How AI Can Help Retailers Stay Ahead of Customer Feedback Surges

Have you ever heard the phrase “the squeaky wheel gets the grease?” Turns out that proverb applies to a lot more than just wheels. It also applies to brands that don’t have tight customer-feedback-to-product loops or customer-feedback-to-operations loops. Many retailers aren’t able  to understand customer feedback, so they respond to the loudest customers (you know, those squeaky wheels) instead of listening to and rectifying the slow-burn issues that impact the most customers. That leaves the majority of their customers feeling like brands aren’t listening to or acting on their feedback.  

Microsoft quantified the scope of this problem: More than half of shoppers (53%) believe their feedback doesn’t go to anyone who can act on it. This is a huge customer challenge for today’s retailers, especially as consumers continue to prove they are anything but forgiving when it comes to customer service expectations - more than six out of 10 consumers stop buying from a company and switch to a competitor after just one poor customer service experience, according to Zendesk.

Working cross functionally with the teams (product/operations) that can fix customer problems systematically is only possible if the brand already has a true understanding of the feedback coming in. And by feedback, we mean ALL of the feedback customers now share through a variety of channels  -  from support inquiries, social media, and app/product reviews, to surveys, forums and more. But all of that feedback is incredibly tough to manage, and even tougher to try to translate and understand across disparate systems. 

To catch up to the customer feedback surge, retailers need to ditch their reliance on manual analysis, human tagging, generalized NLP, and survey-centric customer feedback initiatives. Embracing new technologies that dig deeper and can quickly surface high-level trends and themes across all channels is the only way to make sense of all the valuable customer feedback that is constantly flooding in. Forward-thinking, digital-native retailers like FabFitFun, Instacart and Thrive Market are using advanced artificial intelligence (AI) tools to translate millions of customer feedback data points from various digital sources into easily understandable insights. 

So how can you leverage AI to understand customer feedback and create the feedback loops needed to rectify the slow-burn issues before fires start and customers flee? 

• Categorize and quantify all feedback. Customer feedback should be categorized and quantified. Yes, that even includes all of the unstructured feedback from emails, chats, voice, reviews, and social media channels. This will transform freeform qualitative feedback into quantitative data that looks at the root causes of interactions to identify trends and questions that should be asked in follow-up. Doing this in real-time not only helps brands make better business decisions, but also helps them get more out of their data and cut support costs, complaints and issues. 
• Prioritize the customer-product feedback loop. Retailers need to look at the “why” behind customer feedback and then quantify and prioritize it. Don’t miss out on the opportunity to turn customer feedback into features and revamp your product roadmap according to customer demands and feedback. Often, retailers get tunnel vision about the features they think customers will be most excited about and benefit from, when many times customers will tell you exactly what would matter the most to them. Basing new products/ features/ offers on an AI-driven feedback loop not only helps create the best outcome, but it also ensures that customers feel heard. 
• Get serious about sentiment and specifics. It isn’t enough to have support agents simply label cases. It’s imperative to calibrate sentiment analysis to surface missed trends and unify real-time insights across the board, and via specific channels, in order to make the insight actionable. This approach requires machine learning and AI tools that can create a custom set of labels for unique data sets. Unlike with agents tagging cases, when generally the labels are kept very high-level, AI-driven labeling can get much more specific. Retailers could even properly identify negative sentiment. For example, if customers were having issues logging in due to  password reset emails not going through, identifying that level of specificity  versus just knowing the customer had trouble logging in makes a world of difference. Making things specific and actionable can help retailers solve problems before they impact large numbers of customers.

But remember, introducing AI tools certainly doesn’t mean replacing humans in customer service. Machines can never replace the empathy, human touch or context needed to solve some customer problems. The idea is for machines and humans to work together in tandem to create the best results possible for the customer. For example, machine learning can help capture basic information from a customer and trigger an auto response to get more information ahead of engaging with an agent. This frees up the agent’s time for more strategic uses and automates repetitive tasks. AI gives the 10,000 foot view of where all customer conversations and data can be considered holistically, instead of looking at individual, fragmented sources of information. 

Instead of waiting until the wheels are squeaking loudly to grease them, take note of the soft, repetitive noises. If you address them immediately and do some preventative maintenance, you can keep the car (and your business) running smoothly. Your customers will thank you for it, and remain loyal, when they feel heard – even if their voice isn’t the loudest one. 

– Chris Martinez and Kevin Yang, co-CEOs, Idiomatic