Actionable AI: Driving your customer agenda with sentiment analysis

 Laura Hoyos Salazar

Laura Hoyos Salazar ,
Management Consultant at Sia Partners

In the last few years, Customer Experience (“CX”) has become a driving force of business and a critical differentiator in an increasingly competitive environment, whatever the industry. It influences brand perception and impacts business performance so eminently that most organisations now have distinct teams leading efforts in this space, and C-level executives overseeing customer satisfaction KPIs.

Through significant advancements in Artificial Intelligence (“AI”), sentiment analysis is becoming a popular technique to measure perceptions, identify unmet customer needs, defuse reputation risks, and determine opportunities for improvement to ultimately boost the customer experience. Upholding a high-quality customer engagement is even more critical in today’s COVID-19 era as it helps maintain trust, confidence and loyalty in a disrupted and remote environment.

Sia Partners recently discussed this topic at the Hong Kong General Chamber of Commerce and below are some of the key takeaways from the event.

What is sentiment analysis and how has it evolved over the years?

Defining sentiment analysis

In layman’s terms, sentiment analysis is a way to identify a view or opinion that is expressed through written, visual or spoken communication.

It can be as simple as trying to understand how an individual feels at a given moment, or as complex as assessing how multiple people, groups or even a whole population feel about a particular subject in a given point in time. The latter is where AI comes in particularly handy.

AI, and more specifically, Natural Language Processing (“NLP”) and Machine Learning (“ML”), are techniques that help us process large amounts of data to produce insights (refer to the figures below showcasing how much data is created every minute). Relative to sentiment analysis, it is done by identifying and extracting subjective information from data, analysing it, and classifying it with a distinct sentiment. Most commonly, we see this in the form of opinion polarity, which is classifying a point of view as positive, neutral or negative. The processing of the data is done by NLP, while the classification is done by the continuously evolving ML algorithms.

Examining its evolution

Sentiment analysis has gained an increased interest since the early 2000’s. Traditionally, it was used by Marketers to determine a customer’s opinion on an existing product or service, through first-hand market research and direct exchanges with customers. More recently it has expanded to the point of predicting financial markets behaviours and election results, to name a few. More and more, organisations are realising the powerful advantages that come from this practice and are therefore incorporating it into their organisation-wide strategies, across the whole value chain.

Visual evolution of sentiment analysis

Why is sentiment analysis important and how can it help improve CX?

Emotions are a core part of what makes us human and they play a big role in how we make (purchasing) decisions. By understanding how customers think and feel about a brand or a topic, we are able to gather insights that can in turn help make improvements to many parts of the organisation, including the customer experience. Here, we highlight five key reasons why sentiment analysis is particularly relevant to CX.

1. Elevating customer services

It helps gauge customers’ experience and satisfaction in how they are served. Learning the kinds of interactions that make customers happy and the ones that frustrate them, allows organisations to improve and tailor their services. This is a key enabler to improve key customer metrics like Net Promoter Score (“NPS”).

2. Augmenting product offerings

It can be leveraged to compare and benchmark products against those from competitors. Organisations can learn about the features that customers like the most and those that they don’t, continuously improving products in an agile, incremental way. It can also prove useful to keep an eye on industry trends and identify new products that may be relevant in the mix.

3. Optimising marketing strategies

It can provide powerful insights to boost marketing strategies. It can help organisations define their customer segmentation approach in order to create more tailored experiences and ultimately drive higher campaign ROI.

4. Monitoring brand reputation

It is useful to determine overall brand sentiment from direct interactions with customers, as well as from anything said on social media. This can be handy in turning negative feedback into positive experiences.

5. Staying relevant over time

The way customers feel about a brand can change over time and sentiment analysis allows organisations to keep track of any significant shifts.

How can you get started or accelerate adoption?

All you need to do is to define what you want to learn and determine the kind of data needed. As simple as that. More often than not, we see companies jumping straight into the data to try to interpret it, before even formulating the question they want to address through the exercise; and while this approach may occasionally produce some results, it is short-sighted, fairly random and likely to provide limited insights. Our suggested methodology is to always start with the business question or challenge in mind and let it be the driver of the exercise.

Last words…

A brand is a culmination of touchpoints and experiences, which create a perception in the minds of an audience. Therefore, a brand perception is a customer’s reality. As highlighted in this article, an exponential amount of data can be leveraged to understand customer perceptions and translate them into actionable insights to ultimately provide better customer experiences.

Please feel free to share these in the comments section at the bottom of the page or send us a LinkedIn message if you would like to discuss how we can support you on this journey.

Subscribe to Industry Era Review