Techniques for benchmarking pairing strategies in contact center systems

Abstract

A technique for benchmarking pairing strategies in a contact center system is disclosed. In one particular embodiment, the technique can be implemented as a method for benchmarking pairing strategies in a contact center system, the method comprising: cycling, by at least one processor configured for contact center operations, between at least two pairing strategies; and determining, by the at least one processor, a performance difference between the at least two pairing strategies, wherein at least one contact is paired in a sequence of contact interactions to achieve an end result during an end contact interaction of the sequence of contact interactions.

Description

[0001] This application is a divisional application of patent application entitled "Techniques for Benchmarking Pairing Strategies in Contact Center Systems" having International Application No. PCT / IB2017 / 000570, filed April 18, 2017, and U.S. Application No. 201780002926.6, filed February 27, 2018, which claims priority to U.S. Patent Application No. 15 / 221,698, filed July 28, 2016, which is a continuation-in-part of U.S. Patent Application No. 15 / 131,915, filed July 18, 2016, which is a continuation-in-part of U.S. Patent Application No. 12 / 021,251, filed January 28, 2008, and is a continuation-in-part of U.S. Patent Application No. 14 / 727,271, filed June 1, 2015, which is a continuation of U.S. Patent Application No. 14 / 472,998, filed August 29, 2014, now U.S. Patent No. 9,215,323, and is a continuation of U.S. Patent Application No. 12 / 266,446, filed November 6, 2008, now U.S. Patent No. 8,824,658, each of which is incorporated by reference herein in its entirety as if fully set forth herein.

[0002] Cross Reference to Related Applications

[0003] This international patent application claims priority to U.S. Patent Application No. 15 / 221,698, filed July 28, 2016, which is a continuation-in-part of U.S. Patent Application No. 15 / 131,915, filed July 18, 2016, which is a continuation-in-part of U.S. Patent Application No. 12 / 021,251, filed January 28, 2008, and is a continuation-in-part of U.S. Patent Application No. 14 / 727,271, filed June 1, 2015, which is a continuation of U.S. Patent Application No. 14 / 472,998, filed August 29, 2014, now U.S. Patent No. 9,215,323, and is a continuation of U.S. Patent Application No. 12 / 266,446, filed November 6, 2008, now U.S. Patent No. 8,824,658, each of which is incorporated by reference herein in its entirety as if fully set forth herein. TECHNICAL FIELD

[0004] The present disclosure relates generally to contact centers, and more specifically to techniques for benchmarking pairing strategies in contact center systems. BACKGROUND

[0005] A typical contact center algorithmically assigns contacts arriving at the contact center to agents available to handle those contacts. At times, the contact center can have agents available and waiting to be assigned to inbound or outbound contacts (e.g., telephone calls, Internet chat sessions, emails) or outbound contacts. At other times, the contact center can have contacts waiting in one or more queues of agents to become available for assignment.

[0006] In some typical contact centers, contacts are assigned to agents in an order according to the time of arrival, and agents receive contacts in an order according to the time those agents become available. This strategy can be referred to as a "first-in, first-out," "FIFO," or "round robin" strategy.

[0007] Some contact centers can use a "performance-based routing" or "PBR" approach to order the queues of available agents or occasional contacts. A PBR ordering strategy attempts to maximize the expected outcome of each contact-agent interaction, but often does so without regard for the even use of agents at the contact center.

[0008] When a contact center changes from using one type of pairing strategy (e.g., FIFO) to another type of pairing strategy (e.g., PBR), the overall contact center performance will change over time. It is difficult to measure the amount of performance change due to the use of the new pairing strategy because there can be other factors that cause performance to increase or decrease some over time.

[0009] In light of the foregoing, it can be appreciated that there is a need for a system that can benchmark alternative routing strategies to measure the performance change due to the alternative routing strategy. SUMMARY

[0010] A technique for benchmarking pairing strategies in a contact center system is disclosed. In one particular embodiment, the technique can be implemented as a method for benchmarking pairing strategies in a contact center system, the method comprising: cycling, by at least one processor configured for contact center operations, between at least two pairing strategies; and determining, by the at least one processor, a performance difference between the at least two pairing strategies, wherein at least one contact is paired in a sequence of contact interactions to achieve an end result during a final contact interaction of the sequence of contact interactions.

[0011] According to other aspects of this particular embodiment, the end result indicator for each of the sequence of contact interactions can indicate the end result determined from the final contact interaction of the sequence of contact interactions.

[0012] According to other aspects of this particular embodiment, the pairing strategy indicator for each of the sequence of contact interactions can indicate which of the at least two pairing strategies was used to pair each of the sequence of contact interactions.

[0013] According to other aspects of this particular embodiment, the iteration indicator for each of the sequence of contact interactions can indicate a sequence iteration number at which each of the sequence of contact interactions occurred.

[0014] According to other aspects of this particular embodiment, determining the performance difference can further include determining, by the at least one processor, an intermediate measure of performance for each set of contact interactions indicated by the same sequence iteration number; and determining, by the at least one processor, an overall difference in performance from a combination of the intermediate measures of performance for each set of contact interactions for all sequence iteration numbers.

[0015] According to other aspects of this particular embodiment, the combination is a weighted average of the intermediate measures of performance for each set of contact interactions.

[0016] According to other aspects of this particular embodiment, the method can further include determining, by the at least one processor, each arrival time for the at least one contact; and selecting, by the at least one processor, one of the at least two pairing strategies for each arrival time of the at least one contact based on the each arrival time, regardless of any previous pairing of the at least one contact.

[0017] In another particular embodiment, the technology can be implemented as a system for benchmarking pairing strategies in a contact center system, the system comprising at least one processor configured for contact center operations, wherein the at least one processor is configured to perform the above-described method.

[0018] In another particular embodiment, the technology can be implemented as an article of manufacture for benchmarking pairing strategies in a contact center system, the article of manufacture comprising: a non-transitory processor-readable medium and instructions stored on the medium, wherein the instructions are configured to be readable from the medium by at least one processor configured for contact center operations and thereby cause the at least one processor to operate so as to perform the above-described method.

[0019] The present disclosure will now be described in greater detail in connection with certain particular embodiments thereof. Although the following refers to particular embodiments, it should be understood that the application is not limited thereto. Those skilled in the art with access to the teachings provided herein will recognize additional implementations, modifications, and embodiments, as well as other applications of the present disclosure, and the present disclosure is therefore not limited by the specific disclosure herein provided. BRIEF DESCRIPTION OF DRAWINGS

[0020] For the purposes of this disclosure, like reference numerals in the drawings and the foregoing description shall mean the same or similar elements.

[0021] Figure 1A A schematic representation of a benchmarking sequence is shown in accordance with embodiments of the present disclosure.

[0022] Figure 1B A schematic representation of a reference detection sequence is shown in accordance with an embodiment of the disclosure.

[0023] Figure 2A A schematic representation of a reference detection sequence is shown in accordance with an embodiment of the disclosure.

[0024] Figure 2B A schematic representation of a reference detection sequence is shown in accordance with an embodiment of the disclosure.

[0025] Figure 3A A schematic representation of a reference detection sequence is shown in accordance with an embodiment of the disclosure.

[0026] Figure 3B A schematic representation of a reference detection sequence is shown in accordance with an embodiment of the disclosure.

[0027] Figure 3C A block diagram of a contact center system is shown in accordance with an embodiment of the disclosure.

[0028] Figure 3D A block diagram of a behavior pairing module is shown in accordance with an embodiment of the disclosure.

[0029] Figure 4 A block diagram of a contact center is shown in accordance with an embodiment of the disclosure.

[0030] Figure 5 A flowchart of a reference detection method is shown in accordance with an embodiment of the disclosure.

