Matching using agent/caller sensitivity to performance

• US 9,020,137 B2
• Filed: 09/24/2013
• Issued: 04/28/2015
• Est. Priority Date: 09/24/2012
• Status: Active Grant

• Current Assignee: Afiniti Ltd
• Original Assignee: Satmap International Holdings, Ltd.
• Inventors: Chishti, Zia, Merchant, Akbar A.
• Primary Examiner(s): ADDY, THJUAN KNOWLIN
• Application Number:
  US14/035,506
• Publication Number:
  US 20140086404A1
• Time in Patent Office: 581 Days
• Field of Search: 379/265.12, 379/265.02, 379/265.01, 379/265.06, 379/265.05, 379/265.11, 379/265.03, 705/11
• US Class Current: 379/265.12
• CPC Class Codes: H04M 2203/402 Agent or workforce management
• CPC Class Codes: H04M 3/00 Automatic or semi-automatic exchanges
• CPC Class Codes: H04M 3/5175 Call or contact centers supervision arrangements
• CPC Class Codes: H04M 3/5191 interacting with the Internet
• CPC Class Codes: H04M 3/523 with call distribution or queueing
• CPC Class Codes: H04M 3/5232 Call distribution algorithms
• CPC Class Codes: H04M 3/5233 Operator skill based call distribution
• CPC Class Codes: H04M 5/00 Manual exchanges substation equipment in general H04M1/00

Abstract

A method, system and program product, the method comprising: obtaining for calls in one set of calls a respective pattern representing one or multiple different respective data fields; obtaining performance data for the respective patterns of the calls; performance data for agents in a set of agents; determining pattern performance sensitivity to agent performance comprising the pattern performance data correlated to agent performance data; matching a respective one of the agents from the set of agents to one of the calls based at least in part on the performance data for the one agent and on the pattern performance sensitivity to agent performance for the respective call.

Claims

1. A method of operating a call center, comprising:

obtaining for each call in one set of calls, by one or more computers, a respective pattern representing multiple different respective data fields;

obtaining, by the one or more computers, performance data for the respective patterns of the calls, comprising for each of the patterns a set of data for historical caller-agent matches for that pattern with the data comprising a result and a historical agent performance of a historical agent handling the respective caller-agent match;

correlating, by the one or more computers, the result to the historical agent performance for the caller-agent matches for the respective pattern to obtain a correlation;

obtaining, by the one or more computers, performance data for respective agents in a set of agents;

determining, by the one or more computers, a pattern performance sensitivity to agent performance for each of the respective patterns based at least in part on the respective correlation of the set of data from the historical caller-agent matches; and

matching, by the one or more computers, a respective one of the agents from the set of agents to one of the calls based at least in part on the performance data for the one agent and on the pattern performance sensitivity to agent performance for the pattern of the respective call.

2. The method as defined in claim 1, further comprising:

percentiling or ranking, by the one or more computers, the agents in the set of agents based at least in part on their respective performance data; and

percentiling or ranking, by the one or more computers, the respective patterns for the set of calls based at least in part on their respective pattern performance sensitivity to agent performance,

wherein the matching is based at least in part on the percentile or ranking of the one agent and the percentile or ranking by call performance sensitivity to historical agent performance of the one call.

3. The method as defined in claim 1, wherein the pattern is abstracted in whole or in part, so that a meaning for the field data in the pattern is not known by the system.

4. The method as defined in claim 1, wherein the performance data is based on at least one selected from the group of sales, retention of the caller in a program, call handle time, customer satisfaction, revenue, first call resolution, units sold, and transaction points.

5. The method as defined in claim 1, further comprising:

matching, by the one or more computers, a different set of calls to agents using a different matching algorithm;

comparing, performance data from call-agent matches of the different set of calls using the different matching algorithm with performance data for the one set of calls matched based at least in part on the pattern performance sensitivity to agent performance; and

generating, by the one or more computers, a report or display data for the performance comparing results from using the matching algorithm based at least in part on pattern performance sensitivity to agent performance against performance using the different matching algorithm.

20. The method as defined in claim 2, further comprising:

weighting, by the one or more computers, the ranking or percentile of agent performance and/or call performance and/or call sensitivity to agent performance by one or more parameters.

21. The method as defined in claim 20, wherein the weighting is of agent performance and is based at least in part on a call count for the respective agents during a period.

