Call Survey Formula Customer Talks, AI Thinks
“Are you listening to me?”. I go into auto-pilot and respond “yes dear, that is a good point”. Then my wife goes back to explaining something. I am sure many of us daze off sometime in our life, but that cannot happen when it comes to health care. As I work on On-Call Ma – finding on-demand caregivers, one of the problems I am trying to solve is monitoring after hour calls. More importantly, On-Call Ma must determine if should bring the the call to a supervisor’s attention.
There are many times my wife catches me “multi-tasking”. Personally, I am jealous of her sixth sense that detects and validates if I am listening. One of the ways is by asking “what did I say?” We can also validate if a call went well in the phone system world by doing a survey after the call ends. Usually, we ask at the beginning of the call if it is ok to call back for a survey.
Not All Call Surveys are Created Equal
The less question you ask the higher chance the client will finish the survey. It is important to follow the “don’t make me think” concept because the more work you force into your audience, the more resistance they will have and result in not finishing your survey. We can have tactics such as breaking down the audience in half so we can ask two questions and still get a high survey completion. It is important to not confuse principles with tactics. The principle concept, such as “don’t make me think,” must always be there.
I replied to my wife “you were talking about your students. Wow, you really care about your students.” My wife continues expressing her thoughts, which means she confirms that I am listening. In the phone system world, we take the survey data and create some type of algorithm or call survey formula that gives a number that we can understand. For example, you scored 9/10 on your quiz, and the passing grade is 6/10. This means we have to answer at least 6 questions correctly.
Call Survey Responses Have Their Own Formula
It is important to know that in each industry, the passing grade may be different. A delivery logistics company expects a higher call success rate, compared to a collection agency. One way to find the passing grade in an unknown industry is to get the average result for a few months and use that as a base. Here is an example with a two-question survey: “Would you recommend us to others?” gets a Yes response (pressing 1 for yes) 800/1000. The follow up “Did the representative answer all your questions” with a yes response 850/1000 where 1000 is the total survey completed. Any week if we exceed a 800/1000 ratio, then we are performing as expected or better.
I should admit that my wife usually knows whether I am listening or not. She has mastered all my tactics to cover up my answer. She certainly has higher accuracy than a survey. Artificial Intelligent Tone Analyzer also has higher accuracy than a survey. IBM Watson is an example of a Tone Analyzer. What this product does is use speech to text to convert a call. Then analyze the words to determine if the person has anger, fear, joy, or sadness. As previously explained, we need to come up with a call survey formula that gives us results such as anger is 100/1000 where 1000 is the total calls for 3 months. Then we can see the performance in real-time.
“Why did I marry you, I will never know”, that is the response I get from my wife when I fail my performance test – “what did I say?”. In the business world, we give a bonus based on performance. If the performance is below expectations, then we need to make changes. The same applies to getting a divorce. I am happy to say that I am not divorced :0).