[0031] Figure 6 A block diagram of a reference detection module is depicted in accordance with an embodiment of the disclosure.

[0032] Figure 7A A schematic representation of a reference detection sequence is shown in accordance with an embodiment of the disclosure.

[0033] Figure 7B A schematic representation of a reference detection sequence is shown in accordance with an embodiment of the disclosure.

[0034] Figure 8 A flowchart of a reference detection method is shown in accordance with an embodiment of the disclosure. DETAILED DESCRIPTION

[0035] A typical contact center algorithmically assigns contacts arriving at the contact center to agents that can handle the contacts. At times, the contact center can have agents available and waiting to be assigned to inbound or outbound contacts (e.g., telephone calls, Internet chat sessions, emails) or outbound contacts. At other times, the contact center can have contacts waiting in one or more queues of agents to become available for assignment.

[0036] In some typical contact centers, contacts are assigned to agents in an order based on time of arrival, and agents receive contacts in an order based on when those agents become available. This strategy can be referred to as a "first-in, first-out," "FIFO," or "round robin" strategy. For example, a longest available agent pairing strategy preferentially selects an available agent that has been available the longest time.

[0037] Some contact centers can use a "performance-based routing" or "PBR" approach to order the queue of available agents or occasional contacts. A PBR ordering strategy attempts to maximize the expected outcome of each contact-agent interaction, but often does so without regard for evenly using agents at the contact center. Some variations of PBR can include a highest performance agent pairing strategy (preferentially selecting an available agent with the highest performance) or a highest performance agent pairing strategy for a contact type (preferentially selecting an available agent with the highest performance for the contact type of the pairing).

[0038] For yet another example, some contact centers can use a "behavioral pairing" or "BP" strategy under which contacts and agents are intentionally (preferentially) paired in a way that can assign subsequent contact-agent pairs such that when the sum of the benefits of all assignments under the BP strategy is calculated they can outperform FIFO and PBR strategies. BP is designed to encourage balanced use of agents in skill queues, yet at the same time will allow for improved overall contact center performance to outperform FIFO or PBR approaches. This is a significant achievement because BP works on the same calls and same agents as FIFO or PBR approaches, provides roughly even use of agents as FIFO, and still improves overall contact center performance. BP is described in, for example, U.S. Patent Application No. 14 / 871,658, filed September 30, 2015 (incorporated herein by reference). Additional information about these and other features related to pairing or matching modules (sometimes also referred to as "SATMAP," "routing system," "routing engine," etc.) is described in, for example, U.S. Patent No. 8,879,715 (incorporated herein by reference).

[0039] Some contact centers can use various other possible pairing strategies. For example, in a longest available agent pairing strategy, an agent can be selected that has been waiting (idle) the longest time because the agent's most recent contact interaction (e.g., call) has just ended. In a least occupied agent pairing strategy, an agent can be selected that has the lowest ratio of contact interaction time to waiting or idle time (e.g., time spent on a call vs. time spent off hook on a call). In a least contact interaction agent pairing strategy, an agent can be selected that has the least total contact interactions or calls. In a random selection agent pairing strategy, an agent can be randomly selected (e.g., using a pseudo-random number generator) from available agents. In a sequential tagging agent pairing strategy, agents can be sequentially tagged, and an available agent can be selected that has the next tag in the sequence.

[0040] In situations where multiple contacts are waiting in a queue and an agent is available to connect to one of the contacts in the queue, a variety of pairing strategies can be used. For example, in a FIFO or longest waiting contact pairing strategy, an agent can preferentially pair with a contact that has been waiting the longest in the queue (e.g., the contact at the head of the queue). In a randomly selected contact pairing strategy, an agent can pair with a contact that is randomly selected from all contacts in the queue or a subset thereof. In a priority-based routing or highest priority contact pairing strategy, an agent can pair with a higher priority contact even if a lower priority contact has been waiting in the queue for a longer time.

[0041] A contact center can measure performance according to a variety of metrics. For example, a contact center can measure performance according to one or more of sales revenue, sales conversion rate, customer retention rate, average handling time, customer satisfaction (according to, e.g., customer surveys), etc. Regardless of what metric or combination of metrics a contact center uses to measure performance, or regardless of what pairing strategy (e.g., FIFO, PBR, BP) a contact center uses, performance can vary over time. For example, a contact center's annual performance can vary as a company shrinks or grows over time or introduces new products or contact center promotional campaigns. A contact center's monthly performance can vary as a company's sales cycle, such as a holiday sales season or a period of heavy technical support requests after a new product or upgrade is released. A contact center's daily performance can vary, e.g., if customers are more likely to call on weekends than on weekdays or on Mondays than on Fridays. A contact center's daily performance can also vary. For example, customers can be more likely to call at the beginning of a contact center's day (e.g., 8:00 a.m.) or during lunch (e.g., 12:00 p.m.) or in the evening after typical business hours (e.g., 6:00 p.m.) than at other times during the day. A contact center's hourly performance can also vary. For example, more urgent high-value contacts can be more likely to arrive at the beginning of a contact center's hours (e.g., 9:00 or 9:01) than even slightly later (e.g., 9:05). A contact center's performance can also vary according to the number and skill of agents working at a given time. For example, agents on a 9:00 a.m. - 5:00 p.m. shift can on average perform better than agents on a 5:00 a.m. - 9:00 a.m. shift.

[0042] These examples of variability at certain times of day or over longer periods of time can make it difficult to attribute performance changes within a given time period to a particular pairing strategy. For example, if a contact center uses FIFO routing with an average performance of 20% sales conversion rate for a year, then switches to PBR with an average performance of 30% sales conversion rate the following year, the apparent change in performance is a 50% improvement. However, this contact center can not have a reliable way of knowing what the average performance would have been in the second year if the contact center had continued to use FIFO routing instead of PBR. In the real world, at least some portion of the 50% increase in performance in the second year can be attributable to other factors or variables that are not controlled or measured. For example, the contact center can have retrained its agents or hired agents with higher performance, or the company can have introduced an improved product that is more highly accepted in the market. Thus, due to the challenges associated with measuring the performance increase attributable to a new pairing strategy, a contact center can have difficulty analyzing the internal rate of return or return on investment resulting from a switch to a different pairing strategy.

[0043] In some embodiments, the contact center can periodically switch (or "cycle") between at least two different pairing strategies (e.g., between FIFO and PBR; between PBR and BP; among FIFO, PBR, and BP). Additionally, the outcome of each contact-agent interaction can be recorded along with an identification of the pairing strategy (e.g., FIFO, PBR, or BP) used to assign that particular contact-agent pair. By tracking which interactions produce which outcomes, the contact center can measure performance attributable to a first strategy (e.g., FIFO) and performance attributable to a second strategy (e.g., PBR). In this way, the relative performance of one strategy over another can be benchmarked. Over multiple periods of switching between different pairing strategies, the contact center can more reliably attribute performance gains to one strategy or the other.

[0044] Several benchmarking techniques can achieve precisely measurable performance gains by reducing noise from confounding variables and eliminating bias in favor of one or another pairing strategy. In some embodiments, the benchmarking technique can be time-based ("epoch benchmarking"). In other embodiments, the benchmarking can involve randomization or counting ("inline benchmarking"). In other embodiments, the benchmarking technique can be a hybrid of epoch and inline benchmarking.

[0045] In epoch benchmarking, as explained in more detail below, the frequency of switching (or the duration of the period) can affect the accuracy and fairness (e.g., statistical purity) of the benchmark. For example, assume a period of two years, with a switch between two different strategies each year. In this case, the contact center can use FIFO with a 20% conversion rate in the first year and PBR with a 30% conversion rate in the second year, and measure a gain of 50%. However, this period is too long to eliminate or otherwise control for expected variability in performance. Even a shorter period, such as two months, with a switch in strategy each month, can suffer from similar effects. For example, if FIFO was used in November and PBR was used in December, some of the performance improvement in December can be attributable to holiday sales increases in December rather than PBR itself.

