You can access the Call Category Report directly by clicking the following link: https://pla.delaconcorp.com/siteui/reports/callCategory
For the Call Category Report, the Speech Analytics Module (SAM) places calls into specific call categories using one of two analytical techniques – the Call Categoriser, or Call Tagging. These two techniques use different methodologies to evaluate the transcripts of your calls.
During the setup of your Speech Analytics account, you will be able to define the categories that you want to analyse for. Below is a list of some call category examples:
- Sales – converted or not
- Existing or New Bookings
Each business or business unit will have different call categories depending on the industry and the specific business.
If a call doesn’t fit to any of the predefined categories, it will be placed in the General category.
Generating the report
To generate your report, select the date range you wish to analyse using the 'From' and 'To' date boxes, and select which Service you want to analyse by clicking the ‘Select Services’ button:
You will then need to select which Profile to use. In the case of the Call Category Report, Profiles allow you to select which analytical methodology you will use – the Call Categoriser or Call Tagging:
(To discuss which analytical methodology is best for your business, please speak to your Account Manager)
The Call Categoriser
During the set up of your Speech Analytics account, you will provide Delacon Support with a list of specific keywords and phrases, along with your required categories. Each keyword should relate to one of the categories. Each keyword is assigned a level of importance, related to its relevance to its assigned category.
The SAM can use this information to categorise each call. It will evaluate each transcript and identify each occurrence of each keyword. Using the frequency of use of each keyword in the transcript, and its level of importance, the SAM will calculate the probability of the call falling into a particular category. The higher the probability, the more likely it is that the call belongs to that category.
Often a call may have multiple categories; the SAM has the capability to identify not only the main call category, but also other relevant categories based on its call scoring and analysis system.
The Call Tagger works in a slightly different way. You create one profile per Tagger. During your set up for each Tagger, you will define a list of keywords that relate to it. You will also define how many times each keyword has to be mentioned in a call before it is considered to be a topic of the conversation. For example, for category ABC, you may define the following keywords and required frequencies:
You then define the minimum number of keyword conditions that must be met in order for a call to be placed in the category. In the example above, you can define that in order for a call to be placed in the category ABC, one, two, three, or all of the keyword conditions need to be met. The Tagger will then analyse each call, and count the mentions of each keyword:
|Keyword||Required frequency||Actual frequency||Condition met?|
In the example above, if the overall minimum condition was to meet one, two or three of the individual keyword conditions, then the call would be placed in category ABC. However, if the overall condition was to meet all of the individual keyword conditions, then this particular call would not be placed in category ABC. If a call does not meet your criteria for the category, it will be placed in the General category.
To discuss which analytical methodology is best for your business, please speak to your Account Manager.
You can access the Call Categoriser report directly by clicking the following link: https://pla.delaconcorp.com/siteui/reports/callCategory
To learn how to configure Speech Analytics on your account, read our Speech Analytics Configuration Setup Guide.
To see how Speech Analytics data can be analysed, please read our article on Call Visualisation and Transcription.
To learn more about Keyword Spotting, go here.