WO2001004722A2

WO2001004722A2 - System and method for matching agents with customers

Info

Publication number: WO2001004722A2
Application number: PCT/NZ2000/000122
Authority: WO
Inventors: Paul Michael O'connor
Original assignee: Datamine Limited
Priority date: 1999-07-09
Filing date: 2000-07-10
Publication date: 2001-01-18
Also published as: AU5719800A; WO2001004722A3; AU766555B2

Abstract

The invention provides a system for matching agents with customers, the system comprising a memory in which is maintained an interaction database of interaction data representing interactions between agents and customers, the interaction data comprising a succes rating for each interaction. The system further comprises retrieval means arranged to retrieve from the interaction database interaction data for respective individual agents and prediction means arranged to determine the best match of an individual agent with a customer based on the interaction data retrieved from the interaction database. The invention further provides a method and a computer program for compiling the interaction database, and a method and computer program for matching agents with customers using the interaction database.

Description

SYSTEM AND METHOD FOR MATCHING AGENTS WITH CUSTOMERS

FIELD OF INVENTION

The invention relates to a system and method for matching agents with customers, particularly but not solely designed for matching telemarketing operators with existing and prospective customers. The invention also has application in matching incoming calls for a helpdesk, complaint desk, or promotional campaign with an agent and also in matching agents with customers with face-to-face contact.

BACKGROUND TO INVENTION

Telecommunications systems are widely used for commercial purposes by retailers, wholesalers and more generally vendors for selling services and equipment and for providing customer information with respect to such sales. This use is generally referred to as telemarketing. Telecommunications systems may be used for commercial purposes other than telemarketing, for example call centre applications, in particular helpdesk and customer complaints. They may be used as a means for customers to respond to a promotional campaign initiated by a merchant.

There are some difficulties with use of telecommunications systems in commercial applications such as those described above. For example, in an outbound telephone campaign, the person to whom the call is directed may be unavailable, the call may remain unanswered or someone other than the desired recipient may answer the call.

US patent 5,889,799 to Centerforce Technologies Inc identifies this problem and describes a method and system for optimisation of telephone contact campaigns. In this system, the callers have a stored account of each individual or target to be contacted and a history of attempts to contact the target. The system stores a profile of the target and from this profile, identifies the periods during which the target is most likely to be available. In this way, the likelihood of a successful contact with the target can be increased.

Another difficulty with commercial use of telecommunications systems is that there are often periods of very heavy incoming telephone call volume in situations where a caller is responding to a promotional campaign initiated by a merchant, for example, a television advertisement. The heavy call volume results in callers being put on hold while a finite number of agents deal with incoming calls. This may lead to callers waiting in a queue becoming impatient.

NZ patent specification 299023 to Voice Technology Limited identifies this problem and describes a system in which an incoming call is received and placed in a queue for an appropriate operator. If the operator is currently unavailable, the caller is asked whether they want to leave a message and be called back or wait for the operator to become available. If the caller chooses to leave a message, they are prompted for information, such as a phone number, then they hang up. Once the operator is available, the operator is played the message left by the caller to allow the operator to ring the caller back.

Another difficulty in this area is in matching agents having the appropriate skills with customers requiring those skills. The Voice Technology Limited specification discussed above identifies that some agents may have particular training or experience in dealing with products or services offered to one particular geographical area, but that calls may be received from a range of geographical areas. By identifying the characteristics of a particular call, for example, the geographic region from which the caller is calling, the incoming call can be distributed to an agent having the appropriate skills for answering a call from the identified geographic region.

A further difficulty not addressed by the prior art is that some interactions between agents and customers are more successful than other interactions and this success is not merely determined by a successful connection or by a successful connection with an agent having the appropriate skill set.

