What do you do before purchasing something that costs more than a pack of gum? Whether you want to treat yourself to new sneakers, a laptop, or an overseas tour, processing an order without checking out similar products or offers and reading reviews doesn’t make much sense any more. Thanks to comment sections on eCommerce sites, social nets, review platforms, or dedicated forums, you can learn a ton about a product or service and evaluate whether it’s a good value for money. Other customers, including your potential clients, will do all the above.
People’s desire to engage with businesses and the overall brand perception depends heavily on public opinion. According to a survey by Podium, 93 percent of consumers say that online reviews influence their buying decisions. Users may not give you a chance once they’ve read a few bad reviews. They won’t research whether feedback was fake or not. They’ll choose another option. In this context, organizations that constantly monitor their reputation can timely address issues and improve operations based on feedback. Sentiment analysis allows for effectively measuring people’s attitude towards an organization in the information age.
If this in-depth educational content on using AI in marketing is useful for you, you can subscribe to our Enterprise AI mailing list to be alerted when we release new material.
What is sentiment analysis
Sentiment analysis is a type of text research aka mining. It applies a mix of statistics, natural language processing (NLP), and machine learning to identify and extract subjective information from text files, for instance, a reviewer’s feelings, thoughts, judgments, or assessments about a particular topic, event, or a company and its activities as mentioned above. This analysis type is also known as opinion mining (with a focus on extraction) or affective rating. Some specialists use the terms sentiment classification and extraction as well. Regardless of the name, the goal of sentiment analysis is the same: to know a user or audience opinion on a target object by analyzing a vast amount of text from various sources.
You can analyze text on different levels of detail, and the detail level depends on your goals. For example, you may define an average emotional tone of a group of reviews to know what percentage of customers liked your new clothing collection. If you need to know what visitors like or dislike about a specific garment and why, or whether they compare it with similar items by other brands, you’ll need to analyze each review sentence with a focus on specific aspects and use or specific keywords.
Depending on the scale, two analysis types can be used: coarse-grained and fine-grained. Coarse-grained analysis allows for defining a sentiment on a document or sentence level. And with fine-grained analysis, you can extract a sentiment in each of the sentence parts.
Coarse-grained sentiment analysis: analyzing whole posts/reviews or sentences
This analysis type is done on document and sentence levels. In fact, most specialists use it to analyze sentences rather than whole documents. Coarse-grained SA entails two coherent tasks: subjectivity classification and sentiment detection and classification.
1. Subjectivity classification. First, it’s necessary to determine whether a sentence is objective or subjective. An objective sentence contains some facts about an object or topic: Three strangers are reunited by astonishing coincidence after being born identical triplets, separated at birth, and adopted by three different families.
A subjective sentence, as the name suggests, expresses someone’s attitude regarding a subject: This apartment is wonderful. I enjoy every minute I spend in here.
2. Sentiment detection and classification. The goal of this operation is to define whether a sentence has a sentiment or not and if it does, to determine whether the emotion is positive, negative, or neutral.
Sentiment by polarity. Source: KDnuggets
Sometimes people share their points of view without emotions. For instance, the author of the sentence I think everyone deserves a second chance expresses their subjective opinion. However, it’s hard to understand how exactly the writer feels about everyone. So, the sentence doesn’t express a sentiment and is neutral. Neutral sentences – the ones that lack sentiment – belong to a standalone category that should not be considered as something in-between.
Let’s look at this comment: One of the most surprising and satisfying movies of the year. According to the phrase, the reviewer enjoyed the movie, so this sentence contains a positive sentiment.
And the following review is a clear example of a subjective sentence with negative sentiment: The fact that it’s also clumsily made and rife with mediocre performances seems almost beside the point in the context of how pointless this thing is in the first place.
However, objective sentences can also express a sentiment: I bought this waterproof camera case because it’s meant to be more reliable than a standard one. It’s clear from the context that the case wasn’t what the person expected. The sentence has a negative sentiment, but it’s expressed implicitly.
Sentiment doesn’t depend on subjectivity or objectivity, which can complicate the analysis. But we still need to distinguish sentences with expressed emotions, evaluations, or attitudes from those that don’t contain them to gain valuable insights from feedback data.
Fine-grained sentiment analysis: analyzing sentence by parts
The devil is in the details, as they say. If you need more precise results, you can use fine-grained analysis.
You apply fine-grained analysis on a sub-sentence level and it is meant to identify a target (topic) of a sentiment. A sentence is broken into phrases or clauses, and each part is analyzed in a connection with others. Simply put, you can identify who talks about a product and what exactly a person talks about in their feedback. In addition, it helps understand why a writer evaluates it in a certain way.
The fine-grained analysis is useful, for example, for processing comparative expressions (e.g. Samsung is way better than iPhone) or short social media posts.
Not only does it allow you to understand how people evaluate your product or service, it also identifies which feature or aspect they discuss: A touchpad on my laptop stopped working after 4 months of use. This way, you know exactly what must be improved or reconsidered.