[0046] In some embodiments, to reduce or minimize the effects of performance variability over time, the period can be made much shorter (e.g., less than a day, less than an hour, less than twenty minutes). Figure 1A A benchmarking period of ten units (e.g., ten minutes) is shown. In this case, the contact center can switch between FIFO and PBR every ten minutes. In this way, the contact center can measure performance attributable to FIFO and performance attributable to PBR every ten minutes. In this way, the relative performance of one strategy over another can be benchmarked every ten minutes. Figure 1AIn the diagram, the horizontal axis represents time, and the vertical axis represents whether the first pairing strategy (“1”) or the second pairing strategy (“0”) is used. For the first five minutes (e.g., 9:00-9:05 AM), the first pairing strategy (e.g., BP) can be used. After five minutes, for the remaining five minutes of the ten-minute period (9:05-9:10 AM), the contact center can switch to the second pairing strategy (e.g., FIFO or PBR). At 9:10 AM, the second period can begin, switching back to the first pairing strategy. Figure 1A (Not shown in the image). If the time period is 30 minutes (i.e. Figure 1A If each time unit in the time interval is equal to three minutes, then the first pairing strategy can be used for the first 15 minutes and the second pairing strategy can be used for the second 15 minutes.

[0047] By utilizing short hourly timeframes (10 minutes, 20 minutes, 30 minutes, etc.), benchmarks are unlikely to be biased towards one pairing strategy or another based on long-term variability (such as year-over-year growth or monthly sales cycles). However, other performance variability factors may persist. For example, if the contact center consistently applies [the benchmark] during morning opening hours... Figure 1A During the time period shown, the contact center will consistently use Strategy 1 (BP) for the first five minutes. As explained above, contacts arriving at the contact center when it is open may have a different distribution of type, urgency, value, or type / urgency / value compared to contacts arriving at other times of the hour or day. Therefore, the benchmark may be biased towards a matching strategy that favors the use of the strategy at the start of the day (e.g., 9:00 AM).

[0048] In some embodiments, the order in which the pairing strategy is used within each period can vary in order to reduce or minimize the impact of performance variability, even over short periods. For example, as... Figure 1B As shown, the order in which the pairing strategy is used can be related to... Figure 1A The order shown is reversed. Specifically, the contact center can start with the second pairing strategy (such as FIFO or PBR) for the first five minutes, and then switch to the first pairing strategy (BP) for the next five minutes.

[0049] In some embodiments, to help ensure trust and fairness in the benchmarking system, benchmarking progress can be established and published, or shared in advance with the contact center administration or other users in other ways. In some embodiments, the contact center administration or other users can directly control the benchmarking progress in real time, such as by using a computer program interface to control the cycle duration and sequence of the pairing strategy.

[0050] Embodiments of the present disclosure can use any of a variety of techniques for varying the order in which pairing strategies are used within each time period. For example, a contact center can alternate between starting with the first order shown in Figure 1A and starting with the second order shown in Figure 1B every hour (or every day or every month). In other embodiments, each time period can randomly select an order (e.g., about 50% of the time periods in a given day use the order shown in Figure 1A and about 50% of the time periods in a given day use the order shown in Figure 1B where the distribution of orders among these time periods is uniform and random).

[0051] In the examples of Figure 1A and 1B each pairing strategy is used for the same amount of time within each time period (e.g., every five minutes). In these examples, the "duty cycle" is 50%. However, some pairing strategies are expected to perform better than others, despite other variabilities that affect performance. For example, BP is expected to perform better than FIFO. Thus, a contact center can want to use BP more than FIFO—so that more pairings are made using the higher-performing pairing strategy. As such, a contact center can prefer a higher duty cycle (e.g., 60%, 70%, 80%, 90%, etc.) to represent more time (or contact proportion) being paired using the higher-performing pairing strategy. Figure 2A An example of a ten-minute time period with an 80% duty cycle is shown. For the first eight minutes (e.g., 9:00-9:08), a first pairing strategy (e.g., BP) can be used. After the first eight minutes, the contact center can switch to a second pairing strategy (e.g., FIFO) for the remaining two minutes of the time period (9:08-9:10) before switching back to the first pairing strategy (not shown) again. For another example, if a thirty-minute time period is used, a first pairing strategy can be used for the first twenty-four minutes (e.g., 9:00-9:24) and a second pairing strategy can be used for the next six minutes (e.g., 9:24-9:30).

[0052] As Figure 2BAs shown, the contact center can have six ten-minute periods over the course of an hour. In this example, each ten-minute period has a duty cycle of 80% favoring the first pairing strategy, and the order within each period starts with the preferred first pairing strategy. Over the hour, the contact center can switch pairing strategies twelve times (e.g., at 9:08, 9:10, 9:18, 9:20, 9:28, 9:30, 9:38, 9:40, 9:48, 9:50, 9:58, and 10:00). Over the hour, the first pairing strategy is used for a total of 80% of the time (48 minutes), and the second pairing strategy is used for the other 20% of the time (12 minutes). For a 30-minute period with a duty cycle of 80% (not shown), the contact center can switch pairing strategies four times over the hour (e.g., at 9:24, 9:30, 9:48, and 10:00), and the first pairing strategy is used for a total of 48 minutes and the second pairing strategy is used for 12 minutes.

[0053] In some embodiments, as in the example of FIG. 2, the order in which pairing strategies are used within a period can change (not shown) even though the duty cycle (the percentage of time within a period that a given strategy is used) remains the same. However, for periods that are factors or multiples of 60 minutes (e.g., 10 minutes, 30 minutes), the periods can always or frequently align to the boundaries at the start of each hour (e.g., a new period starts at 9:00, 10:00, 11:00, etc.) regardless of the order in which pairing strategies are used within a period that starts at a given hour. Figure 1B

[0054] In some embodiments, as explained below with respect to Figure 3A -D, selecting periods such as 11 minutes, 37 minutes, some prime number that cannot be factored into 60-minute intervals, or other numbers can increase the number of periods required before a particular pattern repeats. Conversely, the alignment of the periods can drift over hours, days, weeks, etc. before repeating. The duration of each pairing strategy's cycle can not frequently align with a number of hours in a day, a number of days in a week, a number of weeks in a month or year, etc.

[0055] Figure 3A An example of a single non-factor period of 11 minutes with a duty cycle of about 73% is shown, in which the first eight minutes use the first pairing strategy and the last three minutes use the second pairing strategy. Figure 3B Six consecutive cycles are illustrated. For example, at the start of the first hour of the first day of a week (e.g., Monday at 9:00 AM), a first period can begin that is aligned to the start of the hour, the first hour of the day, and the first day of the week. The first period can be from 9:00 AM - 9:11 AM, followed by a second period from 9:11 AM - 9:22 AM, and so on, as in FIG. 2.​Figure 3B and illustrated in Table I below. The sixth period begins at 9:55 and ends at 10:06. The start of the second hour (10:00 AM) occurs during the sixth period and does not align with the beginning of the period. Figure 3C the same six periods are shown, with the horizontal axis labeled in ten minute intervals to further illustrate the intentional intra-hour misalignment. Figure 3B the same six periods are shown, with the horizontal axis labeled in ten minute intervals to further illustrate the intentional intra-hour misalignment.

