Measuring Agent Productivity in an Omnichannel World
As customer call centers have evolved into customer care centers, managers have had to adapt with new strategies for going beyond solving customers’ problems to finding ways to delight them at every touch point. The rewards can be great — customer loyalty and brand advocates — but only if customer care is done right.
Making sure that happens involves a complex melding of best practices in technology, process, and agent performance. One element that brings all three of these factors together is an omnichannel approach.
Ensuring a smooth and consistent experience between channels — phone, email, text, and chat — is the key to success in omnichannel customer care delivery. But to ensure that success, you need to know what’s working, especially in terms of agent efficiency.
In the old voice world you measured the length of phone calls, how many calls per hour, and the percentage of issues resolved during the first call, pretty straightforward, but with agents using multiple channels in a given shift, those measurements no longer tell the whole story. Let’s look at the challenges of measuring agent activity and productivity in an omnichannel world and how you can overcome them:
Challenge #1 – Blending
Within the omnichannel world you might have each agent assigned to just one channel, which keeps measuring their performance simple. However, another approach is having any given agent assigned to multiple channels simultaneously. For example, an agent might take calls, but respond to emails or chats when call volume is slow.
While this approach serves to increase overall efficiency for the care center, a single agent may not be as productive as they would be using just one channel at a time, especially when asked to handle too many channels.
The first step to optimal performance is appropriate performance measurement. For example, let’s say you have two pools of agents handling inbound phone calls:
- Pool #1 handling only calls
- Pool #2 responding to emails in between calls
- Expectations for Pool #2 handle times should be adjusted accordingly, given that these agents need time to shift between channels and tools.
This scenario becomes increasingly complex as you add more channel types and combinations, and you must determine the level of granularity needed to be effective at managing your workforce according to contact load. The best approach is to be as granular as possible on initial setup, and then aggregate metrics as you choose to ignore certain levels of differentiation to simplify ongoing management.
Challenge #2 – Simultaneous Interactions
Simultaneous interaction handling is not recommended. Agents juggling more than one customer at a time are more likely to exhibit lapsed concentration, mistakes in execution, and accidental sharing of private information.
However, there are care centers out there pushing those boundaries, so some agents will be handling multiple channels simultaneously. These agents may be engaged, for example, on a phone call and a chat at the same time.
So, in a five-minute window, if a given agent is talking on the phone to one customer and chatting with another, the appropriate measurement of time spent might be five minutes for each interaction, for a total of ten minutes.
This means that an agent could potentially have more interaction time (say six hours) than their actual shift time (four hours). Managers may need to adjust the processes and tools they use to accurately measure results.
Challenge #3 – Virtual Queuing
When no agents are available, virtual queuing allows customers to schedule a call back when call volume is low or hold their place in the queue and call back when it’s their turn. This method is more convenient for customers, but — to ensure it works most efficiently — care centers must properly attribute the queue time of the interaction even when there is no call physically waiting in the telephony environment.
With falsely short queue times, measurement systems may offer lower numbers than expected as there appear to be more calls handled within acceptable queuing thresholds. It’s also important to note that virtual queuing strategies (and the metrics used to measure their effectiveness) only work if there are down times during which agents can make return calls.
Invest in the Right Technology
The key to accurate measurement in the omnichannel environment is the right technology. Granularity is important because you must be able to group by agent skill set or interaction type combinations. Those base metrics can later be combined to achieve the desired level of reporting.
Therefore, it’s prudent to seek out tools, such as Aria’s Visualizer, that capture and display data at the distinct interaction level. The key is to get reports of all events (including interactions, routing, and meta data) within an environment, so you can later analyze that data in ways that are meaningful to your care center and your company.
You also need to be able to accurately capture presence data, and precisely attribute handle times to each channel method. The right tool for these tasks is one that seamlessly connects your customer relationship management (CRM) system, such as Salesforce, with your workforce management (WFM) system. This type of application can provide accurate metrics like the average queue wait time, or average interaction handle time (including all channels), which help with forecasting and staffing.
Finally, having all channels operating within one system can assist in measuring critical metrics. Commonly, an agent will be working within two or more channels, using a separate tool for each one. Because the systems are separate, so is the data gathering, meaning there could be inconsistencies in how performance is measured. Unified agent desktop technology helps in connecting all channels to one system in order to gather consistent, complete, and useful data.
Bottom line: the challenges of working within an omnichannel environment don’t have to outweigh the benefits. The investment you make in omnichannel operations may come back to you many times over, but only if you can optimize agent performance. The best way to do that is to recognize the challenges, and make sure you have the right tools to address them.