Consumption trend prediction method based on CRM system
Through the consumption trend prediction method based on the CRM system, comprehensively considering factors such as the monthly sales volume, sales price changes, views and return probability of goods, the problem of difficulty in accurately predicting the sales of goods in the existing technology is solved, and more accurate market forecasting and corporate strategy formulation are achieved.
Patent Information
- Application Number
- CN202510276326.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-10
- Publication Date
- 2025-06-27
AI Technical Summary
The prior art is difficult to predict the future transaction volume of commodities through commodity prices, and a single price variable cannot accurately predict commodity sales, and there is a lack of a method to comprehensively consider various factors.
The consumption trend prediction method based on the CRM system is adopted to predict future product sales by obtaining the monthly sales volume, calculating the sales volume increase ratio, obtaining the change curve of sales price with date, obtaining monthly views, calculating the sales volume share and return probability, and comprehensively considering these factors to predict future product sales.
It has achieved more accurate and comprehensive predictions of future commodity sales, helping enterprises formulate reasonable production plans, inventory management and marketing strategies, occupy the market opportunities, and achieve long-term and stable development of enterprises.
Smart Images
Figure CN120218988A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of consumption trend prediction, and particularly relates to a consumption trend prediction method based on a CRM system. Background Art
[0002] The CRM system, namely the customer relationship management system, is a software system specifically used for managing the relationship between an enterprise and its customers. It helps the enterprise better understand customer needs, improve customer satisfaction and loyalty, and thus promote sales growth and business development by integrating multiple functional modules, such as sales automation, marketing automation, customer service and support, etc.
[0003] The CRM system collects, integrates and analyzes the historical sales data of products, price change records, etc. And based on the analysis of these data resources, the CRM system uses advanced data analysis algorithms and models to deeply explore the internal relationship between product prices and transaction amounts. For example, it can analyze the sales situation of products in different price ranges to understand consumers' sensitivity to different prices; at the same time, combined with market trends and competitors' pricing strategies, it predicts the market demand for products at specific prices.
[0004] In the prior art, if only relying on product prices to predict future transaction amounts of products, it will lack credibility, and a single variable cannot accurately obtain the result. Product prices are only part of the factors affecting consumers' purchase decisions. Therefore, there is an urgent need for a consumption trend prediction method based on a CRM system that can comprehensively consider various factors and more accurately and comprehensively predict future product sales situations. In this way, enterprises can grasp market trends in advance, formulate reasonable production plans, inventory management and marketing strategies, so as to gain the upper hand in the fierce market competition and achieve the long-term stable development of the enterprise. Summary of the Invention
[0005] The purpose of the present invention is to provide a consumption trend prediction method based on a CRM system to solve the following technical problems: The purpose of the present invention can be achieved through the following technical solutions: A consumption trend prediction method based on a CRM system includes the following steps: S1: Obtain the monthly sales volume S of a product, and calculate the sales volume growth ratio ΔS = (S now - S last ) / S last , where S now represents the monthly sales volume of the product on the shopping platform in the current month, and S last represents the monthly sales volume of the product on the shopping platform in the previous month. Calculate the average value ΔS ave of the sales volume growth ratio, and use the average value ΔS aveGreater than the preset growth threshold ΔS max The products are marked as key products, and the first expected sales volume S1 of the key products is calculated next = S now *ΔS ave +S now ; S2: Obtain the curve of the sales price G of the key product changing with the date. The expression of the curve is f(R), where R represents the date, and calculate the average value of the sales price , where R1 represents the opening date of the key product, R2 represents the current date, and obtain the sales price G of the key product on the current date now , calculate the price impact coefficient GY = G ave / G now ; S3: Obtain the monthly view volume L of the key product, calculate the sales volume ratio K = S / L * 100%, and obtain the actual conversion volume SK = K * GY after correcting the sales volume ratio K; Obtain the total view volume L of the key product in the current month now , calculate the second expected sales volume S2 next =L now *SK, calculate the comprehensive sales volume SZ = (S1 next +S2 next ) / 2; Obtain the return probability P of the key product = S t / S all , where S t represents the total return volume of the key product, and S all represents the total sales volume of the key product, and calculate the expected return volume , where represents rounding up SZ * P, and calculate the expected sales volume SN of the key product = SZ - T.
