Differences and Conflicts Between Metrics of Response to Customer Service Calls

First Call Resolution (FCR) is a key performance indicator for service call (contact) centres. The indicator pertains to success in achieving a resolution to a problem or an issue raised by a customer in his or her first call made to a contact centre. The metric of FCR success rate (%) usually concerns phone calls, though today we may also include voice chats over the Internet (with a human service agent). However, the resolution criterion competes with another critical criterion of the time duration of a service phone call, related to a metric of Average Handle Time (AHT). The FCR and AHT objectives often stand in conflict between them when setting priorities for the handling of service calls.

While the criterion of FCR has become a leading, more recommended indicator of performance, the time criterion of AHT remains a crucial concern for call centres that cannot be loosely waived off. The AHT is considered more of a cost-driven, company-side indicator whereas the FCR is recognised as a service-driven, customer-side indicator, associated more broadly with the quality of customer experience. Trying to meet objectives with regard to both metrics simultaneously can create tension between FCR and AHT, that is between resolution and time, that service agents may find hard to overcome. Negotiating between resolution and the time taken to achieve it is a challenge that nonetheless has to be dealt with in aim to find a reasonable working solution.

The root of conflict or tension between FCR and AHT is that succeeding in resolving a problem for the customer can take time, first identifying the underlying cause of a problem and then finding a solution so as to fix the problem — that means extending the duration of the phone call, particularly on the first service call, which operators of call centres tend to be adverse to. Yet, for the customers, resolving an issue promptly is most often their primary interest, and is the main determinant of their satisfaction. Putting agents under stress due to which they end the first call too early, without real resolution, is known to lead to likely subsequent ‘returning’ calls by more agitated customers, and increase the overall AHT dedicated to resolving the same customer problem across calls.

Customers are often willing to allow more time for service agents to resolve the issues they presented during their first call, as long as they see progress and believe they can get their problem fixed by the end of that call. But there can be a limit to the time customers are able to afford extending their call even under these conditions. As the call prolongs (say beyond 20 or 30 minutes), the attention of customers is likely to decline and their cognitive and emotional resources get depleted (i.e., they may get nervous, frustrated, and less able to think logically about the matter). Therefore, there is a point in time when it should be completely reasonable and even desirable (for both parties) to terminate the first call, but with a clear schedule for resuming the interaction to conclude resolution of the same issue. The service agent should try and reach the customer as agreed after a given time without waiting for the customer to make the next call.

The FCR criterion should not be observed rigidly any more than the AHT criterion. Moreover, it is quite legitimate to set-up additional measurable goals, such as ensuring that an issue is resolved through a limited number of calls (e.g., up to three) and within a given time period (e.g., 48 hours); the total handling time appropriate may be specified according to the type of issue concerned. Thereof, a resolution objective can be updated and extended to apply beyond the first call, subject to the additional goals, as further explained below.

First, we should consider that customers more frequently turn these days to phone a call centre after they have tried unsuccessfully to resolve their issue through a self-service online tool or utility (e.g., in a website or mobile app, with a utility that may be ‘smart’, AI-enabled in any degree). Thus, the problems that service agents are asked to help with are likely to be the more complicated and challenging that customers could not resolve independently; the time that may be required to resolve such problems by the human service agents is likely to be longer, especially during a first call (i.e., a conflict with an AHT objective). Hence, the handling time criterion or metric should be less strictly enforced, and more flexibility should be allowed for splitting the service interaction into two (or more) calls, until reaching a successful resolution.

Second, there are different types of service calls (e.g., billing, orders & deliveries, technical problems) which vary in level of complexity and difficulty to resolve, and in the average handling time that could be required to achieve a successful outcome. Technical problems are commonly identified as the more likely to be troubling, complicated and time-consuming to fix (e.g., computer-based digital-driven devices and appliances, hardware and software issues). Splitting the interaction into two separate calls may be used to allow the agent to consult other staff or the company’s knowledgebase for a solution as well for a customer to follow some instructions and attempt yet again to resolve the problem on his or her own (if successful, it may result in even greater customer gratification).

