JP5170243B2 - Business support program, business support device, and business support method - Google Patents

Description

  The present invention relates to a business support program, a business support device, and a business support method for supporting a contact business of a contact center that receives an inquiry from a customer.

  Generally, a company is provided with a contact center that receives inquiries from customers such as questions, orders, and complaints. In the contact center, a plurality of operators who respond to inquiries from customers are arranged. The number of operators arranged in the contact center needs to be appropriately determined according to the number of inquiries from customers.

  This is because if the number of operators is small, the waiting time of the customer becomes long, leading to an increase in so-called “abandonment calls” that are abandoned before the operator responds. This makes it impossible to sufficiently respond to inquiries from customers, leading to problems such as reduced customer satisfaction and lost business opportunities.

  On the other hand, when the number of operators is excessive, not only the operator labor cost increases, but also costs such as system maintenance costs and electricity charges increase. As a result, it is necessary to take measures such as reviewing the working conditions and cutting labor costs for the purpose of cost reduction, and as a result, the morale of the operator is lowered.

  For this reason, it is necessary to determine the number of operators to be placed in the contact center based on the number of incoming calls from customers expected to occur in the future. However, the incoming call includes a re-call from the customer who made a call back because the operator could not answer. That is, a plurality of incoming calls may occur for one inquiry from a customer. This makes it difficult to accurately predict the number of incoming calls in the future.

  For this reason, there is a demand for accurately predicting the number of incoming calls in the future by eliminating the re-calls from the incoming calls. Therefore, a technique for estimating the number of customers who actually called (number of new calls) based on the average number of redials of the customer at the sales counter and the number of incoming calls from the customer is disclosed (for example, see Patent Document 1 below). .)

  Also, based on the history of incoming calls from past customers, the corrected number of incoming calls (number of new calls) is calculated for each past time unit excluding the incoming call due to a call back because the operator could not answer. Thus, a technique for predicting the number of incoming calls for each time unit based on the corrected number of incoming calls is disclosed (for example, see Patent Document 2 below).

JP 2007-189329 A JP 2006-254094 A

  However, in the above-described prior art, when the number of incoming calls is estimated after eliminating the re-calls from the incoming calls, the customer hangs up before the operator answers and makes an inquiry to the contact center again. The time interval between abandoned and re-calls is not considered.

  That is, the time interval between the abandoned call occurrence time and the recall occurrence time is ignored, and the recall portion is excluded from the incoming call. However, the time interval from an abandoned call to a re-call often varies depending on the time of occurrence of the abandoned call. For example, if an abandoned call occurs before going to work in the morning or just before the end of the lunch break, the call to the contact center tends to be postponed.

  Therefore, if the time interval between the abandoned call occurrence time and the re-call occurrence time is ignored, the number of incoming calls that will occur in the future cannot be accurately predicted, and it is still necessary to place them in the contact center. There was a problem that it was impossible to grasp the number of operators.

  In order to solve the above-described problems caused by the prior art, the present invention obtains the probability of a re-call from an abandoned call that occurs at an arbitrary time for each time zone that divides the business hours of the contact center. It is an object of the present invention to provide a business support program, a business support device, and a business support method that can accurately predict the number of incoming calls in each time zone.

  In order to solve the above-mentioned problems and achieve the object, the disclosed technology is selected from a plurality of time zones within the business hours of the contact center where the operator terminal that responds to the incoming call from the user terminal is installed via the exchange. To obtain an average time interval specific to any given time period until a re-call from an abandoned call in any given time period occurs, and replay from that abandoned call within the elapsed time since the abandoned call occurred By giving the obtained average time interval to the probability density function representing the probability of occurrence of a call, the probability of occurrence of a re-call from the abandoned call within the elapsed time since the abandoned call of the arbitrary time period occurred The probability density distribution that represents the number of recall calls from the abandoned call in the arbitrary time zone is used to determine the probability of occurrence of a recall call in the time zone. Calculated, and the calculated requirements to output the probability of re-calls per time zone.

  According to this disclosed technique, it is possible to obtain the probability of the occurrence of a re-call from an abandoned call that occurs at an arbitrary time for each time slot that divides the business hours of the contact center.

  According to the business support program, the business support device, and the business support method, by determining the probability of a re-call from an abandoned call that occurred at an arbitrary time for each time slot that divides the contact center business hours, It is possible to predict the number of incoming calls occurring in each time zone with high accuracy.

1 is a system configuration diagram of a contact center according to an embodiment. FIG. It is explanatory drawing showing the correlation of an abandonment call rate and a complete abandonment call rate. It is a block diagram which shows the hardware constitutions of the work assistance apparatus concerning embodiment. It is a block diagram which shows the functional structure of a work support apparatus. It is explanatory drawing which shows the memory content of a data table. It is explanatory drawing which shows the probability density distribution showing the generation | occurrence | production probability of a recall. It is explanatory drawing which shows the memory content of an occurrence probability table. It is explanatory drawing which shows the memory content of the number-of-recalls table. It is explanatory drawing which shows the memory content of the number table of complete abandonment calls. It is explanatory drawing which shows the memory content of a new call number table. It is explanatory drawing which shows the memory content of abandoned call number table. It is explanatory drawing which shows the memory content of the number of prediction abandonment calls. It is explanatory drawing which shows the memory content of a prediction result table. It is explanatory drawing which shows the memory content of a staff number table. It is explanatory drawing which shows the memory content of a shift table. It is a flowchart (the 1) which shows an example of a new call number calculation process sequence. It is a flowchart (the 2) which shows an example of a new call number calculation process sequence. It is a flowchart which shows an example of the specific process sequence of a recall occurrence probability calculation process. It is a flowchart (the 1) which shows an example of an incoming call number calculation processing procedure. It is a flowchart (the 2) which shows an example of an incoming call number calculation processing procedure. It is a flowchart which shows an example of an optimization assistance process procedure.

  Exemplary embodiments of the business support program, the business support apparatus, and the business support method will be described below in detail with reference to the accompanying drawings. In the business support program, the business support device, and the business support method, the number of re-calls generated from an abandoned call is changed from the abandoned call that changes depending on the abandoned call occurrence time to a re-call. Thus, the number of new calls at each time that divides the business hours of the contact center is accurately predicted.

(Contact center system configuration)
First, the system configuration of the contact center according to the embodiment will be described. FIG. 1 is a system configuration diagram of a contact center according to an embodiment. In FIG. 1, a contact center 100 includes a business support apparatus 101, user terminals 102-1 to 102-n, an exchange 103, and operator terminals 104-1 to 104-3, which are a telephone line, the Internet, and a LAN (Local). The network 110 is connected via a network 110 such as an area network (WAN) or a wide area network (WAN).

  The contact center 100 is a department in the company that receives inquiries from customers such as questions, orders, and complaints. In the contact center 100, an operator who responds to an inquiry from a customer is arranged. The operator responds to inquiries from customers using operator terminals 104-1 to 104-3 (three in FIG. 1).

  These operator terminals 104-1 to 104-3 and user terminals 102-1 to 102-n used by customers are connected via the exchange 103. The exchange 103 is a server having a CTI (Computer Telephony Integration) function for linking a telephone and a computer device.

  Here, the basic operation of the contact center 100 will be described. First, a customer who makes an inquiry makes a call to the contact center 100 using the user terminals 102-1 to 102-n (a fixed telephone, a mobile phone, etc.). At this time, the incoming call from the user terminals 102-1 to 102-n once enters the waiting queue of the exchange 103.

  Then, the exchange 103 detects the operator terminals 104-1 to 104-3 that can respond, and assigns the incoming call at the head of the waiting queue to the detected operator terminals 104-1 to 104-3 (corresponding call). At this time, when the number of incoming calls from the customer is larger than the number of operators, the customer waits until the responding operator terminals 104-1 to 104-3 are detected.

  During this waiting time (while in the waiting queue), an incoming call that is hung up because the customer cannot wait for the operator's response is an abandoned call. Some percent of such abandoned calls are recalled at some interval and re-enter the waiting queue of the exchange 103. On the other hand, the remaining several percent that did not go through the re-call are completely abandoned calls that give up inquiries to the contact center 100.

  In general, the quality of the contact center 100 is determined by the low abandoned call rate indicating the ratio of the number of abandoned calls to the number of incoming calls and the short waiting time of the customer. Therefore, it is necessary to maintain a certain abandoned call rate and waiting time by arranging an appropriate number of operators based on the number of incoming calls from customers in the contact center 100.

  The business support device 101 is a computer device that supports the reception business of the contact center 100 that receives an inquiry from a customer. Specifically, the number of incoming calls that will occur in the future is predicted based on the incoming call history from the customer to the contact center 100, and the appropriate number of operators to be placed in the contact center 100 is derived.