[0006] As a further solution of the present invention: In the step S1, another method for screening key products specifically includes: Obtain the collection volume C of the products on the shopping platform, calculate the standard deviation s of the monthly sales volume of the products. If the product collection volume C is greater than the monthly sales volume S and the standard deviation s is greater than the preset judgment value s sta , mark the corresponding products as potential products, obtain the monthly sales volume of the potential products in the past 6 months. If the monthly sales volume shows a monotonically increasing trend with the increase of months, mark the corresponding potential products as key products.
[0007] As a further solution of the present invention: In the step S1, the products with monthly sales volume S < S min are excluded and do not participate in the subsequent steps.
[0008] As a further solution of the present invention: In the step S2, the method for obtaining the selling price G of key commodities specifically includes: Obtain the number M of merchants selling key commodities and the inventory quantity KC of key commodities on the shopping platform, and calculate the selling price , where KC all represents the total inventory quantity of key commodities, KC m represents the inventory quantity of key commodities of the m-th merchant, and G m represents the selling price of key commodities of the m-th merchant.
[0009] As a further solution of the present invention: In the step S2, if R2 - R1 < R min , then stop calculating the average value G of the prices of subsequent commodities ave , where R min represents the preset minimum date.
[0010] As a further solution of the present invention: In the step S3, obtain the browsing duration T of key commodities, preset the minimum judgment duration T min , and do not include the browsing records with browsing duration T ≤ T min in the monthly browsing volume L.
[0011] As a further solution of the present invention: Calculate the standard deviation s of the selling prices of key commodities, and mark the months when the selling price G is outside the interval [G ave - s, G ave + s] as special months, and special months do not participate in the calculation of the expected sales volume SN.
[0012] As a further solution of the present invention: In the step S3, if the return probability P of key commodities is ≥ 50%, then mark them as unstable commodities, and stop calculating the expected sales volume of unstable commodities until the return probability P < 50%.
[0013] Advantages of the present invention: In the present invention, first, it is necessary to comprehensively and accurately obtain the monthly sales volume data of commodities. After obtaining the accurate monthly sales volume data of commodities, next, it is necessary to calculate the sales volume growth ratio of the commodities. The so-called sales volume growth ratio specifically refers to the comparison relationship between the monthly sales volume increase value and the sales volume of the previous month. After obtaining the sales volume growth ratio of each month, it is also necessary to further calculate the average value of these sales volume growth ratios. This average value reflects the average growth level of the commodity sales volume within a certain period, and is of great significance for evaluating the sales trend of the commodity.
[0014] Based on the above calculation results, the commodities whose average sales volume growth ratio is greater than the preset growth threshold are recorded as key commodities. The preset growth threshold is determined based on market rules, industry characteristics, and the strategic goals of the enterprise. For example, in some industries with fierce competition and rapid market changes, the growth threshold may be set relatively low in order to timely discover potential hot commodities; while in some relatively stable markets, the threshold may be appropriately increased.
[0015] For the identified key commodities, the first expected sales volume also needs to be calculated based on their sales volume for the current month. By analyzing the sales data of key commodities over the past period of time, the patterns and trends of sales changes can be summarized, so as to make a preliminary forecast of future sales volume. However, it should be made clear that although this first expected sales volume is based on certain historical data and analysis methods, it is not completely accurate in reality. This is because the market environment is complex and changeable, and is affected by various internal and external factors, such as strategic adjustments by competitors, changes in the macroeconomic situation, and changes in consumer preferences. Therefore, the first expected sales volume is only a reference value, and it needs to be further optimized and adjusted through a series of subsequent steps to make it closer to the actual market situation and provide more reliable support for the company's sales decisions.
