Here’s the bottom line. Speech analytics is a must-have solution if you are willing to increase the effectiveness of your contact center. DMG Report’s estimation shows us that the adoption rate of speech analytics on a worldwide basis was 35% as of 2019, assuming that there were 19,5 million contact center seats at the end of 2018. Close to 7 million seats benefit from this technology.
However, you must have a clear vision of what you’ll do with such a solution like this. Most of the customers can’t get enough of speech analytics because they don’t know what to do and how to do with the valuable data they have in their hands.
I’ll try to give a couple of tips that can be useful if you have, or plan to have speech analytics;
WHAT TO DO
First, you must define what you expect to do with speech analytics. These expectations generally lead you to three basic outcomes;
- Cost optimization
- Increase Agent performance
- Increase customer satisfaction
So, you must think of what to do to reach these outcomes. For example, ATT (Average Talking Time) analysis leads you to take actions for reducing overall ATT averages and these actions’ outcome is cost optimization. Setting up AQM (automated quality management) forms and integrating them with your agents’ performance management leads you to agent performance increase. Thus, you must have a couple of clear and outlined plans before jumping on the speech analytics world.
HOW TO DO IT
To get the most out of speech analytics, you’ve cleared your targets and now you know where to go. To reach your destination, there are 3 steps that you must go thru:
Preparation of data: You have a huge amount of transcribed data, that’s brilliant. You can see every call’s transcription by double clicking on it and yes, it is very fancy. However, the data is still huge, and you must structure your data to make sense of it. Speech analytics products offer different features like queries, topics, categories and by using these features. You can find the complaining customers and see in which call category your customers are complaining the most or see which of your calls categories have the highest ATT averages.
This is important because that you can extract specific types of data sets from thousands, or sometimes millions of interactions.
Getting deeper: Imagine that you’ve found out you have the highest ATT averages in a certain call category. Now what? Most of the speech analytics users waits for the speech analytics systems to hold their hands and take them to the action phase by themselves. In truth, speech analytics only shine a light on your path to action phases. You have to plan and take these steps.
After you have found a lead like above, you can use the root-causes analysis features of speech analytics products to dig out the insights. At this step, speech analytics show you what to find such as the words/sentences that are being used the most. Or point out thee conversational metrics like hold duration averages, silence ratios etc. You are to decide which makes sense or which doesn’t. Because no other product or team knows your operation better than you.
Take action: So, you’ve found the main reason behind the call category which has the highest ATT averages. Let’s say that agent desktop applications work very slow and because of that you’ve seen the average hold duration is high, agents generally use phrases such as “my screen is slow” etc. Now it’s the time to act. That’s where you shake hands with speech analytics and thank it for its services, because now the ball is in your court to fix this problem with your IT department. You have solid proof about the problem, and it is your responsibility to organize related teams on how to fix it. If you take action, observe and see that the average ATT for this certain category has been decreased, congratulations, you’re ready for your next mission with speech analytics.
Author: Fahrettin Yılmaz, Business Analysis & Consultancy Team Leader, Sestek