ICMI is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


Kansas or Oz. The Difference Interaction Analytics Can Have on Performance Management

Measuring agent performance has never been an exact science. In fact, many times the process is onerous, time consuming and doesn’t yield the results you’re looking for. Anyone who has managed agents, attempting to provide the constructive coaching they need to meet their goals and provide the best possible service to our customers will know. Unfortunately, not having the information needed to do so often proved to be a nearly insurmountable challenge. 

Here’s how this process typically works:

Exception or Pattern?

It’s the first of the month and a supervisor, Matthew Ellis, needs to perform an assessment on an agent, Mary Stewart.  Five of Mary’s calls are pulled from the call recording software to review.  The problem is that these are random calls, and could be about any number of issues that Mary typically handles.

In this example, the company is a cable service provider and in the second reviewed call, Mary tries to help a customer who has no picture.  Mary schedules a service technician to go to the customer’s home, without attempting any troubleshooting steps. This goes against company policy and will affect her bonus if she has too many unnecessary “truck rolls.”

However on this call, the customer was impatient and claimed they had this problem before, so no one knows if Mary would have tried to troubleshoot had the caller reacted differently, or if this is indicative of the fact that she really needs more coaching or retraining on this topic.

A Fruitless Process

What’s needed are more examples where a customer calls in with a similar problem to see how Mary handles the issue, to gain a better sense of her skill set in this area.  The only way to find these examples is to set off on what’s essentially a wild goose chase, pulling additional calls from the recorder, listening to a few seconds of each one, and trying to determine if it fits the bill.  But the supervisor has 25 more agents to review and can’t afford to go through this fruitless process, so he gives up. 

When it comes time to review with Mary, they discuss the call, she says it was an anomaly, Matthew offers a few pointers, and honestly, hopes for the best.  Next month rolls around, and the supervisors want to see if Mary’s doing better with these type of calls.  Once again, he is mostly out of luck.  With only random calls to choose from, it’s hit or miss if the ones returned are related to a truck roll.  So as before, an assessment is done on the available calls, without knowing if she’s making progress in this key metric.

And at the end of the quarter, when the numbers come in, it’s seen that Mary does in fact have a higher than average number of truck rolls.  She did need more targeted coaching.  But not only did the supervisor not have good examples of her calls, examples of other agents who were doing it well to use as best practices didn’t exist either.  Finding those calls would have required an even greater hunting expedition.

Here’s how that scenario would have played out had the company been utilizing an interaction analytics system:
Matthew logs into a dashboard and sees his team’s results against key metrics – in this case, the metric in question is number of legitimate dispatch rates. Mary is doing ok, she falls into the mid-range zone, but there’s definitely room for improvement.

No More Wild Goose Chase

But the great thing was, he didn’t have to go on that wild goose chase searching for calls to determine this.  Because the metric for an acceptable number of dispatches had been determined, using an interaction analytics solution automatically generated those results by classifying the calls and identifying the calls where an appointment was set, which indicates a truck was rolled.

Since it was based on 100% of her interactions, the subjectivity was removed and there was no possibility the score was the result of one or two abnormal calls or something not indicative of her overall behavior.

A Priceless Gift

When it comes time to review with Mary, Matthew was armed with the best gift of all -- time. That’s because he didn’t spend hours searching for calls, and he knew that she’s struggling with her dispatch rate, he knew exactly where to focus his coaching. He played her examples of her calls where she got off track, and was even able to pull up best practice examples from coworkers who are doing really well.

The Introduction of Real-Time

But let’s say that as a company, it was discovered that even agents who tried to troubleshoot before sending a truck still ending up sending more trucks than appropriate. After using interaction analytics to isolate the calls, it was determined that they weren’t following the troubleshooting sequence correctly, so it never really stood a shot at working.  So what are the options? Everyone could be retrained, and that would probably improve performance, at least for a while. But it is a somewhat complicated process and it’s easy for the agents to get flustered when dealing with the customers live on the phone.  So a better approach may be to use real-time monitoring and alerting. 

Now that interaction analytics has determined the problem, the solution can be set to look for words that indicate when a customer is calling in for problems that indicate that the troubleshooting sequence should be initiated.  It can then send an alert to that agent’s desktop to remind them to try the troubleshooting before sending a truck, and walk them through the steps.  By first garnering knowledge from post-call analytics, real-time monitoring and alerting becomes much more valuable. 

The Results

Now back to Mary. Because of the personalized coaching, Mary is able to understand how she could do a better job with her troubleshooting and her rates improve.  This can be verified by quickly checking the same metrics next month and seeing the improvement, even before listening to her calls.  And since Matthew saw Mary struggling in this area early in the quarter and delivered the targeted coaching she needed, she was able to meet her goal, qualify for her bonus, and felt like a valued member of the team.  The company benefitted too, since their expensive truck rolls declined.

Final Thoughts

Making the switch to using interaction analytics for performance management is a bit like that moment in the Wizard of Oz when everything switches to color.  Things were ok in black and white, humming along as they had always been.  But having the ability to let the technology do the heavy lifting for you, so you can focus on bringing the human touch of coaching and mentoring to your agents so they can bring better service to your customers, is the color you’ve been missing.