[0016] After completing the above preliminary analysis of key commodities, the next step is to further obtain the curve of sales price changes of key commodities with date. This process involves the collection and organization of a large amount of historical data, and it is necessary to extract the sales price information of each key commodity on different dates from the sales record system of the shopping platform. These data may be interfered by various factors, such as fluctuations in market supply and demand, price adjustments of competitors, changes in raw material costs, and changes in the company's own marketing strategy. In order to ensure the accuracy and reliability of the data, these raw data need to be carefully screened and cleaned to eliminate outliers and erroneous data, so as to obtain an accurate curve that can truly reflect the changes in sales prices of key commodities with date.
[0017] After obtaining a high-quality sales price change curve, the next step is to derive the expression corresponding to the change curve. By analyzing the shape, trend, and key points of the curve, find a function expression that best fits the curve. This expression should be able to accurately describe the quantitative relationship between sales price and date, providing a powerful tool for subsequent calculations and analysis.
[0018] For example, it may be found that the sales prices of some key commodities show seasonal fluctuations, so the expression may contain periodic function components; and for some commodities that are greatly affected by market supply and demand, their expressions may be related to linear or nonlinear demand and supply models.
[0019] The average value of the selling price is obtained by means of integration. This average value reflects the average price level of key commodities over a period of time and has important reference value for enterprises to formulate reasonable pricing strategies. At the same time, we also need to obtain the selling price of key commodities on the current date. This data is real-time and directly reflects the actual amount that consumers need to pay for this commodity in the current market environment. By comparing the current price with the average value of the selling price, we can initially understand the price fluctuations and their potential impact on consumers' purchasing behavior.
[0020] Then the concept of price influencing factors is introduced. It is assumed that the higher the selling price of key commodities on the current date, the smaller the price influence coefficient. This setting is based on the general market consumption psychology and behavioral patterns. In real life, when the price of a commodity rises, consumers tend to be more cautious about their purchase decisions and their desire to buy will decrease accordingly. This is because the higher price may exceed the consumers' budget or make them think that the cost performance of this commodity has decreased, so they turn to look for other substitutes. On the contrary, when the price drops, more people will think it is more cost-effective to buy this commodity and their desire to buy will increase, which conforms to the objective market law.
[0021] To estimate the sales volume of key commodities more comprehensively and accurately, we also need to obtain the monthly page views of key commodities. The monthly page views are an important indicator to measure the degree of attention of a commodity, which reflects consumers' interest and potential demand for this commodity. By analyzing information such as website traffic data and user behavior logs, the number of page views of key commodity pages each month can be counted. Generally speaking, the higher the page views, the higher the exposure of this commodity in the market and the more potential consumers it attracts.
[0022] Based on obtaining the monthly page views, the sales volume ratio needs to be calculated next. The sales volume ratio refers to the ratio of the actual sales volume to the monthly page views, which represents the actual conversion rate of the process from users' attention to purchase. The calculation method is to divide the monthly sales volume of key commodities by the corresponding monthly page views to obtain a numerical value in decimal or percentage form. This value can intuitively reflect how many people among the users who view this commodity finally complete the purchase behavior.
[0023] Based on the conversion rate calculated above, the second expected quantity is further calculated. The second expected quantity is another estimate of the commodity sales volume under the current market environment, considering the actual conversion situation of users from attention to purchase.
[0024] Subsequently, the first expected sales volume and the second expected volume are averaged to obtain the comprehensive sales volume. The comprehensive sales volume comprehensively considers the historical sales pattern and the purchasing intention of users under the current market environment, and thus can more comprehensively reflect the expected sales situation of the commodity in the future for a period of time. Since it contains the expected information of two parts, it can better express the overall trend of the commodity sales volume.
[0025] To estimate the expected sales volume more accurately, the return probability of key commodities is calculated again. The return probability refers to the ratio of the number of returned commodities to the total sales volume, and can be obtained through statistical analysis of historical return data. Finally, the expected sales volume of key commodities is obtained by subtracting the expected return volume from the comprehensive sales volume.