SUMMARY OF INVENTION

In broad terms in one form the invention comprises a system for matching agents with customers, the system comprising a memory in which is maintained an interaction database of interaction data representing interactions between agents and customers, the interaction data comprising a success rating for each interaction; retrieval means arranged to retrieve from the interaction database interaction data for respective individual agents; and prediction means arranged to determine the best match of an individual agent with a customer based on the interaction data retrieved from the interaction database. In another form in broad terms the invention comprises a method of compiling interaction data representing interactions between agents and customers, the method comprising the steps of compiling a list of customers; initiating an interaction between one of the customers and an agent; assigning a success rating to the interaction; and storing in an interaction database interaction data representing the interaction between the customer and the merchant, the interaction data including the success rating for the interaction.

In a further form the invention comprises a method of matching agents with customers comprising the steps of maintaining in a memory an interaction database of interaction data representing interactions between agents and customers, the interaction data including a success rating for each interaction; retrieving from the interaction database interaction data for respective individual agents; and predicting the best match of an individual agent with a customer based on the interaction data retrieved from the interaction database.

In broad terms in a further form the invention comprises a computer program for matching agents with customers, the program comprising an interaction database of interaction data representing interactions between agents and customers maintained in a memory, the interaction data comprising a success rating for each interaction; retrieval means arranged to retrieve from the interaction database interaction data for respective individual agents; and prediction means arranged to determine the best match of an individual agent with a customer based on the interaction data retrieved from the interaction database.

In another form in broad terms the invention comprises a computer program for compiling interaction data representing interactions between agents and customers, the program comprising compiling means arranged to compile a list of customers; interaction means arranged to initiate an interaction between one of the customers and an agent; rating means arranged to assign a success rating to the interaction; and updating means arranged to store in an interaction database interaction data representing the interaction between the customer and the merchant, the interaction data including the success rating for the interaction. BRIEF DESCRIPTION OF THE FIGURES

Preferred forms of the matching system and method will now be described with reference to the accompanying figures in which:

Figure 1 shows a block diagram of a system in which one form of the invention may be implemented;

Figure 2 shows the preferred system architecture of hardware on which the present invention may be implemented;

Figure 3 shows a preferred database schema of the invention;

Figure 4 illustrates a preferred method for compiling interaction data;

Figure 5 shows a preferred method for using the compiled interaction data; and

Figure 6 shows a table of example results of data acquisition and prediction.

DETAILED DESCRIPTION OF PREFERRED FORMS

Figure 1 illustrates a block diagram of the preferred system 10 in which one form of the present invention may be implemented. The system 10 includes one or more agents 20, for example 20A, 20B, 20C and 20D. The agents could comprise a telemarketing organisation promoting goods and/or services. Each agent 20 is preferably connected to a public switched telephone network or PSTN 30. The system 10 also includes one or more customers and/ or prospective customers 40, for example 40A, 40B, 40C and 40D who each have an address in the PSTN 30. Agents 20 communicate with customers 40, also referred to as targets, in an attempt to promote goods and/or services.

The invention 50 preferably comprises a data memory 52 and server 54 operating under suitable operating software and loaded with suitable application software. The invention 50 maintains a customer database 60 in a memory. The customer database 60 contains data representing individual existing customers and/or prospective customers. The invention 50 retrieves from the customer database 60 a fist of customers to which telemarketing agents will attempt to promote their goods and services. Individual agents are provided with details of targets and proceed to communicate with these targets through the PSTN 30. In order to maximise the efficiency of the system, the invention 50 could be interfaced to a predictive dialler 70. The invention 50 provides the predictive dialler 70 with a list of targets together with telephone numbers of the targets. The predictive dialler 70 automatically calls successive customers on the list. Once contact is made with a customer, the dialler 70 connects an agent to the customer.

To further improve efficiency, the predictive dialler 70 may call a prospective customer and then connect the call to the agent predicted by the dialler to be the next available agent. If the agent is available the call is connected immediately. If the agent is not yet available, the call could be disconnected and the target tried again later, or another agent who is available may be connected. It will be appreciated that use of a predictive dialler in this way reduces the time that agents are idle.