[0056] Table I

[0057]

[0058]

[0059] As Figure 3D and shown in Table II below, with the example of a contact center open from 9:00 AM to 5:00 PM (9:00-17:00 hours), the alignment of the periods relative to the nearest hour continues to drift throughout the day. The first period of the first hour (9:00 AM) aligns with the start of the hour (9:00 AM). The first period of the second hour (10:00 AM) begins at 10:06 AM (six minutes after the start of the hour). The first period of the third hour (11:00 AM) begins at 11:01 AM (one minute after the start of the hour). It takes 60 periods over 11 hours for the first period of an hour to align with the start of the hour again. As shown in Table II, a contact center open from 9 to 5 will not align with the hour again until 12:00 PM the next day (1.375 days after 8 hours per day).

[0060] Table II

[0061] Hour Time of the first period of the hour 1 9:00 2 10:06 3 11:01 4 12:07 5 13:02 6 14:08 7 15:03 8 16:09 (next day) 9 9:04 10 10:10 11 11:05 12 12:00

[0062] Table III below shows a sequence of days and the time at which the new period begins at the start of an hour. For example, assuming a week of five days Monday through Friday (with one eight-hour day from 9-5), the sequence would align with 9:00 AM on Monday, 12:00 PM on Tuesday, 3:00 PM (15:00) on Wednesday, 10:00 AM on Friday, and so on. As shown in Table III, it takes 2.2 weeks for a contact center open 5 days per week, 8 hours per day, to align with the start of a day (for example, 9:00 AM on Tuesday 2 weeks later).

[0063] Table III

[0064] Day Next period starts at the beginning of the hour Monday 9:00 Tuesday 12:00 Wednesday 15:00 Friday 10:00 (next week) Monday 13:00 Tuesday 16:00 Thursday 11:00 Friday 14:00 (next week) Tuesday 9:00

[0065] Table IV below shows the sequence of days in a week where a new time period begins on the first day of the week. In this example, assuming a five-day week from Monday to Friday (with each day consisting of eight hours from 9 to 5), the sequence would align with the beginning of Monday of the first week, Tuesday of the third week, Wednesday of the fifth week, and so on. As shown in Table IV, it would take 11 weeks for this contact center to align with the beginning of Monday again.

[0066] Table IV

[0067] Week Next one-day cycle starts at the beginning of the day 1 Monday 3 Tuesday 5 Wednesday 7 Thursday 9 Friday 12 Monday

[0068] Therefore, such as Figures 3A-3D As illustrated in Tables I-IV, selecting non-factorial time periods with boundaries of one hour / day / week / etc. is effective in aligning time periods to “drift” across natural time boundaries over weeks / months / years. Because of this alignment drift, patterns that would confound the relative performance of multiple paired strategies are less likely to emerge. In some embodiments, the selection of non-factorial time periods can be combined with other techniques for reducing the impact of confounding variables on performance, such as randomizing the order of paired strategies within each time period or set of time periods, or otherwise altering the order of paired strategies within each time period or set of time periods.

[0069] In some embodiments, the contact center may determine which matching strategy to use based on the time it takes to make a matching request for a contact. For example, suppose the contact center uses... Figure 1A The example (a ten-minute timeframe with a 50% duty cycle, the first half starting with BP and the second half starting with FIFO) benchmarks BP and FIFO. If the contact center requests pairing at 9:04 AM, the pairing time falls within the first half of the timeframe, so the BP strategy can be used. If the contact center requests pairing at 9:06 AM, the pairing time falls within the second half of the timeframe, so the FIFO strategy can be used.

[0070] In other embodiments, the contact center can determine which pairing strategy to use based on the arrival time of the contact. For example, suppose the contact center performs benchmark detection for both BP and FIFO as in the previous example. If the first contact arrives at 9:04 AM, the arrival time falls within the first half of the time period, so the BP strategy can be used for the contact. Pairing can still be performed using the BP strategy even if the first contact has to wait in the queue for two minutes and does not request pairing until 9:06 AM. Furthermore, if the second contact arrives at 9:05 AM while the first contact is still waiting in the queue, the second contact can be designated for FIFO pairing. Therefore, at 9:06 AM, the selection of contacts under behavioral pairing may be limited to those in the queue that arrived during the BP period of that time period, and in this example, only the first contact that arrived is available.

[0071] In embodiments of time-based benchmark detection where a connection arrives at a boundary between time periods or between paired measurements within a time period, the system may have a predetermined tie-breaking strategy. For example, the boundary may be defined as "at or before the aforementioned time" or "at or after the aforementioned time," and so on. For instance, if a time period is defined as associated with strategy "A" from 9:00 to 9:08 and strategy "B" from 9:08 to 9:10, this might mean that a connection must arrive at or after 9:00 but before 9:08 (e.g., 9:07.99) to be considered within the first part of that time period.

[0072] In some embodiments, an inline benchmark detection technique can be used, in which a pairing strategy can be selected contact-by-contact. For example, suppose a first pairing method (e.g., FIFO) should be used to pair approximately 50% of the contacts arriving at the contact center, and a second pairing method (e.g., BP) should be used to pair the remaining 50%.

[0073] In some embodiments, each connection may be randomly assigned with a 50% probability to be paired using one method or another. In other embodiments, connections may be assigned sequentially based on a specific time period. For example, the first 5 (or 10, or 20, etc.) connections may be assigned to a FIFO strategy, and the next 5 (or 10, or 20, etc.) connections may be assigned to a BP strategy. Other percentages and proportions, such as 60% (or 80%, etc.), may also be paired with a BP strategy, and another 40% (or 20%, etc.) may be paired with a FIFO strategy.

[0074] Sometimes, a contact may return to the contact center multiple times (e.g., a callback). Specifically, some contacts may require multiple "contacts" (e.g., multiple interactions with one or more contact center agents) to resolve a problem. In these cases, it is desirable to ensure that the same matching strategy is used to match contacts each time they return to the contact center. If the same matching strategy is used for every contact, benchmarking techniques will ensure that this single matching strategy is associated with the end result (e.g., a solution) of multiple contact-agent interactions. In other cases, it is desirable to change the matching strategy each time a contact returns to the contact center so that each matching strategy has an equal chance of being used during the matching process of resolving the contact's needs and generating the end result. In still other cases, it is desirable to select a matching strategy regardless of whether the contact has contacted the contact center multiple times regarding the same issue.

[0075] In some embodiments, determining whether a repeat contact should be assigned to the same (or different) pairing strategy may depend on other factors. For example, there may be time constraints so that the contact must return to the contact center within a specified time period (e.g., within an hour, within a day, within a week) of the previous pairing strategy to be considered. In other embodiments, the pairing strategy used in the first interaction may be considered, regardless of how much time has passed since the first interaction.

[0076] In another example, repeated contacts might be limited to specific skill queues or customer needs. Consider calling the contact center and requesting a connection with a customer service agent regarding a billing issue. The contact is disconnected, and a few minutes later, a follow-up call is made requesting a technical support representative to discuss technical difficulties encountered during the connection. In this scenario, the second call can be considered a new issue rather than a second "contact" related to a billing problem. In this second call, it can be determined that the pairing strategy used in the first call is irrelevant to the second call. In other embodiments, regardless of why the contact returns to the contact center, the pairing strategy used in the first call can be considered.

[0077] exist Figure 4 The paper describes a method that takes into account previous pairings used for inline benchmark detection techniques. Figure 4 A flowchart of a benchmark detection method 400 according to an embodiment of the present disclosure is shown. The benchmark detection method 400 may begin at block 410.

[0078] In box 410, an identifier may be used to identify or otherwise determine the contact (e.g., the caller). In this example, the caller's "billing phone number" or "BTN" may be used. This example assumes that the caller uses the same BTN for every call. In other embodiments, other identifiers of the contact (e.g., customer identification number, Internet Protocol (IP) address) may be used instead. After identifying the caller's BTN (or other contact identifier), the benchmark detection method 400 may proceed to box 420.

