<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=1028521297200037&amp;ev=PageView&amp;noscript=1">

PR Reporting

The Advantages and Limits of PR Sentiment Analysis

October 14, 2014 03:33 PM
0 shares
AJ Bruno: Founder, President

AJ Bruno: Founder, President

AJ co-founded TrendKite in 2012 and oversees all aspects of the organization’s sales operations.

All POSTS

shutterstock_176964464Whoever said that any press is good press was an idiot.

The examples of bad publicity having a negative impact on business performance are too many to count. Sea World, for example, has seen a decline in attendance, visits and stock price since the documentary film, Blackfish, brought to light killer whale training and husbandry practices that many Americans find deeply troubling. For Sea World, Blackfish was not good press.  Malaysia Airlines may have to rebrand after seeing a bookings decline of 33% following two tragic disasters. Airplane crashes are not good press. So, if not all press is good press, it stands to reason that it is as important to understand what is being said about your brand as it is to understand how often your brand is mentioned.

What is PR Sentiment Analysis?

Sentiment analysis, sometimes called “opinion mining,” is the practice of assigning an attitude to each brand mention. Usually the designations of, positive, negative, and neutral are used.  PR software uses natural language processing, computational linguistics and text analysis to uncover the subjective outlook of the writer.

Advantages

PR sentiment analysis gives brands an efficient way to look at big-picture trends across all media channels. PR professionals and agencies are able to determine how news impacts the overall perception of the brand and which stories are most effective at generating positive sentiment. Performing sentiment analysis for competitors is also useful in setting benchmarks and probing for competitor weaknesses.   

The growing importance of social networks and product review websites has brought renewed attention to the practice. Online reputation can make or break a brand’s fortunes.  Sentiment analysis can help filter through mass quantities of content and identify the specific content that needs attention from a reputation management prospective.

Limitations

Of course, software isn’t 100% accurate at determining sentiment.  In fact, it isn’t unusual for two humans reading the same content to disagree about the writer’s mindset. So while sentiment analysis is very effective at recognizing trends and identifying outliers, it is still up to humans to parse the fine nuances of human language. Context also plays a big role in understanding a writer’s feelings on a subject. “That movie was bad!” is definitely negative sentiment from a 50 year old film critic, but it might be glowing praise from a 17 year old boy. 

Sentiment analysis is an important tool for monitoring public relations and reputation management success. When used thoughtfully, it can help measure past performance and improve future results. 

Recommendations

When selecting a PR Analytics tool or evaluating the sentiment capabilities of your current provider, there are a few questions to ask.

  • Language Support: If you business is global, it is critical to have broad language coverage.  Often, regional coverage or issues will drive sentiment, so seeing a huge spike in a specific language or region is a good indicator of a problem to tackle.  Your provider should be analyzing between 15 and 20 languages to cover the major regions of the world adequately.
  • "Named Entity Extraction": If you have spearheaded a charitable campaign for natural disaster relief, the "sentiment of the article" will be negative, but your brand will be viewed in a positive light.  Headline or overall article sentiment is important, but understanding how you, your product, or competitor is portrayed is even more important.
  • Accuracy: The more content there is, the more accurate the sentiment should be.  Sentiment on a tweet, at 140 characters, won't get much better than 70%.  However, blog posts and news articles should be highly accurate.
  • Tunability: Does your provider learn over time and surface tools for you to flag or override inaccurate sentiment analysis?
AJ Bruno: Founder, President

AJ Bruno: Founder, President

AJ co-founded TrendKite in 2012 and oversees all aspects of the organization’s sales operations.

All POSTS

comments