The system 10 further comprises an agent database 80 which contains data representing individual agents and an interaction database 90 which contains data representing interactions between agents and customers. It will be appreciated that customer database 60, agent database 80 and interaction database 90 may be implemented as separate databases or as a single database. For simplicity, the invention will be described with reference to separate databases. The databases 60, 80 and 90 could be maintained by the telemarketing organisation or could be maintained by a third party.

The databases 60, 80 and 90 could be stored in a single memory. Alternatively, the databases may be stored in separate memories on one or more separate servers and accessed by dedicated or dial-up telecommunications facilities using email, electronic data interchange (EDI) and/ or communications via internet websites, or stored on CD roms, floppy disks, tape drives, or other storage media. Alternatively, the databases 60, 80 and 90 could be accessed using terminal emulation or Telnet facilities.

Figure 2 shows the preferred system architecture of the invention 50. The computer system 100 typically comprises a central processor 102, a main memory 104 for example RAM and an input/ output controller 106. The computer system 100 further comprises peripherals such as a keyboard 108, a pointing device 110 for example a mouse, track ball or touch pad, a display or screen device 112, a mass storage memory 1 14 for example a hard disk, floppy disk or optical disc, and an output device 116 for example a printer. The system 100 could also include a network interface card or controller 1 18 and/or a modem 120. The individual components of the system 100 could communicate through a system bus 122.

Figure 3 illustrates a preferred database schema for implementation of the invention. The schema shown is in the form of a relational database schema having a number of linked tables. It is envisaged that the same data could be implemented in other forms, for example, an object-oriented environment using objects and attributes.

The preferred schema for the customer database is indicated at 200. Each customer record includes at least a customer identifier 202 which is preferably unique to the customer.

The customer database could also include characteristics of individual customers as shown at 204. These characteristics could include, for example, geographic location, professional status, family size, age, gender, marital status, ethnicity, education and vocation. The characteristics could also include psychodynamic or psychographic characteristics.

The characteristics could be obtained from either internal or external data. For example, where the customer is an existing customer of an agent, the agent may have customer characteristics stored in one or more databases as internal data. These characteristics could include the characteristics described above. Additional characteristics could be obtained by the agent from external data sources, for example, market surveys and census meshblock data. It is possible that the agent may have access to financial or other data for a customer and so the stored characteristics could include profile data, for example data representing financial transactions in which the customer is involved or data representing the use a customer makes of telecommunications or internet services.

A preferred schema for the agent database is shown at 220. The data includes at least an agent identifier 222. The agent schema 220 could also include the network address 224 of an individual agent. Where the agent is a telemarketer operating a telephone, the network address 224 could include the telephone number of the agent or the telephone extension of the agent within the organisation. In addition to the network address, the agent schema 220 could include characteristics about the individual agent, similar to the customer characteristics 204.

Interaction schema 240 represents interactions between agents and customers. The schema includes at least an agent identifier 242 and customer identifier 244. Where there may be more than one interaction between an agent and a customer, the interaction schema 240 could include an interaction identifier 248.

The schema 240 includes a success rating 250. The preferred success rating comprises an ordinal value representing a level of success of the interaction. For example, where an agent attempts to sell life insurance cover to a customer, the success rating could indicate whether or not the agent was successful in selling life insurance to the customer. It is envisaged that the success rating may be extended to represent, for example, situations where an agent attempts to sell one level of cover to a customer and the result of the interaction is that the agent sells a different level of cover to that originally intended.

The schema 240 could also include other data representing the interaction, for example, the duration of the interaction between the agent and the customer. An agent may take into account the time spent on a particular customer in assessing the success rating of the interaction.

It is envisaged that references to a customer include references to a prospective customer. Where a customer is not an existing customer of an agent but is instead a prospective customer, the customer schema 200 could still include customer characteristics obtained from external data sources. For example, the geographic location of the customer may be obtained from the area code of the customer telephone number. This geographic location forms part of a meshblock, and data on meshblocks is available from census data.

