JP2008242531A - Sales landing prediction program and enclosure - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To search for the fluctuation range of predicted sales. <P>SOLUTION: A data storage part 22 receives and stores subject data to which an estimated sum and the probability of the future reception of the order of each subject are preliminarily input about a plurality of order reception scheduled subjects. A random number application part 28 applies random numbers to subjects included in the subject data. A decision part 32 decides whether to receive the order of the subjects by comparing the random numbers applied to each subject and the probability. A sales calculation part 36 calculates predicted sales by totaling the estimated sums of all the subjects whose order reception is decided to be possible. A trial control part 30 repeatedly executes trial configured of the application of random numbers and the decision of order reception. An analyzing part 40 calculates the probability distribution of the predicted sales based on the several numbers of trail results. When the order reception scheduled subjects are divided into a first group and a second group having mutual correlation concerning the success/failure of order reception, the random number application part 28 makes the random numbers applied to the subjects included in the first group and the second group have the same correlation as that between the groups. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for predicting overall sales when an order probability and sales at the time of order are given for a plurality of items.

Businesses, especially departments in charge of sales, need to forecast sales for the current and next fiscal years. In that case, not only a predicted value is given for each person in charge or business project, but also a predicted value for the entire department or the entire company is required for performance judgment and disclosure to investors. Conventionally, such sales forecasts are often calculated by simply summing the predicted values for each business case reported from the person in charge. For example, in the sales prediction device described in Patent Document 1, the sales prediction value is calculated by adding the sales calculated by the first and second sales calculation processing means.
JP 7-21115 A

  However, only a single predicted value can be obtained only by summing up individual sales as in Patent Document 1. From the standpoint of decision-makers, there is a demand for various quantitative analyzes by determining the probability distribution of sales forecast values.

  The present invention has been made in view of such a situation, and the purpose of the present invention is to manage an entire organization when managing a plurality of matters that determine whether or not sales can be acquired by success or failure of an order in a certain organization. The purpose is to provide a technology for predicting the sales that can be obtained.

  One embodiment of the present invention is a sales landing forecast program. This program has a retention function that receives and stores in a memory the case data in which the estimated amount and the probability of receiving each case in the future are input in advance for a plurality of scheduled orders belonging to the first group, and the items included in the item data Random number assignment function that assigns random numbers generated between the minimum and maximum probabilities of the probability and the random number assigned to each case and the probability are compared to determine whether or not the case can be ordered Judgment function, calculation function to calculate the estimated sales of the first group by summing up the estimated amounts of all cases determined to be accepted, and trial control function to repeatedly execute trials consisting of random number assignment and order determination And a distribution creation function for obtaining a probability distribution of the predicted sales of the first group based on a plurality of trial results. When the average and standard deviation of the probability distribution of the predicted sales are obtained for a plurality of scheduled orders belonging to the second group, the random number assignment function assigns one random number to the second group in each trial. The calculation function calculates the estimated amount of the second group corresponding to the random number given in each trial, assuming that the distribution of the predicted sales of the planned order received belonging to the second group is a normal distribution. The estimated sales of the two groups and the predicted sales of the first group are totaled to obtain the total predicted sales of the first group and the second group.

  According to this aspect, if only the average value and the standard deviation of the forecasted sales are given for the second group without the estimated amount and order probability data of the individual case, the simulation is performed together with the individual case of the first group. The total sales of the first group and the second group can be predicted. That is, it is not necessary to input data for individual cases for the second group.

  Another aspect of the present invention is also a sales landing prediction program. This program has a retention function that receives and stores in the memory the case data in which the estimated amount and the probability of receiving each case in the future are input in advance, and the probability for the cases included in the case data A random number assignment function that assigns a random number generated between the minimum and maximum values, a determination function that determines whether or not an order can be accepted by comparing the random number and probability assigned to each case, Based on the result of multiple trials, a calculation function that calculates the estimated sales by summing the estimated amounts of all projects that are judged to be possible, a trial control function that repeatedly executes trials consisting of assigning random numbers and judging orders The computer creates a distribution creation function that calculates the probability distribution of the predicted sales. When the planned order is divided into the first group and the second group that are correlated with each other regarding the success or failure of the order, the random number assigning function is applied to the first random number and the second group assigned to the matter included in the first group. The same correlation coefficient as that between groups is given between the second random numbers given to the included cases.