[0079] In box 420, a pseudo-random number generator (PRNG) using a BTN (or other contact identifier) ​​can be seeded. After the PRNG is seeded using a BTN, the benchmark detection method 400 can proceed to box 430.

[0080] In box 430, pseudo-random numbers can be generated for the connection using the seeded PRNG. Because the seed will be the same for a given connection whenever the connection returns to the connection center, the pseudo-random numbers generated for a given connection will also be the same each time. After generating the pseudo-random numbers, the benchmark detection method 400 can proceed to box 440.

[0081] In box 440, a pairing strategy (e.g., BP or FIFO) can be selected for a given relationship based on the generated pseudo-random number. For example, if 50% of the relationships should be paired using BP and the other 50% should be paired using FIFO, the PRG can be configured to generate either 1 or 0. If the generated pseudo-random number is 1, the relationship can be specified for BP pairing. If the generated pseudo-random number is 0, the relationship can be specified for FIFO pairing.

[0082] In this manner, contacts will always be paired using the same strategy whenever they return to the contact center. The PRNG is seeded each time with the same seed (e.g., the contact's BTN), so the PRNG will generate the same pseudo-random number for the contact each time. Therefore, the benchmark detection method 400 can select the same pairing strategy for each contact. In this manner, previous pairings can be considered without relying on a database or other storage device to determine whether or how contacts were previously paired. In this manner, the benchmark detection method 400 is state-independent in terms of whether or how contacts were previously paired. After selecting a pairing strategy for a contact, the benchmark detection method 400 can proceed to box 450.

[0083] In box 450, a selected pairing strategy can be used to pair a contact with an available agent. Once a contact has been paired with an available agent, components of the contact center system (such as switches or routers) can connect the contact to the agent. After (or during) a contact-agent interaction, the agent can create a record of the interaction result. For example, in a sales queue, an agent can create an order for a contact. In a technical support queue, an agent can create or modify a service ticket. The contact center system can also record information related to the interaction, such as the call time and duration, the contact's BTN or other identifier, the agent's identifier, and other data. At this point, the benchmarking method can proceed to box 460.

[0084] In box 460, the identifier of the selected pairing strategy can be associated with the contact-agent interaction record created in box 450. In some embodiments, this can occur simultaneously with the creation of the record. For example, when the contact center system records the time and duration of a call, it can also record whether the call was paired using a BP or FIFO pairing strategy. In other embodiments, another module can create a separate record for the pairing. This module can record the time of the pairing, the contact and agent identifiers, the pairing strategy used (e.g., BP or FIFO), and any other data that may help to subsequently match the pairing record with a record of the caller-agent interaction result. At a later time, the pairing record can be matched with the caller-agent interaction record so that the pairing strategy information can be associated with the result in one or more records (or both). After box 460, the benchmark detection method 400 can terminate. In some embodiments, the benchmark detection method 400 can return to box 410, waiting for another contact to arrive.

[0085] exist Figure 5 The text describes another approach that considers combining previous pairings with chronology benchmark detection techniques. This type of technique can be considered "hybrid inline-chronology benchmark detection". Figure 5 A flowchart of a benchmark detection method 500 according to an embodiment of the present disclosure is shown. The benchmark detection method 500 may begin at block 510.

[0086] In box 510, a contact (e.g., "contact n") arrives at the contact center at a specific time t. Benchmark detection method 500 can proceed to box 520.

[0087] In box 520, it can be determined whether the contact has been previously paired; that is, whether the contact is returning to the contact center for subsequent contact or interaction. Various techniques can be used to make this determination. For example, a benchmark detection system can use contact identifiers (such as BTN or customer ID) in a database to look up contact records to determine whether and when a previous contact contacted the contact center. Using appropriate techniques, a benchmark detection system can determine if a contact has been previously paired, and in some embodiments, determine whether and how the previous pairing affects the current pairing.

[0088] In some embodiments, the benchmark detection system may preferably use the same pairing strategy to pair contacts each time a contact returns to the contact center. Therefore, if contact n was previously paired using pairing strategy "A" (e.g., BP), the benchmark detection method 500 may proceed to block 560 to subsequently pair again using pairing strategy A. Similarly, if contact n was previously paired using pairing strategy "B" (e.g., FIFO), the benchmark detection method 500 may proceed to block 570 to subsequently pair again using pairing strategy B.

[0089] However, if it is determined in block 520 that the connection n was not previously paired (or in some embodiments, any previous pairing will not affect the current pairing), the benchmark method 500 can continue to use the epoch benchmark detection at block 550.

[0090] In box 550, time can be used to determine which pairing strategy is used for contact n. In this example, arrival time t can be used. If contact n arrives during the time period when the benchmark detection system uses strategy A for pairing, then benchmark detection method 500 proceeds to box 560 to subsequently use strategy A for pairing. Similarly, if contact n arrives during the time period when the benchmark detection system uses strategy B for pairing, then benchmark detection method 500 can proceed to box 570 to subsequently use strategy B for pairing.

[0091] In boxes 560 and 570, pairing strategies A or B can be used to pair a contact with an available agent, respectively. In some embodiments, more than two pairing strategies can be used (e.g., using previous pairings of A, B, C, etc., or using A, B, C, etc. for periodic baseline detection over a time period). Once paired, the contact can be routed or otherwise connected to an available agent within the contact center system. Refer to baseline detection method 400 above (…). Figure 4 As described above, agents can create contact-agent interaction records, and the contact center system can also create or modify these records. The benchmark testing method can proceed to frame 580.

[0092] In box 580, the identifier for the selected pairing strategy (e.g., A or B) can be associated with a record created in box 560 or 570. As described above with reference to benchmark detection method 400, this association can occur simultaneously with the creation of a contact-agent interaction record, or it can subsequently be matched with other records created by the benchmark detection module or other modules. After box 580, benchmark detection method 500 can terminate. In some embodiments, benchmark detection method 500 can return to box 510, awaiting another contact.

[0093] In benchmarking methods such as 400 and 500, by correlating pairing strategies with results, the results associated with each pairing strategy can be measured (e.g., averaged, summed), and the relative performance of each pairing strategy can be measured (e.g., relative overall performance growth attributable to pairing using BP instead of FIFO). This benchmarking data can be used for various purposes. For example, the data can be used to assess the strength of one pairing module relative to another. As another example, the data can be used to enhance the strength of the BP module by providing “BP on” and “BP ​​off” (e.g., FIFO) contact-agent interaction logs to improve the artificial intelligence in the system. As yet another example, the data can be used for billing. Because the added value of one pairing strategy relative to another can be measured accurately and fairly, this benchmarking data can be used in performance-based business models where customers pay a pairing strategy provider a percentage of the actual metric increased by using the provider's pairing strategies (e.g., when BP is on rather than off).

[0094] Specifically, in some embodiments, relevant outcome data can be used to determine the economic value or growth associated with using one pairing strategy rather than another. In some embodiments, the economic value or growth can be used to determine compensation to suppliers or other service providers that offer modules for creating higher performance pairing strategies that generate economic value. For example, if a contact center benchmarks a business partner (BP) against a first-in-first-out (FIFO) strategy and determines that the BP performs an average of 5% better than the FIFO over a given time period (e.g., a day, a week, a month, etc.), the BP supplier may receive compensation corresponding to the 5% increase in value through the BP (e.g., 5% additional sales revenue or 5% additional cost savings, etc.). Under this business model, the contact center owner can forgo capital expenditures or supplier fees, paying the supplier only for the period during which the supplier demonstrates the value-added performance of the contact center.

