How do I calculate the Agent non-response metric in Amazon Connect?

3 minute read
2

I want to know how to calculate the Agent non-response metric in Amazon Connect.

Resolution

The Agent non-response metric increases when an agent doesn't accept an incoming contact, or the caller disconnects the call. Amazon Connect tracks the Agent non-response metric as a real-time metric and a historical metric.

You can use the contact trace records (CTRs) and contact ID to manually calculate the Agent non-response metric. The contact ID must have the queue name for the queue that you're calculating the Agent non-response metric for. The following attributes indicate that the agent missed the contact:

  • The Agent connection attempt value is greater than zero.
  • The initiation method is either Inbound, Transfer, or Callback.
  • No value is present under the Agent column, such as, a Connected to Agent timestamp or Agent interaction duration.

For more information, see Contact records data model.

The following scenarios increase Agent non-response metrics:

  • Each time Amazon Connect routes a contact to an agent, but the agent doesn't answer the call. If the agent doesn't answer the call, then Amazon Connect routes the call to another agent to handle. Because agents can miss a single contact multiple times (including by the same agent), you count it multiple times.
  • A caller disconnects when in a customer queue flow. For example, if the caller reaches a loop prompt and disconnects before the agent connects, the Agent non-response metric increases. By default, when Amazon Connect routes a call to an agent, the agent has 20 seconds to accept or decline the incoming contact.

Example scenario

For example, Amazon Connect transfers an inbound call to Queue A. Because Amazon Connect transfers the call to the queue, the customer queue flow runs. Two agents are available in Queue A, Agent A and Agent B.

A call initiates to Agent A for 20 seconds, but Agent A misses the call. The call then initiates to Agent B for 20 seconds, but Agent B misses the call. There are now no agents to route the call to, and the customer queue flow starts over. Then, Agent A becomes Available, and the call initiates to Agent A for 10 seconds, but the caller disconnects before Agent A accepts the call.

The Agent non-response metric value for this scenario is three.

The Agent non-response metric calculates a value of three because both Agent A and Agent B miss the initiated call. Agent A is available again for the initiated call, but the caller disconnects before the agent answers the call. Because the disconnected call rang for the agent, you calculate it as an Agent non-response.


Related information

How do I identify who disconnected a call in my Amazon Connect contact center?

How do I connect agents in my Amazon Connect contact center to incoming calls automatically?

AWS OFFICIAL
AWS OFFICIALUpdated a year ago
4 Comments

Is there a way to determine in historical reporting the makeup of the non-response category, i.e. How many did not answer and it rolled to the next person in the Missed state vs how many the customer hung up in the first few seconds of ringing before the agent was able to answer.

Our agents are getting dinged stats wise for not answering these calls, and some of them where the customer hung up while it was ringing the agent increases their non-response. So we would like a way to see Missed vs Customer Hung up.

Phill
replied a year ago

Thank you for your comment. We'll review and update the Knowledge Center article as needed.

profile pictureAWS
MODERATOR
replied a year ago

"The following scenarios increase Agent non-response metrics: -A caller disconnects when in a customer queue flow. For example, if the caller reaches a loop prompt and disconnects before the agent connects, the Agent non-response metric increases. By default, when Amazon Connect routes a call to an agent, the agent has 20 seconds to accept or decline the incoming contact."

--This is bad design, in my opinion. There are many scenarios in which a caller might decide to randomly hang up even though they are suddenly now ringing to an agent (I am guilty of doing this myself a few times).

Granted, it might be a low percentage of overall calls, but it does happen. Contact center admins and supervisors will need to know about this exception so they do not punish agents unnecessarily -- as many supervisors expect agent non response to be as close to zero as possible. ... and an agent armed with knowledge of this could even exploit it by ignoring calls and then telling the supervisor that a lot of their callers hung up before they could answer the call. Priority to fix this: very low in the grand scheme of things ... but it would definitely be nice if there were an extra column entitled "caller hung up while ringing" or something similar. ... and then we admins wouldn't need to explain this exception to supervisors anymore.

PBode
replied 10 months ago

Thank you for your comment. We'll review and update the Knowledge Center article as needed.

profile pictureAWS
MODERATOR
replied 10 months ago