22. The method as defined in claim 20, wherein the weighting is of agent performance and is based at least in part on a call handle time total or an average handle time for the agents during a period.

23. The method as defined in claim 20, wherein the weighting is of call performance and is based at least in part on call count of the call pattern or an estimated call handle time for the respective calls in the set of calls.

24. The method as defined in claim 1, wherein the determining a pattern performance sensitivity to agent performance step comprises:

grouping, by the one or more computers, the historical agents into at least two groups comprising one group with agent performance in a particular range and a different group with agent performance in a lower range;

determining, by the one or more computers, an average measure for the performance of the historical agents in the one group and an average measure for the historical agents in the different group;

determining, by the one or more computers, a difference between the average measure for the one group and the average measure for the different group; and

comparing, by the one or more computers, the difference to a predetermined number or to a sensitivity number determined on the fly.

6. A method of operating a call center, comprising:

obtaining for each call in one set of calls, by one or more computers, a respective pattern representing multiple different respective data fields;

obtaining, by the one or more computers, performance data for the respective patterns of the calls;

obtaining, by the one or more computers, performance data for respective agents in a set of agents;

determining or obtaining, by the one or more computers, pattern performance sensitivity to agent performance for each of the patterns, computed for each pattern based at least in part on a correlation of results and historical agent performance for historical caller-agent matches for the respective pattern;

grouping, by the one or more computers, the patterns for the one set of calls into at least two groups comprising one group with performance data in a first range and a different group with performance data in a second range;

matching for the one group of the calls, by the one or more computers, a respective one of the agents from the set of agents to one of the calls in the one group based at least in part on the performance data for the one agent and the performance data for the pattern of the one call in the one group; and

matching for the different group of the calls, by the one or more computers, a different agent from the set of agents to one of the calls in the different group based at least in part on the performance data for the different agent and the pattern performance sensitivity to agent performance for the respective pattern for the call in the different group.

7. The method as defined in claim 6, wherein the determining pattern performance sensitivity to agent performance comprises correlating call performance data in the data set to agent performance of the agents handling the calls in a data set.

8. The method as defined in claim 6, further comprising:

percentiling or ranking, by the one or more computers, the agents in the set of agents based at least in part on their respective performance data;

percentiling or ranking, by the one or more computers, the respective patterns for the one set of calls based at least in part on their respective performance data; and

percentiling or ranking, by the one or more computers, the respective patterns for the one set of calls based at least in part on their pattern performance sensitivity to agent performance;

wherein the matching for the one group is based at least in part on the percentile or ranking of the respective pattern of the one call in the one group and the percentile or ranking of the one agent, and

wherein the matching for the different group is based at least in part on the percentile or ranking of the different agent and the percentile or ranking by call performance sensitivity to agent performance of the one call in the different group.

9. The method as defined in claim 6, further comprising:

percentiling or ranking, by the one or more computers, the respective patterns of the one set of calls based at least in part on their respective performance data in a data set;

wherein the grouping is based at least in part on the performance percentiles or rankings of the patterns, with the patterns in one of the groups having lower percentiles or rankings than patterns in the different group.

10. The method as defined in claim 6, wherein the matching steps are performed concurrently.

11. The method as defined in claim 6, wherein the matching steps are performed consecutively or with partial overlap in time.

12. The method as defined in claim 6, further comprising:

percentiling or ranking, by the one or more computers, the respective patterns of the one set of calls based at least in part on their respective performance data in a data set;

wherein the grouping is based at least in part on the percentiles or rankings of the patterns, with the patterns in one of the groups having lower percentiles or rankings than patterns in the different group, and

wherein a percentile or ranking break point between the one group and the different group is determined based on one or more criteria.

13. The method as defined in claim 6, further comprising:

matching, by the one or more computers, a different set of calls to agents using a different matching algorithm;

comparing, performance data from call-agent matches of the different set of calls using the different matching algorithm with performance data for the one set of calls using a combination of the matching of the one group of call patterns to agents and the matching of the different group of call patterns to agents; and

generating, by the one or more computers, a report or display data for the performance comparing results from using the combination of matching algorithms against performance using the different matching algorithm.

