How to Use Sentiment Analysis to Transform Your Digital Marketing Strategy
Sentiment analysis is a fascinating concept.
Brands use it to better understand customer reactions, behaviors, and opinions toward their products, services, reputation, and more. The goal of sentiment analysis is to increase customer acquisition, retention, and satisfaction. Moreover, it helps put the right brand messaging in front of the most interested eyes.
Before the digital age, gauging and understanding sentiment was an incredibly cumbersome process. It typically involved sending out surveys manually, going to the streets and asking people, or gathering focus groups in one place at one time. The big data-infused model of sentiment analysis we know today hit its stride on the political scene in 2010. Since then, it has morphed into a key tactic in marketing plans. These days, most of the grunt work is automated.
However, even with all of the advances in areas like martech, voice search, conversational commerce on social media, virtual assistants, and big data analytics, understanding how to actually use sentiment analysis to improve the bottom line is a complicated task.
Here are a few key approaches to help you get the value you need.
Know the Terms and Phrases That Indicate Intent
Most businesses today (hopefully) don’t even begin their digital branding and marketing efforts without a list of keywords relevant to their industry and a plan on how to target their audiences. You should have a good idea of the terms and variations that bring you traffic to your website, when used in conjunction with your brand and products. If you run an auto repair shop, people are likely finding you on the web through terms such as: body shop near me, auto repair, replace brake pads, etc.
Google Search Console gives you a great, fairly accurate idea of what’s bringing people to your website:
In terms of sentiment analysis, to gain actionable insight, you need to know how people are using these keywords in a way that indicates interest and engagement potential. Now, this is perhaps the biggest gray area in sentiment analysis, because not all positive sentiment equates to sales. Just because there are a lot of positive words around luxury cars doesn’t necessarily mean people are about to buy.
However, there are certain terms and phrases that signal people have entered your buyer’s journey. Let’s say you run an SEO agency and one of the terms you’re tracking for sentiment analysis is "Google update." If you notice that a lot of people are searching for things like “what to do after a google algorithm update?” or “how to recover from a google penalty?” it’s a good indicator that they might need your services at the moment; you should target them accordingly.
Spot Patterns in Product Reviews
At its core, sentiment analysis is a game of pinpointing patterns and reading between the lines. Simply put, the more genuine and meaningful feedback you get on your product, the better insights you will gain into your customers.
Of course, gathering such high-quality feedback is easier planned than executed; especially for newer or smaller companies. Only 10% of customers will review or rate a business after a purchase, while half of consumers will leave a review only some of the time. However, the number of reviews jump significantly to 68% when a company asks the customer directly to leave one.
In order to find fruitful, up-to-date patterns, you need to make it a marketing process to consistently seek out new reviews. Then, you’ll want to start by searching for common adjectives. These should include words like:
- great, simple, easy,
- or awful, difficult, poor, etc.
In the above image, there are a good amount of reviews that include the word “great” for this product. Looking at the context around this term, we notice recurring patterns around components, like features and usability, and “not so” great opinions on customer service.
Finding recurring themes in customer sentiment will give you a better picture into the positive and negative aspects of your business or product. These can indicate the level of trust people have in your brand and how likely they are to give you a recommendation. When you are looking for patterns, try to come up with several adjectives that shed light on both sides of the spectrum.
- What words are commonly used to describe their experience?
- Is there an issue that forces multiple people to leave negative reviews?
- What part delights them the most?
- What’s preventing you from solving common problems?
- Which products or solutions are users comparing yours to?
Look to Social Media for Unabashed (Unfiltered) Opinions
Oftentimes, social media is one of the best places to get raw opinions, where people don’t hold back — both in positive and negative lights. Knowing how people feel in an unfiltered environment can be a great way to tell which parts of your business are working very well — and not so well.
A social listening platform is an important tool to keep in your portfolio for monitoring online mentions and gathering important datasets. Tools like Mention, Talkwalker, and Brand24, not only keep an ear on social mentions, but also turn these comments and hashtags into valuable customer analytics to help your marketing team understand your customers even better.
For instance, the online gaming developer Wargaming used brand monitoring techniques to analyze its customer’s desires and see which products performed best. The company tracked its users’ social media conversations to see what they were looking for, what parts of the games they liked or disliked, and any suggestions they offered for improvements.
Similarly, you can use a social listening tool to combine all your brand mentions into one database, giving your marketing team a bird’s eye view of audience sentiment on social platforms and identify areas to work on.
While gathering this sentiment is good, the most important thing is knowing what to do with it. About 83% of customers who make a social mention of a brand — specifically, a negative one — expect a response within a day, and 18% want one immediately. Unfortunately, a majority of these mentions go unanswered, which can really impact a brand’s image. By utilizing an effective real-time social listening program, you can not only stay on top of social buzz, you can intervene and reply to any negative sentiment right away.
Some of the next steps will be fairly obvious, especially when you’re dealing with negative feedback. For instance, if your customer sentiment from social listening reveals that people are having trouble updating their software or there are issues with the product itself, this indicates that some redesign is necessary. However, don’t get too comfortable when you are getting positive reactions — these tend to trick companies into thinking that no improvements are needed.
This kind of feedback can support a stronger marketing strategy. Let’s say your business sells pool supplies. While your customers may not be tweeting about your great chlorine chemicals, they are more likely talking about the fun pool floaties and games your website sells. Therefore, it would be helpful to highlight these fun accessories, as well, by listing them more prominently on your page and even including UGC to promote them.
Use Predictive Analysis to Spot Trends and Automate Actions
Now that you have all these valuable insights, you need to know how you can use them to shape your current and future business strategies.
Plugging your sentiment analysis into a predictive model is crucial for spotting trends, getting a feel for how opinions are progressing, and determining your next steps. Predictive analytics use machine learning and AI technology to not only gather, but analyze loads of consumer data and make accurate projections. These systems gauge historical behavioral data to help determine the best plan of action in the future.
In fact, customer segmentation and targeting (which is the logical next step after you analyze your audience’s sentiments) is one of the areas where applying AI and predictive analytics has the highest chance of working well for business.
In order to develop an optimal predictive model for sentiment analysis, ask yourself:
- What do you want to know?
- What is the expected outcome? What do you think your customers are thinking?
- What actions will you take to improve overall sentiment when you get the answers? How will you automate these actions?
- What are the success metrics for these actions?
Chances are, your customers are already telling you what you need to make improvements to your business. By gathering as much data as possible on customer sentiment, your marketing team can understand just what needs to be done to provide a better experience, tweak campaigns accordingly, and acquire and retain more customers in the process.
Be sure you know what to data to collect, how to mine it, and how to apply it to keep raking in the revenue.
Rohan Ayyar is the regional marketing manager for India at SEMrush. His blog, The Marketing Mashup, covers digital marketing from the perspective of B2B, B2C, lead generation, mobile marketing, SEO, social media, content marketing, database marketing including predictive analytics, and conversion rate optimization. In addition, he'll look at emerging marketing technology and how marketers can use it. Reach Ayyar at firstname.lastname@example.org.