If you regularly analyse customer data, you’ll know how challenging it is to interpret them.
Businesses receive tons of customer data through phone calls and focus groups. Many also don’t know how to make sense of the data because speech simply can’t be nicely fit into spreadsheets or dashboards.
That’s where speech analytics come in.
NeoSound Intelligence specializes in using artificial intelligence and machine learning to understand what people say and how they say it and turn this into data for you.
One of our key breakthroughs in 2019 has been the ability to classify speech fragments into categories. Our clients have used this for market analysis, focus group polls, and customer call analysis.
Let’s take Greta Thunburg’s UN speech for example. She was named TIME’s Person of the Year 2019 for inspiring a global movement to fight climate change.
We did a little experiment where we ran her speech through NeoSound’s algorithm to pick out phrases in Greta’s speech that were pronounced with the highest amount of emotion which indicated points of emphasis. These phrases were then classified with a model of psychological needs that was developed by our partner Evgeny Zakharov, a psychology specialist and business coach.
The model is based on three key needs that drive a person in their daily life: Safety, Love and Respect. The phrases were mapped to personal values, which are linked to the psychological needs. For example, Life and Care were values categorised under the need for Security.
The word cloud below shows all the phrases identified. The green utterances are classified as the need for Security whereas the orange phrases are the need for Respect.
In her speech, Greta made 19 references to the need for Security and 17 references to Respect, including the ‘How dare you’ phrase, pronounced 4 times. Most speeches have references to all the three needs, but there was no Love from Greta here. The need for Security was expressed through word phrases such as “People are suffering”, “Science has been crystal clear”, and “evil”. Meanwhile, the repeated “How dare you” phrases, “will not let you”, “I refuse to believe” were calls indicating the need for Respect.
Greta went on and used utterances such as “Why do I have to deal with this” indicating resentment; “You should…” showed a certain amount of anger; and finally “I’ve been deceived… how dare you!” expressing hate.
Dramatic, and certainly Greta has revealed more hate and resentment but no love here.
Her use of language showed she was deceived, with people saying one thing and doing another one. Her speech gave the impression of a child whose expectations of a happy and carefree childhood were not met, and the resentment leads to her desire to restore justice to get what she was promised.
It’s not surprising that her speech got many strong reactions, both positive and negative.
If this is the depth of insights that can be gathered from one speech, think of all the possibilities you could achieve with speech analytics. You could analyse any number of speeches, customer calls, focus group responses. You’ll understand the customer’s needs in the categories that are important to you and make sure your business is making the right response across all touchpoints. Psychological needs are just one example of a model and you can run any other model you want to analyse your customers’ needs. For example, an automotive insurer can analyse customer conversations for the most frequent categories of complaints such as coverage, price and liabilities. Aggregating all the insights together would help you gain a whole new level of understanding of what your customers really want, so you can serve them better.
Greta’s speech was a cool experiment for us, and we’re testing more of Greta’s other speeches and expanding into speeches of other prominent figures such as Donald Trump. Let us know in the comments what speeches you would like us to analyse and share.
If this article was interesting, recommend and share it with all your friends and business associates. We had a lot of fun doing this and we are looking forward to playing with all the possibilities of speech analytics.