Artificial intelligence enables retailers to interpret customer comments, enriching structured data such as Net Promoter Scores and star ratings. And employing AI-driven text analytics during the holiday season allows retailers to react quickly to customer feedback.

Ying Chen, chief product and marketing officer, Luminoso

Now that we’re just a few weeks away from the busiest shopping day of the year, most retailers have finished up their holiday season planning and are entering execution mode. But while retailers’ product, marketing, and operations decisions have largely been made, there’s still time to explore one area where they can gain an edge: customer feedback analysis.

Between now and the end of December, retailers who can quickly uncover what customers are saying will emerge from the holiday season ahead of competitors. Why? They’ll be able to react to and address needs from real-time insights around customer complaints and praise.

With text analytics, retailers can easily look at these comments from all customers—not just those who are the most and least satisfied.

In recent years, tech companies have launched a variety of efforts looking to apply artificial intelligence (AI) to help retailers perform real-time customer feedback analysis on mixed data sets, or those containing both numeric and text data related to Net Promoter Score (NPS), customer satisfaction (CSAT), and simple star ratings.

Because of the complexity and importance of these types of mixed data analyses, some retailers have adopted simplified approaches with predefined text response options for customers to choose from, rather than allowing for unstructured feedback. Although this means retail leaders can get some kind of qualitative feedback from customers, they miss out on the richness of text conversations that sheds light on what customers really think, and previously unknown topics and issues.


Enter AI-powered text analytics, which can uncover insights from conversational customer feedback data without data training, setup, or maintenance. Here are a few of the ways that AI-powered text analytics can allow retailers to better understand customer feedback in real time, allowing them to move more quickly to both address customer concerns and redouble efforts in areas that are drawing significant praise.

1. Unpack the “why” behind NPS and reviews

Over the last decade, NPS has become the standard for measuring customer relationships. In many cases, text analytics is the perfect complement to NPS. There are many skeptics who argue the metric is flawed because it disregards key feedback from a portion of respondents, since it’s calculated by subtracting the percentage of detractors from the percentage of promoters. But many organizations ask respondents to provide open-ended explanations, asking a question along the lines of, “Care to tell us why you’ve given us this score?” Being able to analyze the complete mixed dataset means seeing a fuller picture of what customers think about products and offerings.

With text analytics, retailers can easily look at these comments from all customers—not just those who are the most and least satisfied. This empowers them to determine patterns and trends among all customers. Promoters, for example, might leave open-ended feedback about changes they’d like to see in a product or service. And customers who left a neutral score might give positive and negative feedback—both of which are valuable.

Text analytics software allows retailers to quickly see correlations between the feedback with scores, enabling retailers to better understand what issues drive NPS up or down. At Luminoso, we have been focused on making this process easier for retailers by doing this analysis in combination rather than splitting out the score from the feedback.

Additionally, being able to understand these drivers of scores over time enables retailers to understand if the improvements being made are having an impact and or identify concepts they may not be aware of.


2. Understand why customers are reaching out—and improve first-contact resolution

Beyond NPS, text analytics is highly useful for customer service interactions. These applications can examine transcripts from call center conversations, emails, or live chats as a way of performing a “retrospective” on why interactions occur, and how to prevent future problems.

Consider a simplified example: A retailer would like to analyze thousands of live chat sessions with customers over a period of time, with a focus on minimizing returns of a specific product, like a coffee maker. Text analytics software can focus on an area of interest by identifying all conversations in any way referencing “returns” and “coffee maker”—including misspellings or synonyms of both phrases.

From here, the text analytics application can help identify the top words or phrases associated with coffee maker returns. The retailer would quickly be able to see if the product is being returned because the glass carafe was broken upon arrival, or it wouldn’t turn on.

Retailers that can create a pipeline of feedback by analyzing these types of unstructured feedback are able to address root causes, prevent issues, and fast-track their ability to respond to customer needs. Other innovative use cases also include the application of AI text analytics as a way of improving the effectiveness of chatbots and virtual assistants to better understand what customers expect in their interactions with these channels.

3. Get ahead of the market: orange is the new black

Retailers typically only analyze feedback they receive from product purchases or call center interactions annually or semiannually, which makes it difficult to truly understand the ever-changing needs of customers. And around the holiday season, when feedback volume is often highest and most critical, retailers cannot wait until year-end recaps to figure out what’s working, or what could have been addressed.


Text analytics helps retailers analyze, understand, and act on data from a variety of real-time customer feedback channels, such as third-party online reviews, blogs and forums, and comments on social media. Using text analytics to layer survey data with real-time feedback from these channels allows retailers to get a steady pulse on customer satisfaction and seize opportunities as they arise.

Most importantly, retailers need not fear they’ll miss out on emerging, unknown customer trends. With the ability to analyze data from all channels, text analytics gives customer-facing organizations the ability to see a clearer picture of what makes customers happy, in addition to problems and issues that are looming on the horizon.

With great power comes great ability

I have one final piece of advice for retailers this holiday season: Don’t underestimate the power of qualitative customer feedback. Rather than trying to conquer understanding all data across all channels, focus on rich areas of feedback that can give you immediate value such as service tickets, reviews, and surveys. Going beyond your numbers might help you find something you didn’t expect.

Luminoso provides AI-powered text analytics applications designed to enable retailers to derive insights from unstructured customer feedback.