  According to this aspect, a random number pair having the same correlation coefficient as the correlation coefficient of the first group and the second group is generated, and one of the random number pairs is assigned to the cases included in the first group. The other is given to the cases included in the second group. By doing so, a more realistic simulation result can be obtained than when random numbers having no correlation are assigned to the cases included in the first group and the second group.

  It should be noted that any combination of the above-described components and a representation of the present invention by a method, apparatus, system, recording medium, and computer program are also effective as an aspect of the present invention.

  According to the present invention, it is possible to predict the sales amount that can be acquired in the entire organization when managing a plurality of matters that determine whether or not sales can be acquired depending on the success or failure of an order in a certain organization.

Embodiment 1 FIG.
One embodiment of the present invention is a sales landing prediction technique for predicting sales of an entire organization when an estimated amount and an order forecast are given for a plurality of cases handled by a certain organization.

  First, a conventional general landing prediction method will be described. As shown in FIG. 5, it is assumed that there is a table in which the order probability 96 and the estimated amount 94 are input for a plurality of business projects. Conventionally, an expected value is often obtained in order to predict sales at the end of the organization. That is, the expected order value 98 is calculated for all cases by multiplying the estimated amount by the order probability. Then, the expected sales amount is obtained by summing up these expected orders.

  However, this method only shows a single value for the forecasted sales, how much is expected to increase or decrease depending on the order status during the period, and what is the probability of achieving the target sales? I can't get any information. For decision-makers who need to manage the progress of multiple projects and predict performance, it is preferable to be able to forecast sales that include more detailed and useful information for decision-making.

  Therefore, in the present embodiment, the sales probability distribution is obtained by applying a well-known Monte Carlo simulation to sales prediction. As a result, detailed information such as the average value and standard deviation of the predicted sales can be obtained, and the decision maker can make a quantitative judgment.

  FIG. 1 is a configuration diagram of a sales landing prediction system 10 according to an embodiment of the present invention. In the following, in a division within a company that has multiple teams under its control, the person in charge of each team reports the sales forecast for the current period to the manager, and the manager aggregates the sales forecast and estimates the estimated sales for the current period. It will be explained on the assumption.

  The person-in-charge terminal 12 is a terminal for the person-in-charge to input item data such as an item name, an estimated amount, and an order probability that are expected to be received during the current term. The manager terminal 20 is a terminal used by a business manager or a sales officer.

  The person-in-charge terminal 12 and the manager terminal 20 are connected via a network 14 such as an in-house LAN (Local Area Network). The person-in-charge terminal 12 and the manager terminal 20 are, for example, personal computers, and include at least a display as a display device and input devices such as a keyboard and a mouse.

  The manager terminal 20 receives the case data from a plurality of person-in-charge terminals 12 and executes simulation to obtain the predicted sales. In FIG. 1, the manager terminal 20 is shown as a functional block diagram. These configurations can be realized in terms of hardware by a CPU, memory, or other LSI of an arbitrary computer, and can be realized in terms of software by a program loaded in the memory. Those skilled in the art will understand that these functional blocks can be realized in various forms by hardware only, software only, or a combination thereof.

  The data holding unit 22 receives and holds item data for a plurality of planned orders for orders from the person-in-charge terminal 12.

  FIG. 2 is a table showing an example of item data input by each person in charge. A column 62 represents a number assigned to each case. Columns 64 and 66 represent customer names and project names, respectively. A column 68 represents the probability of receiving an order with a predetermined accuracy code. The column 70 is filled with the name of the dependency destination when there is a dependency condition among a plurality of items. This dependency condition will be described later. A column 72 represents an estimated amount of each case. In FIG. 2, it can be seen that the X team expects orders for three projects X01 to X03.

