How to Analyse Conversations and Make Better Business Decisions

A lie detector test is used by crime agencies to get to the truth of what an undercover agent is saying. The conversation is recorded and a graph spikes when lies are told. Using a voice first approach to your business and ensuring that people agree to appropriate privacy statements means you can start recording conversations and, then, get to grips with what is being said in the business.

Just as a company owns its email and phone records so too should it have ownership of the proprietary content of voice mails, voice notes and voice messages. We are, after all, on working time and work is paying the bills. In a world that when and where we work is changing, voice messaging is justifiable for check-ins with remote workers and better connected teams.

Every business has an organisational structure which perhaps is updated annually and shows who reports to whom. What it doesn’t show is who speaks to whom and how often. No doubt some people will talk more than others and will have solid lines of communication rather than dotted lines. Furthermore, some people will ignore the organisational hierarchy and go over managers’ heads to have conversations. Depending on your company culture, leap frogging the hierarchy will have varying degrees of acceptability. I personally believe everyone should be able to talk to anyone from the cleaner on the shop floor to the CEO and Chairman.

To start with, understanding who is conversing the most and listening the most is useful for active participation in your business. This does not speak to the quality of the conversation, but the frequency of listening and speaking gives you some level of understanding around organisational relationships.

Now a person could be listening and talking a lot, it could be office politics or gossip, and this is where word counts come into the analytics. What words are being used, are they negative words such as can’t and won’t or are they positive like we will or we shall. How many open ended questions are being asked can be discovered by measuring the frequency of What, Why, Who, Where and How in conversations.

Analytics to detect the mood of a conversation and hence its energy can also be reviewed. So a high energy conversation using positive words and lots of open ended questions is what we are aiming for if we are to increase the quality of conversations in our business. Furthermore, users who are managers should be listening for a greater duration of time. Then, they are speaking and the analytics can show the listening to speaking ratios which should be 2 to 1. Twice as much listening as talking.

Wordle analytics can also be used to determine which customer names are coming up the most. This might then indicate a big opportunity which has landed or perhaps a customer complaint, which needs to be resolved. Likewise, location and names of cities can be analysed to determine what the big issues are and drill down into conversations from a bird’s eye view. Clicking on ‘New York’ lists all the conversations with ‘New York’ in the sentence can be played back for further investigation. Green words depict positive conversations, amber for indifferent conversations and red words for conversations which are negative in nature, word composition and tone.

How much a conversation branches can flag problem identification or opportunity uncovering. Conversations which do not branch and which are one line statements might simply be the CEO reading the news to be broadcast one way rather than a highly engaged brainstorm by employees. The number of likes, listens and replies can also surface more interesting conversations recorded.

Just as word counts are used by students writing essays, word counts of conversations can be utilised to coach staff into more effective communication. People who need to create multiple recordings of say 20 second clips might have a problem staying on point or being concise. These people can then be given the appropriate coaching and feedback.

On a more serious note, you can protect your business from malice by using indexes of words with fraudulent or other serious consequences, for example, password sharing which goes against security policies. Indeed people will start logging onto applications using face and voice id in the future to protect your business.

Recording conversations and analysing their content needs to be seen as a self improvement tool for employees, a quicker way to remember and administrate a business and a tool for management to make smarter decisions. There will be nay sayers who will flag conversation around privacy and personal freedom but the hope is that the benefits to the individual and organisation to foster better more meaningful conversations will surpass these concerns. To do this, the voice first technology has to have the WOW factor. As artificial intelligence advances, we will come to a point where the machines give us scripts to suggest the questions we should pose and will coach us on how to say them! If this means we are more effective in our jobs and we get paid more, people will be happy with this and put the privacy concerns to one side. Conversations recorded will also help to avoid disagreements about what was agreed because playing back audio will bring clarity to conversations just as a pond filter clears out the dirt from murky waters!


Sean GilliganAuthor Bio

Sean Gilligan is a UK based entrepreneur and author of the book “Flexible”. Sean for the last 15 years has run Webanywhere in Chicago, Leeds and Katowice Poland. Sean is number 67 in the Worldwide Listing of Corporate Learning Movers & Shakers 2018. Sean is a bootstrapper and has not taken on outside capital to grow Webanywhere in 3 countries and has recently founded Ventures Anywhere his start up arm which includes Sound Branch.