We’ve all heard of sentiment analysis, but what exactly is it and what can it do for your brand, your business, and how can you get started with it?
What is Sentiment Analysis?
Sentiment analysis relates to analyzing content such as social media comments, customer feedback, employee feedback, and even facial expressions in images to render sentiment orientation. These sentiments can be as broad as just saying that the specific content is from a “detractor” or “promoter,” or it can be as detailed as listing out all the emotions within the content.
Predicting fine-grained sentiment in images
Why is Sentiment Analysis Important in Business?
While, on the surface, sentiment analysis can seem like a fancy class project, in actuality, it has many uses in business. Let’s look at some sentiment analysis examples applied to business problems.
- You can aggregate customer sentiments from free-form feedback data and determine if your customers are primarily promoters or detractors. You can then take corrective measures to gradually rebuild trust with the detractors and turn them into promoters.
- You can keep your online platform clean and free from bullies by detecting hateful and inappropriate comments.
- You can determine which employees are demotivated or about to quit based on their outlook from recent feedback, peer reviews, and manager feedback and provide a constructive path ahead for employees to succeed at the company.
Overall, as you can see from these sentiment analysis examples, sentiment analysis is a versatile tool that can help you better understand employees and customers, keep platforms safe, provide customers with a better shopping and product selection experience, and learn from competitor brands.
More importantly, when you combine sentiment analysis with other AI-driven technologies such as text summarization, you can get deeper, more powerful insights.
How are Businesses Using Sentiment Analysis? (Real-World Examples)
Now that we know what sentiment analysis can help accomplish, let’s see how three companies are using sentiment analysis for a specific business purpose.
Great Wolf Lodge (GWL), a chain of resorts and indoor water parks, has expanded its broad digital strategy by using AI to classify customer comments based on sentiment. They developed what they call the Great Wolf Lodge’s Artificial Intelligence Lexicographer (GAIL).
GWL capitalizes on the concept of net promoter score (NPS) to gauge the experience of individual customers.
Instead of using an NPS score to determine customer satisfaction, GAIL determines if customers are net promoters, detractors, or neutral parties based on the free-text responses posted in monthly customer surveys. This is analogous to predicting if the customer sentiment is positive, negative, or neutral. GAIL essentially “reads” the comments and generates an opinion.
Through this effort, the company hopes to understand its guests better and improve the customer experience. For example, by analyzing comments by detractors, Great Wolf Lodge would know areas in their service that need improvement.
Analyzing this unstructured data manually would take far too long for humans. However, GAIL can parse this data in seconds and determine whether the author is a net promoter, detractor, or neutral party.
Meta—with nearly 1.7 billion daily active users—naturally has content posted on the platform that violates its rules. Among this negative content is hate speech. Defining and detecting hate speech is one of the biggest political and technical challenges for Meta and similar platforms. Detecting hate speech is a type of sentiment analysis problem focused on content with overall negative implications.
Humans review the AI-flagged posts in the same way as posts reported by users. In fact, the platform removed 9.6 million pieces of content flagged as hate speech in the first quarter of 2020 alone. While the sentiment models alone may not be sufficient to control hate speech on the platform, the tool does capture a massive number of spam posts, significantly reducing the amount of manual work by humans.
The volume of AI-based hate speech removal on Facebook. Source: Wired
Detecting which content contains hate speech is a complicated problem. AI algorithms must understand the subtle meanings in text and nuances in expressions, analyze the cultural context, and then determine whether it’s offensive without incorrectly penalizing harmless content.
Example hate speech. Source: arxiv.org
When the morning juice market weakened, Ocean Spray, an agricultural cooperative of cranberry and grapefruit growers, sought a new strategy to improve sales. Ocean Spray first needed to understand consumer sentiment and behaviors around cranberry juice better so that they could innovate.
Typically such innovation is done with the help of small focus groups of 10-15 people. However, Ocean Spray decided to leverage AI-driven analysis of thousands of online conversations, such as user reviews and tweets around cranberry juice, to really listen at scale.
Plus, instead of just classifying content like what Meta does, Ocean Spray leveraged themes and opinion summaries to understand consumer sentiment around specific topics. Through this analysis, Ocean Spray understood how consumers were using cranberry juice in real life, giving them ideas on how best to innovate and fill gaps in the marketplace.
The research surfaced unexpected customer behaviors. For example, they found that women enjoyed cranberry juice as a substitute drink without the alcohol in place of cocktails. Such insights helped them launch two new beverage lines, boosting revenues and helping them get out of an over-saturated segment of the market.
How to Get Started with Sentiment Analysis
As you’ve seen in this article, sentiment analysis has many nuances—you can detect sentiments in a sentence, paragraphs of text, and even from facial expressions in images. Further, you have various ways to leverage sentiment information—from using it for new product innovation to improving the customer experience.
To get started with sentiment analysis, you first need to understand your business application. Consider these questions:
- What would you like to know about your brand, customers, or employees?
- How granular should the information be?
- Do you just need sentiment information, or textual themes and summaries?
- Are you planning to integrate the solution into your dashboards or perform an independent analysis?
Let’s take an example. Say you need to understand the general sentiment in your company’s support conversations. You want to learn the ongoing “tone” and “mood” of your customers. Further, you want to visualize this within your dashboards. In such a case, you’d need to employ an emotion classifier to generate predictions on relevant conversations. You can then leverage those sentiments in your dashboards for downstream analysis.
Depending on your sentiment analysis problem, in some cases, you’d have to custom build the classifiers. But for others, you can leverage off-the-shelf tools such as Google’s Natural Language API or the Perspective API.
Often, for a multi-faceted analysis, you’d have to combine off-the-shelf tools with custom pipelines and analysis to help you answer all questions for optimal decision-making. This is what one of my clients did. They combined insights from an independent off-the-shelf text analytics tool such as Netbase (extremely pricey, by the way) with custom-built pipelines for a complete market research analysis.
There are endless possibilities in how you can employ these sentiment analysis tools. But remember to let the application guide the solutions that you’ll employ.
Now, over to you. What sentiment analysis applications come to mind after reading this article? What tools will you use for your analysis?
Keep Learning From Me:
- Join my AI Integrated newsletter, which clears the AI confusion and teaches you how to successfully integrate AI to achieve profitability and growth in your business.
- Read The Business Case for AI to learn applications, strategies, and best practices to be successful with AI (select companies using the book: government agencies, automakers like Mercedes Benz, beverage makers, and e-commerce companies such as Flipkart).