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.


Chat Metrics For Those of Us That Don't Like Chat Support

I have a confession for you.  As a customer, I love the convenience of chat support.  As a customer service director, I hate it.  Sorry, I’m just being honest.  While chat support is a convenient means of receiving customer service, it can be a real headache to plan and schedule for. 

Comparing chat support to phone support, and the planning that goes along with it, is like comparing apples to oranges.  Where the average phone call might last eight minutes, the average chat might last twenty.  Consider also the fact that agents can handle multiple chats at once, and you see where it can get confusing.  How do you properly plan for this and then rest assured that your agents are being well utilized?

In our contact center at Phone.com, our chat volume is such that we typically get away with dedicating one agent to it.  It is, however, tricky when they take their breaks or we experience enough volume to necessitate a second agent.  Pulling a second person out of the call queue can pose a problem when trying to maintain service levels.

Here are some of the most useful metrics I have found to evaluate and manage chat support.  Please keep in mind that I am more of a student than an expert, and I reserve the right to ask for help at the end of this post.

Average Handle Time

In monitoring the handle times of our agents, we have found a much wider variance among our team compared to phone support.  You may observe that your faster typists have shorter handle times.  You may also find that some agents are more adept at balancing multiple chat conversations.  A good chat agent should be able to handle multiple chats concurrently without neglecting any one conversation.

Average Speed Of Answer

This metric is much trickier for chat support than it is for phone support.  Not only do you have to ensure that chats are answered in a timely manner, but you also have to monitor how many chats your agents answer at once.  In our contact center, we often rely on our primary chat agent to call for back up once they are handling three chats, with more waiting in the queue.  This can be a bit subjective at times and require a lot of communication amongst the team.  You also want to guard against two agents each working on one really long chat while your call queue backs up.  Never forget that chats can be transferred.

Concurrent Chat Time

When monitoring the utilization of our agents, we had to write a custom report for our system to see the actual time spent chatting.  If an agent takes four, ten-minute chats over a twenty-minute stretch, I care more that they spent twenty minutes concurrently chatting.  If you were instead to say four, ten-minute chats equals forty minutes of chatting; your utilization metrics will be flawed.  This also gives you an idea of how many chats your agents are handling concurrently on average.

Offline Chat Requests

If your chat system is anything like ours, when no agents are online for chat support, customers have the option to leave a message.  Those messages are logged in a completely separate area in our chat system.  Nonetheless, the requests need to be factored into your overall chat volume to accurately schedule agents.

Finally, never forget about quality when it comes to chat support.  You should have a survey at the end of each chat and you should absolutely be reading and taking action on every one of those.  Couple this with listening to feedback from your agents, and you should have everything you need to offer awesome chat support. 

Now, I told you earlier that I am not the expert when it comes to chat metrics.  Are there any metrics that I missed?  What metrics have you found most useful for chat?  If you tell me, I promise to hate chat support a little bit less.

More Resources