  Apparently, even for a single project, a certain amount of money is certain to receive an order, but the amount that can be added may be unknown. For example, out of a total of 1 million yen, an order is surely expected for 500,000 yen, but it is unclear whether an order can be received beyond that. In such a case, it is preferable to input the order-accurate part and the other part as separate matters.

  FIG. 3 shows an example of the accuracy code. In this example, the probability that an order can be received is classified into six codes. “P” indicates 100%, that is, the order of the project is confirmed. “A”, “B”, “C”, and “D” indicate that orders can be received with a probability of 80%, 50%, 20%, and 5%, respectively. “X” indicates 0%, that is, the order of the project has failed. The person in charge selects an accuracy code that is considered to be the closest to the prospect of orders received for each case and enters it in the case data. Of course, the order probability may be directly input as a numerical value without using the accuracy code.

  Returning to FIG. 1, the tabulation unit 24 creates item list data by combining item data from a plurality of persons in charge held in the data holding unit 22.

  The simulation execution unit 26 receives the case list data from the totaling unit 24 and executes a simulation for obtaining the predicted sales using the Monte Carlo method. The Monte Carlo method is well known as a method for approximately obtaining the probability. For example, if a simulation is performed n times and an event occurs m times, the probability of the event occurring is approximated by m / n. A rough approximation is obtained when the number of trials is small, and a better approximation is obtained as the number of trials is increased.

  The random number generator 34 generates a random number between the minimum value and the maximum value of the order probability given by the person in charge, usually between 0 and less than 1. The accuracy of the Monte Carlo method is determined by the nature of the random number used, but in the present embodiment, a pseudo-random number is used. Instead, hardware that generates physical random numbers may be used.

  The random number assigning unit 28 assigns the random numbers generated by the random number generating unit 34 one by one to each case included in the case data. However, the random number assigning unit 28 operates so as to generate the same random number for both cases when the dependency condition between cases, that is, when an order of one case is a condition of another case order. To do.

  The determination unit 32 compares the random number assigned for each item by the random number assigning unit 28 with the order probability input in advance by the person in charge, and determines whether or not each item can be ordered. This determination is made based on the random number and the order probability. If the order probability is greater than or equal to a random number, it is determined that the corresponding case has been accepted. If the order probability is less than a random number, it is determined that the corresponding case has not been accepted. More specifically, for example, when a random number of 0.80 or less is assigned to a case with an accuracy code of A (80%), it is determined that the case has been accepted. If a random number greater than 0.80 is assigned, it is determined that the order has not been received. In this way, success and failure of orders are determined digitally for individual cases.

  The trial control unit 30 repeatedly executes trials consisting of assignment of random numbers and determination of orders for all cases. By repeating this trial many times, for example, 1000 times or more, the accuracy of the probability distribution becomes reliable.

  The sales calculation unit 36 sums the estimated amounts of all the cases determined to be accepted in each trial and obtains the predicted sales in the trial. The sales calculator 36 calculates the predicted sales for all trials.

  The distribution creation unit 42 creates a graph representing the probability distribution of the predicted sales in a plurality of trials using the predicted sales calculated by the sales calculation unit 36. In addition, the distribution creation unit 42 calculates numerical values such as the average value and standard deviation of the probability distribution, and the achievement probability of the current period target. The created graph is output so that it can be viewed on the display of the manager terminal 20 or the person-in-charge terminal 12.

FIG. 4 is a flowchart illustrating a sales landing forecast process according to one embodiment of the present invention.
First, each person in charge inputs an estimated amount and accuracy of an order scheduled item at the person in charge terminal 12 (S10). The totaling unit 24 of the manager terminal 20 collects the input item data and creates item list data including all information of the item scheduled to be received in the current period (S12). The random number assigning unit 28 assigns random numbers to all cases (S14). The determination unit 32 compares the random number assigned to each case with the accuracy, and determines that the order is successful if the accuracy is greater than or equal to the pseudo random number, and the order is failed if the accuracy is less than the pseudo random number (S16).

