What is sentiment, and how can I use it?

Sentiment is a performance metric that can help you understand the nature and quality of monitored media mentions, whether it’s gauging the overall sentiment of a story or diving deeper into how specific entities are being represented.

Medianet uses rigorously tested AI modelling to provide automated sentiment analysis within the platform for all print, online and broadcast media mentions*. We provide sentiment analysis in two distinct ways: Article-level sentiment and Entity-level sentiment. 

Article-level sentiment

What is it? 

Article-level sentiment analyses the overall tone of an entire piece of content. The sentiment rating applies to the article overall and is not an assessment of your organisation or the specific keywords that inform the search. Articles are given either a Positive, Neutral, or Negative rating.

How can I use it? 

Article-level sentiment is useful for understanding the broader media tone around themes or issues that make up coverage. For example:

  • Track how the overall tone of media coverage shifts over time.
  • Evaluate how the sentiment of coverage at the article level aligns or differs from your own assessment of your organisation’s sentiment.
  • Use pre-set Dashboards within the platform to get a quick overview of the sentiment distribution across a media campaign or search query.

Where can I find it?

  • Saved search results and Detailed view: When viewing an individual mention, the sentiment is displayed for the entire content piece, giving you a quick overview of how the story is framed.
  • Dashboards: A visual representation that breaks down the sentiment of mentions found in the search results, making it easy to monitor shifts in tone across different media outlets and topics.

Entity-level sentiment

What is it? 

Entity-level sentiment goes deeper than the overall article sentiment and provides the sentiment of the top three entities mentioned within it. Entities can be organisations, people or places. Our algorithm first identifies the top three entities based on the number of references within a piece of content and the contextual importance of the references. It then assesses the sentiment towards that entity specifically. 

An article might mention several organisations, and the sentiment for each one could vary. For example, one organisation might be portrayed positively, while another might receive negative coverage within the same story.

Entity-level sentiment for your organisation will appear in the Detailed View only if it ranks among the top three entities in the article. For instance, if your organisation is mentioned prominently several times, its sentiment rating is likely to display. However, if mentioned only once, it may not meet the threshold for display.

How can I use it? 

Entity-level sentiment provides a more granular view of media coverage:

The three entities most prominently mentioned in an article or broadcast mention will be displayed. This may indicate the entities that are driving the media narrative or identify the entities that are the focus of negative attention. 

  • It can allow you to see how your brand or competitors are being represented within the broader media narrative.
  • Understand the contextual differences in how each entity is perceived, offering valuable insight into brand reputation and competitive analysis.

Where can I find it?

  • Detailed view: Sentiment ratings for the top 3 entities mentioned in the content are displayed, allowing for detailed comparison between organisations.

 

*Article-level sentiment and Entity-level sentiment for broadcast mentions are under development and are currently in the platform in beta. Sentiment analysis for social media posts is not currently available.