[0026] In summary, the present invention realizes a consumption trend prediction method based on the CRM system, which can predict the future commodity sales situation more accurately and comprehensively, and helps enterprises formulate reasonable production plans, inventory management and marketing strategies. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] The present invention will be further described below with reference to the accompanying drawings.
[0028] Figure 1 It is a schematic flowchart of a consumption trend prediction method based on the CRM system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0029] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.
[0030] Please refer to Figure 1 As shown, the present invention is a consumption trend prediction method based on the CRM system, including the following steps: S1: Obtain the monthly sales volume S of the commodity, calculate the sales volume increase ratio ΔS = (S now - S last ) / S last , where S now represents the monthly sales volume of the commodity on the shopping platform in the current month, S last represents the monthly sales volume of the commodity on the shopping platform in the previous month, calculate the average value ΔS ave of the sales volume increase ratio, and mark the commodity with the average value ΔS ave greater than the preset growth threshold ΔS max as a key commodity, and calculate the first expected sales volume S1 next = S now*ΔS ave +S now ; S2: Obtain the curve of the sales price G of key commodities changing with the date. The expression of the curve is f(R), where R represents the date, and calculate the average value of the sales price , where R1 represents the opening date of the key commodity, R2 represents the current date, and obtain the sales price G of the key commodity on the current date now , calculate the price impact coefficient GY = G ave / G now ; S3: Obtain the monthly view volume L of key commodities, calculate the sales volume ratio K = S / L * 100%, and obtain the actual conversion volume SK = K * GY after correcting the sales volume ratio K; Obtain the total view volume L of key commodities in the current month now , calculate the second expected sales volume S2 next =L now *SK, calculate the comprehensive sales volume SZ = (S1 next +S2 next ) / 2; Obtain the return probability P = S t / S all , where S t represents the total return volume of key commodities, S all represents the total sales volume of key commodities, and calculate the expected return volume , where represents rounding up SZ * P, and calculate the expected sales volume SN of key commodities = SZ - T.
[0031] It should be noted that first, it is necessary to comprehensively and accurately obtain the monthly sales volume data of commodities. After obtaining the accurate monthly sales volume data of commodities, next, it is necessary to calculate the sales volume growth ratio of commodities. The sales volume growth ratio mentioned here specifically refers to the comparison relationship between the monthly sales volume increase value and the previous month's sales volume. After obtaining the sales volume growth ratio of each month, it is also necessary to further calculate the average value of these sales volume growth ratios. This average value reflects the average growth level of commodity sales volume within a certain period and is of great significance for evaluating the sales trend of commodities.
[0032] Based on the above calculation results, mark the commodities with the average value of the sales volume growth ratio greater than the preset growth threshold as key commodities. The preset growth threshold is determined comprehensively based on various factors such as market rules, industry characteristics, and the strategic goals of the enterprise. For example, in some industries with fierce competition and rapid market changes, the growth threshold may be set relatively low to promptly discover potential popular commodities; while in some relatively stable markets, the threshold may be appropriately increased.
[0033] For the identified key products, it is also necessary to calculate the first expected sales volume based on their sales volume in the current month. By analyzing the sales data of key products over a certain period of time, the patterns and trends of their sales changes are summarized, so as to make a preliminary prediction of future sales volume. However, it should be clear that although this first expected sales volume is obtained based on certain historical data and analysis methods, it is not completely accurate in reality. Because the market environment is complex and changeable, affected by various internal and external factors, such as the strategic adjustments of competitors, changes in the macroeconomic situation, and changes in consumer preferences. Therefore, the first expected sales volume is only a reference value and needs to be further optimized and adjusted through a series of subsequent steps to make it closer to the actual market situation and provide more reliable support for the enterprise's sales decisions.