  The sales calculation unit 36 sums the estimated amounts of the cases that have been successfully received by the determination unit 32, and obtains the predicted sales (S18). The trial control unit 30 determines whether or not the processing from S14 to S18 has been repeated a predetermined number of trials (S20). If the predetermined number of times has not been reached (N in S20), the same processing is repeated. When the predetermined number of trials is reached (Y in S20), the predicted sales in all trials are totaled to obtain a probability distribution of the predicted sales (S22).

  Hereinafter, an analysis example according to the present embodiment will be described.

Analysis Example 1
FIG. 5 is a table showing an example of item list data aggregated by the aggregation unit 24. A column 82 represents the names of teams that handle each case. That is, the table 80 includes three teams of X, Y, and Z. Columns 84 to 94 correspond to columns 62 to 72 of the table 60 input by each person in charge. In FIG. 5, “Y01” and “Z01” are entered in the condition column 92. This indicates that the case Y02 can be received only when the order of the case Y01 is successful, and the case Z02 can be received only when the order of the case Z01 is successful.

FIG. 6 shows an example of a result when a simulation is performed on the case list data of FIG. In FIG. 6, only the team name, item number, accuracy, and estimated amount are extracted from the data in FIG. 5, and an “order probability” column 102 that represents the accuracy with a numerical value is added.
The results of trials executed a plurality of times on the case list data are represented as trials 1, 2, 3,. “Random number” columns 104, 108, and 112 are random numbers assigned to each case by the random number assigning unit 28. The “order sales” columns 106, 110, and 114 are estimated amounts of cases determined to be successful as a result of comparing the order probability and random numbers for each case by the determination unit 32. In each trial, the total order sales of the items determined to be successful orders are shown as predicted sales 118.

  As an example, the determination by the determination unit 32 will be described for the row 116 specified by the item number X02. The accuracy of this case is B, so the order probability is “0.5”. In the first trial, a random number “0.376” is assigned. The determination unit 32 compares the order probability “0.5” with the random number “0.376” and determines that the order has been received because the order probability is higher. As a result, the estimated amount “100” is counted as order sales. In the second trial, a random number “0.614” is assigned. This time, since the order probability is smaller than the random number, the determination unit 32 determines that this case cannot be ordered. Therefore, the order sales are counted as “0”. Thereafter, the same determination is repeated in each trial.

FIG. 7 is a graph showing the distribution of predicted sales when the trial is repeated 1000 times. As a result of the simulation, it is understood that the predicted sales have a spread close to a normal distribution.
FIG. 8A shows the average value and the standard deviation of the predicted sales amount at the time of enforcement 1000 times. Assuming that this distribution is a normal distribution, it can be seen that the average ± 1σ is in the range of 745 million yen to 1201 million yen. FIG. 8B shows the predicted sales 126 when the predicted sales are arranged in descending order and counted from the maximum value. The maximum value of the predicted sales is “1552 million yen”, the tenth value from the top is “1451 million yen”, and the minimum value is “296 million yen”. If the target sales of this department is 1000 million yen, it can be seen from the data shown in FIG. 7 that the achievement probability is 46.5%.

  As described above, when the sales are predicted according to the present embodiment, it is possible to quantitatively grasp the variation that the predicted sales can take, and it is also possible to obtain information such as how much the target sales can be achieved. .

Analysis Example 2
As described above, some orders are subject to dependency conditions. In other words, the second project can be ordered only when the first project has been successfully ordered. Specific examples include a new order and a maintenance contract related to the product. In this case, if an order cannot be received, a maintenance contract will not be concluded. In such a case, if only an order for a maintenance contract item succeeds and an order for a new item fails in each trial during the simulation, the trial becomes an impossible trial. Therefore, in the analysis example 2, a simulation in which the dependency condition is considered so that such a trial cannot occur will be described.

  FIG. 9 shows an example of a result when a simulation is performed in consideration of the dependency condition on the case list data of FIG. As shown in FIG. 5, the cases Y02 and Z02 are in a relationship depending on Y01 and Z01, respectively. Therefore, when the simulation is performed in consideration of the dependency condition, the trial results are different for Y02 and Z02. Portions having different results from those in FIG. 6 are shown surrounded by thick frames 152 and 154.

