Struggling with manual call logging in the call centre? Now Speech Analytics can help you do the job

As a customer service organisation, you most likely tell your agents to log and categorise their calls providing management with data-based insight into the customer service. If you do, you will know, that those numbers don’t come easy, but requires a constant attention from management, and extends wrap up time.

So, you have to ask yourself questions like: Is it really worth the effort? Or more precisely: Is the benefit of having those numbers worth the cost of extended wrap up time?

Logging 80 % of 80.000 calls still leaves 16.000 calls out of sight

This dilemma recently became very much vibrant to me, while visiting a contact centre. Here, the 150 agents desperately struggle to log as many as impressing 80 % of their 80.000 monthly calls. 

The contact centre is in fact a BPO (business process organisation), handling customer service for a client. And naturally this client has an urgent need for insight in the customer service, they are paying the contact centre for. Still even logging 80 % of the calls leaves 16.000 calls uncategorised – and completely out of sight of the client.

Agents struggle to find the right category

Consequently, the agents are logging the calls in a complex structure of more than 700 different categories and subcategories, defined by the client, attempting to get a deep insight in why the customers are calling.

In fact, the quantity of 80 % logged and categorised calls was a progress from 60 % last year – a temporary low peak caused by an extraordinary high volume of new hires in the contact centre and a focus on onboarding at the expense of call logging. 

Obviously, one has to recognise the contact centre for the energy they are putting in this.

The quality people are doubting the numbers themselves

Nonetheless I noticed a rising questioning amongst the persons I talked to in the contact centre of the value of the numbers and data derived from the call logging. Simply because they are perfectly aware of the harsh reality behind getting them. You might put it like, the client asks for numbers, and that’s what the client gets. And maybe not that much more. Let me just unfold my discussions with the contact centre a little bit:

1. Real world customer calls are not one dimensional
   Real world conversations of holds a multiple of themes, thus also calls to a contact centre. But given that a logging system is only one dimensional, agents must choose only one. That leaves a lot of insights out of sight, if you are only looking on the numbers afterwards.
2. Let two people listen to the same call and you will get two different views
   The human factor is as always, a bias. Let two people listen to the same call, and they would most likely log them differently. Logging a call, deciding which category among the 775 is appropriate for a conversation lasting say 10 minutes, requires some human interpretation. And that fact alone questions the validity of the manual call logging.
3. If you force agents to log their calls, they will find a way to do it
   I once heard about a contact centre wondering why it was always the first choices in the logging system, that got the most hits. Then they changed the order of the choices, and guess what, still the first choices got the most hits. Agents are normally under the pressure of time, so if you force them to log the call, they will surely do it, regardless of the correctness of the result.

So, do you have any alternatives to the manual call logging?

Of course, modern contact centre technology is starting to find new ways to automate the manual call logging. At the contact centre I facilitated a test alongside with the quality managers, trying to replace at least some of the call logging utilising the technology called Speech Analytics.

Speech Analytics can categorise 100 %

Given, that you record the customer calls, this technology, just recently prepared for Nordic languages, transcribes all the customer calls and provides the contact centre with almost endless possibilities for categorising as many calls, as you record. If you record 100 % of the calls, Speech Analytics can help you categorise all 100 %. 

Not a one-to-one alternative – but probably a lot more

I am not suggesting, that Speech is an easy one-to-one alternative for the complex call logging, as the one I met at the contact centre.

Speech Analytics can mirror some of your categories, though you must approach the categorising of calls a bit differently.

But you will get something else: A multi-dimensional categorising, providing you with the insights lost in the one-dimensional manual logging. You will potentially get every call categorised. And it will be done automatically in precisely the same fashion every time and without involving the agents – thus lowering the wrap up time and taking away the human bias of logging and interpreting – consequently judging all calls alike.