  At this time, the contact center 100 takes into consideration the recall rate that indicates the ratio of the abandoned call to the recall, which changes depending on the person's life pattern, and the time interval from the abandoned call to the occurrence of the recall. Accurately predict the number of incoming calls that occur during each time period that divides your business hours.

(Correlation between abandoned call rate and complete abandoned call rate)
Here, the correlation between the abandoned call rate indicating the ratio of the number of abandoned calls to the number of incoming calls and the complete abandoned call rate indicating the ratio of the number of completely abandoned calls to the number of new calls will be described. The new call is the first incoming call to the contact center 100 regarding a certain inquiry.

  FIG. 2 is an explanatory diagram showing the correlation between the abandoned call rate and the complete abandoned call rate. In FIG. 2, a graph 200 showing the correlation between the abandoned call rate and the complete abandoned call rate is displayed. Here, it is assumed that the number of operators arranged in the contact center 100 is 90 and there is a goal of reducing the abandoned call rate by 10%.

  The abandoned call rate can be lowered by increasing the number of operators. For this reason, to reduce the abandoned call rate by 10%, it seems necessary to increase the number of operators by at least 10% (9 people). However, as shown in the graph 200, for example, in order to reduce the abandoned call rate from 22% to 12%, the complete abandoned call rate may be reduced from 10% to 5%.

  The total abandonment call rate can be reduced from 10% to 5% by increasing the number of operators by 5 (an increase of 5.6%). For this reason, considering that the goal is cleared simply by paying attention to the abandoned call rate, four (= 9-5 people) surplus personnel are generated. This increases the cost for the contact center 100.

  That is, the substantial quality of the contact center 100 depends on the complete abandonment call rate. For this reason, in order to evaluate the substantial quality of the contact center 100 and to allocate an appropriate number of operators to the contact center 100, the complete abandonment call rate (total number of abandonment calls / number of new calls) is accurately determined. It is necessary to grasp.

  Therefore, in the present embodiment, the number of new calls and the number of completely abandoned calls for each time period within the business hours of the contact center 100 are considered in consideration of the re-call occurrence mechanism that changes depending on a person's life pattern and the like. Is proposed, and a method for deriving an appropriate number of operators to be placed in the contact center 100 is proposed.

(Hardware configuration of business support device)
FIG. 3 is a block diagram of a hardware configuration of the business support apparatus according to the embodiment. In FIG. 3, a business support apparatus 101 includes a CPU (Central Processing Unit) 301, a ROM (Read-Only Memory) 302, a RAM (Random Access Memory) 303, a magnetic disk drive 304, a magnetic disk 305, and an optical disk. A drive 306, an optical disk 307, a display 308, an I / F (Interface) 309, a keyboard 310, a mouse 311, a scanner 312, and a printer 313 are provided. Each component is connected by a bus 300.

  Here, the CPU 301 governs overall control of the business support apparatus 101. The ROM 302 stores a program such as a boot program. The RAM 303 is used as a work area for the CPU 301. The magnetic disk drive 304 controls the reading / writing of the data with respect to the magnetic disk 305 according to control of CPU301. The magnetic disk 305 stores data written under the control of the magnetic disk drive 304.

  The optical disk drive 306 controls the reading / writing of the data with respect to the optical disk 307 according to control of CPU301. The optical disk 307 stores data written under the control of the optical disk drive 306, and causes the computer to read data stored on the optical disk 307.

  The display 308 displays data such as a document, an image, and function information as well as a cursor, an icon, or a tool box. As this display 308, for example, a CRT, a TFT liquid crystal display, a plasma display, or the like can be adopted.

  An interface (hereinafter abbreviated as “I / F”) 309 is connected to a network 110 such as a LAN, a WAN, or the Internet through a communication line, and is connected to another device via the network 110. The I / F 309 controls an internal interface with the network 110 and controls data input / output from an external device. For example, a modem or a LAN adapter can be adopted as the I / F 309.

  The keyboard 310 includes keys for inputting characters, numbers, various instructions, and the like, and inputs data. Moreover, a touch panel type input pad or a numeric keypad may be used. The mouse 311 performs cursor movement, range selection, window movement, size change, and the like. A trackball or a joystick may be used as long as they have the same function as a pointing device.

  The scanner 312 optically reads an image and takes in the image data into the business support apparatus 101. The scanner 312 may have an OCR (Optical Character Reader) function. The printer 313 prints image data and document data. As the printer 313, for example, a laser printer or an ink jet printer can be employed.

(Functional configuration of business support device)
Next, a functional configuration of the business support apparatus 101 will be described. FIG. 4 is a block diagram illustrating a functional configuration of the business support apparatus. The business support apparatus 101 includes an acquisition unit 401, a specification unit 402, a calculation unit 403 (first calculation unit 405, second calculation unit 406, third calculation unit 407, fourth calculation unit 408, fifth The calculation unit 409, the sixth calculation unit 410), and the output unit 404 are included.

  Specifically, the functions (acquisition unit 401 to sixth calculation unit 410) serving as the control unit are stored in a storage area such as the ROM 302, the RAM 303, the magnetic disk 305, and the optical disk 307 illustrated in FIG. The function is realized by causing the CPU 301 to execute the program or by the I / F 309. Further, the connection destination function indicated by the arrow in FIG. 4 reads output data from the connection source function from the storage area and causes the CPU 301 to execute a program related to the function.

  The acquisition unit 401 includes a plurality of times within the business hours of the contact center 100 where the operator terminals 104-1 to 104-3 that respond to incoming calls from the user terminals 102-1 to 102-n via the exchange 103 are installed. It has a function of obtaining an average time interval unique to any time zone until a re-call from abandoned calls in any time zone selected from the time zone occurs.

  The business hours of the contact center 100 are reception times for receiving inquiries from customers. Further, the time zone within business hours is a time zone divided by dividing the business time of the contact center 100 at an arbitrary time interval. For example, it is assumed that the business hours of the contact center 100 are 24 hours per day, and the time zones are divided in units of one hour.

  In this case, each time zone from 0 o'clock to 23 o'clock is a time zone within business hours. The time zone in the 0:00 range is a time zone from 0:00 to 1:00 (excluding 1:00). Further, in this specification, the business hours of the contact center 100 are divided in units of one hour, but this time interval is arbitrary, and may be divided in units of 1 minute, 5 minutes, or 30 minutes, for example.

  Further, the abandoned call is an incoming call that is abandoned before an operator answers among incoming calls received by the exchange 103. The average time interval unique to an arbitrary time zone is an average value of each time interval until an abandoned call that occurs in an arbitrary time zone and a re-call from the abandoned call.

  For example, it is assumed that abandoned calls 1, 2, and 3 occur in the time zone of 0:00. At this time, the time interval until the recall 1 from the abandoned call 1 occurs is 100 [minutes], the time interval until the recall 2 from the abandoned call 2 occurs is 150 [minutes], It is assumed that the time interval until the recall 3 occurs is 200 [minutes]. In this case, the average time interval peculiar to the time zone of 0 o'clock is “(100 + 150 + 200) / 3 = 150 [minutes]”.

  The average time interval peculiar to each time zone is calculated based on, for example, a log recording the incoming history from the user terminals 102-1 to 102-n accumulated in the exchange 103. This log includes the arrival time of the incoming call received by the exchange 103, whether or not the operator has responded to the incoming call, the operator connection time of the corresponding call (abandoned time in the case of an abandoned call), the originator When the telephone number is notified, the telephone number is recorded.

  For example, the average time interval of re-calls from abandoned calls that occurred in each time zone can be calculated using the caller's telephone number stored in the log accumulated in the exchange 103 as a clue. In addition, the “recall rate” for each time zone, which will be described later, can be calculated using the telephone number of the caller as a key. Note that the calculation method for calculating various information from the telephone number of the caller is a known technique, and therefore description thereof is omitted here (for example, see Patent Document 2 described above).

  Here, the average time interval specific to an arbitrary time zone may be directly input to the job support apparatus 101 or may be acquired from an external computer apparatus. Furthermore, it is good also as calculating the average time interval peculiar to arbitrary time zones by acquiring the log accumulated by accessing the exchange 103.

  Hereinafter, in this specification, an arbitrary time zone i (i = 0, 1,..., 23) is treated as the time i. For example, the occurrence times of abandoned calls that occurred in the 7 o'clock time zone (7:00 to 8:00 (excluding 8:00)) are all 7 o'clock. Further, a time zone j (j = 0, 1,..., 23) of a re-call generated from an abandoned call at time i is handled as time j. For example, it is assumed that all the occurrence times of the re-calls occurring in the 8 o'clock time zone (8:00 to 9:00 (not including 9:00)) are 8 o'clock.

