From sentiment to empathy: understanding how customers feel

How AI is helping brands treat consumers with empathy at scale.

Chat with MarTechBot

Last week, the cross-channel marketing platform Iterable announced the launch of Brand Affinity, a new cross-channel personalization tool with customer sentiment analysis at its core. In simple terms, by using AI to score customer sentiment, the solution allows brands to create sentiment-based segments which can be appropriately messaged at scale.

In 2020, like never before, marketers need to be sensitive to customer sentiment: lives have been disrupted, priorities have changed, purchase patterns have changed, and those consumers with spending power are shopping across an endless digital landscape, not just in the high street.

It seemed a good time to look at how innovation is helping brands understand how we feel. First we turned to Iterable’s VP, Product Bela Stepanova to learn more about Brand Affinity.

A positive state of mind

As a platform seeking to offer connected customer experiences across multiple channels, Iterable serves brands — like Ipsy, DoorDash and Curology — which engage with their consumers in a variety of digital settings rather than through a static ecommerce website. “The majority of our customers are looking to launch a great conversation across channels, so it fluidly flows from one channel to another,” Stepanova confirmed.

Personalization solutions are widely offered, of course, but Brand Affinity places an emphasis on empathy. “When I think about customer communications and customer experience empathy is everything. When a business is small, it’s very easy to tell what customer sentiment is, because they know each customer. Once you get to millions of customers, and there’s billions of signals coming in at you, that’s not something we can deal with without AI.”

Iterable’s AI crunches this large volume of data to analyze sentiment, and how it changes over time, for each individual. Sentiment labels are attached to customer records, varying across a spectrum from negative to loyal. Marketers can then use that continuously updated field in the record to segment audiences. Messaging is tailored to sentiment-defined segments rather than individual customers.

Brand Affinity is aimed at going beyond predictive churn tools by identifying customers in positive as well as negative states of mind. “There are multiple use cases,” said Stepanova. “For customers who have a negative sentiment, they’re likely to churn; but you could also look at programs across the board of sentiment, beyond churn risk.”

Happy customers, for example, may be receptive to up-selling or cross-selling. For example, one Iterable client using Brand Affinity in beta had success in identifying users ready to convert from loyal to donor on non-profit websites. “By targeting programs to the right audience, and understanding the loyal people who are ready to become donors, they’ve been able to increase conversion 2X.”

Iterable, founded in 2013 by Justin Zhu and Andrew Boni, former Twitter and Google engineers respectively, raised $60 million in Series D funding in December 2019.

Looking at the whole picture

Matt Nolan, Pega’s Senior Director of Product Marketing for Marketing, AI, and Decision Sciences, takes the long view on where marketing has come from and where it’s going. When two people are in conversation, especially face-to-face, they each use all kinds of unconsciously received signals to understand how the other person is feeling. “Most software is like a person with no intuition: it can’t see somebody else, so we have to train it.”

Marketers have been working on this challenge, with increasingly sophisticated tools, since the early days of direct marketing, and the filtering of mail lists using business rules. This was the beginning of segmentation. What was state of the art 15 years ago, he said, was next-best-product models, which identified groups of customers most likely to be interested in a product depending on their buying patterns.

Ten years ago: “Let’s look at all the different conversations we might want to have with a customer. Not just the products we want to sell, but other conversations that might add more value — things like, are there service actions we can take pro-actively? Are there ways we can nurture him to build a better relationship? Do we need to retain him, and what amount of money makes sense to do that?”

This was the next-best-action approach. “Where the market is now,” Nolan continued, “is empathy; trying to really understand the customer, put yourself in their shoes. What’s changed is the availability and velocity of data. There’s a constant stream of information from the customer to the brand that gives them signals that can be interpreted to figure out if there’s a way they can help you, support you, or market to you. It’s not perfect, but it’s the best that we can do right now.”

Is sentiment analysis part of this route to understanding? “Absolutely,” said Nolan, “and we do have a natural language processing package that sits within our Decision Hub, and [our] customers it, especially in a customer service setting. You still have to train that stuff, and historically that process has been slow.”

Pega’s models are designed less to interpret a customer’s emotional state as what they are trying to do. “The problem with pure sentiment models is that most of the time they’re very myopic; they lack the rest of the context for that person’s situation. And how are sentiment models actionable? Whether someone is happy or sad is great to know, but it’s hard to take action on that.”

Asked about Pega, Stepanova said: “In the real-time interaction space, a lot of [what they do] is focused on customer case management, how an individual conversation happened, and we are looking at communications at scale when it comes to sentiment.”

Messaging the ‘in-betweenies’

To drill deeper, we turned to Seth Grimes, President and Principle Consultant at Alta Plana, past organizer of a series of Sentiment Analysis Symposia, and more recently the convener of the CX Emotion conference in London.

“Companies have been doing sentiment analysis commercially, with credible offerings, for over ten years, ” he said. “The most successful use cases that I’ve seen have been specifically in customer experience rather than in marketing. Customer experience is more like after the fact, studying reviews and surveys and so on, and also looking at customer service use cases. For forward-looking market research, the idea of segmenting by sentiment — that’s actually a new one to me.”

Based on research looking at Net Promoter Scores, Grimes acknowledges evidence that it’s effective and worthwhile to message “in between people” rather than trying to reverse negative attitudes. “Maybe if you can identify people who are positives, you can find up-sell and cross-sell opportunities.” This was consistent with Iterable’s proffer as described by Stepanova.

In addition, baking a machine learning component into sentiment analysis would, said Grimes, differentiate an offering like Iterable’s from companies who still use rule-based methods (in machine learning, the model self-corrects, possibly with some human supervision).

Empathy is the watchword

Whether we’re talking about consumer sentiment or emotion, or brands empathizing with their customers, we’re going to be talking about it a lot more in 2021. The question goes beyond what the next-best marketing message is, to how brands find the right tone of voice in general.

“Over the last ten months,” said Nolan, “the whole world has shifted. Brands have been building analytics equity over the course of years. All of a sudden the world changes, and all of our behavior changes, and it’s like those models are worth nothing — at least for a period of time.”

The challenge is one of engaging with people whose whole world view might this year have changed. “Unless you have the resources for it,” said Nolan, “it’s tough to figure out what to say and when to say it.”

Correction: An earlier version of this post referred to a Series D funding round by Iterable in December 2020; the round took place in December 2019.


About the author

Kim Davis
Staff
Kim Davis is currently editor at large at MarTech. Born in London, but a New Yorker for almost three decades, Kim started covering enterprise software ten years ago. His experience encompasses SaaS for the enterprise, digital- ad data-driven urban planning, and applications of SaaS, digital technology, and data in the marketing space. He first wrote about marketing technology as editor of Haymarket’s The Hub, a dedicated marketing tech website, which subsequently became a channel on the established direct marketing brand DMN. Kim joined DMN proper in 2016, as a senior editor, becoming Executive Editor, then Editor-in-Chief a position he held until January 2020. Shortly thereafter he joined Third Door Media as Editorial Director at MarTech.

Kim was Associate Editor at a New York Times hyper-local news site, The Local: East Village, and has previously worked as an editor of an academic publication, and as a music journalist. He has written hundreds of New York restaurant reviews for a personal blog, and has been an occasional guest contributor to Eater.

Get the must-read newsletter for marketers.