14. A system of operating a call center, comprising:

one or more computers configured with the following components;

a call data extractor component for obtaining for each call in a set of calls a respective pattern representing multiple different respective data fields;

a call pattern performance extractor for obtaining performance data for the respective patterns of the calls, comprising for each of the patterns a set of data for historical caller-agent matches for that pattern with the data comprising a result and a historical agent performance of a historical agent handling the respective caller-agent match;

an agent performance extractor component for obtaining performance data for respective agents in a set of agents;

a call sensitivity to agent performance correlation engine configured for determining pattern performance sensitivity to the agent performance comprising;

correlating, by the one or more computers, the result to the historical agent performance for the caller-agent matches for the respective pattern to obtain a correlation;

determining, by the one or more computers, a pattern performance sensitivity to agent performance for each of the respective patterns based at least in part on the respective correlation of the set of data from the historical caller-agent matches; and

a matching engine configured for matching a respective one of the agents from the set of agents to one of the calls based at least in part on the performance data for the one agent and on the pattern performance sensitivity to agent performance for the pattern of the respective call.

15. The system as defined in claim 14, further comprising:

a first percentile or ranking engine configured for percentiling or ranking the agents in the set of agents based at least in part on their respective performance data; and

a second percentile or ranking engine configured for percentiling or ranking the respective patterns for the set of calls based at least in part on their respective pattern performance sensitivity to agent performance;

wherein the matching engine is configured to perform matching based at least in part on the percentile or ranking of the one agent and the percentile or ranking by call performance sensitivity to agent performance of the one call.

25. The system as defined in claim 14, wherein the call sensitivity to agent performance correlation engine is configured for:

grouping, by the one or more computers, the historical agents into at least two groups comprising one group with agent performance in a particular range and a different group with agent performance in a lower range;

determining, by the one or more computers, an average measure for the performance of the historical agents in the one group and an average measure for the historical agents in the different group;

determining, by the one or more computers, a difference between the average measure for the one group and the average measure for the different group; and

comparing, by the one or more computers, the difference to a predetermined number or to a sensitivity number determined on the fly.

16. A system of operating a call center, comprising:

one or more computers configured with the following components;

a call data extractor component for obtaining for each call in one set of calls a respective pattern representing multiple different respective data fields;

a call pattern performance extractor for obtaining performance data for the respective patterns of the calls;

an agent performance extractor component for obtaining performance data for respective agents in a set of agents;

a call sensitivity to agent performance correlation engine configured for determining or obtaining pattern performance sensitivity to the agent performance for each of the patterns, computed for each pattern based at least in part on a correlation of results and historical agent performance for historical caller-agent matches for the respective pattern;

a grouping engine configured for grouping the patterns for the one set of calls into at least two groups comprising one group with performance data in a first range and a different group with performance data in a second range;

a first matching engine configured for matching a respective one of the agents from the set of agents to one of the calls in the one group based at least in part on the performance data for the one agent and the performance data for the pattern of the one call in the one group; and

a second matching engine configured for matching a different agent from the set of agents to one of the calls in the different group based at least in part on the performance data for the different agent and the pattern performance sensitivity to agent performance for the respective pattern for the call in the different group.

17. The system as defined in claim 16, further comprising:

a first percentile or ranking engine configured for percentiling or ranking the agents in the set of agents based at least in part on their respective performance data;

a second percentile or ranking engine configured for percentiling or ranking the respective patterns for the one set of calls based at least in part on their respective performance data;

a third percentile or ranking engine configured for percentiling or ranking the respective patterns for the one set of calls based at least in part on their pattern performance sensitivity to agent performance;

wherein the first matching engine is configured for matching the one group based at least in part on the percentile or ranking of the respective pattern of the one call in the one group and the percentile or ranking of the one agent, and

wherein the second matching engine is configured for matching based at least in part on the percentile or ranking of the different agent and the percentile or ranking by call performance sensitivity to agent performance of the one call in the different group.

18. The system as defined in claim 16, further comprising:

a percentile or ranking engine configured for percentiling or ranking the respective patterns of the one set of calls based at least in part on their respective performance data in a data set;

wherein the grouping engine is configured for grouping based at least in part on the performance percentiles or rankings of the patterns, with the patterns in the one group having lower percentiles or rankings than patterns in the different group.

19. The system as defined in claim 18, wherein the percentile or ranking engine is configured for setting a percentile or ranking break point between the one group and the different group based on one or more criteria.