How to improve Customer Experience by Vocal Assistant

The business world had to adapt the way through which consumers are engaged. A new form of technology arose during the past few years: vocal assistant. Firms have started to take advantage of it, finding in vocal assistants a new channel of sales for their e-commerce. 

Thanks to voice commerce, the customer experience changes radically: the whole process is faster, simpler and it can be enriched in many ways. It can also improve the consumer journey by deeply personalizing processes and offering enhanced client insights to marketers. Companies should figure out both benefits and drawbacks of such innovation. 

We have recently started a collaboration with the LUISS Guido Carli University, participating in some classes of the Bachelor’s Degree course in “Management & Computer Science” held by prof. Francisco Villaroel Ordenes.

The challenge wanted to test our young analysts on a real challenge that they will face up in one of their first job experiences. The goal was to solve a problem for its management, using the data offered by KPI6 on the market, on the players, and above all on consumers.

Students:

The goal of this challenge was to help managers improve CX by identifying all the influencing elements of voice commerce, which can even be exploited in order to increase their value creation.

For gathering data, we just had to write down the right queries to be inserted in the KPI6 search engine. We collected the Twitter posts containing the most important keywords, concerning voice assistants and voice commerce (also in form of hashtags). We then collected all the tweets with a tag on the official Alexa and Google Nest pages.

The first part of our work aims at finding out which are the negative and the positive most correlated topics, with a high engagement. This was performed on the KNIME Analytics Platform. We used the data obtained from the KPI6 queries exported in a csv file. For any post we are interested in text, number of likes and sentiment score. The latter is a feature provided by the KPI6 analysis. Specifically, a value associated to each post; whose range varies from 0 (very negative) to 100 (very positive).

We preprocessed the text for a text mining, doing several procedures to keep only usable and frequent words (using bigrams) and performing a topic extraction on them through an LDA. After that, we used the values of belonging to the topic for the existing words within them, in order to perform two negative binomial regressions and looking at how much the presence of every topic was statistically relevant to predict the number of likes. The outcomes showed that the most discussed positive topics are Amazon & Google, apps & skills, cities, and music. While, the most discussed negative one is radio.

The other part, which consists of a descriptive analysis with the goal of understanding the voice commerce market just like all the markets influenced by it, was directly done on KPI6. We first looked at the sentiment analysis and the emotions generated by the posts, clearly noting that most posts are positive.

The predominant emotion is admiration, whilst the negative ones are mostly related to disapproval. Secondly, we looked at the most used words, hashtags and bigrams; discovering that most of them included the products’ names (especially regarding music, radio and apps).

We could also observe that Alexa (exceeding “Big-G” by far) is the most discussed one. We then decided to make a comparison among the sentiment analyses of four leading tech companies: Microsoft, Amazon, Google and Apple. We noticed that Amazon is the most appreciated, probably because of the lower cost and the wider compatibility of its products. 

Lastly, we have analyzed the topics again, in order to get which markets are more connected with vocal assistance. Most of posts are about art and entertainment, in which music and cinema are the most common ones. Another interesting and quite discussed topic is the one related to shopping; whose category focuses entirely on supermarkets. 

Get in touch with us to learn about the products of the KPI6 suite