"People will forget what you said, people will forget what you did, but people will never forget how you made them feel." - Maya Angelou
"After yet another frustrating experience with the famed e-commerce ‘prime’ goliath, I’m now convinced that my polite complaint emails aren’t taken seriously. Maybe I should spare the pleasantries and give them a piece of my mind! After all, aren’t they always going on and on about ‘customer obsession’? Clearly, either these guys aren’t smart enough or they simply don’t care!".
Such negative customer sentiments can rapidly influence future customer journeys: a study quoted by Mckinsey & Co shows that 25% of customers will defect to a competitor's product or brand after just ONE bad experience with a retailer. On the other hand, for a shrewd retailer trying to gain an edge in his/her segment through data-driven intelligence, the capability to identify and exploit such sentiments is a powerful tool.
As a matter of fact not many retailers are well versed in Sentiment Analysis (SA), a technique that uses text analytics algorithms to classify the overall "sentiment" of the text content. SA helps brands classify customer sentiments, such as - satisfied, happy, or annoyed with the product or about the service provided by parsing both structured and unstructured data. Even neutral sentiments can be useful to the retailer - in some cases, objective and unbiased reviews may be provided without expressing any sentiment, providing useful insights. Thus, the possibilities are endless for the retailer in this relatively unexplored field.
Examples of sentiment analysis in the retail industry
The use of sentiment analysis in retail can be better understood from the following examples:
- Operational improvements
- Product quality improvements
- Competitive intelligence
- Brand reputation threat analysis
The strategic differentiation of these use cases vis-à-vis their tactical value is shown in Fig 1.
Fig 1: Long-term edge v/s immediately actionable
The tactical value clearly outweighs the strategic differentiation in these use cases. This means that actionable insights may affect the top line & bottom line in the short term. This makes SA a valuable tool for players in the retail industry.
Here is a closer look at the use cases:
- Operational improvements:
Customer sentiment analysis can provide a clear direction for operational improvements. For example: retailers can leverage SA for conducting analysis of service-related reviews, and attribute a score to each of them:
Such a scoring gives a clear indication of the magnitude of the customer sentiment and the urgency that needs to be assigned to it.
2. Product quality improvements:
SA can provide very specific insights across product categories and sub-categories, often down to the level of individual features of the product.
Example: A.com, a major US retailer tried to assess the customer sentiments across a chosen category (patio, in this case). As seen below, the insights can be drilled down to the level of specific attributes of a product (chair) within the patio category:
In this example, SA pinpoints the most unpopular features (legs, rust, crack, screws, etc.) of a specific product (chair) within the patio category. The retailer can then take the necessary steps to address this issue.
3. Competitive intelligence :
Competitive intelligence is defined as gathering and analyzing information regarding a competitor's performance, capabilities, and offerings.
Extending the previous example of A.com, here is how SA was used to compare the product quality of the patio category against that of a competitor B.com:
As shown, the dashboard gives a detailed comparison of the products across the two sites. It is even possible to pinpoint which features of the products are superior for each retailer.
4. Brand reputation threat analysis
A brand's reputation is affected by customer sentiments expressed through various channels, especially via social media. Monitoring the social media is essential to identify and address threats to the brand reputation, such as a damaging post by a customer, a video highlighting a defective product purchased, etc. For example, this is how a customer described a video posted by him about a product purchased from Target:
“Video taken of an Archer Farms brand by Target. It is a Maple Walnut flavored cereal bar. Expiration date is 4 months from the present. You will find three worms crawling on the cereal bar and there may be more hiding in the nooks and crannies of it…”
At the same time, SA can also help identify the brand influencers. A study by Twitter found that purchase decisions were closely correlated to the posts from brand influencers. By racking the sentiment on social media, retailers can track the various brand influencers talking about their products and can go to the extent of engaging with their fans as well.
Future of Sentiment Analysis
The quest for Artificial General Intelligence (AGI) has resulted in several breakthroughs in the field of ML (Machine Learning). Using SA in tandem with these new techniques can turn it from a mere diagnostic analysis tool to a proactive capability. For example, an AI model can proactively monitor the customer sentiments, thereby enabling the retailer to address the quality issues faster than before.
Although SA will provide a certain degree of competitive advantage in any industry, early adoption could prove to be crucial in the retail industry. Considering the intense competition over shrinking margins across all segments, SA may soon turn into an indispensable ‘hygiene factor’ for retailers of all sizes.