  Here, a data table input to the business support apparatus 101 will be described. FIG. 5 is an explanatory diagram showing the contents stored in the data table. In FIG. 5, the data table 500 includes the number of incoming calls, the abandoned call rate, the re-call rate, and the average time interval for each time i within the business hours of the contact center 100.

  The number of incoming calls C (i) is the number of incoming calls from the user terminals 102-1 to 102-n received by the exchange 103 at time i. For example, if 100 incoming calls are received by the exchange 103 at 0:00, the number of incoming calls C (0) at 0:00 is “C (0) = 100”.

  The abandoned call rate A (i) is a ratio of the abandoned calls abandoned before the operator's response among the incoming calls at time i. For example, when 100 incoming calls are received by the exchange 103 at 0 o'clock and 10 of them are abandoned calls, the abandoned call rate A (0) at 0 o'clock is “A (0) = 10/100 = 0. .1 ”.

  The recall rate R (i) is the ratio of abandoned calls that occurred at time i to re-calls. For example, if 10 abandoned calls occur at 0 and 5 of them are re-called, the re-call rate R (0) at 0 is “R (0) = 5/10 = 0.5”. Become. The average time interval D (i) is the average time interval of recalls from abandoned calls that occurred at time i.

  This data table 500 is input to the work support apparatus 101 by the user operating the keyboard 310 and the mouse 311 shown in FIG. The input data table 500 is stored in a storage area such as the ROM 302, the RAM 303, the magnetic disk 305, and the optical disk 307. In this case, the acquisition unit 401 reads an average time interval specific to any time i from the data table 500 stored in the storage area.

  The specifying unit 402 has a probability density function representing a probability of occurrence of a re-call from the abandoned call in an elapsed time since the abandoned call abandoned before the response of the operator terminals 104-1 to 104-3. Probability density that represents the probability of the occurrence of a re-call from the abandoned call within a specific elapsed time from the occurrence of the abandoned call at time i by giving the average time interval specific to time i acquired by the acquiring unit 401 Has the function of specifying the distribution.

  Here, the probability of occurrence of a recall is a rate of occurrence of a recall within a specific elapsed time of a recall that occurs from an abandoned call. The probability of occurrence of a recall varies in inverse proportion to the approximately 1.5th power of the time interval between the occurrence of an abandoned call and the occurrence of a recall.

  Specifically, for example, the following equation (1) can be used as a probability density function representing the probability of occurrence of a recall. The following equation (1) is a function that approximates the occurrence probability of a recall that changes inversely proportional to the 1.5th power of the time interval between the occurrence time of the abandoned call and the occurrence time of the recall. Where Q (u, v) is the probability of occurrence of recall, u is a parameter specified from the time of occurrence of abandoned call, v is a parameter specified from the time of occurrence of recall, and D (i) is specific to time i Is the average time interval between recalls.

  In the above equation (1), a (D (i)) is a function for defining the minimum value of the time interval between the abandoned call occurrence time i and the recall occurrence time j. 2). Where γ is a constant.

    a (D (i)) = γ × D (i) (2)

  The constant γ in the above equation (2) can be obtained by regression analysis based on the recall value for each contact center 100. For example, when the seriousness of the inquiry to the contact center 100 is high, and even if it is an abandoned call, it is immediately re-called (for example, pre-release ticket reservation), the value of the constant γ is small (for example, Γ = 0.04).

  The adjustment term (vu, D (i)) in the above equation (1) is a function for defining the characteristics of the contact center 100, and can be expressed by, for example, the following equation (3). . However, α and β are constants.

Adjustment term (v−u, D (i)) = {1−exp (−α × (v−u) / D (i))}
Xexp (−β × (v−u) / D (i))
... (3)

  The constants α and β in the above formula (3) can be obtained by regression analysis based on the recall value for each contact center 100. The constant α changes according to the severity of the inquiry to the contact center 100. For example, if the severity is high and the call is abandoned frequently, it is often close to a re-call. On the other hand, the constant β is almost the same value (for example, β = 0.04) in all the contact centers 100.

  Further, the normalization term (D (i)) in the above formula (1) is a function for setting the sum of all regions of the probability density distribution representing the probability of recall to “1 (100%)”. .

  In this case, the identification unit 402 substitutes the average time interval D (i) specific to any time i acquired by the acquisition unit 401 into the above formula (1), so that the abandonment call at the time i occurs. It is possible to specify a probability density distribution representing the probability of the occurrence of a re-call from the abandoned call in the elapsed time.

  The probability density function (for example, the above equation (1)) may be stored in a storage area such as the ROM 302, RAM 303, magnetic disk 305, and optical disk 307 shown in FIG. It is good also as acquiring from.

  Here, the probability density distribution representing the probability of occurrence of a recall will be described. FIG. 6 is an explanatory diagram showing a probability density distribution representing the probability of occurrence of a recall. In FIG. 6, a probability density distribution 600 is an example of a graph representing the probability of a recall from an abandoned call that occurred at a certain time.

  In the probability density distribution 600, the occurrence probability of a recall, the time interval [minutes] between the occurrence time of an abandoned call and the occurrence time of a recall from the abandoned call are expressed in association with each other. For example, the probability of a re-call occurring in one minute 60 minutes after the occurrence time of the abandoned call is “6 × 10 −4”. The probability density distribution 600 is a log-log graph in which both vertical and horizontal axes are logarithmic scales.

  The first calculation unit 405 uses the probability density distribution specified by the specification unit 402 to calculate the occurrence probability of a re-call from an abandoned call at an arbitrary time i for each time within the business hours of the contact center 100. It has a function. Specifically, for example, the occurrence probability of a recall at each time is obtained by obtaining the occurrence probability of a recall within the elapsed time from the abandoned call that occurred at time i using the above formula (1). .

  Here, an outline of a calculation process for calculating the probability of occurrence of a recall will be described. Here, the probability density distribution specified by the specifying unit 402 will be described as the probability density distribution 600 shown in FIG. For example, it is assumed that the probability of occurrence of a re-call that occurred at 1 o'clock after 60 minutes has elapsed among re-calls from an abandoned call that occurred at 0 o'clock.

  First, the time interval X (X = 1-0 = 1 hour until the call is re-required, which is obtained by subtracting the call release time “0” from the re-call occurrence time “1”, is negative if the day is crossed. In this case, 24 is added to correct the time interval). Next, the time interval on the graph shown in FIG. 6 is between “X” and “X + 1” using an equation in which the average time interval D (0) peculiar to 0:00 is substituted into the above equation (1). The area S1 of the probability density distribution 600 is obtained.

  That is, the area S1 of the probability density distribution 600 between 60 minutes (= 1 hour) after the abandoned call occurrence time and 120 minutes (= 2 hours) has elapsed. Note that “1” to be added to the time interval X depends on the time unit for dividing the business time of the contact center 100, and is “1” because it is divided in units of one hour here.

  Thereafter, the minute interval ΔX is added to obtain “X = X + ΔX”, and the area S2 of the probability density distribution 600 between the time intervals “X” to “X + 1” is obtained as described above. Note that the minute interval ΔX (1 minute to 59 minutes) can be arbitrarily set. Here, it is assumed that ΔX is set to “20 minutes”.

  In this case, the area S2 of the probability density distribution 600 between 80 minutes after the abandoned call occurrence time and 140 minutes later is obtained. Then, until the time interval X becomes equal to or greater than the value obtained by subtracting the abandoned call occurrence time from the recall occurrence time and adding “1”, the minute interval ΔX is added, and the area of the probability density distribution 600 is repeated. Ask.

  Here, when the time interval X is 120 minutes (= 2 hours), it is equal to or more than the value obtained by subtracting the occurrence time “0” of the abandoned call from the occurrence time “1” of the recall and adding “1”. When the area S3 of the probability density distribution 600 from 100 minutes to 120 minutes after the abandonment call occurrence time is obtained, the calculation of the area ends. Thereafter, an average value of a total of three areas S1, S2, and S3 is obtained. Then, this average value is set as the probability of occurrence of a re-call occurring at 1 o'clock among the re-calls generated from the abandoned call at 0 o'clock.

  The average value of the area described above, that is, the probability of occurrence of a recall at time j that occurs from an abandoned call at time i can be approximately calculated using the following equation (4). However, the occurrence probability of a re-call at time j that occurs from an abandoned call at time i is denoted as P (i, j, D (i)).

  Here, a case where the occurrence probability P (0, 1, D (0)) at 1 o'clock when a re-call from an abandoned call occurring at 0 o'clock occurs will be described. First, i = 0, j = 1, D (i) = D (0) is substituted into the above equation (4) to determine the integration range of the parameters u and v. Then, the recall probability P (0, 1, D (0)) is obtained by calculating the integral of the parameters u and v for each integration range. Similarly, the occurrence probability of a recall at each time from an abandoned call at 0:00 (i = 0) is obtained by obtaining the occurrence probability of a recall at another time (j = 0, 2 to 23). Can do.

