By: Nic Ray, CEO of DataEQ
Marketers around the world, no matter what sector they are in, will agree that how customers view their brand is of paramount importance.
Understanding how your customers are struggling with your products, or the areas where competitors are outshining you, can deliver great value – not just from a customer experience (CX) and research and development (R&D) standpoint, but for improving marketing campaigns, refining messaging and brand reputation too.
Traditionally, survey-based metrics, such as Net Promoter Score (NPS), have provided companies with a historical view of customer experience and how people perceive their brand. However, with many customers not willing to waste their time filling out NPS surveys – or only doing so when they’ve had a particularly positive or negative experience – these approaches may only be scratching the surface of a brand’s true reputation.
New data sources
Luckily, new sources of customer data, such as social media interactions, are providing a solution. Millions of customers are either criticising or complimenting products and services online, every minute. The arrival of ‘social sentiment analysis’ is enabling marketers to harness the value of this honest, unsolicited feedback in real time, and the result is a goldmine of powerful insights.
While some scepticism still exists around the accuracy of social sentiment analysis – and tech’s ability to make sense of every-day human language and slang – recent advancements are changing the way this information is reviewed and processed, providing stronger, more robust data.
The power of sentiment
Simply put, social media sentiment analysis is the science of understanding how people really feel based on what they say and how they say it online. It relies on artificial intelligence (AI) and analytical techniques to extract data around emotion and opinion from large volumes of text. Accurate sentiment analysis can highlight companies’ critical blind spots and pinpoint where they are falling short with their customers.
For example, DataEQ recently carried out a social media sentiment analysis of the UK’s energy suppliers, revealing that most companies were failing to properly support their customers online during the cost-of-living crisis. UK utilities scored an alarming negative 53.6% on public social media sentiment, and it’s estimated that, of the 1.3 million vulnerable customers who reach out for help from energy suppliers in a year, only half of them are likely to receive a public response.
So, as a marketer working within UK utilities, what does this tell you? Well, the social sentiment analysis flagged that UK energy brands – across the board – must pay closer attention to their online conversation and ensure that their social customer service teams are equipped resolve online complaints successfully. By leveraging social sentiment data, marketing and CX teams can work together to create a more customer-centric experience, and ensure their community is satisfied and positively engaged with the brand online. For any omni-channel marketing efforts that span a range of social media channels, the benefits of a happy and engaged audience can be a key differentiator when it comes to campaign performance.
To gain this level of insight from sentiment analysis, establishing sentiment polarity is key. That is, analysing the use of emotive phrases such as ‘love’ and ‘hate’ to develop a sense of the climate of feeling. In fact, the correct assignment of sentiment – whether consumers’ feelings are broadly positive or negative toward a given topic – is the most important aspect of any sentiment analysis campaign.
As sophisticated as AI has become, the complexity of human interactions still confounds machines. Tools that use natural language processing (NLP) algorithms exclusively for sentiment analytics achieve, at best, 70% accuracy when determining sentiment in English.
Layering an element of human insight over the analytical work performed by machines is the best way to achieve a valid understanding of how customers perceive your brand. In other words, allowing real people – a crowd – to refine and review the work carried out by AI. By augmenting an AI sentiment classifier with this crowd of human data classifiers, it is possible to achieve 95% or greater accuracy across various languages.
While this process may sound complicated, all data verification can happen in near real time. The result is a highly accurate, fully structured sentiment analysis available in live dashboards within five minutes or less. For marketers, this makes accurate sentiment analysis an invaluable tool for monitoring a brand’s reputation, in real time, to drive data-led decisions that will increase customer loyalty and satisfaction.
The bottom line
With issues such as the cost-of-living crisis impacting all industries today – from retail and tourism to finance and transport – how brands engage with their customers online has never been more important.
Metrics such as NPS often fail to capture the ‘bigger picture’ of a brand’s reputation, not only when it comes to real-time data, but also with regards to the competitor landscape. By tapping into new data sources, and harnessing the power of social media conversations, marketers are able to gain a better understanding of how customers really perceive their brand.
The insights derived from an accurate social sentiment analysis – powered by AI and human intelligence – will drive better decision-making for marketers, allowing them to foster a loyal online community, take their digital campaigns to new heights, and transform their brand’s reputation.
About Nic Ray
Nic Ray is the CEO of DataEQ, a data business that specialises in creating high-quality, actionable data from unstructured online customer interactions. Using a unique blend of AI and human intelligence, DataEQ filters social media conversation to optimise social customer service, generate new CX insights, manage risk, and improve conduct reporting.