The capability to define sentiment intensity is another advantage of fine-grained analysis. In addition to three sentiment scores (negative, neutral, and positive), you can use very positive and very negative categories.
How to conduct sentiment analysis: approaches and tools
Sentiment analysis allows you to look at your operations from a customer point of view. But how do you extract that knowledge from user-generated data?
Data collection and preparation. First, you need to gather all relevant brand mentions in one document. Consider selection criteria – should these mentions be time-limited, use only one language, come from a specific location, etc. Then data must be prepared for analysis: one has to read it, delete all non-textual content, fix grammar mistakes or typos, exclude all irrelevant content like information about reviewers, etc. Once we have data prepared, we can analyze it and extract sentiment from it. You can learn more about data preparation in our story.
As dozens or even hundreds of thousands of mentions may require analysis, the best practice is to automate this tedious work with software.
Using off-the-shelf tools and APIs. Various customer experience software (e.g. InMoment, Clarabridge) collect feedback from numerous sources, alert on mentions in real-time, analyze text, and visualize results. Text analysis platforms (e.g. DiscoverText, IBM Watson Natural Language Understanding, Google Cloud Natural Language, or Microsoft Text Analytics API) have sentiment analysis in their feature set.
InMoment provides five products that together make a customer experience optimization platform. One of them, Voice of a Customer, allows businesses to collect and analyze customer feedback in a text, video, and voice forms. The number of data sources is sufficient and includes surveys, social media, CRM, etc. Developers provide users with real-time notifications, custom dashboards, and various reporting options.
Clarabridge is a customer experience management (CEM) platform. It pulls and analyzes text from chats, survey platforms, blogs, forums, and review sites. Users can also gain insights from emails, employee and agent notes, call recordings and Interactive Voice Response (IVR) surveys: The system can convert them into text. They provide social media listening as well. The system considers industry and source, understanding the meaning and context of every comment. Sentiment analysis results display on an 11-point scale. Users can modify sentiment scores to be more business-specific if needed.
DiscoverText is a cloud-based collaborative text analytics system for researchers, entrepreneurs, and governments. Capterra users note the solution is great for importing/retrieving, filtering, and analyzing data from various sources, including Twitter, SurveyMonkey, emails, and spreadsheets. Sentiment analysis is one of numerous text analysis techniques of DiscoverText.
IBM Watson Natural Language Understanding is a set of advanced text analytics systems. Analyzing text with this service, users can extract such metadata as concepts, entities, keywords, as well as categories and relationships. It also allows for defining industry and domain to which a text belongs, semantic roles of sentence parts, a writer’s emotions and sentiment change along the document. IBM Watson Natural Language Understanding currently supports analysis in 13 languages. Tools for developers are also provided, so they can build their solutions (e.g. chatbots) using IBM Watson services.
Microsoft Text Analytics API users can extract key phrases, entities (e.g. people, companies, or locations), sentiment, as well as define in which among 120 supported languages their text is written. The Sentiment Analysis API returns results using a sentiment score from 0 (negative) to 1 (positive). As of today, the software can detect sentiment in English, Spanish, German, and French texts. Developers specify that the analysis be done on the whole document and advise using documents consisting of one or two sentences to achieve a higher accuracy.
That’s how Microsoft Text Analytics API analyzes a review for The Nun movie. It has detected the English language with a 100 percent confidence, and the sentiment is measured in percentages. Analysis results are returned in a JSON format as well.
Sentiment analysis results by Microsoft Text Analytics API
Google Cloud Natural Language API will extract sentiment from emails, text documents, news articles, social media, and blog posts. Its use includes extracting insights from audio files, scanned documents, and documents in other languages when combined with other cloud services.
Learn more about cloud machine learning platforms in our dedicated article.
Developers offer users the opportunity to try the service right away and see what it can do. Here is an example of a customer review of headphones from Amazon.
Sentiment analysis results by Google Cloud Natural Language API
The tool assigns a sentiment score and magnitude for every sentence, making it easy to see what a customer liked or disliked most, as well as distinguish sentiment sentences from non-sentiment sentences.
Hiring a data science team for domain-specific tasks. Commercial software may be less accurate when analyzing texts from such domains as healthcare or finance. In 2011, researchers Loughran and McDonald found out that three-fourths of negative words aren’t negative if used in financial contexts. For these cases, you can cooperate with a data science team to develop a solution that fits your industry.
Use cases of sentiment analysis
Various industries use sentiment analysis. While the areas of sentiment analysis application are interconnected, they are all about enhancing performance via analysis of shifts in public opinion.
If the Internet was a mountain river, then analyzing user-generated content on social media and other platforms is like fishing during trout-spawning season. People enjoy sharing their points of view regarding the latest news, local and global events, and their experience as customers. Twitter and Facebook are favorite places for daily comment wars and spirited (to put it mildly!) conversations. News about celebrities, entrepreneurs, and global companies draw thousands of users within a couple of hours after being published on Reddit. Media giants like Time, The Economist, CNBC, as well as millions of blogs, forums, and review platforms flourish with content on various topics.