  Various processes by the acquisition unit 401, the specifying unit 402, and the first calculation unit 405 are repeatedly executed until there is no unselected time selected from a plurality of times (here, 0:00 to 23:00), for example. The Here, an occurrence probability table that stores, for each time, the occurrence probability of a re-call from an abandoned call that occurred at each time within the business hours of the contact center 100 calculated by the first calculation unit 405 will be described.

  FIG. 7 is an explanatory diagram showing the contents stored in the occurrence probability table. In FIG. 7, the occurrence probability table 700 shows the occurrence probability P (i, j, D (i)) of the re-call from the abandoned call for each occurrence time j of the abandoned call. Is remembered. For example, the probability of occurrence of a recall from an abandoned call that occurred at 1 o'clock at 23:00 is P (1, 23, D (1)).

  The output unit 404 has a function of outputting the occurrence probability of a recall at time j from an abandoned call that occurred at an arbitrary time i calculated by the first calculation unit 405. Specifically, the occurrence probability table 700 shown in FIG. 7 is output. As a result, the probability of a re-call from an abandoned call that occurred at an arbitrary time can be recognized for each time that divides the business hours of the contact center 100.

  The output format by the output unit 404 includes, for example, display on the display 308, print output to the printer 313, and transmission to an external device by the I / F 309. Alternatively, the data may be stored in a storage area such as the RAM 303, the magnetic disk 305, and the optical disk 307.

  In addition, the acquisition unit 401 has a function of acquiring the number of abandoned calls that occurred at an arbitrary time i and the re-call rate specific to the time i that indicates the rate of occurrence of re-calls from abandoned calls. The number of abandoned calls at an arbitrary time i and the recall rate specific to the time i may be directly input to the business support device 101, for example, or acquired from an external computer device (for example, the exchange 103). Also good.

  In addition, the number of abandoned calls that occurred at time i can be obtained by, for example, multiplying the number of incoming calls that occurred at time i by the abandoned call rate specific to time i. More specifically, the number of incoming calls C (i) at time i and the abandoned call rate A (i) are read from the data table 500 shown in FIG. Ask.

  For example, when obtaining the number of abandoned calls at 0 o'clock, the number of incoming calls C (0) and the abandoned call rate A (0) are read from the data table 500 and multiplied to obtain the number of abandoned calls “AC ( 0) = C (0) × A (0) ”. However, the number of abandoned calls at time i is AC (i).

  The recall rate specific to time i may be read from the data table 500. For example, when acquiring a recall rate specific to 0 o'clock, the recall rate R (0) is read from the data table 500.

  The second calculation unit 406 calculates the number of abandoned calls AC (i) at the time i acquired by the acquisition unit 401, the recall rate R (i) specific to the time i, and the first calculation unit 405. Based on the occurrence probability P (i, j, D (i)) of the re-call at time j, the number of re-calls generated from the abandoned call at time i is calculated for each time.

  Specifically, first, the total number of re-calls generated from abandoned calls is obtained by multiplying the number of abandoned calls AC (i) and the recall rate R (i). Hereinafter, RC (i) is the total number of recalls that occur from abandoned calls at time i. For example, the total number of recalls RC (0) generated from an abandoned call at 0:00 is the product of the number of abandoned calls AC (0) and the recall rate R (0) [RC (0) = AC (0) × R (0) ”.

  Thereafter, by multiplying the total number RC (i) of the recalls by the occurrence probability P (i, j, D (i)) of the time j of the recall, the time j generated from the abandoned call at the time i Calculate the number of recalls. Hereinafter, the number of re-calls at time j generated from abandoned calls at time i is R (i, j).

  For example, the number of recalls RC (0,1) at one hour of a recall that occurs from an abandoned call at 0:00 is the total number of recalls RC (0) and the probability of occurrence of a recall P (0,1, D ( 0)) and “RC (0,1) = RC (0) × P (0,1, D (0))”.

  Here, the recall number table that stores the calculation result calculated by the second calculation unit 406 will be described. FIG. 8 is an explanatory diagram showing the stored contents of the recall number table. In FIG. 8, the number-of-recalls table 800 stores, for each abandoned call occurrence time i, the number of recalls RC (i, j) at the occurrence time j of the recall generated from the abandoned call. For example, the number of 0:00 recalls from abandoned calls that occurred at 23:00 is RC (23,0).

  The output unit 404 has a function of outputting the number of re-calls RC (i, j) at time j generated from the abandoned call at time i calculated by the second calculation unit 406. Specifically, the recall number table 800 shown in FIG. 8 is output. As a result, the number of re-calls from abandoned calls that occurred at an arbitrary time can be recognized for each time that divides the business hours of the contact center 100.

  The third calculation unit 407 uses the number of abandoned calls AC (i) at time i and the recall rate R (i) to calculate the number of completely abandoned calls that did not satisfy the recall among the abandoned calls at time i. It has a function. Specifically, the number of completely abandoned calls PAC (i) is obtained by multiplying the number of abandoned calls AC (i) at time i by {1-R (i)}. Hereinafter, the number of completely abandoned calls at time i is PAC (i). For example, the number of completely abandoned calls at 0 o'clock PAC (0) can be obtained from the number of calls abandoned at 0 o'clock AC (i) by “AC (i) × (1−R (i))”.

  Here, a complete abandoned call count table that stores the total abandoned call count calculated by the third calculation unit 407 will be described. FIG. 9 is an explanatory diagram showing the contents stored in the complete abandoned call count table. In FIG. 9, the total abandoned call count table 900 stores the number of completely abandoned calls that occur at each time within the business hours of the contact center 100. For example, the number of completely abandoned calls occurring at 23:00 is PAC (23).

  The output unit 404 has a function of outputting the number of complete abandoned calls PAC (i) at the time i calculated by the third calculation unit 407. Specifically, the complete abandoned call number table 900 shown in FIG. 9 is output. This makes it possible to recognize the number of completely abandoned calls that occurred at an arbitrary time.

  The fourth calculation unit 408 has a function of calculating the number of new calls that occurred at time i by subtracting the re-call that occurs at time i from the abandoned call at each time from the incoming call at time i. Specifically, first, referring to the data table 500, the number of incoming calls C (i) at the time i, and the number of re-calls RC generated at the time i from the abandoned call at each time j with reference to the data table 800. The number of new calls NC (i) at time i is calculated using (j, i). More specifically, the number of new calls at time i can be obtained using the following equation (5). However, NC (i) is the number of new calls at time i.

NC (i) = C (i) − (RC (0, i) + RC (1, i) +... + RC (23, i))
... (5)

  Here, a new call number table that stores the number of new calls at each time will be described. FIG. 10 is an explanatory diagram showing the contents stored in the new call count table. In FIG. 10, the new call count table 1000 stores the number of new calls that occur at the time for each time within the business hours of the contact center 100. For example, the number of new calls that occur at 23:00 is NC (23).

  The output unit 404 has a function of outputting the number of new calls NC (i) at the time i calculated by the fourth calculation unit 408. Specifically, the new call number table 1000 shown in FIG. 10 is output. This makes it possible to recognize the number of new calls that have occurred at an arbitrary time.

  The number of new calls is determined by subtracting the number of recalled calls from incoming calls taking into account the recall rate at each time and the time interval between the abandoned call occurrence time i and the recall occurrence time j. Is.

  Next, a method for predicting the future number of incoming calls that occur at each time using the number of new calls at each time and the target abandoned call rate that should be satisfied at each time will be described. This number of incoming calls includes the number of abandoned calls generated from a re-call from an abandoned call at each time when a target number of abandoned calls occurs as an upper limit.

  First, the acquisition unit 401 sets a target value of the response rate of the operator terminals 104-1 to 104-3 for an incoming call at time i by a ratio of the number of abandoned calls to the number of incoming calls at time i. It has a function to obtain the abandoned call rate. This target abandonment call rate may be directly input to the business support apparatus 101 or may be acquired from an external computer apparatus. Hereinafter, TA (i) is the target abandoned call rate specific to time i.

  In addition, the third calculation unit 407 includes the target abandoned call rate TA (i) specific to the time i acquired by the acquisition unit 401 and the time generated from the abandoned call at each time calculated by the second calculation unit 406. Prediction of abandoned calls that are predicted to occur with respect to the number of incoming calls at time i using the number of re-calls of i and the number of new calls NC (i) at time i calculated by the fourth calculation unit 408 It has a function to calculate the number of abandoned calls. Note that the number of re-calls at the time i generated from the abandoned call at each time is “RC (0, i) + RC (1, i) +... + RC (23, i)”.

  Specifically, first, the target abandoned call rate at time i is obtained by multiplying the target abandoned call rate TA (i) specific to time i by the number of new calls NC (i) at time i. Hereinafter, the target number of abandoned calls at time i is TAC (i).