[0034] After completing the above preliminary analysis of key products, next, it is necessary to further obtain the curve of the sales price of key products changing with the date. This process involves the collection and collation of a large amount of historical data, and it is necessary to extract the sales price information of each key product on different dates from the sales record system of the shopping platform. These data may be interfered by various factors, such as fluctuations in market supply and demand relations, price adjustments of competitors, changes in raw material costs, and changes in the enterprise's own marketing strategies. In order to ensure the accuracy and reliability of the data, these original data need to be carefully screened and cleaned to eliminate outliers and incorrect data, so as to obtain a precise curve that can truly reflect the change of the sales price of key products with the date.
[0035] After obtaining a high-quality curve of sales price changes, the next step is to obtain the corresponding expression of this curve. By analyzing the shape, trend, and key points of the curve, a function expression that best fits the curve is found. This expression should be able to accurately describe the quantitative relationship between the sales price and the date, providing a powerful tool for subsequent calculations and analyses.
[0036] For example, it may be found that the sales prices of some key products show regular seasonal fluctuations, so the expression may contain components of periodic functions; for some products that are greatly affected by market supply and demand relations, their expressions may be related to linear or non-linear demand-supply models.
[0037] The average selling price is obtained by means of integration. This average value reflects the average price level of key commodities over a period of time and has important reference value for enterprises to formulate reasonable pricing strategies. At the same time, we also need to obtain the selling price of key commodities on the current date. This data is real-time and directly reflects the actual amount that consumers need to pay for this commodity in the current market environment. By comparing the current price with the average selling price, we can initially understand the price fluctuations and their potential impact on consumers' purchasing behavior.
[0038] Then the concept of price influencing factors is introduced. It is assumed that the higher the selling price of key commodities on the current date, the smaller the price influence coefficient. This setting is based on the general market consumption psychology and behavioral patterns. In real life, when the price of a commodity rises, consumers tend to be more cautious about their purchase decisions and their desire to buy will decrease accordingly. This is because the higher price may exceed the consumers' budget or make them think that the cost performance of the commodity has decreased, so they turn to look for other substitutes. On the contrary, when the price drops, more people will think it is more cost-effective to buy this commodity and their desire to buy will increase, which conforms to the objective market law.
[0039] To more comprehensively and accurately estimate the sales volume of key commodities, we also need to obtain the monthly page views of key commodities. The monthly page views are an important indicator to measure the degree of attention of a commodity, which reflects consumers' interest and potential demand for this commodity. By analyzing information such as website traffic data and user behavior logs, the number of page views of key commodity pages can be counted each month. Generally speaking, the higher the page views, the higher the exposure of this commodity in the market and the more potential consumers it attracts.
[0040] Based on obtaining the monthly page views, the sales volume ratio is calculated next. The sales volume ratio refers to the ratio of the actual sales volume to the monthly page views, which represents the actual conversion rate of users from attention to purchase. The calculation method is to divide the monthly sales volume of key commodities by the corresponding monthly page views to obtain a value in decimal or percentage form. This value can intuitively reflect how many people among the users who view this commodity finally complete the purchase behavior.
[0041] Based on the conversion rate calculated above, the second expected quantity is further calculated. The second expected quantity is another estimate of the commodity sales volume under the current market environment, considering the actual conversion situation of users from attention to purchase.
[0042] Subsequently, the first expected sales volume and the second expected volume are averaged to obtain the comprehensive sales volume. The comprehensive sales volume comprehensively considers the historical sales pattern and the purchasing willingness of users in the current market environment, and thus can more comprehensively reflect the expected sales situation of the commodity in the future for a period of time. Since it contains the expected information of two parts, it can better express the overall trend of the commodity sales volume.
[0043] To estimate the expected sales volume more accurately, the return probability of key commodities is calculated again. The return probability refers to the ratio of the number of returned commodities to the total sales volume, which can be obtained through statistical analysis of historical return data. Finally, the expected sales volume of key commodities is obtained by subtracting the expected return volume from the comprehensive sales volume.
