents the number of periods included in the historical operation data, and m represents the total number of key metrics and multiple attributes;

[0092] The correlation calculation module is configured to calculate the variance between the key metrics and each associated attribute, and calculate the correlation coefficient between each attribute and the key metrics using the least squares method;

[0093] The normalization module is configured to normalize the correlation coefficient to obtain the correlation between each attribute and the key metrics.

[0094] Further, the above data processing device further includes:

[0095] The statistics module is configured to calculate and output the deviation between the key metric value predicted by the index prediction model for the current statistical period and a predetermined value.

[0096] For the specific limitations of the data processing device, reference can be made to the limitations on the data processing method in the above text, which will not be elaborated here. Each module in the above data processing device can be implemented in whole or in part through software, hardware, and their combinations. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form so that the processor can call and execute the operations corresponding to the above modules.

[0097] This embodiment also provides a computer device, see Figure 3, which includes at least one processor and at least one memory. Among them, a computer program is stored in the memory. When the computer program is executed by the processor, the processor is caused to execute the steps of the above-mentioned method for allocating KPIs for the vehicle enterprise activity operation based on the sliding sales funnel; in this embodiment, the types of the processor and the memory are not specifically limited. For example, the processor can be a microprocessor, a digital information processor, a programmable logic system on a chip, etc.; the memory can be a volatile memory, a non-volatile memory, or a combination thereof, etc.

[0098] This computer device can also communicate with one or more external devices (such as a keyboard, a pointing terminal, a display, etc.), and can also communicate with one or more terminals that enable a user to interact with this computer device, and / or communicate with any terminal (such as a network card, a modem, etc.) that enables this computer device to communicate with one or more other computing terminals. Such communication can be carried out through an input / output (I / O) interface. And, the computer device can also communicate with one or more networks (such as a Local Area Network (LAN), a Wide Area Network (WAN), and / or a public network, such as the Internet) through a network adapter.

[0099] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned method embodiments. Among them, any reference to a memory, a storage, a database, or other media used in the various embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memories (ROMs), programmable ROMs (PROMs), electrically programmable ROMs (EPROMs), electrically erasable programmable ROMs (EEPROMs), or flash memories. Volatile memories can include random access memories (RAMs) or external cache memories. By way of illustration and not limitation, RAMs are available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0100] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0101] It is easy for those skilled in the art to understand that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for allocating KPIs for automotive enterprise activity operations based on a sliding sales funnel, characterized in that Including: Obtain historical data for multiple statistical periods, construct a funnel table using the AARRR model, and input the historical data into the funnel table in units of periods; The historical data includes user attribute data, timestamped user behavior data, and timestamped operation behavior data generated during the digital operation process; Calculate the correlation between key indicators and multiple associated attributes based on the historical data, and obtain several attributes with the highest correlation with the key indicators; Extract key indicators with a window length of a set period and multiple associated attributes from the funnel table as training data to construct an indicator prediction model for the current statistical period; predict the key indicator values for the current statistical period according to the indicator prediction model; Obtain the actual key indicator values for the current statistical period and fill them into the funnel table. Slide the window of the funnel table forward by one statistical period, extract the key indicator values with the set period and including the current period, and send them to the model construction unit to construct an indicator prediction model for the next statistical period and predict the key indicator values for the next statistical period until the activity deadline is reached; The calculation of the correlation between key indicators and multiple associated attributes based on the historical data includes constructing a relationship matrix based on the key indicators and multiple associated attributes. The relationship matrix is an [m, n + 1]-dimensional matrix, where n represents the number of periods included in the historical operation data, and m represents the total number of key indicators and multiple attributes; calculate the variance between each key indicator and its associated attribute, and use the least squares method to calculate the correlation coefficient between each attribute and the key indicator; Standardize the correlation coefficient to obtain the correlation between each attribute and the key indicator.

2. The method for allocating KPIs for vehicle enterprise activity operation based on a sliding sales funnel according to claim 1, wherein, It also includes: Calculate the deviation between the key indicator values predicted by the indicator prediction model for the current statistical period and the predetermined values and output it.

3. A vehicle enterprise activity operation KPI allocation device based on a sliding sales funnel, characterized in that Including: A collection unit for obtaining historical data for multiple statistical periods, constructing a funnel table using the AARRR model, and inputting the historical data into the funnel table in units of periods; The historical data includes user attribute data, timestamped user behavior data, and timestamped operation behavior data generated during the digital operation process; An association unit for calculating the correlation between key indicators and multiple associated attributes based on the historical data and obtaining several attributes with the highest correlation with the key indicators; A model construction unit for extracting key indicators with a window length of a set period and multiple associated attributes from the funnel table as training data to construct an indicator prediction model for the current statistical period; predicting the key indicator values for the current statistical period according to the indicator prediction model; A sliding unit for obtaining the actual key indicator values for the current statistical period and filling them into the funnel table, sliding the window of the funnel table forward by one statistical period, extracting the key indicator values with the set period and including the current period and sending them to the model construction unit to construct an indicator prediction model for the next statistical period and predict the key indicator values for the next statistical period until the activity deadline is reached; The association unit includes a matrix construction module configured to construct a relationship matrix based on key metrics and a plurality of attributes associated therewith. The relationship matrix is an [m, n+1]-dimensional matrix, where n represents the number of periods included in the historical operation data, and m represents the total number of key metrics and a plurality of attributes; a correlation calculation module configured to calculate the variance between each key metric and the attributes associated therewith, and calculate the correlation coefficient between each attribute and the key metric using the least squares method; and a normalization module configured to normalize the correlation coefficient to obtain the correlation degree between each attribute and the key metric.

4. The vehicle enterprise activity operation KPI allocation device based on a sliding sales funnel according to claim 3, wherein It further includes: a statistics module configured to calculate and output the deviation between the key metric value predicted by the index prediction model in the current statistical period and a predetermined value.

5. A computer device, characterized in that, It includes at least one processing unit and at least one storage unit. The storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit is caused to execute the steps of the method according to any one of claims 1 or 2.

6. A computer-readable medium, characterized in that, It stores a computer program executable by a computer device, and when the computer program runs on the computer device, the computer device is caused to execute the steps of the method according to any one of claims 1 or 2.

Images