  Next, the number of abandoned calls at time j generated from the re-call of the abandoned call that occurred at time i is obtained. Specifically, the number of abandoned calls at time j generated from the recall can be obtained using the following equation (6). However, RAC (i, j) is the number of abandoned calls at time j generated from the re-call of the abandoned call that occurred at time i.

RAC (i, j) = TAC (i) × R (i) × P (i, j, D (i))
÷ {1-R (i) × P (i, j, D (i))} (6)

  Here, the abandoned call count table that stores the number of abandoned calls RAC (i, j) at time j generated from the re-call of the abandoned call that occurred at time i will be described. FIG. 11 is an explanatory diagram showing the contents stored in the abandoned call count table. In FIG. 11, the abandoned call count table 1100 stores the abandoned call count RAC (i, j) at time j generated from the re-call of the abandoned call for each abandoned call occurrence time i.

  For example, the number of abandoned calls at 0:00 resulting from a re-call of an abandoned call that occurred at 0:00 is RAC (0, 0). The abandoned call number table 1100 is realized by storage areas such as the RAM 303, the magnetic disk 305, and the optical disk 307 shown in FIG.

  Thereafter, the call is generated at time i by using the target call abandoned call number TAC (i) at time i and the number of abandoned calls RAC (i, j) at time j generated from the re-call of the call abandoned at time i. Determine the number of predicted abandoned calls that are expected. Hereinafter, the number of predicted abandoned calls at time i is JAC (i).

  For example, the predicted number of abandoned calls JAC (i) at 0:00 is the sum of the target number of abandoned calls TAC (0) and the number of abandoned calls at 0:00 generated from the re-calls of abandoned calls generated at each time [RAC (0 , 0) + RAC (1, 0) +... + RAC (23, 0) ”.

  Here, a predicted abandoned call number table that stores the calculation result calculated by the third calculation unit 407 will be described. FIG. 12 is an explanatory diagram showing the stored contents of the predicted abandoned call count table. In FIG. 12, the predicted abandoned call count table 1200 stores the number of predicted abandoned calls that are predicted to occur at each time within the business hours of the contact center 100. For example, the predicted number of abandoned calls expected to occur at 23:00 is JAC (23).

  The output unit 404 has a function of outputting the number of predicted abandoned calls at time i calculated by the third calculation unit 407. Specifically, the predicted abandoned call number table 1200 shown in FIG. 12 is output. Thereby, it is possible to recognize the number of predicted abandoned calls that are predicted to occur at an arbitrary time.

  The fifth calculation unit 409 includes the number of new calls NC (i) at time i calculated by the fourth calculation unit 408 and the number of predicted abandoned calls JAC (i) at time i calculated by the third calculation unit 407. ) To calculate the predicted number of incoming calls for which an incoming call is predicted at time i.

  Specifically, by adding the number of new calls NC (i) at time i and the predicted number of calls abandoned at time i JAC (i), the prediction of an incoming call that is predicted to arrive at time i is predicted. Find the number of incoming calls. Hereinafter, EC (i) is the number of predicted incoming calls at time i. For example, the predicted number of incoming calls EC (0) at 0:00 is “EC (0) = NC (0) + JAC (0)”.

  Also, the first calculation unit 405 includes the predicted number of abandoned calls JAC (i) at time i calculated by the third calculation unit 407 and the predicted number of incoming calls at time i calculated by the fifth calculation unit 409. EC (i) is used to calculate a predicted abandoned call rate indicating the ratio of abandoned calls to incoming calls at time i.

  Specifically, the predicted abandoned call rate at time i is obtained by dividing the predicted number of abandoned calls JAC (i) at time i by the estimated number of incoming calls EC (i) at time i. Hereinafter, the predicted abandonment call rate at time i is EA (i). For example, the predicted abandonment call rate EA (0) at 0:00 is “EA (0) = JAC (0) / EC (0)”.

  Here, the predicted number of incoming calls EC (i) at time i calculated by the fifth calculation unit 409 and the predicted abandoned call rate EA (i) at time i calculated by the first calculation unit 405 are stored. The prediction result table to be described will be described. FIG. 13 is an explanatory diagram of the storage contents of the prediction result table.

  In FIG. 13, the predicted result table 1300 stores the predicted number of incoming calls and the predicted abandoned call rate for each time within the business hours of the contact center 100. For example, the predicted number of incoming calls that are predicted to arrive at 0 is EC (0), and the predicted abandoned call rate for abandoned calls is EA (0).

  Further, the output unit 404 outputs the predicted number of incoming calls EC (i) at the time i calculated by the fifth calculation unit 409 and the predicted abandoned call rate EA (i) at the time i calculated by the first calculation unit 405. ). Specifically, the prediction result table 1300 shown in FIG. 13 is output. As a result, it is possible to recognize the predicted number of incoming calls that are predicted to occur at an arbitrary time and the predicted abandoned call rate that will be an abandoned call.

(Optimization of the number of operators)
Next, a method for optimizing the number of operators arranged in the contact center 100 will be described. Here, first, using the control conditions (abandoned call rate, operation rate, waiting time, etc.) to be satisfied by the contact center 100 and the number of new calls at each time (new call number table 1000 shown in FIG. 10), The number of operators placed in the contact center 100 is calculated.

  Thereafter, using the estimated number of incoming calls at each time (prediction result table 1300 shown in FIG. 13), an index value (abandoned call rate, operation rate) for evaluating the quality of the contact center 100 in which the above-mentioned number of operators is arranged. Rate, wait time, etc.). Then, the index value is compared with the control condition to determine whether or not the number of operators arranged in the contact center 100 is correct.

  In other words, the number of operators tentatively determined from the number of new calls excluding the number of re-calls is used to predict the number of incoming calls after placement, evaluate the index value at that time, and assign the appropriate number of operators. Check if it has been done. Hereinafter, a specific implementation method for optimizing the number of operators will be described.

  First, the sixth calculation unit 410 calculates the number of operators to be placed in the contact center 100 for each time based on the number of new calls at each time calculated by the fourth calculation unit 408. Specifically, by giving the constraints that the contact center 100 should satisfy, the operator's operator at each time based on the number of new calls at each time using the Erlang B formula, Erlang C formula, Erlang A formula, simulation, etc. Calculate the number of personnel. Hereinafter, the number of operators at time i is assumed to be B (i).

  The constraint conditions are conditions that define the abandoned call rate, waiting time, operating rate, and the like at each time that the contact center 100 should satisfy. The waiting time is the time from when the customer calls the contact center 100 until the operator responds. The operation rate is the ratio of the time during which the operator is engaged in customer service in each time zone (the time when the customer responds to inquiries from the customer regarding the working hours).

  Specific examples of the control conditions include “abandoned call rate: 5% or less, operation rate: 60% or more, waiting time: 10 seconds or less”, and the like. The technique for calculating the number of operator personnel from the number of incoming calls at each time (the number of new calls in this embodiment) using the Erlang B formula, Erlang C formula, Erlang A formula, simulation, or the like is a known technique. The description is omitted here.

  Various data required for Erlang B, Erlang C, Erlang A, simulations, etc. (average service time for operators to respond to customer inquiries, average patience time for waiting time that customers can withstand without giving up) ) Is input to the work support apparatus 101 by the user operating the keyboard 310 and the mouse 311 shown in FIG.

  Here, a personnel number table that stores the number of operators at each time calculated by the sixth calculator 410 will be described. FIG. 14 is an explanatory diagram showing the stored contents of the personnel number table. In FIG. 14, the number of personnel table 1400 stores the number of personnel of operators to be arranged for each time within the business hours of the contact center 100. For example, the number of operators to be placed in the contact center 100 at 0 is B (0).

  The output unit 404 has a function of outputting the operator number B (i) of the operator at time i calculated by the sixth calculation unit 410. Specifically, the personnel number table 1400 shown in FIG. 14 is output. As a result, the number of operator personnel to be arranged at an arbitrary time within the business hours of the contact center 100 can be recognized.

  The sixth calculation unit 410 has a function of calculating the number of operators of each shift for each time based on the number of operators at each time and a shift table that defines the working hours of the operators. The shift table is electronic data that defines the working hours of the operator for each shift. In the shift table, for example, working hours such as shift 1 “9: 0 to 12:00” and shift 2 “13: 0 to 17:00” are defined for each shift. This shift table is input to the business support apparatus 101 when the user operates the keyboard 310 and the mouse 311 shown in FIG.

  More specifically, for example, the sixth calculation unit 410 can convert the number of operators at each time into the number of personnel for each shift by optimization using an existing linear programming method. Hereinafter, the number of operators at time i of shift k is assumed to be H (k, i). However, in H (k, i), the non-working time such as a break time must be 0, and the values of the working time must be the same. Here, a shift table that stores the number of operators of each shift calculated by the sixth calculation unit 410 will be described.

