Given the themes covered in my last few posts, the topic of service levels seems very timely. During the course of recent research, the rationale around service level metrics keeps raising new questions, and it’s due time to explore them.

I’ll begin with the service level concept and who is best served by the commonly used metrics. Service level defines the target of performance, primarily in terms of contact center operations. Defining good performance and managing good performance, of course, are two different things, and this is where the challenges begin.

Contact centers are under constant pressure to run lean on agents, and this relates to service level in two ways. First, there is the fundamental need to balance head count to deliver the optimal level of service at an acceptable operational cost. Secondly, to keep those costs in line, there is a growing trend to outsource and/or offshore to access a pool of cheaper agents who typically are not fully dedicated to your customers.

Service level metrics have traditionally been operations-based, meaning that decisions around staffing – both how many agents and where they are based – are driven by what’s best for the contact center. The classic 80/20 service level target is a prime example of this: ensure that 80% of calls are answered in 20 seconds or less. That metric speaks well to efficiency, but hardly reflects the outcome of the overall experience, especially in the eyes of the customer.

In that regard, the 80/20 service level is in the same category as other standard measures of operational performance, such as TTR – time to resolution, AHT – average handle time, FCR – first call resolution, or agent adherence. Essentially, these metrics are within the control of those managing or supervising contact center agents. This also means these are the metrics they themselves are evaluated on, upon which rest their prospects for advancement and improved remuneration. As such, it’s no wonder why these metrics matter so much to supervisors.

The bigger issue is that the ultimate impact of contact center performance on business performance is increasingly being driven by factors beyond the control of supervisors. As customer expectations grow more demanding, it becomes easier to question the importance of answering calls within 20 seconds.  Can you really justify this level for all customers at all times? Is speed of answering really the most important thing? If speaking to a live agent is really the starting point for providing quality service, does it matter if the agent picks up the call, or if instead, the agent calls the customer?

Factors beyond the control of supervisors largely rest with the agents, and how they handle customer inquiries. Supervisors can certainly monitor calls for quality assurance, but customer satisfaction will be the outcome of agent-customer interactions. Supervisors can easily help their own cause by delivering solid operational metrics, but that is no guarantee of what matters most to the business – customer satisfaction, customer retention, share of wallet, intent to buy in the future, willingness to recommend, etc.

A strong case can be made for delivering best-in-class service when wait times exceed 20 seconds. The length of time waiting is quickly forgotten if the agent handles the situation effectively. During peak times when call volumes are beyond what the agents can handle, the pressure will be intense to answer all calls within 20 seconds. If that’s the overarching metric, agents will do that for schedule adherence, but odds are they will deliver a below-par quality experience for the customer. Is that really the best way to run a contact center?

In today’s environment, there are more tools and options to get the best of both worlds – solid operational metrics for the contact center, along with solid results for the business. Service level performance has always been very telecom-centric, and based on the traditional mode of inbound calling. As we move into a world of multi-channel communication, that becomes a limiting model for both agents and customers. This holds true on two levels.

First, reasonable wait times will vary by mode; 20 seconds may be acceptable for inbound calls, but email can be longer, while text-based modes will definitely be shorter. As such, just because calls aren’t being answered within 20 seconds, doesn’t mean the agent isn’t doing a good job. In this regard, current service level metrics are not only inadequate, but can be very misleading. Either service level needs to be re-defined for a multi-channel world, or new metrics should be added to provide a more complete picture of operational performance.

The second aspect of multi-channel communication is additive and further amplifies the need for change. Multi-channel is not just discrete – agents can use multiple modes concurrently to deliver service exactly the way the customer wants to interact. As social media finds its way into the contact center – either though the front door or through the back door via Marketing – one mode can be used to reinforce another, often in real time.
 
Virtual queuing is a prime example of this. During peak times when the 80/20 level cannot be reached, customers could be given this option for a callback to them when the next agent is free to do so. This is often an acceptable tradeoff since they no longer need to wait on hold. In contact centers with the right reporting and analytics tools, the uncertainty around when the callback will happen can be effectively managed. Once the callback request has been entered in the queue, the customer could be notified by text or chat or IM as to the expected wait time. Furthermore, to ensure the right agent calls back, they could use this channel to provide some specifics about their issue.

Regardless of how long the callback takes, the customer can still have a positive outcome, so long as expectations are properly set and managed. With multi-channel options, this is feasible in ways that a telecom-centric model could not support. Not only that, but virtual queuing provides real leverage for supervisors to right-size staff levels. The cost savings from a slightly reduced head count will easily outweigh the cost of supporting multi-channel tools as described in this example. This isn’t the place for a cost-benefit analysis, but a call-to-action is in order. If your thinking around service level was formed before VoIP and Unified Communications came along, it’s time for a re-fresh. Agents don’t think this way anymore, nor do you customers, and neither should you.