• It may be additionally advisable to distinguish between customers who refer first hand to a service agent for help (i.e., more dependent on service assistance and support) and customers who turn to a service agent in the contact centre after failing to resolve the problem online by self-service digital tools (i.e., they might arrive more frustrated or agitated to the service call); the way a customer is treated may have to be adapted to his or her situation when making the call (e.g., the stage or state of service journey).

Talk time is part of handle time during a service call in which the customer and service agent actually spend in conversation between them. Sometimes the metric of average talk time is used as equivalent or interchangeable with average handle time (AHT), but as argued by technology and consulting firm TTEC (formerly ‘TeleTech’) this is a mistaken view (TTEC, 4 February 2022). In another part of the time during a phone call, handling the case may require the service agent to perform actions without talking with the customer, such as looking for information in the profile of the customer, seeking relevant background information and guidance in the company’s knowledge management system (knowledgebase) for resolving the issue, consulting with a colleague or supervisor, and taking actual steps to remedy the problem or issue raised; and after the call, the agent may have to perform ‘wrap-up work’. Hence, it is suggested that total handle time of a service call encompasses talk time, hold time, and post-call time. (Note: it may be added that occasionally it is the customer who is assigned to perform some actions, such as on one’s computer, without talking).

In its blog article, TTEC recommends to reduce the focus on the average handle time or talk time, though it proposes still some ways for saving on handle time, particularly decreasing hold time (e.g., automating routine and mundane tasks such as data verification, training agents). Focus in evaluating the performance of service associates in the contact centre should be placed more actively on rates of successful resolution (as in FCR) and customer satisfaction.

• The blog of TTEC suggests that the average hold time versus average talk time (or ‘handle time conversation’) becomes less relevant. Yet, there might be value in using a metric of ratio of talk time relative to hold time. The talking time of conversation should be dedicated to hearing an explanation of the customer about the troubling issue, and subsequently for guiding the customer or de-briefing about the solution the agent obtained. Hence, it is important to keep an indication whether the agent has given sufficient personal attention in interacting directly with the customer.

Nevertheless, when measuring customer satisfaction, it is advisable to address two dimensions of the service interaction: the outcome and the process (course) of experience. The outcome pertains primarily to whether the issue was successfully resolved, especially if achieved during a first call (i.e., FCR). Satisfaction from the course of experience of interacting with a service agent may address how pleasant it went through (e.g., if the customer was treated patiently, politely, and concretely). A pleasant, comforting experience may soften the disappointment from not resolving the issue on the first call (especially if follow-up handling of the issue is guaranteed), whereas a negative, annoying experience can spoil the satisfaction from getting the issue resolved (not all customers are forgiving of bad treatment no matter the outcome).

As new technologies entered the traditional call centres (i.e., digital, automated, AI-enabled), the scope and functions of these centres have changed, and created new capabilities. In particular the class of chat facilities, by text or voice, with a human or virtual (smart chatbot) agent, transformed call centres into contact centres, and their technological evolution is on-going. Consequently new and modified considerations have to be addressed. For instance, as in cases of transferring a call from a service agent to a supervisor or more senior manager (escalation), one has to account for ‘escalation’ from a virtual chatbot to a human chat (or phone conversation). These are multi-mode and multi-agent situations with effects on time and resolution metrics. Human agents and customers may utilise similar intelligent (AI-enabled) tools, explicitly or implicitly, to assist them in resolving issues. Furthermore, text-based chats allow a human service agent (and moreover a robotic agent) to maintain interactions with several customers in parallel; this is possible partly because it gives more freedom and flexibility to both the agent and customer for performing some actions with longer intervals between conversational exchanges. Hence, time becomes less crucial and the efficacy of the AHT metric is reduced for these chats.

The form and utilisation of metrics such as average handle time (AHT) and first call resolution (FCR) may have to be modified and updated as their relevance and efficacy changes with service practices, and new metrics may have to be introduced. It is essential to guide and train human service agents on how to work in accordance with these criteria and their metrics, even when the guidelines are not straightforward and may entail some combinations between the criteria. It should be for the good of companies. their employees (agents), and not least their customers.