  FIG. 15 is an explanatory diagram showing the contents stored in the shift table. In FIG. 15, the shift table 1500 stores the number of operators to be arranged for each time for each shift within the business hours of the contact center 100. For example, the number of operators at 0 o'clock in shift 2 is H (2, 0).

  In addition, the sixth calculation unit 410 calculates the index for each time based on the total number of personnel of all shifts 1 to k at each time and the number of new calls at each time calculated by the fourth calculation unit 408. Calculate the value. The index value is the abandoned call rate, operating rate, waiting time, etc. at each time when the quality of the contact center 100 is evaluated.

  Specifically, first, the total number of personnel of all shifts 1 to k at each time is obtained. Hereinafter, the sum of the number of personnel for all shifts 1 to k at time i is H (i). For example, the total number of personnel of all shifts 1 to k at 0 is H (0) [H (0) = H (1,0) + H (2,0) +... + H (k, 0) ].

  After this, using Erlang B, Erlang C, Erlang A, simulation, etc., based on the total number of shifts 1 to k at each time and the number of new calls at each time The index value of is calculated.

  Next, using the index value calculated by the sixth calculation unit 410, an index value when an incoming call including a re-call is generated is obtained. Specifically, for example, the abandoned call rate calculated by the sixth calculator 410 (the index value when the re-call portion is excluded) is used as the target abandoned call rate, and the number of predicted incoming calls at time i using the above-described method. EC (i) and the predicted abandoned call rate EA (i) at time i are calculated.

  Thereafter, the sixth calculation unit 410 calculates an operation rate and a waiting time in the calculated predicted abandoned call rate by using the Erlang B formula, Erlang C formula, Erlang A formula, simulation, and the like. Here, the predicted abandoned call rate, the operating rate and the waiting time at the predicted abandoned call rate are index values when an incoming call including a re-call is generated.

  Finally, it is determined whether or not an operator with an appropriate number of personnel is arranged by comparing an index value when an incoming call including a re-call is generated and a control condition that the contact center 100 should satisfy. . Here, when the index value does not satisfy the control condition, the constraint condition given when the sixth calculation unit 410 calculates the number of operators is changed, and the number of operators is calculated again.

  In the present embodiment, the contact center 100 that performs a telephone answering service in response to an inquiry from a customer has been described. However, the present invention is not limited to this. For example, it may be a contact center that accepts inquiries from customers by chat (corresponding to an incoming call) in which character-based conversation is performed in real time via the network 110.

(Business support processing procedure of business support device)
Next, a job support processing procedure of the job support apparatus 101 according to the present embodiment will be described. First, a new call number calculation processing procedure for calculating the number of new calls at each time in consideration of the re-call rate at each time within the business hours of the contact center 100 and the time interval until the re-call occurs will be described. To do. 16 and 17 are flowcharts illustrating an example of a new call count calculation processing procedure.

  In the flowchart of FIG. 16, first, it is determined whether or not an input of the data table 500 has been accepted (step S1601). Here, it waits for the data table 500 to be input (step S1601: No), and when it is input (step S1601: Yes), the calculation unit 403 refers to the data table 500 and performs the operation of the contact center 100. The number of abandoned calls at each time within the time is calculated (step S1602), and the abandoned call occurrence time i is initialized to “0” (step S1603).

  Then, the re-call occurrence time j from the abandoned call at time i is set to “j = i” (step S1604), and the re-call occurrence probability for calculating the occurrence probability at time j of the re-call generated from the abandoned call at time i. Calculation processing is executed (step S1605).

  Thereafter, the calculation unit 403 calculates the number of recalls at the time j of the recalled call from the abandoned call at the time i (step S1606), and increments “1” to the occurrence time j of the recall (step S1607). . “% 24” means that when the calculation result exceeds 24, it returns to 0 (... → 23 → 0 → 1 →...). That is, if “j + 1 = 24” is obtained as a result of incrementing “1” at the occurrence time j of the recall, “j = 0” is set.

  Next, it is determined whether or not “j = i” (step S1608). If “j ≠ i” (step S1608: No), the process returns to step S1605. On the other hand, in the case of “j = i” (step S1608: Yes), the calculation unit 403 uses the number of abandoned calls at time i and the number of recalls at time j to re-call among the abandoned calls at time i. The number of completely abandoned calls at each time j that is not matched is calculated (step S1609).

  Thereafter, “1” is incremented at the generation time i of the abandoned call (step S1610), and it is determined whether or not “i = 24” is obtained (step S1611). If “i ≠ 24” (step S1611: NO), the process returns to step S1604. On the other hand, when “i = 24” (step S1611: Yes), the process proceeds to step S1612 shown in FIG.

  In the flowchart of FIG. 17, first, the abandoned call occurrence time i is initialized to “0” (step S1612), and the calculation unit 403 uses the number of incoming calls at time i and the abandoned call rate at time i. The number of corresponding calls at time i is calculated (step S1613). Then, the calculation unit 403 calculates the number of new calls at time i using the number of corresponding calls at time i and the number of completely abandoned calls at time i (step S1614).

  Thereafter, “1” is incremented at the occurrence time i of the abandoned call (step S1615), and it is determined whether or not “i = 24” is obtained (step S1616). If “i ≠ 24” (step S1616: NO), the process returns to step S1613. On the other hand, in the case of “i = 24” (step S1616: Yes), the output unit 404 outputs the number of new calls at each time (step S1617), and the series of processing according to this flowchart ends.

  In this way, the number of recalls of recalls occurring at each time is determined in consideration of the recall rate at each time that divides the business hours of the contact center 100 and the time interval from the abandoned call until the recall occurs. By calculating, the number of new calls at each time can be accurately predicted.

  Next, an example of a specific processing procedure of the re-call occurrence probability calculation processing in step S1605 in FIG. 16 will be described. FIG. 18 is a flowchart illustrating an example of a specific processing procedure of the recall occurrence probability calculation processing. In the flowchart of FIG. 18, first, the specifying unit 402 specifies a probability density distribution representing the probability of occurrence of a recall by substituting the average time interval peculiar to time i into the above equation (1) (step S1801).

  Thereafter, the calculation unit 403 calculates the time interval X from the occurrence time i of the abandoned call and the occurrence time j of the recall (step S1802), and using the probability density distribution specified in step S1801, The area of the probability density distribution from the occurrence time i to the time (X + 1) after the lapse of X hours is calculated (step S1803).

  Next, the minute interval ΔX is added to the time interval X (step S1804), and it is determined whether or not “X <[(j−i + 24)% 24 + 1]” is satisfied (step S1805). If “X <[(j−i + 24)% 24 + 1]” (step S1805: YES), the process returns to step S1803.

  On the other hand, in the case of “X ≧ [(j−i + 24)% 24 + 1]” (step S1805: No), the average area value calculated in step S1803 is calculated as the recall occurrence probability (step S1806). The process proceeds to step S1606 shown in FIG.

  In this way, the probability of a re-call from an abandoned call that occurred at an arbitrary time is calculated using the probability density distribution that represents the probability of a re-call that changes according to the elapsed time after the occurrence of the abandoned call. It can be calculated for each time that divides business hours.

  Next, a procedure for calculating the number of incoming calls for predicting the number of incoming calls at each time within the business hours of the contact center 100 will be described. 19 and 20 are flowcharts showing an example of the procedure for calculating the number of incoming calls. In the flowchart of FIG. 19, first, it is determined whether or not an input of a target abandoned call rate specific to each time has been received (step S1901).

  Here, it waits for the input of the target abandonment call rate specific to each time (step S1901: No), and when it is input (step S1901: Yes), the calculation unit 403 causes the target abandonment call rate specific to each time to be received. And the data table 500, the target number of abandoned calls at each time is calculated (step S1902).

  Thereafter, the abandoned call occurrence time i is initialized to “0” (step S1903), and the recall occurrence time j from the abandoned call at time i is set to “j = i” (step S1904). Then, the calculation unit 403 calculates the number of abandoned calls at time j generated from the re-call of the abandoned call that occurred at time i using the above equation (6) (step S1905).

  Next, “1” is incremented at the re-call occurrence time j (step S1906), and it is determined whether or not “j = i” is satisfied (step S1907). In step S 1906, if “j + 1 = 24” is obtained as a result of incrementing “1” at the re-call occurrence time j, “j = 0” is set.

  If “j ≠ i” (step S1907: NO), the process returns to step S1905. On the other hand, if “j = i” (step S1907: Yes), “1” is incremented at the abandoned call occurrence time i (step S1908), and it is determined whether or not “i = 24” is obtained (step S1908). S1909).