  The random number assigning unit 28 assigns the same random number as the random number assigned to the dependency destination to the case having the dependency relationship. Therefore, the random number assigned to the case Y02 is equal to the random number assigned to the case Y01, and the random number assigned to the case Z02 is equal to the random number assigned to the case Z01. As a result, the determination of the success / failure of the order by the determination unit 32 changes.

Specifically, in the second trial of the matter Y02, the determination unit 32 determines that the order has been successfully received for the matter Y02 as a result of the assigned random number being the same as Y01 and falling below the order probability 0.2. Therefore, the estimated amount “135” is recorded in the order sales column. On the contrary, in the third trial, as a result that the given random number is the same as Y01 and exceeds the order probability 0.2, the determination unit 32 determines that the order for the case Y02 has failed. Therefore, the estimated amount is not recorded in the order sales column.
Similarly, in the first trial of the matter Z02, the determination unit 32 determines that the order received for Z02 has been successful as a result of the assigned random number being the same as Y01. As described above, the determination result by the determination unit 32 is changed by considering the dependency condition. As a result, the predicted sales, which is the total of sales orders, also changes.

It will be understood that since the same random number is assigned to both cases that are dependent, there is no contradictory result that only the second case is ordered and the first case is not. Also in FIG. 9, it can be seen that when the order sales are recorded in Y02 or Z02, the order sales are also recorded in the dependence-destination case.
It should be noted that for both cases that are in the dependency relationship, the person in charge needs to set the accuracy so as to satisfy the relationship of (the accuracy of the first case)> (the accuracy of the second case). This is naturally understood because the second case cannot be ordered unless the first case can be ordered.

FIG. 10 is a graph showing the distribution of predicted sales when the trial is repeated 1000 times in consideration of the dependency condition. FIG. 11A shows the average value and standard deviation of this graph. FIG. 11B shows the predicted sales when the predicted sales are arranged in descending order and counted from the maximum value, divided into 164 when the dependency condition is not considered and 166 when the dependency condition is considered.
As shown in the figure, when the dependency condition is taken into account, the distribution changes from the case where the dependency condition is not taken into consideration. Compared to the case where the conditions are not taken into account, the average value is almost the same, but since the standard deviation is increased, it can be seen that the variation in the predicted sales is increased.
As described above, it is possible to derive a more realistic prediction result by executing the simulation in consideration of the dependency condition.

  As described above, according to the first embodiment, when sales of a plurality of projects are predicted, a probability distribution of predicted sales can be obtained instead of a single value such as an expected value. Therefore, it is possible to easily know the fluctuation range of the predicted sales, the probability that the sales target can be achieved, and the like. Furthermore, since it is possible to predict that it can be achieved with a probability of 80% if it is 500 million yen, and 10% if it is 800 million yen, announcements that are more in line with the organization's substance in financial forecasts etc. become able to. In addition, when the success or failure of an order for an item included in the list data is confirmed, the distribution of the predicted sales can be obtained again simply by changing the accuracy of the item. Can make predictions.

  Moreover, even when there are dependency conditions between a plurality of matters, it is possible to perform a simulation without performing exceptional processing only by performing an operation of giving them the same random number. By considering the dependency condition, it is possible to make a sales forecast that is more realistic.

Embodiment 2. FIG.
In the second embodiment, a method for predicting the total sales of a plurality of organizations using a probability distribution of predicted sales totaled in other organizations will be described.
As an example, it is assumed that there is a consolidated subsidiary to which sales are to be added in addition to the above three teams X to Z. A table 190 in FIG. 12 shows data related to the probability distribution of the predicted sales of the subsidiary. According to the table 190, there are an AA company and a BB company as the companies 192 to be consolidated, and an average 194 and a standard deviation 196 of respective predicted sales are obtained in advance. In addition, the probability distribution of the predicted sales in the AA company and the BB company can be obtained by the method described in the first embodiment.

