Key Metrics to Analyze the Gap Between Self- and Live-Agent Service
9 May, 2011
By: Eric CamulliA recent Ovum survey of 4000 consumers shows that 'wait times' and 'routing' are still the biggest issues for contact centers today.
Even with the continued growth of self-service in the call center, some customers continue to need, or prefer to speak with, a live agent. What happens when they transfer from self to live-agent service is often a weak spot or “gap” in the continuum of customer touch-points with your call center. This is especially true with the increasing use of voice and web self-service strategies. Best practices today call for bridging that gap when it exists, typically with a virtual queuing solution. But how do you know if and when your call center has this problem? That is the issue addressed here.
Many call centers rely on their Average Speed of Answer (ASA) metric when reviewing the Automatic Call Distributor (ACD) data from their phone system. However, glancing at a daily or weekly total can mask long hold times that occur during lunch periods, training sessions, Monday mornings, or whenever conditions such as a marketing promotion, a recall, inclement weather or other events stimulate a high volume of inbound calls or a decrease in the number of resources available to handle customer demand. To determine if you have a problem with customer hold, better metrics to look at closely are Maximum Hold Time (MHT) also called Max Delay or Longest Call Waiting, as well as Abandons, and Average Handle Time (AHT) during these intervals.
Maximum Hold Time. Let’s take the example of a call center that has a 30-second ASA. However, when the MHT is reviewed, it is found to be 23 minutes. Additional analysis indicates that even though 90% of calls are being handled quickly, long hold times are occurring about 10% of the time in various pockets of time throughout the day. The result is that while ASA looks acceptable or even very good, 10% of inbound callers are extremely frustrated with the customer service they are receiving.
These callers are likely to report their negative experience with your customer service within their sphere of influence -- to family, friends and co-workers -- either in person or through any number of today’s commonly-used social media channels such as Facebook, Twitter and blogs.
Even with only 10% of callers dissatisfied, the potential damage to your company’s image and cost of loss of customer loyalty can be huge. As a result, it’s important to manage your resources in a manner that allows better service for the 10% of customers who are experiencing this significant pain-point -- without over-staffing 90% of the time. Maximum Hold Time is an important indicator that there is a hold problem that should be addressed with an appropriate solution.
Abandons. Abandons are another key indicator of a hold problem. What better way for customers to express their frustration with you than to hang up moments before you were about to answer their call? The average speed to abandon is typically around two minutes for most queues, however, consumers have come to expect longer hold times when calling for technical support and will hang on even longer before giving up. If it’s a sales queue, the tolerance for waiting is less, usually around 60 seconds, before the feelings of buyer’s remorse begin to creep into the mind of the customer. In some cases, waiting on hold for 60 seconds may be all the excuse that’s needed to abandon the purchase completely or try somewhere else. In some situations, the customer must call back because switching brands is difficult for them. However, the time it takes for customers to perform the process all over again is frustrating and this is exhibited in their behavior once they reach an agent through their tone and attitude. The length of the call increases because of the consumer’s need to vent this frustration.
Average Handle Times. The increased length in talk time between customer and customer service representative is especially a problem in technical support and retention queues where complex problems are being resolved. Longer talk times lower agent productivity because they handle fewer calls in an interval of time, backing up the queue even further. This exacerbates the problem and makes it a key indicator of hold problems. Why?
Frustrated callers take time to soothe. Often, agents need to spend several seconds at the beginning of such a call allowing the customer to vent about their wait on hold and to calm down. Further, “loaded gun calls,” which sometimes result when a caller is angry about hold time, may increase the immediate post-call period (often referred to as wrap-time) which is considered part of the AHT. This happens when it is necessary for an agent who has just handled a high-stress call to take a moment to breathe and re-group before accepting the next call.
Voice of the Customer. In addition to looking at these metrics, another way to determine if you have a problem with hold time is to survey your customers or conduct sessions in which you listen in on calls from the point of entry into the system all the way through to conclusion. You also can see what customers are saying about you on social media sites such as Twitter and Facebook, or by receiving Google or Yahoo! Alerts aimed at picking up blogger comments about your company. Our company, Virtual Hold Technology, monitors the Twitter feed “on hold with” at http://search.twitter.com/search?q=%22on+hold+with%22. This tweet-stream is often laced with profanity, which is indicative of consumer sentiment regarding waiting on hold. Or you may choose to dial into your system as a customer would to experience, like your customers, what happens when they transfer out of self-service into an agent queue.
Solutions. Most call centers will set a goal for each queue type for customer hold time, based on what their data indicates is considered a reasonable wait for that particular customer type. For example, the goal for a call center that handles only sales queues may be to answer 70% of calls with less than 30 seconds of hold time. This 70/30 scenario is fairly common. This particular call center might require 80 agents to achieve its 70/30 goal, while it could require 200 agents to service 100% of inbound callers with no more than 30 seconds of hold time.
But what about that 30 percent? We’ve already noted that even a 10% customer dissatisfaction rate can be enormously costly to a company. Virtual queuing is a safety net for managing call spikes without staffing to peaks. It bridges the gap when customers transfer from self-service to live-service so they don’t fall into a holding trough while waiting to speak to a live agent. This not only reduces the toll minutes your company pays for customers waiting on hold, but also allows call centers to find and stay in their staffing sweet-spot and control labor costs.
In most real-use situations, virtual queuing isn’t offered to every caller. It may be offered only to the 20%-25% of callers who are likely experience long hold times. Nonetheless, it will still provide a significant Return on Investment (ROI) for the call center.
A recent Wellness analysis and report consolidating the results of 19 Department of Labor call centers indicates that about 67% of callers offered Virtual Hold opted to receive a return call rather than wait on hold. Peak ASA improved 66% to 3.8 minutes, down from 11.1 minutes without it. Peak Service Level improved 63% while peak Abandon rates improved by 61%. These metrics indicate significant improvements in call center performance.