[0044] In another preferred embodiment of the present invention, another method for screening key commodities specifically includes: Obtain the collection volume C of the commodity in the shopping platform, calculate the standard deviation s of the monthly sales volume of the commodity. If the commodity collection volume C is greater than the monthly sales volume S and the standard deviation s is greater than the preset judgment value s sta , mark the corresponding commodity as a potential commodity, obtain the monthly sales volume of the potential commodity in the past 6 months. If the monthly sales volume shows a monotonically increasing trend with the increase of months, mark the corresponding potential commodity as a key commodity.
[0045] It is worth noting that during the process of commodity evaluation and screening, it is necessary to clarify that if a commodity does not meet the growth threshold, it does not simply mean that the commodity has no potential. The actual situation is often more complex. One possibility is that the commodity is in the initial stage of the product and has relatively low popularity. In such a case, its sales growth may not have been fully demonstrated, but this does not mean that it has no broad development space and market potential in the future.
[0046] The collection volume of the commodity is an important reference index. When consumers have strong interest in a certain commodity, they often choose to add the commodity to the collection list. The amount of the collection volume reflects the attention degree and potential demand of consumers for the commodity. At the same time, the change in the sales volume in the past six months is also one of the key factors. Through in-depth analysis of the sales volume data during this period, we can observe the sales trend of the commodity. If a commodity shows a steady upward trend in sales volume in the past six months, even if its current sales volume has not reached the set growth threshold, it may have great development potential. This may be because the commodity is in the market promotion period, and as the effects of brand promotion and word-of-mouth spread gradually appear, its sales volume is expected to continue to grow in the future.
[0047] Comprehensively considering the collection volume of the commodity and the change in the sales volume in the past six months can better screen out key commodities and avoid omissions. This comprehensive evaluation method not only focuses on the current performance of the commodity, but also pays more attention to its potential development power and market prospect.
[0048] In another preferred embodiment of the present invention, products with monthly sales volume S < S min are excluded and do not participate in the subsequent steps.
[0049] It can be understood that when the amount of data is too small, the results of prediction are often accompanied by great errors, which may have a serious negative impact on decision-making and business development.
[0050] In another preferred embodiment of the present invention, the method for obtaining the selling price G of key products specifically includes: Obtaining the number M of merchants selling key products and the inventory quantity KC of key products on the shopping platform, and calculating the selling price , where KC all represents the total inventory quantity of key products, KC m represents the inventory quantity of key products of the m-th merchant, and G m represents the selling price of key products of the m-th merchant.
[0051] It should be noted that the selling prices of different merchants may vary. To simplify the calculation, weights are calculated based on the inventory quantities of the merchants to obtain the selling price, achieving the purpose of simplifying the calculation.
[0052] In another preferred embodiment of the present invention, if R2 - R1 < R min , then stop calculating the average value G of the prices of subsequent products ave , where R min represents the preset minimum date.
[0053] It should be noted that here it is stated that products with a short release date are not included in the scope of prediction. Since the feedback on such products is not comprehensive due to their short market time, subsequent calculations are stopped.
[0054] In another preferred embodiment of the present invention, obtaining the browsing duration T of key products, presetting the minimum judgment duration T min , and not including browsing records with browsing duration T ≤ T min in the monthly browsing volume L.
[0055] It can be understood that the purpose here is to exclude browsing behaviors caused by accidental touches by users, thereby improving the accuracy of prediction.
[0056] In another preferred embodiment of the present invention, calculating the standard deviation s of the selling price of key products, and marking months with selling price G outside the interval [G ave - s, G ave + s] as special months, and special months do not participate in the calculation of the expected sales volume SN.
[0057] It is worth noting that, excluding errors, there may be significant price fluctuations when the product is first launched. Eliminate the data with a large fluctuation range to ensure the accuracy of the results.
[0058] In another preferred embodiment of the present invention, if the return probability P of the key product is ≥50%, it is marked as an unstable product, and the calculation of the expected sales volume of the unstable product is stopped until the return probability P < 50%.