[0095] Figure 6A block diagram of a contact center system 600 according to an embodiment of the present disclosure is shown. The description herein describes network elements, computers, and / or components for simulating a contact center system that may include one or more modules. As used herein, the term "module" can be understood to mean computing software, firmware, hardware, and / or various combinations thereof. However, these modules should not be construed as software that is not implemented on hardware, firmware, or recorded on a processor-readable recordable storage medium (i.e., a module is not software itself). It should be noted that these modules are exemplary. These modules may be combined, integrated, stand-alone, and / or replicated to support a variety of applications. Furthermore, in lieu of or in addition to the functions performed on a particular module, the functions performed on a particular module described herein may be performed on one or more other modules and / or by one or more other devices. Furthermore, these modules may be implemented across multiple devices and / or other components, either locally or remotely. Additionally, these modules may be removed from one device and added to another, and / or may be contained in both devices.

[0096] like Figure 6 As shown, the contact center system 600 may include a central switch 610. The central switch 610 may receive incoming contacts (e.g., callers) or support outbound connections to contacts via a telecommunications network (not shown). The central switch 610 may include contact routing hardware and software to facilitate contact routing within one or more contact centers or to one or more PBX / ACD or other queuing or switching components within a contact center.

[0097] In the contact center system 600, if there is only one contact center, or if there is only one PBX / ACD routing component, then the central switch 610 may not be necessary. If more than one contact center is part of the contact center system 600, then each contact center may include at least one contact center switch (e.g., contact center switches 620A and 620B). Contact center switches 620A and 620B may be communicatively coupled to the central switch 610.

[0098] Each contact center switch in each contact center can be communicatively coupled to multiple agents (or "pools"). Each contact center switch can support a certain number of agents (or "seats") logging in simultaneously. At any given time, a logged-in agent is available and waiting to connect to a contact, or a logged-in agent may be unavailable for any of a variety of reasons (such as connecting to another contact, performing certain post-call functions such as recording information related to the call, or taking a break).

[0099] existFigure 6 In the example, central switch 610 routes connections to one of two contact centers via contact center switches 620A and 620B, respectively. Each of contact center switches 620A and 620B is shown to have two agents. Agents 630A and 630B can log in to contact center switch 620A, and agents 630C and 630D can log in to contact center switch 620B.

[0100] The contact center system 600 can also be communicatively coupled with, for example, integrated services from third-party vendors. Figure 6 In the example, the benchmark detection module 640 may be communicatively coupled to one or more switches (such as central switch 610, contact center switch 620A, or contact center switch 620B) in the switching system of the contact center system 600. In some embodiments, the switches of the contact center system 600 may be communicatively coupled to multiple benchmark detection modules. In some embodiments, the benchmark detection module 640 may be embedded in a component of the contact center system (e.g., embedded in a switch or otherwise integrated with a switch). The benchmark detection module 640 may receive information related to agents logged into the switch (e.g., agents 630A and 630B) and related to incoming calls from another switch (e.g., central switch 610), or in some embodiments, from a network (e.g., the Internet or a telecommunications network) (not shown).

[0101] The contact center may include multiple pairing modules (e.g., BP modules and FIFO modules) (not shown), and these one or more pairing modules may be provided by one or more different vendors. In some embodiments, the one or more pairing modules may be components of the benchmark detection module 640 or one or more switches such as the central switch 610 or contact center switches 620A and 620B. In some embodiments, the benchmark detection module may determine which pairing module can handle pairing for a particular contact. For example, the benchmark detection module may alternate between being able to pair via the BP module and being able to pair using the FIFO module. In other embodiments, a pairing module (e.g., the BP module) may be configured to simulate other pairing strategies. For example, the benchmark detection module, or a benchmark detection component integrated with the BP component in the BP module, may determine whether the BP module can use BP pairing or simulated FIFO pairing for a particular contact. In this case, "BP on" may refer to the time when the BP module applies the BP pairing strategy. "BP off" may refer to other times when the BP module applies a different pairing strategy (e.g., FIFO).

[0102] In some embodiments, regardless of whether the pairing strategy is handled by a separate module, or if some pairing strategies are simulated within a single pairing module, the single pairing module can be configured to monitor and store information related to pairings performed under any or all pairing strategies. For example, the BP module can observe and record data related to FIFO pairings performed by the FIFO module, or the BP module can observe and record data related to simulated FIFO pairings performed by the BP module operating in FIFO simulation mode.

[0103] The embodiments disclosed herein are not limited to benchmarking only two pairing strategies. Rather, benchmarking can be performed on two or more pairing strategies. Figure 7A and Figure 7B Examples of benchmark detection systems for three pairing strategies (e.g., benchmark detection for FIFO, PBR, and BP) are depicted.

[0104] Figure 7A A schematic representation of a benchmark detection sequence according to an embodiment of the present disclosure is shown. In this periodic benchmark detection example, the period is 15 time units, and each pairing strategy is used for one-third of the time (5 units). Figure 7A Two complete time periods are shown, cycling twice between pairing strategies "2", "1", and "0" within 30 units of time. For example, FIFO can be used from 9:00 AM to 9:10 AM; PBR can be used from 9:10 AM to 9:20 AM; and BP can be used from 9:20 AM to 9:30 AM. This FIFO-PBR-BP pattern is repeated in the second time period.

[0105] Figure 7B A schematic representation of a benchmark detection sequence according to an embodiment of the present disclosure is shown. In this period benchmark detection example, a complete period is 30 time units. Two-thirds of the time uses a preferred pairing strategy "2" (e.g., BP), and each one-sixth of the time uses other pairing strategies "1" and "0" (e.g., FIFO and PBR). In this example, pairing strategies "1" and "0" are alternately turned on whenever strategy "2" is off. For example, the pattern could be BP-FIFO-BP-PBR. In addition... Figure 7A and Figure 7B Beyond the example, many other patterns are also possible for switching between multiple pairing strategies.

[0106] In some embodiments, contact center management or other users may prefer a "stable period" or other neutral zone. For example, consider a contact center benchmarking BP and FIFO pairing strategies. When the system transitions from BP to FIFO (and vice versa), contact center management may be concerned that the effectiveness of one pairing strategy might somehow impact the performance of the other. To mitigate fairness-related concerns, a stable period could be added.

[0107] One technique for achieving stable time periods could be to exclude the first part of the contact-agent interaction results after switching pairing strategies. For example, suppose the contact center performs benchmark detection on BP and FIFO with a 50% duty cycle over a 30-minute time period. In the above embodiments (e.g.) Figure 1A and 1B During the initial 30-minute period, BP will be active for 15 minutes, followed by FIFO for 15 minutes, and all contact-agent interactions within this 30-minute period will be included in the baseline detection metric. BP will then be active for, for example, 10 minutes during a stable period. After 10 minutes, the system will switch back to FIFO. However, the first 10 minutes will be considered a stable period, and FIFO pairings performed during this period will be excluded from the baseline. The final 10 minutes of this period will continue to use FIFO for pairing, and these FIFO pairings will be included in the baseline.

[0108] exist Figure 7A The pattern is described below. In this example, instead of describing the switching between the three pairing strategies "2", "1", and "0", "1" can represent a stable period. Pairing strategy "2" (e.g., BP) can be enabled for the first five time units. After five time units, BP can be disabled, and another pairing strategy (e.g., FIFO) can be used for the remaining ten time units. The next five units ("1") may be excluded as part of the stable period, and the five time units following ("0") can be included as part of the FIFO baseline detection period.

[0109] In some embodiments, the settling period can be longer or shorter. In some embodiments, the settling period can also be used in the FIFO to BP transition, in addition to the BP to FIFO transition (or any transition between two different pairing strategies) or in addition to the BP to FIFO transition (or any transition between two different pairing strategies).

