Emotional Intelligence Part 4: Supervisors Guide to Agent Trends
How are your agent trend statistics? Are you happy with your agents performances in their customer interactions? Indeed, do you even record and take notice of analytics relating to your agents performance?
There are a variety of statistics you can generate to give you a guide to your agent trends.
An obvious first guide to your agents performance is their work attendance records. Unless you operate some form of remote working system, allowing your agents to operate from home, there will be a clear correlation between your agents work attendance patterns and their work effectiveness. If their work attendance pattern is sporadic, it is most likely that the quality of their work will also be patchy.
Another basic means to gauge your agents efficiency is to compare their ATH (or Average Handle Time) to determine how long it takes to handle customer calls. You could compare different agents performances across the group, or you could look for historical trends in this statistic for each particular agent. For instance, if Agent A is managing five calls per hour, then on the surface at least, she is outperforming Agent Bs four calls per hour. However, if Agent B had only managed two calls per hour on average for the previous quarter, then this would have shown a positive agent trend.
Many call centers also survey their clients to gain feedback to gauge how well they believe an agent handled their calls. These are of course very subjective, as different clients will look at the concept of good customer service in differing ways, and unfortunately represent only a tiny sample of a call centers total operation. Furthermore, when customers rate provide an agent rating of 5 out of 10 or 3 out of 5, what can you actually do with that?
These basic measures are obviously an oversimplification of your agents actual performance. Quality has been notoriously difficult to quantify objectively. Getting to real performance can be difficult to measure, particularly in a busy call center. There can also be quite a time lag before agent trends become obvious. The time lag alone may in fact negate some benefits of remedial action you choose to take since the speed at which negative experiences are communicated are almost instantaneous with todays social media.
Predictive Voice Analytics can help a modern busy call center to spot agent trends quickly and alert supervisors of issues on a micro level before they become obvious through traditional means of call monitoring. Machine-learning platforms such as RankMiners Predictive Analytics can quickly become a busy supervisors best friend, by correlating emotions and behaviors to the business outcomes that matter for your company. RankMiners predictive models not only analyze every call, but will predict future outcomes and prescribe your operations next best action. You no longer have to rely on testing a small sample of calls. You can focus on the interaction that truly matters the one between your customer and your agent rather that relying on statistics that are easy to measure.
When you use RankMiners Agent Insight, CEOs and call center supervisors are able to get to the emotions and behaviors that drive customer experience, and are able to view that information across 100% of their teams performance. With highly actionable information, your management team can drill down to which teams, and which individual staff members, are helping your business and which are hurting your business so that you can intervene with intention and improve overall customer experience.
Keys to RankMiners Predictive Analytics Platform is that companies can use its machine-learning algorithms across multiple applications: from automating Agent Quality Monitoring, to increasing Agent Retention to measuring Customer Satisfaction. Supervisors can quickly see how the emotions and behaviors that drive customer experience are impacting their business. They can also address emerging trends before those trends become issues and cause potential damage to your business.