Sentiment analysis is a method used to analyze the positive and negative emotions associated with a particular series of words. The technology has evolved in scope and availability in recent years, and the impacts on marketing and SEO are only just beginning to show themselves.
As sentiment analysis creeps into use by both search engines and marketers, it will have growing impact on the SEO industry. Here are four ways those changes are happening right now.
1. Impact On Featured Snippets
As early as 2008, Google had already filed a patent for an algorithm that would ensure “each snippet comprises a plurality of sentiments about the entity,” which, in theory, would keep the emotional content of the snippets relatively balanced. While this patent is meant to apply to all search snippets, it was likely put in place, or at least taken into consideration, during the development of featured snippets.
Nevertheless, there has been a continuing issue with featured snippets that Danny Sullivan spoke on in January of 2018. As he pointed out:
...people who search for “are reptiles good pets” should get the same featured snippet as “are reptiles bad pets” since they are seeking the same information: how do reptiles rate as pets? However, the featured snippets we serve contradict each other.
The difference in those queries is one that can be addressed with sentiment analysis. Since both queries are asking about the sentiment associated with reptiles as pets, good sentiment analysis would be able to return identical featured snippets for both.
Sullivan confirmed that Google was working on solving the problem by compiling information from multiple sources:
"There are often legitimate diverse perspectives offered by publishers, and we want to provide users visibility and access into those perspectives from multiple sources,” Matthew Gray, the software engineer who leads the featured snippets team, told me.
As announced at the Microsoft AI event in December 2017, we believe that your search engine should inform you when there are different viewpoints to answer a question you have...When you issue a search query like “is coffee good for you”...we build clusters over the passages to determine similarity and sentiment using deep recurrent neural network (Deep RNN) models. Lastly, we rank the most relevant passages from each cluster based on sentiment analysis...to deliver you the most relevant results from web sources.
Without question, inclusion in featured snippets is good SEO since it places you in the most prominent location of the search results. How does sentiment analysis factor into this?
For the search engines to use a piece of information within a featured snippet, it already needs to be authoritative, high quality, and relevant, as well as easy to extract. For it to be a good fit for the snippet based on sentiment analysis, it also needs to use emotionally charged words expressing a different sentiment than what is already included in the snippets.
2. Deciphering The Influence Of Customer Reviews
Everybody knows that star ratings are an important influence on customer behavior, but that doesn't mean we shouldn't care about the content of the reviews themselves.
According to a study conducted by ConsumerAffairs, the content of reviews matters more than the star ratings. In fact, the influence of the review content dramatically outweighs the influence of the star ratings, to the point that customers would actually prefer to buy a product with a 1-star rating featuring great reviews than a product with a 5-star rating featuring bad reviews.
The study measured whether reviews were “good” or “bad” by analyzing the quality and sentiment of the reviews.
Sentiment analysis tools allow you to analyze your reviews to determine if they are favorable of your product. This gives you a more reliable estimate of how consumers will react to your reviews than the star ratings will on their own.
Some of the tools on the market are already incorporating sentiment analysis into their review platforms. TrustPilot, for example, has incorporated sentiment analysis, and their reviews on Capterra clearly indicate that this is working out well for them.
With all of these tools on the market, it would be surprising if the search engines weren't using similar technologies to measure the sentiment of reviews themselves, especially considering that algorithms capable of estimating the quality of content have been in place since Panda hit the industry in 2011.
While it's unlikely Google could be reading the sentiment of reviews in real time, the technology to apply it separately from the indexation process certainly exists.
In fact, Google even has a tutorial for Google Cloud users that teaches them how to implement sentiment analysis.
While we as SEOs certainly don't have direct influence over the reviews people will leave of products, we can have indirect influences by monitoring reviews for negative sentiment and responding accordingly, as well as helping implement processes to ensure that satisfied customers are more likely to leave reviews.
3. As A Supplement To Backlinks
At the 2017 State of Search, Gary Illyes made statements that many interpreted as indicating Google was using sentiment analysis off-site as a supplement to backlinks. Illyes later clarified that he wasn't talking strictly about sentiment analysis, but he was indicating that Google was looking at brand mentions and evaluating the context as a way of better understanding what the brand was about.
Whether or not Google currently incorporates sentiment analysis into brand mentions, it's clear Google is using off-site mentions, even without links, and it's only a matter of time before sentiment is incorporated into that, if it isn't already.
A sound SEO strategy will anticipate this and take action in some of the following ways:
- Be proactive with campaigns that encourage user-generated content from your most popular fans.
- Involve outreach that isn't just about placing links, but about getting influencers involved and developing a positive view of your business.
- Respond to negative feedback positively, in a timely manner, with a solution that your target audience would find acceptable.
- Have a solid customer service staff in place to help upset customers with issues before they feel the need to take their frustrations online.
- Develop a proactive PR strategy that continuously generates positive news stories and sentiment, as opposed to a reactive PR approach that deals with negative press only when it arises.
- Be involved in conversations about topics related to your brand in order to build positive sentiment with your target audiences, including those among them who will never become customers.
All of the above will have a positive impact on your branding and search engine performance even if Google has not yet implemented sentiment analysis into brand mentions.
4. To Measure Social Media Impact Deeper Than Standard Engagement Metrics
Aside from the traffic that they refer to your site, which should be a high priority, the success of a social media campaign or strategy is typically measured using standard engagement metrics such as how many people saw a post, what proportion of them left a comment, liked it, or shared it, how many followers it created, and so on.
The problem with these metrics is that they don't necessarily tell you about the overall perception your campaign created. The number of “likes” may be dwarfed by the number of people put off by your post, for example.
This is where sentiment analysis tools become very useful. Here are a few examples of some tools currently on the market that are incorporating sentiment analysis:
- Trackur: Similar to Google Alerts, this tool notifies you when your brand gets mentioned, but it also includes sentiment analysis and influence scores to measure how positive the mentions are.
- Unamo: A brand monitoring tool that, in addition to sentiment analysis, provides detailed analytics and insights as well as competitor information.
- Brand24: Another brand monitoring tool with solid sentiment analysis tools in place.
These kinds of tools allow you to identify which brand mentions are positive, which are negative, and the overall sentiment associated with your brand at any given time based on the collection of these types of mentions.
While very useful, it's important to keep in mind the limitations of sentiment analysis:
- It cannot fully incorporate context. This is important because words which might, for example, typically have a negative connotation, might have a positive connotation in some contexts. For example, the word “burn,” is typically negative, but if it were used in the context of burning fat, it would typically be seen as positive.
- It can't interpret sarcasm. This is something to watch out for, and why it's important to read a fair sampling of the brand mentions in any circumstances.
As sentiment analysis tools become increasingly available, the SEO industry cannot help but be affected by them. Keep an eye on the four changes we discussed here, and look for opportunities in the search landscape as these technologies continue to impact it.