  In addition to the X to Z teams, when trying to predict the overall sales including the sales of AA and BB companies, the probability distribution of the predicted sales is already required for AA and BB companies. It is more convenient to use this result. In the second embodiment, assuming that the forecast sales of the AA company and the BB company have a normal distribution, the detailed sales data about the forecast sales of the AA company and the BB company is not used. The simulation can be performed at the same level as the case of Z team.

  FIG. 13 is a table obtained by adding data of the AA company and the BB company to the table 100 of FIG. The AA company data is shown in row 202 and the BB company data is shown in row 204. However, the accuracy, the order probability, and the estimated amount are not entered, and an average 206 and a standard deviation 208 are entered instead.

In each trial, the random number assigning unit 28 assigns random numbers to the AA company and the BB company in the same manner as other cases. Then, instead of comparing the random number and the order probability by the determination unit 32, the sales calculation unit 36 calculates the order sales S of the AA company and the BB company by the following formula.
S = μ + N (rand) · σ (1)
Here, rand is a random number assigned by the random number assigning unit 28, and μ and σ are the average and standard deviation of the predicted sales distribution of the AA company or BB company, respectively. N () is an inverse function of the cumulative distribution function of the standard normal distribution having an average of 0 and a standard deviation of 1. According to the equation (1), random numbers of 0 or more and less than 1 can be corrected to a normal distribution with μ as an average and σ as a standard deviation.

  In each trial, the order sales 212, 216, 220 corresponding to the random numbers 210, 214, 218 are obtained using the formula (1), so that the distribution of the order sales of the AA company and the BB company after multiple executions is normal. Distribution.

  FIG. 14 shows the probability distribution of the predicted sales when the trial is repeated 1000 times. As shown in the figure, the distribution of the predicted sales of the normal case, that is, the case in charge of the X to Z team, and the distribution of the order sales of the AA company and the BB company have shapes close to normal distributions, respectively.

  The sales calculation unit 36 calculates the total predicted sales by adding up the predicted sales of the normal case and the order sales of the AA company and the BB company. According to FIG. 14, the overall predicted sales amount also has a normal distribution.

  FIG. 15 shows the average and standard deviations 232 to 238 of the probability distribution of the normal case, the AA company, the BB company, and the total predicted sales of these. From this, it can be seen that the average and standard deviation of the probability distributions of the AA company and the BB company are close to the average and standard deviation of the data initially given.

  As described above, according to the second embodiment, even if only the average and standard deviation of the probability distribution of the predicted sales are given for the consolidated subsidiaries, the total sales combined with the predicted sales of each item data High can be obtained by probability distribution. Therefore, it is not necessary to collect individual project data at a subsidiary and perform a simulation by the Monte Carlo method.

  If there are many sub-organizations and you want to calculate the forecast sales for all of them, instead of collecting all the project data that the sub-organizations are in charge and performing a simulation, the forecast sales for each sub-organization What is necessary is just to obtain | require the average value and standard deviation of a distribution, and add them later and the total estimated sales amount. In general, when each of independent random variables U and V follows a normal distribution, the temporary combination aU + bV (a and b are arbitrary real numbers other than 0) follows a normal distribution, but the second embodiment applies this theorem. It can also be said to be a thing.

Embodiment 3 FIG.
In the third embodiment, a prediction method will be described in the case where the cases for which the predicted sales are to be calculated have a correlation rather than an independent relationship.

  In general, the distribution in two groups is not completely independent in an analysis, and usually has some correlation. For example, if both department A and department B are in the same industry, for example, a sales department that deals with the financial industry, the sales of department A and department B should have a positive correlation. In addition, if department A and department B are in an industry where business results tend to be opposite, for example, sales departments dealing with the financial industry and distribution industry, respectively, sales of department A and department B are negative. There should be a correlation. In such a case, if a random number is simply assigned to the item in each department and compared with the order probability, the above correlation is not taken into account at all, and the predicted value becomes inaccurate.

  Therefore, in the third embodiment, a pair of random numbers having a correlation is generated and assigned to cases included in two groups having a correlation with each other.