Figure 4 illustrates a preferred method for compiling interaction data for the interaction database 90. As indicated at 300, a list of targets is compiled. These targets may be retrieved from customer database 60 and/or from other sources, for example, telephone directories.

As indicated at 302, the first target is identified. Preferably, the first target is the customer appearing at the top of the list and is preferably contacted automatically by the predictive dialler 70. As indicated at 304, the invention connects the next available agent to the customer. The system could use the predictive dialler 70 to predict the individual operator who will next be available for an interaction with the customer.

An interaction record is created based on the interaction between the agent and the customer, the interaction record including a success rating. As indicated at 306, the interaction database is updated with this new data.

The invention then checks whether all the targets in the original fist have been contacted, as indicated at 308. Targets are preferably selected in sequential order from the target list. As targets are identified and interact with an agent, the target could be removed from the list or could be assigned a value indicating that the customer has already been the subject of an interaction with an agent. If all targets have been contacted the process is complete. Otherwise, the next target is contacted as indicated at 302.

The process described above could be performed by an organisation employing the agents or could be performed by a third party.

Following the data acquisition outlined in Figure 4, the system is arranged to use the stored interaction data as will be described with reference to Figure 5. As indicated at 400, a target list is compiled in a manner similar to that described in Figure 4.

The invention identifies an available operator as indicated at 402. It is envisaged that a predictive dialler 70 may predict which agent will next become available for interaction with a customer.

As indicated at 404, the invention then retrieves from the interaction database interaction data for the agent identified at step 402.

From the data retrieved from the interaction database, the invention identifies the characteristics of customer with which the agent typically has had the highest success. Preferably, this information is identified from the success rating in the interaction data.

By identifying the characteristics of the customers in the target list and comparing those characteristics with the characteristics of customer for which the agent has had the most success, the system can predict the best match of target customer to agent, as indicated at 406. The comparison and prediction could use any one of a number of known techniques suitable for the purpose, for example logistic regression.

Once the best match has been predicted, the appropriate customer is contacted as indicated at 408, then the agent is connected to that customer at 410.

The invention may be arranged to update the interaction database with the results of the interaction indicated at 412 in a similar manner to that described with reference to Figure 4.

As indicated at 414, the invention then checks to see whether all targets have been contacted, in a similar way as described with reference to Figure 4. If there are still targets remaining to be contacted the invention identifies the next available agent.

Figure 6 illustrates the results of data acquisition and prediction described above with reference to Figures 4 and 5 in a typical case. The results of the analysis are illustrated in the form of a table. The data in the table could be calculated at runtime, or could be stored temporarily or permanently in a memory.

As shown in Figure 6, the invention has identified as a percentage chance the predicted success rate of individual agents 500 with existing or prospective customer targets Tim Jones 502, Bob Smith 504, Angela Dee 506 and Frank Hayden 508.

The invention may predict that agent 510 is about to finish their present interaction with a customer. Agent 510 has the highest chance of success with customer 508, out of all the customers. Agent 510 also has the highest chance, out of all of the other agents, of making a successful sale to customer 508. The system may, on this basis, predict that customer 508 would be the best match for agent 510.

The next agent about to finish their call could be agent 512. Agent 512 has the highest chance of success with customer 504 out of all the customers. Agent 512 also has the highest chance, out of all of the other agents, of making a successful sale to customer 504. Therefore, the system may estabhsh that customer 504 is the best match for agent 512. Agent 514 may be the next to finish their interaction. Agent 514 has the highest chance of success with customer 506 out of all the customers. However, since agents 516 and 518 have a much higher probability of success with customer 506, the invention may elect not to match customer 506 with agent 514. The invention would instead search the target list until a target is located with which agent 514 has a higher chance of success than agent 516 or agent 518, and match customer 506 with agent 516 or agent 518.

It is envisaged that the invention could be arranged to connect customers with agents with which the agent's chance of success exceeds a minimum threshold level, for example, 60%. This parameter could be overridden in cases where none of the agent's success rates exceed the threshold level. In this situation, the invention would estabhsh a connection with the customer when the agent having the highest chance of success with that customer is available, or about to become available.

