How to Collect In-Store Retail Analytics on a Massive Scale
Updated: Aug 9
Upgrade your consumer insights with sentiment analysis
How many times have you heard the phrase, “I’m fine”?
How about, “I’m fine”?
We all know that tone and body language play a large part in how we communicate with others. For example, “I’m fine” from your significant other might mean that they are actually content or it might mean that you are sleeping on the couch tonight.
Facial expressions play an even bigger part in how we feel. In the US, smiling is typically an indicator of positive sentiment (happiness). When you can capture how hundreds of customers are feeling in a retail store over time, this becomes an extremely powerful tool.
What is sentiment analysis?
Sentiment analysis “refers to the use of natural language processing, text analysis, computational linguistics, and biometrics to systematically identify, extract, quantify, and study affective states and subjective information.”¹
In other words, advancements in computer vision allow us to analyze crowds of faces to extract how happy people are feeling. The data is then used to drive business decisions such as how to optimize the in-store customer experience or where on the floor to allocate more staff. There are many tools which also analyze and process sentiment from text, but let’s focus on the biometrics piece (i.e., facial analysis).
Our micro-expressions say a lot about how we feel
How does it work?
In order to capture people’s feelings and behavior on a massive scale, you need a device that can detect and analyze hundreds of faces at the same time. The easiest way to do this is by using a high-resolution camera to film your space, which then connects to an edge processing unit to compute the analytics. The edge processing unit acts like a super-computer which processes the video feed locally and then sends the aggregate statistics (e.g., positive sentiment) to the cloud.
Secure data processing by Zenus
From there, you can view a live dashboard that tells you 200 people were 93% happy at 4pm in the jeans section.
As a retailer, you might want to understand the engagement levels of customers at different areas of the store. Are people less happy in the back of the store versus the front? How do people feel when they pass by different visual merchandising displays? What about a slight change in lighting?
To achieve accurate results, it is important to use a facial analysis system that has been trained using industry benchmarks as well as real world data.
What emotions can you recognize?
The typical human face has 43 different muscles that can be activated around the eyes, nose, mouth, jaw, chin, and brow to make thousands of different expressions.²
There are seven universal emotions: anger, contempt, disgust, fear, joy, sadness, and surprise. While it would be interesting to measure the angriness level of shoppers, the six emotions (aside from happiness) are not as universal and much more subjective. They contain many more nuances which can be difficult for a sentiment analysis tool to predict with confidence.
For example, when I am extremely angry, I tend to cry (curse of being a Cancer). The sentiment analysis tool might mistake this for sadness, even though I have the fury of the Hulk.
To keep things simple while still extracting important insights, stick with a system that measures positive sentiment easily and accurately.
How is it used?
Now that you know what sentiment analysis is, let’s talk about how it can be used in a retail setting.
Understand how many people displayed positive sentiment over time
Here are some common examples:
If you’re a big Fortune 500 brand, you may want to test the effectiveness of promotions before launching nation-wide. With sentiment analysis, you can measure how successful the promotion is on a small scale.
If you’re a smaller sized business, you might work with big name retailers and can use Happy Maps to differentiate yourself from competitors and upsell display/retail space.
According to Review42, “…opening a new store increases traffic to that retailer’s website by an average of 37%.” Brands can create a better experience for customers in-store, which will cascade to how well e-commerce performs.
If you’re a marketing agency that has a retail brand as a client, you can utilize sentiment analysis (on top of sales data) to prove the ROI obtained from your branded activations, events, and promotions.
Okay, but what about masks?
A well-trained algorithm can detect sentiment even with masks! As the quote mentions above, “…the typical human face has 43 different muscles that can be activated around the eyes, nose, mouth, jaw, chin, and brow.” It’s true that the eyes are the window to the soul.
Sentiment analysis allows you to automate the collection of consumer insights like never before. Retailers get an instant bird’s-eye-view of shopper engagement levels across multiple locations. While it is a great tool to save time and make decisions with, it is merely a piece of the puzzle. Once you combine engagement data with foot traffic, dwell time, and demographics (age/sex), then you’ll be able to see the full picture of the consumer’s experience in physical stores.