  FIG. 16 is a diagram for explaining the assignment of random numbers when it is assumed that two teams for which sales are to be calculated, that is, the M team and the N team are dealing with a business in a correlated industry. It is assumed that it is known in advance that there is a relationship represented by the correlation coefficient ρ between the sales of the case handled by the M team and the sales of the case handled by the N team. The correlation coefficient ρ may be estimated from past sales data, or may be estimated from the linkage of business performance between industries.

In each trial, the random number assigning unit 28 assigns a random number U to the case in charge of the M team and assigns a random number V to the case in charge of the N team. The random number U and the random number V are random numbers that are correlated with each other. A random number (U, V) having a correlation coefficient ρ can be generated by the following equation.
U = X (2)
V = ρ · X + √ (1−ρ ² ) · Y (3)
However, X and Y are independent standard normal random numbers. Processing after assigning a random number to each case is the same as in the first embodiment. That is, for each item, the random number and the order probability are compared to determine the success or failure of the item.

  When the project data can be classified into two groups that are correlated with each other as in the case of different industries, a pair of random numbers having the same correlation coefficient as the correlation coefficient between the groups is generated. By giving to a group, the success or failure tendency of orders for each case in each trial becomes close to reality. As a result, a probability distribution of predicted sales more realistically can be obtained.

The present invention has been described based on some embodiments. It is understood by those skilled in the art that these embodiments are exemplifications, and that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are also within the scope of the present invention. By the way.
It should also be understood by those skilled in the art that the functions to be fulfilled by the constituent elements described in the claims are realized by the individual functional blocks shown in the present embodiment or their linkage.

  In the embodiment, the forecast of sales has been described. However, the present invention has a plurality of other items that can be digitally changed depending on whether or not a certain item has been achieved. It can also be applied when required. For example, the present invention can be applied to the forecast of the total sales of all departments and subsidiaries in the company and the forecast of the purchase cost.

1 is a configuration diagram of a sales landing prediction system according to an embodiment of the present invention. It is a figure which shows an example of the matter data input by each person in charge. It is a figure which shows an example of a probability code | symbol. 6 is a flowchart illustrating a sales landing forecast process according to an embodiment of the present invention. It is a figure which shows an example of item list data. It is a figure which shows an example when a simulation is implemented with respect to the case list data of FIG. It is a graph showing distribution of the predicted sales when a trial is repeated 1000 times. (A) shows the average value and standard deviation of the graph of FIG. 7, and (b) is a table showing the predicted sales when the predicted sales are arranged in descending order and counted from the maximum value. FIG. 6 is a diagram illustrating an example of a result when a simulation is performed on the item list data of FIG. 5 in consideration of a dependency condition. It is a graph showing distribution of the predicted sales when the trial is repeated 1000 times in consideration of the dependency condition. (A) shows the average value and standard deviation of the graph of FIG. 10, and (b) is a table showing the predicted sales when the predicted sales are arranged in descending order and counted from the maximum value. It is a figure which shows the result of the forecast sales distribution of the consolidated subsidiary which should calculate | require sales by adding together. It is the table which added the data of the subsidiary company of a connection object to the table of FIG. It is a figure which shows distribution of the predicted sales when a trial is repeated 1000 times. It is a table which shows the average value and standard deviation of the graph of FIG. It is a figure explaining the random number provided to each case, when there are two teams dealing with a mutually correlated case.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 Sales landing prediction system, 12 Person in charge terminal, 14 Network, 20 Manager terminal, 22 Data holding part, 24 Counting part, 26 Simulation execution part, 28 Random number assignment part, 30 Trial control part, 32 Judgment part, 34 Random number generation part 36 Sales calculation part, 42 Distribution creation part.