[0110] As mentioned above, some contacts may require multiple "contacts" (e.g., multiple interactions with one or more contact center agents) to resolve the need for the contact. For example, an individual may call or otherwise contact a mortgage bank multiple times. The first call may be purely introductory or informational, the second call may be to evaluate different loan offers, and the third call may be to terminate (accept) or reject a loan offer. Similarly, some technical support and customer service requests may require more than one contact interaction to resolve.

[0111] In some cases, it is desirable to select a pairing strategy without considering whether a contact repeatedly contacts the same issue with the contact center. Instead, benchmarking techniques can be configured to take each pairing strategy into account. For example, each pairing strategy might be weighted in a way that fairly considers the extent to which it might contribute to a series of contact interactions that lead to the final outcome or solution for the contact request.

[0112] Figure 8 A flowchart of a benchmark detection method 800 according to an embodiment of the present disclosure is shown. The benchmark detection method 800 may begin at block 810.

[0113] In box 810, a connection (e.g., "connection n") arrives at the connection center at a specific time t and in a specific iteration i. For example, if this time is the first time connection n contacts the connection center for this specific need, it will be specified as the first iteration (i.e., i = 1), or the time could be the second (i = 2), the third (i = 3), ..., the nth call. Benchmark detection method 800 can proceed to box 820.

[0114] In box 820, time can be used to determine which pairing strategy is used for contact n. In these embodiments, even if this arrives subsequently such that i > 1, the pairing strategy selected for any previous pairing does not need to affect the current pairing. In this example, arrival time t can be used. If contact n arrives during the time period during which the benchmark detection system uses strategy A for pairing, then benchmark detection method 800 can proceed to box 830 to subsequently pair using strategy A. Similarly, if contact n arrives during the time period during which the benchmark detection system uses strategy B for pairing, then benchmark detection method 800 can proceed to box 840 to subsequently pair using strategy B. In some embodiments, pairing strategy A can be a behavioral pairing strategy (behavioral pairing "on"), and pairing strategy B can be a different pairing strategy such as FIFO or performance-based routing (behavioral pairing "off").

[0115] In boxes 830 and 840, pairing strategies A and B can be used respectively to pair contacts with available agents. In some embodiments, more than two pairing strategies can be used. Once paired, the contact can be routed to or otherwise connected to available agents within the contact center system. See reference benchmark detection method 400 above (…). Figure 4 ) and 500 Figure 5 As described, agents can create records of contact-agent interactions, and the contact center system can also create or modify these records. The benchmarking method can proceed to box 850.

[0116] In box 850, the identifier of the selected pairing strategy (e.g., A or B) can be associated with the identifier of the current iteration i of contact n within the record (record set) created in boxes 830 or 840. As described above with reference to benchmark detection methods 400 and 500, this association can occur simultaneously with the creation of the contact-agent interaction record, or it can be matched later with other records created by the benchmark detection module or other modules.

[0117] In some embodiments, the benchmark detection method 800 may proceed to block 860. At block 860, it may be determined whether the need for contact n has been resolved. If the need for contact n has not been resolved, or in other embodiments where the outcome of a given call may match at a later time, the benchmark detection method 800 may terminate or return to block 810 for the next contact to arrive. If the need for contact n has been resolved, the benchmark detection method 800 may proceed to block 870.

[0118] In box 870, the final outcome (e.g., a mortgage being terminated or rejected, a sale being completed or canceled) can be associated with the record of each of the i iterations of contact n. Table V shows an exemplary set of records for each of the four contacts W, X, Y, and Z that contacted the contact center system three times before reaching a final solution. In the real world, there may be many more contacts, and the number of contact interactions required to resolve an individual contact's needs may vary from contact to contact, ranging from one contact interaction to three or more.

[0119] Table V

[0120] Iteration i Contact W Contact X Contact Y Contact Z 1 A B A B 2 B B A A 3 A A B B Result Sale Sale Sale No Sale

[0121] As shown in Table V, and as described in reference box 850 above, the identifiers of the relationships (W, X, Y, or Z) and iterations (1, 2, or 3) are associated with the selected pairing strategy (A or B) used for the interaction of a given relationship. For relationship W, iteration 1 is paired with strategy A, iteration 2 with strategy B, and iteration 3 with strategy A. For relationship X, iteration 1 is paired with strategy B, iteration 2 with strategy B, and iteration 3 with strategy A. For relationship Y, iteration 1 is paired with strategy A, iteration 2 with strategy A, and iteration 3 with strategy B. For relationship Z, iteration 1 is paired with strategy B, iteration 2 with strategy A, and iteration 3 with strategy B.

[0122] As shown in Table V and as described in reference box 870 above, the final outcome of a contact can be associated with the record of each iteration. Contacts W, X, and Y completed a sale. Contact Z did not complete a sale. In some embodiments, a "no sale" determination can be made if a contact explicitly indicates that it does not intend to complete a sale. In other embodiments, the contact center system can make a "no sale" determination after a predetermined number of iterations without a sale, or after a predetermined amount of time has elapsed since the most recent iteration without a sale. In some cases, such as when a contact resolves a technical support or customer service need, the final solution may be a customer satisfaction survey result or score.

[0123] After block 870, the benchmark detection method 800 may terminate. In some embodiments, the benchmark detection method 800 may return to block 810, waiting for another contact to arrive (e.g., a different contact or a contact n with new requirements).

[0124] Performance differences between pairing strategies can be determined based on continuous hourly, daily, weekly, etc. In some embodiments, performance differences for each iteration can be stratified (performance differences for all contact interactions in iteration i=1, all contact interactions in iteration i=2, etc.). Although a final solution may not have been reached after the first contact interaction (when i=1), the benchmark for the first stratification can be measured using the final results previously associated with the first contact interaction in box 870 or matched with other contact center system records at subsequent times.

[0125] Using the example shown in Table V, for the baseline of the first tier (i=1), in the first iteration, strategy A is used to pair two connections (W and Y), and both of them have a final result that leads to a sale, which achieves a 100% conversion rate. In the first iteration, strategy B is used to pair the other two connections (X and Z), and only one of them (X) has a final result that leads to a sale, which achieves a 50% conversion rate.

[0126] Similarly, for the baseline of the second tier (i=2), in the second iteration, strategy A was used to pair the two relationships (Y and Z), and only one of them (X) had a final result that led to a sale, which achieved a 50% conversion rate. In the second iteration, strategy B was used to pair the other two relationships (W and X), and both of them had a final result that led to a sale, which achieved a 100% conversion rate.

[0127] Finally, for the baseline of the third tier (i=3), in the third iteration, strategy A was used to pair two connections (W and X) with both having a final result that leads to a sale, achieving a 100% conversion rate. In the third iteration, strategy B was used to pair the other two connections (Y and Z) with only one (Y) having a final result that leads to a sale, achieving a 50% conversion rate.

[0128] These stratifications and conversion rates are shown in Table VI below:

[0129] Table VI

[0130] Iteration / Layer i Strategy A Strategy B 1 100％ 50％ 2 50％ 100％ 3 100％ 50％

[0131] After determining the performance or performance difference between each paired strategy in each iteration (based on the final result), the performance differences can be combined. In some embodiments, the performance differences can be averaged. In the example of Table V, strategy A has an average conversion rate of approximately 83%, and strategy B has an average conversion rate of approximately 67%. Strategy A performs nearly 24% better than strategy B. In some embodiments, the performance differences are standardized based on the total number of contact interactions for each strategy in each iteration.

[0132] In some embodiments, some iterations may be weighted more or less than others. For example, the first contact interaction may be determined (e.g., by contact center system administrators or other business leaders) to be the most critical for determining whether or to what extent the desired end result will be achieved. In other cases, the final contact interaction may be determined to be the most critical. In these cases, the more critical layers may be weighted more heavily compared to the less critical layers in the final benchmark results.