  If “i ≠ 24” (step S1909: No), the process returns to step S1904. On the other hand, if “i = 24” (step S1909: Yes), the calculation unit 403 causes the time i Using the target number of abandoned calls and the number of abandoned calls at time j generated from the re-call of the abandoned call at time i, the predicted number of abandoned calls expected to occur at time i is calculated (step S1910). Control proceeds to step S1911 shown in FIG.

  In the flowchart of FIG. 20, first, the abandoned call occurrence time i is initialized to “0” (step S <b> 1911), and the calculation unit 403 calculates the number of new calls at time i and the predicted number of abandoned calls at time i. Is used to calculate the predicted number of incoming calls that are expected to arrive at time i (step S1912).

  Then, the calculation unit 403 calculates a predicted abandoned call rate indicating the ratio of abandoned calls to incoming calls at time i using the predicted number of abandoned calls at time i and the estimated number of incoming calls (step S1913). Thereafter, “1” is incremented at the occurrence time i of the abandoned call (step S1914), and it is determined whether or not “i = 24” is obtained (step S1915).

  If “i ≠ 24” (step S1915: NO), the process returns to step S1912. On the other hand, in the case of “i = 24” (step S1915: Yes), the output unit 404 outputs the predicted number of incoming calls and the predicted abandoned call rate at each time (step S1916), and performs a series of processing according to this flowchart. finish.

  In this way, when the abandoned call of the target number of calls at each time occurs, the number of abandoned calls generated at each time in the future by calculating the number of abandoned calls generated from the re-call of the abandoned call. Can be accurately predicted.

  Next, an optimization support processing procedure for supporting optimization of the number of operators arranged at each time within the business hours of the contact center 100 will be described. FIG. 21 is a flowchart illustrating an example of the optimization support processing procedure. In the flowchart of FIG. 21, first, it is determined whether or not an input of a constraint condition and a shift table has been accepted (step S2101).

  Here, it waits for the input of the constraint condition and the shift table (step 2201: No), and when it is input (step S2101: Yes), the calculation unit 403 sets the constraint condition and the number of new calls at each time. Based on this, the number of operators to be placed in the contact center 100 is calculated for each time (step S2102).

  After that, the calculation unit 403 calculates the number of operators of each shift for each time based on the number of operators at each time and the shift table (step S2103), and the number of all shift personnel at each time. Is calculated (step S2104).

  Then, the calculation unit 403 calculates an index value for each time based on the total number of personnel for all shifts at each time and the number of new calls at each time (step S2105). Next, using the calculated index value, an index value when an incoming call including a re-call is generated is calculated (step S2106). Finally, the output unit 404 outputs an index value when an incoming call including a re-call is generated (step S2107), and the series of processes according to this flowchart is terminated.

  Thus, by calculating the index value when an incoming call including a re-call is made in the contact center 100 where the number of operators corresponding to the number of new calls at each time is arranged, the staff assignment satisfies the control condition. You can check whether or not.

  As described above, according to the present embodiment, each time is considered in consideration of the recall rate at each time that divides the business time of the contact center 100 and the time interval from the abandoned call until the re-call occurs. It is possible to calculate the probability of occurrence of a re-call that occurs in In addition, by calculating the number of re-calls that occur at each time from the probability of re-calls at each time, the number of new calls that occur at each time that divides the business hours of the contact center 100 can be accurately predicted. Can do.

  In other words, considering the recall mechanism that changes depending on the person's life pattern (especially before morning work, lunch break, evening) Recalls can be appropriately excluded from incoming calls at the time.

  In addition, the number of calls abandoned at each time in the future can be calculated by calculating the number of calls abandoned due to the abandoned call re-call when the number of calls abandoned at the target number of calls abandoned. It can be predicted accurately.

  Further, it is possible to calculate an index value when an incoming call including a re-call in the contact center 100 in which the number of operators obtained from the number of new calls at each time is arranged is generated. Then, by comparing the index value with the control condition, it is possible to check whether or not the personnel assignment satisfies the control condition, and the number of operators to be placed in the contact center 100 is optimized. Can do.

  The business support method described in the present embodiment can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. This program is recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, and a DVD, and is executed by being read from the recording medium by the computer. The program may be a medium that can be distributed via a network such as the Internet.

  The following additional notes are disclosed with respect to the embodiment described above.

(Appendix 1) Computer
A probability density function representing the probability of a re-call from an abandoned call within the elapsed time since the abandoned call abandoned before the operator terminal responds to the incoming call from the user terminal via the exchange. Storage means for storing,
Obtain an average time interval unique to any time period until a re-call occurs from an abandoned call in any time period selected from a plurality of time periods within the business hours of the contact center where the operator terminal is installed Acquisition means,
By giving the average time interval acquired by the acquisition means to the probability density function stored by the storage means, the abandoned call from the abandoned call within the elapsed time since the occurrence of the abandoned call of the arbitrary time zone is generated. A specifying means for specifying a probability density distribution representing the probability of occurrence of a recall;
A calculation means for calculating, for each time period, a probability of occurrence of a re-call from an abandoned call in the arbitrary time period using the probability density distribution specified by the specifying means;
Output means for outputting the probability of occurrence of a recall for each time period calculated by the calculating means;
Business support program characterized by functioning as

(Appendix 2)
A second number for calculating the number of re-calls from the abandoned call in the arbitrary time zone based on the occurrence probability for each time zone calculated by the calculation means (hereinafter referred to as “first calculation means”). Function as a calculation means,
The acquisition means includes
Furthermore, the number of calls of abandoned calls that occurred in the arbitrary time zone, and the recall rate specific to the arbitrary time zone that indicates the rate of occurrence of re-calls from the abandoned call,
The second calculation means includes:
Based on the number of abandoned calls in any time zone acquired by the acquisition means, the recall rate specific to any time zone, and the occurrence probability for each time zone calculated by the first calculation means , Calculate the number of re-calls from abandoned calls in any time zone for each time zone,
The output means includes
The business support program according to appendix 1, wherein the number of recalled calls for each time period generated from an abandoned call in an arbitrary time period calculated by the second calculating unit is output.

(Appendix 3)
By using the number of abandoned calls in the arbitrary time zone and the number of recall calls for each time zone calculated by the second calculating means, the call is re-called out of the abandoned calls in the arbitrary time zone. It functions as a third calculation means for calculating the number of completely abandoned calls that were not met,
The output means includes
The business support program according to appendix 2, wherein the number of calls of a complete abandonment call in an arbitrary time zone calculated by the third calculation means is output.

(Appendix 4)
A call of a new call that occurred in the arbitrary time period using the number of calls of the incoming call in the arbitrary time period and the number of re-calls in the arbitrary time period calculated by the second calculating means Function as a fourth calculating means for calculating the number;
The output means includes
The business support program according to appendix 3, wherein the number of calls of a new call in an arbitrary time period calculated by the fourth calculation means is output.

(Supplementary Note 5) The acquisition means includes:
The target specific to the arbitrary time zone, wherein the target value of the response rate of the operator terminal for the incoming call in the arbitrary time zone is defined by the ratio of the number of abandoned calls to the number of incoming calls in the arbitrary time zone Get abandoned call rate,
The third calculation means includes:
The target abandoned call rate specific to an arbitrary time zone acquired by the acquiring unit, the number of calls of a new call in an arbitrary time zone calculated by the fourth calculating unit, and the second calculating unit Predicted calls of abandoned calls that are expected to occur for the number of incoming call calls in any time period using the number of recall calls in any time period generated from abandoned calls in each time period Calculate the number,
The output means includes
The business support program according to appendix 4, wherein the predicted number of abandoned calls in an arbitrary time zone calculated by the third calculating means is output.

(Appendix 6)
Using the number of calls of a new call in an arbitrary time zone calculated by the fourth calculation means and the predicted number of abandoned calls in an arbitrary time zone calculated by the third calculation means, the arbitrary call Functioning as a fifth calculating means for calculating the predicted number of incoming calls that are expected to arrive during
The output means includes
The business support program according to appendix 5, wherein the predicted number of incoming calls in an arbitrary time period calculated by the fifth calculation means is output.

(Supplementary Note 7) The first calculation means includes:
Using the predicted number of abandoned calls in any time period calculated by the third calculating means and the estimated number of incoming call calls in any time period calculated by the fifth calculating means, Calculate the expected abandoned call rate, which shows the ratio of the number of abandoned calls to incoming calls at any time,
The output means includes
The business support program according to appendix 6, wherein a predicted abandoned call rate in an arbitrary time zone calculated by the first calculation means is output.