Claims (4)

A holding function for receiving the matter data in which the estimated amount and the probability of being able to accept each matter in the future are received in advance for a plurality of planned orders that belong to the first group,
A random number assignment function for assigning a random number generated between the minimum value and the maximum value of the probability for the case included in the case data;
A determination function that determines whether or not an order can be accepted by comparing the random number assigned to each case and the probability;
A calculation function that calculates the estimated sales of the first group by summing the estimated amounts of all the cases determined to be accepted,
Trial control function that repeatedly executes trials consisting of random number assignment and order determination;
Based on the results of a plurality of trials, a distribution creation function for obtaining a probability distribution of the predicted sales of the first group,
Is a sales landing forecasting program that makes computers perform
When the average and standard deviation of the probability distribution of the predicted sales are obtained for a plurality of planned orders that belong to the second group, the random number assigning function assigns one random number to the second group in each trial. ,
The calculation function calculates the estimated amount of the second group corresponding to the random number given in each trial, assuming that the distribution of the predicted sales amount of the planned sales order belonging to the second group is a normal distribution, A sales landing forecasting program characterized by summing up the estimated sales amount of the second group and the predicted sales amount of the first group to obtain the total predicted sales amount of the first group and the second group. A holding function for receiving the matter data in which the estimated amount and the probability of accepting each matter in the future are input in advance for a plurality of orders to be received and holding it in the memory;
A random number assignment function for assigning a random number generated between the minimum value and the maximum value of the probability for the case included in the case data;
A determination function that determines whether or not an order can be accepted by comparing the random number assigned to each case and the probability;
A calculation function that calculates the estimated sales by adding the estimated amounts of all the cases that are determined to be accepted,
Trial control function that repeatedly executes trials consisting of random number assignment and order determination;
A distribution creation function that calculates the probability distribution of forecast sales based on the results of multiple trials,
Is a sales landing forecasting program that makes computers perform
When the order-scheduled item is divided into a first group and a second group that are correlated with each other regarding the success or failure of the order, the random number assigning function includes a first random number and a second group that are assigned to the items included in the first group. A sales landing forecasting program characterized in that the same correlation coefficient as that between groups is given between the second random numbers given to the cases included in the item. A data holding unit that receives case data in which estimated amounts and probabilities of receiving orders in the future are input in advance for a plurality of scheduled orders belonging to the first group, and holds the data in a memory;
A random number assigning unit for assigning a random number generated between the minimum value and the maximum value of the probability for the case included in the case data;
A determination unit that determines whether or not the order can be accepted by comparing the random number assigned to each case and the probability;
A sales calculator that calculates the estimated sales of the first group by summing the estimated amounts of all the cases determined to be accepted;
A trial control unit that repeatedly executes trials including assignment of random numbers and determination of orders;
A distribution creation unit for obtaining a probability distribution of the predicted sales of the first group based on a plurality of trial results;
With
When the average and standard deviation of the probability distribution of the predicted sales are obtained for a plurality of scheduled orders belonging to the second group, the random number assigning unit assigns one random number to the second group in each trial. ,
The sales calculation unit calculates the estimated amount of the second group corresponding to the random number assigned in each trial, assuming that the distribution of the predicted sales of the planned sales order belonging to the second group is a normal distribution. Then, the estimated sales of the second group and the predicted sales of the first group are totaled to obtain the total predicted sales of the first group and the second group. A data holding unit that receives matter data in which an estimated amount and a probability of accepting each matter in the future are input in advance for a plurality of planned orders, and stores them in a memory;
A random number assigning unit for assigning a random number generated between the minimum value and the maximum value of the probability for the case included in the case data;
A determination unit that determines whether or not the order can be accepted by comparing the random number assigned to each case and the probability;
A sales calculation unit that calculates the estimated sales by adding the estimated amounts of all the cases determined to be accepted,
A trial control unit that repeatedly executes trials including assignment of random numbers and determination of orders;
A distribution creation unit that obtains a probability distribution of predicted sales based on the results of multiple trials;
With
When the order-scheduled case is divided into a first group and a second group that are correlated with each other regarding the success or failure of the order, the random number assigning unit includes a first random number and a second group that are assigned to the cases included in the first group The sales landing forecasting apparatus, wherein the same correlation coefficient as that between the groups is given between the second random numbers given to the cases included in.

Images