Guy Marson, co-Founder of data science and intelligence marketing company Profusion, discusses how understanding the voice of the customer can benefit businesses

The sheer volume of Voice of the Customer (VoC) products available today is explained by the importance businesses now place on being able to understand what their customers say and feel. However, these solutions aren’t created equal. With so many channels of communication open to customers and a number of ways to interpret and prioritise what they say, it is far from easy to determine what insights should be acted upon.

Today’s businesses sit on a mountain of data that they can use to deduce what their customers think of a product or brand, what’s important to them and their aspirations and needs. This includes contact centre data, interactions consumers have in store, and how they talk about your brand online through social media, blogs and the increasingly commonplace video blogs – vlogs.

Traditional VoC technology has long relied on determining the sentiment behind contact centre chats and calls. Text can be categorised into angry, sad, happy or other states using formal naming conventions, machine learning, and teaching a computer how certain words and concepts are related.

If all this sounds complicated, that’s because it is. Teaching a computer to analyse the way we communicate is no easy task. It is made much more complex by our different languages, slang, misspellings and, of course, sarcasm. Adding yet another level of complexity is the modern day habit of littering text with emojis, or, indeed, communicating exclusively through emoji. As a result, sentiment analysis is not a perfect, one-size-fits-all tool for determining your customers’ opinions.

Combining sentiment analysis with text mining and topic modelling techniques can help make sense of what your customers are saying, in whatever language or form they may be using. Text mining determines patterns in what your customers write to (or about) you. Meanwhile, topic modelling determines the main subjects your customers are talking about and groups them accordingly.

In other words, whereas with sentiment analysis you have to teach a computer the connections between different words, combining this with topic modelling and text mining – and a little machine learning – allows the computer to work these out for itself.

For instance, in working out the sentiment behind a body of text that contains the words ‘kills’ ‘bacteria’ ‘rarely’ and ‘fails,’ sentiment analysis alone will have determined the phrase to be negative. However, looking at the phrase in its entirety – “rarely fails to kill bacteria’ you can deduce that it is, in fact, positive. VoC technology using sentiment analysis, topic modelling and text mining will have revealed this.

Data from contact centres can be analysed to determine how well a certain product or service is doing in the eyes of your customers. If people have been contacting you to complain about a new or re-vamped product, you have a chance to offer alternative products, change the product design, or find another solution to their complaints. Likewise, receiving positive feedback on a product can tell you to increase production, marketing spend and what attributes you should perhaps consider in your next product design.

Hearing your customers in their own words will help you understand what makes them tick. This, in turn, can inform how you market to them and what products and services they may also be interested in. Conversely, this data can also be used to tell you how well your marketing and sales promotions are working and where you could improve.

Receiving feedback at all stages of the customer journey can also aid you in identifying pain points for your consumers and ways to fix this.

Finally, analysing contact centre interactions with your customers will provide your customer service staff with on-going feedback on their performance, customer satisfaction and training.

Obtaining the benefits of VoC programmes isn’t without some legwork however. To quickly profit from VoC technology, you need to have clear objectives for the programme, which aligns with your business aims. You will need to determine what information you have which you would like analysed. Data sources such as chat transcripts, survey data, transactional data, product data and even event calendars and planned product launches can all be analysed through VoC technology.

Knowing how your customers think, what they love and what they hate about you has widespread applications across your entire business. In an increasingly competitive business environment, companies who listen up will find they hold the attention and loyalty of their customers far better than their deaf competitors.

Interesting links:

Post Views: 904