Three Levels of Social Media Sentiment Analysis - it goes far beyond just Brand Sentiment

Social media platforms are our equivalent to the modern-day town square. Users catch up on the news, share laughs, and have real-time discussions about the products they use in their everyday life. They complain when an in-app video freezes or an airline loses their bags. So there’s no shortage of data you can scour to understand how customers feel about your products. They’re willingly sharing it on social media for everyone in the world to see.

A typical social media sentiment analysis uses natural language processing (NLP) and machine learning to categorize the prevailing emotion in a user’s post about your company. Posts are rated as positive, negative, or neutral. The NLP or listening tool can rate your overall brand or product sentiment from there.

Social media sentiment analysis tools can help you understand how users feel about your brand and products, but you shouldn’t stop there. Simply finding out if customer feelings are positive or negative won’t give your product team enough to act on.

But this base layer of sentiment analysis, which consists of social listening, will only get you so far. Let’s say your analysis shows your company has received 63% positive mentions, 24% negative mentions, and 13% neutral mentions in the last three months. 

You’re now confident in your customer sentiment, but these metrics don’t offer anything actionable. There aren’t any insights you can use to turn the negative sentiment around for that 24% — nearly a quarter of your social media audience. An effective social media analysis lends tangible ways for your product team to change your customer experience and improve your sentiment score.

Go deeper to understand what’s driving sentiment and turn it around

True sentiment analysis should also seek out the reasons behind consumers’ feelings — positive or negative — and provide insights for you to incorporate in product development.

Basic sentiment analysis can give you a pulse of how users feel about your brand and products. Still, when you don’t consider the reasons for customer feedback, you risk making all the wrong inferences of the true “why” behind positive and negative sentiment.

The qualitative feedback left by users on social media contains trove of insights in both positive and negative mentions, provided you are able to do two things -

• Go beyond just brand sentiment and dig deeper into not only specific keywords, but also semantically summarized insights or ‘reasons’ for the feedback.
• You can actually quantify both keywords and ‘reasons’ to find spikes/dips to help prioritize what to act on.

Successful social media sentiment analysis tracks both keywords and motivation

Using an artificial intelligence (AI)-based analysis tool, you can track social media sentiment at three different levels: for your overall brand, for individual keywords, and by breaking it down into actionable reasons. Here are the three levels of social media sentiment analysis -

Level 1 - Capture overall sentiment

Track audience sentiment across all social media mentions you’re receiving to see if users are happy with your product or experience any unexpected issues.

‍

‍

This is only the first layer of sentiment analysis you’ll conduct. It’s another form of social listening; you’re just applying it directly to your mentions and replies instead of brand sentiment as a whole. Still, it’s helpful for understanding broad trends and seeing immediate changes in real-time, such as after a product release.

Any social listening tool, like Brandwatch, Hootsuite Insights, and Talkwalker, can report on the general sentiment around your brand on social media. But you should focus on product mentions and feedback specifically to equip your product team with valuable insights.

Level 2 - Track sentiment by keywords and topics

Your users will naturally bring up their favorite features and the peskiest bugs in your social mentions. You can track the keywords and sentiment around each keyword to understand specific pain points and opportunities.

In the example below, we are looking at some specific keywords and the distribution of sentiment for each of those keywords. Which keywords get talked about the most positively for the product, and which most negatively impact the overall sentiment?

‍

‍

A customer insights person can see that Application Performance(crashes/glitches) have a large percentage of negative mentions compared to positive mentions.

Level 3 - Track granular `reasons` for social media feedback

Level 3 is when you get quantified and specific reasons for feedback and can quickly and confidently prioritize what needs to be fixed.

‍

We can see here that “unable to register” is spiking month over month on the trendline, along with “want to customize background”. This is even more specific than “login - negative sentiment” and far more specific than just “negative sentiment tweet”.

Source feedback proactively

It’s always a good idea to check in with users and ask them directly what they want to see from your product. Take the findings of your sentiment analysis and use that to shape your direct outreach to users since you understand what they’re looking for.

Talk about how you’re gearing up for your next product planning or launch or you want to hear what needs improvement. Figma’s VP of Product Tweeted out a great example of how to do this by asking his audience what should be included in 2023 planning.

Close the loop and update customers when you’ve acted on their feedback

Users aren’t shy about telling you when they’re experiencing problems, so don’t keep it a secret once you’ve finally fixed them. Let your customers know about product releases that address their concerns and ease their pain points. They’ll appreciate you’ve been listening carefully to their feedback and you’re building products with their needs in mind.

Measuring customer satisfaction is a continuous process made more accessible with automation. Book a demo with Enterpret to see it in action.