Cogitations
Data analysis of call centres | Data analysis of call centres |
I have a friend who runs a call centre, and I wondered if there would be any benefit to him in having text files of every call his staff make or take. This was after reading about the latest advances in voice recognition technology.
A text file is a searchable database, and I already had some thoughts about what you could do with that with regard to analysing the content and style of postings on online forums. Would there be any benefit in doing the same thing with customer service calls?
Well, for starters, if you're dealing with outbound telemarketing you can have a program that checks for the %age similarity between what the telemarketer is saying and what the script says. This will automatically alert the management if someone is straying from what they are supposed to be saying.
Maybe more importantly (especially if you don't like telemarketers) are the things you can do with inbound customer service calls. Currently, a lot of companies outsource their customer service to people who are paid to deal with as many calls as they can as quickly as possible. The objective is to keep the calls as short as possible and to minimise the cost to the company. The people who answer the calls are under pressure to 'resolve' the issue rather than 'escalate' it to someone else, and their reports are usually not read by anyone.
This means that few companies use their customer service lines to gather data about what their customers are telling them - they don't listen to the feedback from the marketplace. There is usually no mechanism for the company to find out the specifics of customer service calls.
But imagine a fairly stupid program that simply scans the text of each and every call and lists all the words that are used. We might see, for instance, that the phrase "doesn't fit" appears repeatedly and so we take a look to see the context and discover that there's a small design flaw with the product that is causing a problem. The customer service call only deals with the immediate problem, but the customer will tell us what the cause of the problem is - if we listen.
I'm sure there are many other ways that call centres could benefit from analysing what is said to and by their staff, and another angle on this concerns a slightly different use for the voice recognition software.
This program has to be trained to recognise different accents etc. So it must be doing some analysis of the voice, and presumably it can be tweaked to recognise the emotional state of the caller. Could you use this to say "operator A doesn't follow the script and approximately 30% of the way through the call the customer is no longer engaged"? Or "operator B seems to be good at dealing with angry customers, and tends to use certain phrases to achieve that", and so on?
|
| < Prev | Next > |
|---|