Where the agent database has maintained in it agent characteristics, an organisation could identify the most desirable agent characteristics for a particular customer base. In this way a company may use the invention to assist in recruitment and management of its agent employees.

It is envisaged that the same technique described above could be applied to situations other than telemarketing. For example, the area of application may be a helpdesk or complaint desk in which customers initiate interactions with agents. Incoming customer calls would be routed by a call router to the most appropriate agent, based on the techniques described above. Characteristics of each customer could be retrieved from a customer database using the number from which the customer is calling or an identifier supplied by the customer when placing the call. This identifier could include a membership or account number entered by the customer using a touch tone telephone.

The success rating for a helpdesk or complaint desk could include the duration of the interaction and/ or the frequency of repeat calls. The success rating could additionally or alternatively be measured by customer satisfaction based on, for example, surveys or follow up contact with the customer.

A merchant may also initiate a promotional campaign soliciting contact from customers, which could be managed in the same way. The invention could also be applied to internet rather than telecommunication interactions. Referring to Figure 3, the customer identifier 202 could include the customer's internet address, and agent identifier 222 and/or agent network address 224 could include the agent's internet address.

The same system could also be applied to face-to-face contact situations with a customer and a merchant. It will be appreciated that such an arrangement would be most suitable for high value sales in which a customer may be directed to the most appropriate agent for a sale, or where the most appropriate agent may be directed to a customer.

The foregoing describes the invention including preferred forms thereof. Alterations and modifications as will be obvious to those skilled in the art are intended to be incorporated within the scope hereof, as defined by the accompanying claims.

Claims

1. A system for matching agents with customers, the system comprising: a memory in which is maintained an interaction database of interaction data representing interactions between agents and customers, the interaction data comprising a success rating for each interaction; retrieval means arranged to retrieve from the interaction database interaction data for respective individual agents; and prediction means arranged to determine the best match of an individual agent with a customer based on the interaction data retrieved from the interaction database.

2. A system as claimed in claim 1 wherein the prediction means is arranged to determine the best match by identifying the characteristics of an individual customer, identifying from the success rating in the interaction data the characteristics of customer with which respective agents have had the highest success, and comparing the characteristics of the individual customer with the characteristics of customer with which each individual agent has had the highest success.

3. A system as claimed in claim 1 or claim 2 further comprising a customer memory in which is maintained a customer database of customer data representing individual customers including characteristics of individual customers, the system arranged to retrieve characteristics of individual customers from the customer database.

4. A system as claimed in any one of claims 1 to 3 further comprising an agent memory in which is maintained an agent database of agent data representing individual agents, the system arranged to retrieve agent data from the agent database.

5. A system as claimed in any one of the preceding claims wherein the interactions maintained in the interaction database comprise communications between agents and customers provided by one or more telecommunications service providers.

6. A system as claimed in any one of the preceding claims wherein the interactions represent communications from agents to customers initiated by the agents.

7. A system as claimed in any one of claims 1 to 5 wherein the interactions represent communications from customers to agents initiated by the customers.

8. A method of compiling interaction data representing interactions between agents and customers, the method comprising the steps of: compiling a Hst of customers; initiating an interaction between one of the customers and an agent; assigning a success rating to the interaction; and storing in an interaction database interaction data representing the interaction between the customer and the merchant, the interaction data including the success rating for the interaction.

9. A method as claimed in claim 8 further comprising the steps of maintaining in a memory a customer database of customer data representing individual customers; and compiling at least part of the Hst of customers from the customer database.

10. A method as claimed in claim 9 wherein the customer data includes characteristics of individual customers, the method further comprising the step of retrieving characteristics of individual customers from the customer database.

11. A method as claimed in any one of claims 8 to 10 wherein an interaction is initiated for each customer in the Hst of customers.

12. A method as claimed in any one of claims 8 to 1 1 wherein an interaction is initiated between a customer and the next available agent.