[0133] It should be noted that the behavior pairing in the contact center system according to this disclosure, as described above, may involve, to some extent, the processing of input data and the generation of output data. This input data processing and output data generation can be implemented in hardware and software. For example, specific electronic components may be employed in the behavior pairing module or similar or related circuits described below to implement functions associated with behavior pairing in the contact center system according to this disclosure, as described above. Alternatively, one or more processors according to instructions may implement functions associated with behavior pairing in the contact center system according to this disclosure, as described above. If this is the case, these instructions may be stored on one or more non-transitory processor-readable storage media (e.g., a disk or other storage medium) or transmitted to one or more processors via one or more signals embodied in one or more carrier waves, all within the scope of this disclosure.

[0134] The scope of this disclosure is not limited to the specific embodiments described herein. In fact, various other embodiments and modifications of this disclosure will be apparent to those skilled in the art from the foregoing description and the accompanying drawings, in addition to those described herein. Therefore, such other embodiments and modifications are intended to fall within the scope of this disclosure. Furthermore, while this disclosure has been described in the context of at least one specific implementation in at least one specific environment for at least one particular purpose, those skilled in the art will recognize that its usefulness is not limited thereto and that this disclosure can be advantageously implemented in any number of environments for any number of purposes. Therefore, the claims set forth below should be interpreted in light of the full breadth and spirit of the disclosure described herein.

Claims

1. A method comprising: During the first time period, at least one computer processor, communicatively coupled to and configured to operate in the contact center system, pairs the first contact with the first agent using a first pairing strategy. The at least one computer processor records the results of the first contact-agent pairing during the first time period using the first pairing strategy; During a second time period later than the first time period, the at least one computer processor pairs the second contact with the second agent using a second pairing strategy different from the first pairing strategy. The results of the second contact-agent pairing during the second time period are recorded by the at least one computer processor using the second pairing strategy; During a third time period later than the second time period, the at least one computer processor pairs the third contact with the third agent using a third pairing strategy different from the second pairing strategy. The results of the third contact-agent pairing during the third time period are recorded by the at least one computer processor using the third pairing strategy; During a fourth time period later than the third time period, the at least one computer processor uses a fourth pairing strategy, different from the first pairing strategy and the third pairing strategy, to pair the fourth contact with the fourth agent. The results of the fourth contact-agent pairing during the fourth time period are recorded by the at least one computer processor using the fourth pairing strategy; The at least one computer processor determines a first performance metric associated with the first matching strategy based on the results of the first contact-agent pairing during the first time period; The at least one computer processor determines a second performance metric associated with the second matching strategy based on the results of the second contact-agent pairing during the second time period; The at least one computer processor determines a third performance metric associated with the third matching strategy based on the results of the third contact-agent pairing during the third time period; The at least one computer processor determines a fourth performance metric associated with the fourth matching strategy based on the results of the fourth contact-agent pairing during the fourth time period; as well as The at least one computer processor outputs data capable of comparing the first performance metric, the second performance metric, the third performance metric, and the fourth performance metric; The output is based on minimizing the effects other than those of the first pairing strategy, the second pairing strategy, the third pairing strategy, and the fourth pairing strategy.

2. The method according to claim 1, wherein, The first time period, the second time period, the third time period, and the fourth time period occur within a period of less than 24 hours.

3. The method according to claim 1, wherein, The pairing strategy switching is based on minimizing the effects other than those of the first pairing strategy, the second pairing strategy, the third pairing strategy, and the fourth pairing strategy.

4. The method according to claim 1, wherein, Minimizing the impact is based on selecting a time period of less than one hour.

5. The method according to claim 1, wherein, Minimize the impact based on the contact phone number or a portion thereof.

6. The method according to claim 1, wherein, Minimize the impact based on the contact identifier or a part of the contact identifier.

7. The method according to claim 1, wherein, Minimizing the impact is based on random or pseudo-random processes.

8. A system comprising: At least one computer processor communicatively coupled to and configured to operate in the contact center system, wherein the at least one computer processor is further configured to: During the first time period, the first matching strategy is used to match the first contact with the first agent. Record the results of the first contact-agent pairing during the first time period using the first pairing strategy; During a second time period later than the first time period, a second pairing strategy, different from the first pairing strategy, is used to pair the second contact with the second agent. Record the results of the second contact-agent pairing during the second time period using the second pairing strategy; During a third time period later than the second time period, a third pairing strategy, different from the second pairing strategy, is used to pair the third contact with the third agent. Record the results of the third contact-agent pairing during the third time period using the third pairing strategy; During a fourth time period later than the third time period, a fourth pairing strategy, different from the first pairing strategy and the third pairing strategy, is used to pair the fourth contact with the fourth agent. Record the results of the fourth contact-agent pairing during the fourth time period using the fourth pairing strategy; A first performance metric associated with the first matching strategy is determined based on the results of the first contact-agent pairing during the first time period. A second performance metric associated with the second matching strategy is determined based on the results of the second contact-agent pairing during the second time period; A third performance metric associated with the third matching strategy is determined based on the results of the third contact-agent pairing during the third time period. A fourth performance metric associated with the fourth matching strategy is determined based on the results of the fourth contact-agent pairing during the fourth time period; and The output is data that can compare the first performance metric, the second performance metric, the third performance metric, and the fourth performance metric; The output is based on minimizing the effects other than those of the first pairing strategy, the second pairing strategy, the third pairing strategy, and the fourth pairing strategy.

9. The system according to claim 8, wherein, The first time period, the second time period, the third time period, and the fourth time period occur within a period of less than 24 hours.

10. The system according to claim 8, wherein, The pairing strategy switching is based on minimizing the effects of the first pairing strategy, the second pairing strategy, the third pairing strategy, and the fourth pairing strategy.

11. The system according to claim 8, wherein, Minimizing the impact is based on selecting a time period of less than one hour.

12. The system according to claim 8, wherein, Minimize the impact based on the contact phone number or a portion thereof.

13. The system according to claim 8, wherein, Minimize the impact based on the contact identifier or a portion of the contact identifier.

14. The system according to claim 8, wherein, Minimizing the impact is based on random or pseudo-random processes.

15. An article comprising: Non-transitory computer processor readable medium; as well as Instructions stored on the medium; The instructions are configured to be readable from the medium by at least one computer processor communicatively coupled to and operating in the contact center system, thereby enabling the at least one computer processor to operate in order to: During the first time period, the first matching strategy is used to match the first contact with the first agent. Record the results of the first contact-agent pairing during the first time period using the first pairing strategy; During a second time period later than the first time period, a second pairing strategy, different from the first pairing strategy, is used to pair the second contact with the second agent. Record the results of the second contact-agent pairing during the second time period using the second pairing strategy; During a third time period later than the second time period, a third pairing strategy, different from the second pairing strategy, is used to pair the third contact with the third agent. Record the results of the third contact-agent pairing during the third time period using the third pairing strategy; During a fourth time period later than the third time period, a fourth pairing strategy, different from the first pairing strategy and the third pairing strategy, is used to pair the fourth contact with the fourth agent. Record the results of the fourth contact-agent pairing during the fourth time period using the fourth pairing strategy; A first performance metric associated with the first matching strategy is determined based on the results of the first contact-agent pairing during the first time period. A second performance metric associated with the second matching strategy is determined based on the results of the second contact-agent pairing during the second time period; A third performance metric associated with the third matching strategy is determined based on the results of the third contact-agent pairing during the third time period. A fourth performance metric associated with the fourth matching strategy is determined based on the results of the fourth contact-agent pairing during the fourth time period; and The output is data that can compare the first performance metric, the second performance metric, the third performance metric, and the fourth performance metric; The output is based on minimizing the effects other than those of the first pairing strategy, the second pairing strategy, the third pairing strategy, and the fourth pairing strategy.

Images