(Appendix 8)
Based on the control conditions to be satisfied by the contact center and the number of new calls in each time zone calculated by the fourth calculating means, the number of personnel for each time zone to be placed in the contact center is calculated. Sixth calculating means for
Based on the calculation result calculated by the sixth calculation means, function as seventh calculation means for calculating the abandoned call rate of the contact center for each time period,
The third calculation means includes:
The contact center abandoned call rate calculated by the seventh calculating means, the number of new calls in each time zone, and the abandoned calls in each time zone calculated by the second calculating means The number of abandoned calls that are predicted to occur with respect to the number of incoming calls in each time period is calculated using the number of re-calls for each time period. Business support program.

(Supplementary Note 9) The fifth calculation means includes:
Using the number of new calls in each time zone and the predicted number of abandoned calls in each time zone calculated by the third calculation means, the number of incoming calls predicted to arrive in each time zone is as follows. Calculate the estimated number of calls,
The first calculation means includes
Using the predicted number of abandoned calls in each time period calculated by the third calculating means and the estimated number of incoming calls in each time period calculated by the fifth calculating means, A predicted abandoned call rate indicating the ratio of the number of abandoned calls to incoming calls in the band, an operating rate indicating the percentage of time that the operator is engaged in customer service operations in each time period, and the customer's ratio to the contact center Calculate the waiting time at the time of inquiry,
The output means includes
9. The business support program according to appendix 8, wherein the predicted abandonment call rate, the operation rate, and the waiting time for each time period calculated by the first calculation means are output.

(Supplementary note 10) A computer-readable recording medium in which the business support program according to any one of supplementary notes 1 to 9 is recorded.

(Supplementary Note 11) Represents the probability of the occurrence of a re-call from the abandoned call within the elapsed time since the abandoned call abandoned before the operator terminal responds to the incoming call from the user terminal via the exchange. Storage means for storing a probability density function;
Obtain an average time interval unique to any time period until a re-call occurs from an abandoned call in any time period selected from a plurality of time periods within the business hours of the contact center where the operator terminal is installed Acquisition means to
By giving the average time interval acquired by the acquisition means to the probability density function stored by the storage means, the abandoned call from the abandoned call within the elapsed time since the occurrence of the abandoned call of the arbitrary time zone is generated. A specifying means for specifying a probability density distribution representing the probability of occurrence of a recall;
Calculating means for calculating the occurrence probability of a re-call from an abandoned call in the arbitrary time period for each time period using the probability density distribution specified by the specifying means;
Output means for outputting the probability of occurrence of a recall for each time period calculated by the calculating means;
A business support apparatus comprising:

(Additional remark 12) The computer provided with the control means and the memory | storage means,
Recall from an abandoned call in any time zone selected from a plurality of time zones within the business hours of the contact center where an operator terminal that responds to an incoming call from a user terminal is installed by the control means. An acquisition step of acquiring an average time interval specific to the arbitrary time zone until occurrence of
By giving the average time interval acquired by the acquisition step to the probability density function representing the probability of the occurrence of a re-call from the abandoned call within the elapsed time from the occurrence of the abandoned call by the control means, A specific step of specifying a probability density distribution representing a probability of occurrence of a re-call from the abandoned call within an elapsed time after the abandoned call of the arbitrary time period occurs;
A calculation step of calculating, for each time period, a probability of occurrence of a re-call from the abandoned call in the arbitrary time period, using the probability density distribution specified in the specifying step by the control unit;
An output step of outputting the occurrence probability of the recall for each time period calculated by the calculation step;
A business support method characterized by executing

DESCRIPTION OF SYMBOLS 100 Contact center 101 Business support apparatus 102-1 to 102-n User terminal 103 Exchange machine 104-1 to 104-3 Operator terminal 401 Acquisition part 402 Identification part 403 Calculation part 404 Output part 405 First calculation part 406 Second calculation Unit 407 third calculation unit 408 fourth calculation unit 409 fifth calculation unit 410 sixth calculation unit

Claims (8)

Computer
A probability density function representing the probability of a re-call from an abandoned call within the elapsed time since the abandoned call abandoned before the operator terminal responds to the incoming call from the user terminal via the exchange. Storage means for storing,
Obtain an average time interval unique to any time period until a re-call occurs from an abandoned call in any time period selected from a plurality of time periods within the business hours of the contact center where the operator terminal is installed Acquisition means,
By giving the average time interval acquired by the acquisition means to the probability density function stored by the storage means, the abandoned call from the abandoned call within the elapsed time since the occurrence of the abandoned call of the arbitrary time zone is generated. A specifying means for specifying a probability density distribution representing the probability of occurrence of a recall;
A calculation means for calculating, for each time period, a probability of occurrence of a re-call from an abandoned call in the arbitrary time period using the probability density distribution specified by the specifying means;
Output means for outputting the probability of occurrence of a recall for each time period calculated by the calculating means;
Business support program characterized by functioning as The computer,
A second number for calculating the number of re-calls from the abandoned call in the arbitrary time zone based on the occurrence probability for each time zone calculated by the calculation means (hereinafter referred to as “first calculation means”). Function as a calculation means,
The acquisition means includes
Furthermore, the number of calls of abandoned calls that occurred in the arbitrary time zone, and the recall rate specific to the arbitrary time zone that indicates the rate of occurrence of re-calls from the abandoned call,
The second calculation means includes:
Based on the number of abandoned calls in any time zone acquired by the acquisition means, the recall rate specific to any time zone, and the occurrence probability for each time zone calculated by the first calculation means , Calculate the number of re-calls from abandoned calls in any time zone for each time zone,
The output means includes
2. The business support program according to claim 1, wherein the number of recalled calls for each time period generated from an abandoned call in an arbitrary time period calculated by the second calculating unit is output. The computer,
By using the number of abandoned calls in the arbitrary time zone and the number of recall calls for each time zone calculated by the second calculating means, the call is re-called out of the abandoned calls in the arbitrary time zone. It functions as a third calculation means for calculating the number of completely abandoned calls that were not met,
The output means includes
The business support program according to claim 2, wherein the number of calls of a complete abandonment call in an arbitrary time period calculated by the third calculation means is output. The computer,
A call of a new call that occurred in the arbitrary time period using the number of calls of the incoming call in the arbitrary time period and the number of re-calls in the arbitrary time period calculated by the second calculating means Function as a fourth calculating means for calculating the number;
The output means includes
4. The business support program according to claim 3, wherein the number of new calls in an arbitrary time zone calculated by the fourth calculation means is output. The acquisition means includes
The target specific to the arbitrary time zone, wherein the target value of the response rate of the operator terminal for the incoming call in the arbitrary time zone is defined by the ratio of the number of abandoned calls to the number of incoming calls in the arbitrary time zone Get abandoned call rate,
The third calculation means includes:
The target abandoned call rate specific to an arbitrary time zone acquired by the acquiring unit, the number of calls of a new call in an arbitrary time zone calculated by the fourth calculating unit, and the second calculating unit Predicted calls of abandoned calls that are expected to occur for the number of incoming call calls in any time period using the number of recall calls in any time period generated from abandoned calls in each time period Calculate the number,
The output means includes
5. The business support program according to claim 4, wherein the predicted number of abandoned calls in an arbitrary time period calculated by the third calculation means is output. The computer,
Using the number of calls of a new call in an arbitrary time zone calculated by the fourth calculation means and the predicted number of abandoned calls in an arbitrary time zone calculated by the third calculation means, the arbitrary call Functioning as a fifth calculating means for calculating the predicted number of incoming calls that are expected to arrive during
The output means includes
6. The business support program according to claim 5, wherein the predicted number of calls of an incoming call in an arbitrary time period calculated by the fifth calculation means is output. The first calculation means includes
Using the predicted number of abandoned calls in any time period calculated by the third calculating means and the estimated number of incoming call calls in any time period calculated by the fifth calculating means, Calculate the expected abandoned call rate, which shows the ratio of the number of abandoned calls to incoming calls at any time,
The output means includes
7. The business support program according to claim 6, wherein a predicted abandoned call rate in an arbitrary time zone calculated by the first calculation means is output. A probability density function representing the probability of a re-call from an abandoned call within the elapsed time since the abandoned call abandoned before the operator terminal responds to the incoming call from the user terminal via the exchange. Storage means for storing;
Obtain an average time interval unique to any time period until a re-call occurs from an abandoned call in any time period selected from a plurality of time periods within the business hours of the contact center where the operator terminal is installed Acquisition means to
By giving the average time interval acquired by the acquisition means to the probability density function stored by the storage means, the abandoned call from the abandoned call within the elapsed time since the occurrence of the abandoned call of the arbitrary time zone is generated. A specifying means for specifying a probability density distribution representing the probability of occurrence of a recall;
Calculating means for calculating the occurrence probability of a re-call from an abandoned call in the arbitrary time period for each time period using the probability density distribution specified by the specifying means;
Output means for outputting the probability of occurrence of a recall for each time period calculated by the calculating means;
A business support apparatus comprising:

Images