13. A method of matching agents with customers comprising the steps of: maintaining in a memory an interaction database of interaction data representing interactions between agents and customers, the interaction data including a success rating for each interaction; retrieving from the interaction database interaction data for respective individual agents; and predicting the best match of an individual agent with a customer based on the interaction data retrieved from the interaction database.

14. A method as claimed in claim 13 wherein the step of predicting the best match further comprises the steps of identifying the characteristics of an individual customer; identifying from the success rating in the interaction data the characteristics of customer with which respective agents have had the highest success; and comparing the characteristics of the individual customer with the characteristics of customer with which each individual agent has had the highest success.

15. A method as claimed in claim 13 or claim 14 further comprising the steps of maintaining in a customer memory a customer database of customer data representing individual customers including characteristics of individual customers; and retrieving characteristics of individual customers from the customer database.

16. A method as claimed in any one of claims 13 to 15 further comprising the steps of maintaining in an agent memory an agent database of agent data representing individual agents; and retrieving agent data from the agent database.

17. A method as claimed in any one of claims 13 to 16 wherein the interactions maintained in the interaction database comprise communications between agents and customers provided by one or more telecommunications service providers.

18. A method as claimed in any one claims 13 to 17 wherein the interactions represent communications from agents to customers initiated by the agents.

19. A method as claimed in any one of claims 13 to 17 wherein the interactions represent communications from customers to agents initiated by the customers.

20. A computer program for matching agents with customers, the program comprising: an interaction database of interaction data representing interactions between agents and customers maintained in a memory, the interaction data comprising a success rating for each interaction; retrieval means arranged to retrieve from the interaction database interaction data for respective individual agents; and prediction means arranged to determine the best match of an individual agent with a customer based on the interaction data retrieved from the interaction database.

21. A computer program as claimed in claim 20 wherein the prediction means is arranged to determine the best match by identifying the characteristics of an individual customer, identifying from the success rating in the interaction data the characteristics of customer with which respective agents have had the highest success, and comparing the characteristics of the individual customer with the characteristics of customer with which each individual agent has had the highest success.

22. A computer program as claimed in claim 20 or claim 21 further comprising a customer database of customer data representing individual customers including characteristics of individual customers maintained in a customer memory; the program arranged to retrieve characteristics of individual customers from the customer database.

23. A computer program as claimed in any one of claims 20 to 22 further comprising an agent database of agent data representing individual agents maintained in an agent memory, the program arranged to retrieve agent data from the agent database.

24. A computer program as claimed in any one of claims 20 to 23 wherein the interactions maintained in the interaction database comprise communications between agents and customers provided by one or more telecommunications service providers.

25. A computer program as claimed in any one of claims 20 to 24 wherein the interactions represent communications from agents to customers initiated by the agents.

26. A computer program as claimed in any one of claims 20 to 24 wherein the interactions represent communications from customers to agents initiated by the customers.

27. A computer program as claimed in any one of claims 20 to 26 embodied on a computer readable medium.

28. A computer program for compiling interaction data representing interactions between agents and customers, the program comprising: compiling means arranged to compile a Hst of customers; interaction means arranged to initiate an interaction between one of the customers and an agent; rating means arranged to assign a success rating to the interaction; and updating means arranged to store in an interaction database interaction data representing the interaction between the customer and the merchant, the interaction data including the success rating for the interaction.

29. A computer program as claimed in claim 28 further comprising a customer database of customer data representing individual customers maintained in a memory; the program arranged to compile at least part of the Hst of customers from the customer database.

30. A computer program as claimed in claim 29 wherein the customer data includes characteristics of individual customers, the program further comprising retrieval means arranged to retrieve characteristics of individual customers from the customer database.

31. A computer program as claimed in any one of claims 28 to 30 wherein an interaction is initiated for each customer in the Hst of customers.

32. A computer program as claimed in any one of claims 28 to 31 wherein an interaction is initiated between a customer and the next available agent.

33. A computer program as claimed in any one of claims 28 to 32 embodied on a computer readable medium.