Why not use these data sources to monitor what people think and say about your organization and why they perceive you this way? Sentiment analysis of brand mentions allows you to keep current with your credibility within the industry, identify emerging or potential reputational crises, to quickly respond to them. You can compare this month’s results and those from the previous quarter, for instance, and find out how your brand image has changed during this time.
It’s not only important to know social opinion about your organization, but also to define who is talking about you. Measuring mention tone can also help define whether industry influencers are mention your brand and in what context. And what’s more exciting, sentiment analysis software does all of the above in real time and across all channels.
There is one thing for sure you and your competitors have in common – a target audience. You can track and research how society evaluates competitors just as you analyze their attitude towards your business. What do customers value most about other industry players? Is there anything competitors lack or do wrong? Which channels do clients use to engage with other companies? Use this knowledge to improve your communication and marketing strategies, overall service, and provide services and products customers would appreciate.
Competitive analysis that involves sentiment analysis can also help you understand your weaknesses and strengths and maybe find ways to stand out.
Flame detection and customer service prioritization
Hospitality brands, financial institutions, retailers, transportation companies, and other businesses use sentiment classification to optimize customer care department work. With text analysis platforms like IBM Watson Natural Language Understanding or MonkeyLearn, users can automate the classification of incoming customer support messages by polarity, topic, aspect, and priority. Then queries are sent to dedicated teams and specialists. Since it’s better to put out a spark before it turns into a flame, new messages from the least happy and most angry customers are processed first. Satalytics, for example, groups feedback by device, customer journey stage, and new or repeat customers.
Every entrepreneur dies to see fans standing in lines waiting for stores to open, so they can run inside, grab that new product, and become one of the first proud owners in the world. How to bring the desired product to market? The only approach is to ask people what they want. Successful companies build a minimum viable product (MVP), gather early feedback, continuously improving a product even after its release. Feedback data comes from surveys, social media, and forums, and interaction with customer support. Questions like how to define which customer groups to ask, analyze this ocean of data, and classify reviews arise.
That’s when sentiment analysis comes in handy. It allows for learning about product advantages and drawbacks. For example, a student from Oklahoma State University has analyzed Amazon reviews about two Samsung phone models (Galaxy S6 & Galaxy S7) and two devices from Apple (iPhone 6 & iPhone 7) to find out why customers preferred one brand over another. He found out that users whose priorities are a reliable battery and a good screen choose Samsung phones. And customers who are more interested design and camera buy iPhones.
Filtering comments by topic and sentiment, you can also find out which features are necessary and which must be eliminated. Armed with sentiment analysis results, a product development team will know exactly how to deliver a product that customers would buy and enjoy.
Market research and insights into industry trends
As we said before, social media sites and forums are sources of information on any topic. People discuss news and products, write about their values, dreams, everyday needs, and events. And they do this voluntarily 24/7.
Sentiment analysis solves the problem of processing large volumes of unstructured data. Using this type of text analysis, marketers track and study consumer behavior patterns in real time to predict future trends and help management make informed decisions. Another benefit of sentiment analysis is that it doesn’t require heavy investment and allows for gathering reliable and valid data since its user-generated.
Workforce analytics/employee engagement monitoring
Some organizations go beyond using sentiment analysis for market research or customer experience evaluation, applying it internally for HR-related processes. These companies measure employee satisfaction, detect factors that discourage team members and eventually reduce company performance. Specialists automate the analysis of employee surveys with SA software, which allows them to address problems and concerns faster. Human resource managers can detect and track the general tone of responses, group results by departments and keywords, and check whether employee sentiment has changed over time or not.
Sentiment analysis takes employee mood monitoring to the next level with real-time monitoring capabilities. For instance, team members can fill out survey forms with a single request to rate their workplace conditions every month. They can also analyze their posts in social media to find a possible connection between their state of mind and work lives.
Sentiment analysis allows businesses to harness tremendous amounts of free data to understand customer needs and attitude towards their brand. Organizations monitor online conversations to improve products and services and maintain their reputation. The analysis takes customer care to the next level. Customer support systems with incorporated SA classify incoming queries by urgency, allowing employees to help the most demanding customers first. Sentiment analysis is a powerful tool for workforce analytics as well.
Consider these steps if you decided to leverage sentiment analysis in your operations:
- Collect feedback data
- Make sure data is qualitative enough for analysis
- Look for ready-made software and APIs
- Hire a data science team if you’re working in a specific industry like healthcare, finance, or transportation.
This article was originally published on AltexSoft and re-published to TOPBOTS with permission from the author.
Enjoy this article? Sign up for more updates on using AI in marketing.
We’ll let you know when we release more technical education.