[0059] It is worth noting that if the return probability of the product is too high, accurate prediction is impossible at this time. Suspend the prediction until the return probability P < 50% to ensure rigor.
[0060] The above has described in detail an embodiment of the present invention, but the content described is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the present invention application should still fall within the patent coverage scope of the present invention.
Claims
1. A consumption trend prediction method based on a CRM system, characterized in that: The following steps are involved: S1: Get the monthly sales volume S of the product and calculate the sales volume growth rate ΔS = (S now -S last ) / S last , where S now Represents the monthly sales volume of the goods on the shopping platform that month, S last Represents the monthly sales volume of the products on the shopping platform last month, and calculates the mean value of the sales volume increase ratio ΔS ave , the mean ΔS ave Greater than the preset growth threshold ΔS max The products are recorded as key products, and the first expected sales volume S1 of key products is calculated. next = S now *ΔS ave +S now ; S2: Obtain the change curve of the sales price G of the key commodity with the date. The expression of the change curve is f(R), where R represents the date. Calculate the mean of the sales price , where R1 represents the sales date of the key product, and R2 represents the current date. Get the sales price G of the key product on the current date now , calculate the price impact coefficient GY=G ave / G now ; S3: Get the monthly page views L of key products, calculate the sales volume share K=S / L*100%, and correct the sales volume share K to get the actual conversion volume SK=K*GY; Get the total number of views of the key products of the month L now , calculate the second expected sales volume S2 next =L now *SK, calculate the comprehensive sales volume SZ = (S1 next +S2 next ) / 2; Get the return probability of key products P=S t / S all , where S t Represents the total return amount of key products, S all Represents the total sales volume of key products and calculates the expected return volume ,in, It represents rounding up SZ*P to calculate the expected sales volume of the key product SN=SZ-T.
2. The consumption trend prediction method based on CRM system according to claim 1, characterized in that: In the step S1, another method for screening key commodities specifically includes: Get the collection volume C of the product on the shopping platform, calculate the standard deviation s of the monthly sales volume of the product, if the collection volume C of the product is greater than the monthly sales volume S and the standard deviation s is greater than the preset judgment value s sta , mark the corresponding product as a potential product, obtain the monthly sales volume of the potential product in the past 6 months, and if the monthly sales volume shows a monotonically increasing trend with the increase of months, mark the corresponding potential product as a key product.
3. The consumption trend prediction method based on CRM system according to claim 1 is characterized in that: In step S1, the monthly sales volume S<S min The products will be removed and will not participate in the subsequent steps.
4. The consumption trend prediction method based on CRM system according to claim 1 is characterized in that: In step S2, the method for obtaining the sales price G of the key commodity specifically includes: Get the number M of merchants selling key products on the shopping platform and the inventory KC of key products, and calculate the sales price , among which, KC all Represents the total inventory of key commodities, KC m represents the inventory of key products of each merchant at the mth level, G m Represents the sales price of the key products of each m-th merchant.
5. The consumption trend prediction method based on CRM system according to claim 1 is characterized in that: In step S2, if R2-R1<R min , then stop calculating the mean value G of subsequent commodity prices ave , where R min Represents the preset minimum date.
6. The consumption trend prediction method based on CRM system according to claim 1, characterized in that: In step S3, the browsing time T of the key product is obtained, and the minimum judgment time T is preset. min , set browsing time T≤T min The browsing history is not included in the monthly browsing volume L.
7. The consumption trend prediction method based on CRM system according to claim 1, characterized in that: Calculate the standard deviation s of the sales price of key commodities and set the sales price G in the interval [G ave -s, G ave +s] are marked as special months, which are not included in the calculation of expected sales volume SN.
8. The consumption trend prediction method based on CRM system according to claim 1 is characterized in that: In the step S3, if the return probability P of the key commodity is ≥ 50%, it is marked as an unstable commodity, and the calculation of the expected sales volume of the unstable commodity is stopped until the return probability P is less than 50%.