You can’t improve what you can’t measure, the saying goes. Software tools have largely solved the problem of information capture and measurement. But information itself is largely useless – unless it is interpreted and then used to address business objectives.
Contact centre managers know the drill. Looking to drive continuous improvement, steer training and unearth new sales opportunities, managers turn to Interaction Analytics systems to get answers to burning questions from data in recorded telephone conversations, emails, web chats, and even social media.
However, managers come up short when they don’t have the time or skills to interpret the information and use it to change behaviours and processes impacting business objectives.
It’s one thing to invest in an analytics tool and quite another to do the job well.
What’s going on?
Unlike, say, consumer apps, which are ready to go from the moment they’re downloaded and opened, analytics tools are more like CRM – an operating architecture and templated processes that must be fitted to users and the wider business.
Simply processing phone calls, emails, social media or webchat will only add to the mountain of data.
The key is to follow a managed approach that combines speech analytics and contextual metadata to organise, analyse and, most critically, operationalise discovery and change. It’s a dynamic process requiring ongoing refinement, targeting and assessment. Only then can answers that drive performance improvements be found.
Approaches to speech analytics
There are two technology approaches to speech analytics: transcription and phonetic analysis.
Transcription uses language models to render text from spoken words and phrases. Text is analysed alongside other sources (i.e. email, chat, instant messaging and social media), uncovering emerging topics, quantifying related phrases, and categorising interactions according to topic.
Phonetic indexing is based on phonemes – actual sounds that make up language. Combined with analytical processing, phonetics is the quickest way to analyse information contained in the audio files.
Both approaches have pros and cons. Transcription is slower and must be reapplied to audio for every search. However, it provides insight without predefined words and phrases. Phonetics, on the other hand, is extremely quick, more accurate, and searches can be re-run from the phoneme database.
A combined approach delivers the best results, with transcription driving the formulation of a hypothesis, and phonetics gathering empirical data to quantify the size and shape of the problem, and the impact of improvements.
Doing a good job of analytics
Interaction Analytics done well will increase contact centre revenue, reduce costs, and improve customer experience.
On the revenue front, analytics tools identify critical agent skills driving successful sales or collections. They also help precisely target training and coaching to focus on those key skills. And they continuously monitor conversations to gauge skills development.
Operating costs tumble when analytics tools are used to identify and remedy the root cause of high Average Handle Times (AHT). They also pave the way for automation, by identifying types of contacts ripe for self-service methods.
And when contact centre managers eliminate key causes of dissatisfaction, lifting the rate of first contact resolution (FCR), customers will notice.
Contact Pyrios to learn how these powerful tools will make a difference to your business.
Author: Paddy Neill