Retour haut de page

Sentiment analysis is one of the current hot topics in the data and analytics world, with more and more tools and algorithms out there attempting to understand the intent behind the written word rather than just the definition of the words. Platforms like IBM’s Watson (which I will always hold a grudge against for purchasing my beloved AlchemyAPI), MeaningCloud and a number of packages for programming languages like R and Python are becoming more and more popular as everyone races to offer the complete text analysis solution.

Personally, text classification and sentiment analysis have been a part of my approach to SEO since around 2012, but this was largely limited to keyword research and competitor analysis. With search engines incorporating more of this, I felt it was time to up my game, hence the sentiment analysis project I’ve recently begun on my personal site.

What Are The SEO Implications?

Bing is currently utilising this to improve the quality of their featured snippet results (called Multi-Perspective Answers), allowing them to give better-quality instant answers. Google’s version of this, both in terms of featured snippets and the answer box, is often criticised for providing answers that are biased by the way the question is asked, rather than taking the intent behind it and providing the most relevant answers.

It would be naive to think that Google hasn’t been working sentiment analysis into RankBrain, especially with their Natural Language API being out there in the wild, and they are likely waiting until RankBrain has gathered and analysed enough data to make sentiment analysis a core part of the algorithm. What are the implications of this? They could be fairly wide-ranging.

Online Branding Considerations

We know that Google likes websites and online businesses to be a brand and, like in the real world, online brands can have positive and negative sentiments. I wouldn’t be at all surprised if backlinks which talk negatively around a brand begin to provide less authority in search engine results – possibly even being discounted overall, encouraging brands to stay in their customers’ good graces.

It’s not dissimilar from the offline world, really, but if Google is to go in this direction, they will need to have a very robust sentiment analysis process and much better spam detection. This will not be difficult for unscrupulous competitors to exploit.

So Is It The Next Big Thing?

The key thing to understand about search engines – Google in particular – is that they’ve always wanted to understand content the way humans understand it so that they can provide the best results possible. Better results lead to more searches, which leads to more advertising revenue, after all.

Will better understanding of textural topics and sentiment become an important part of SEO in the coming years? I’m absolutely certain that it will, but will it be the next big thing in search? I’m not so sure about that. I suspect it will be a gradual rollout, with a lot of testing and adaptation and it will eventually find its way into the algorithm. As for the weighting it has in the algorithm? We’ll have to wait and see, but I wouldn’t be shocked if it was a fairly important factor.

Whether your brand’s sentiment becomes a significant ranking factor or not, I definitely think now is the time to ensure that your reputation is good with your customers and your peers. SEO aside, it’s just good business and, in the end, that’s what we’re all here for.

23 April 2018