Back to insights
Article • 27/07/2020

A guide to sentiment analysis

Sentiment Analysis is a vital tool for public relations (PR), media relations and communication professionals – not just to ensure they are making the most out of media monitoring, but also in strategic planning.

It allows a measure of the mood around a brand, sector, or industry that goes way beyond a basic count of mentions.

Sentiment analysis enables the assessment of the positivity of news, coverage, and reaction. It provides a gauge of feelings held, and responses provoked to help demonstrate the success of the activity, shape future plans, and initiate action.

What is sentiment analysis?

Sentiment Analysis is the process of understanding the meaning behind the words spoken online and in the media and powers effective media monitoring.

It allows PR professionals to grade whether articles, social media posts, or broadcasts are substantially positive, negative, or neutral.

You can attempt sentiment analysis manually, but doing so effectively and comprehensively in a world of never-ending online, broadcast, print, and social media output is time-consuming. This can mean that urgent responsive action is delayed or becomes too late to implement.

Artificial intelligence (AI) tools have automated sentiment analysis to allow it to be achieved across many forms of media and online output for many topics extremely quickly and effectively.

Via AI software and products, sentiment analysis tools can be used to sort vast quantities of published and broadcast reports and comments by topic into ‘positive,’ ‘negative,’ and ‘neutral’.

The table below offers a simplified idea of the insight given:

Comment Sentiment
‘Sentiment analysis is amazing!’ Positive
‘I hate sentiment analysis’ Negative
‘A definition of sentiment analysis’ Neutral

Sentiment analysis is just one phrase within the ever-growing glossary of PR and media intelligence terms.

What is sentiment analysis used for?

Sentiment analysis is particularly useful for PR, media relations, and communications teams to gauge whether what is said about a brand, competitors, and industry is positive, negative, or neutral.

Too much neutral chatter may imply missed opportunities to promote the activity of your business or organisation.

Positive mentions indicate that current activities are proving successful, which is good for demonstrating results, but also contributes to future planning.

Knowing how popular an initiative has been can help in planning how to manage comms around any changes, for example if a bank has to end a particular offer for its customers. Monitoring the sentiment around the offer can tell you how popular it’s been.

Negative mentions can show a need for a change of communications or business direction—or merely a need to respond to manage the situation and influence the conversation.

What are the benefits of sentiment analysis?

Sentiment analysis helps to demonstrate genuine value in PR campaigns, to shape future campaigns, manage reputation, gain insight into competitors and keep your finger on the pulse during crisis management situations.

Benefits include:

  • assessing the value and outcome of PR and communications activities and enabling meaningful reporting
  • shaping PR campaigns and tone of input into wider conversations
  • informing PR campaigns to ensure they have a real impact
  • offering insight into competitor performance
  • providing an ability for PR teams to support the growth and development of their business or organisation beyond merely ‘spreading the word’ about it and its activities
  • opportunities to enhance public perception by enabling quick and effective responses to negative feedback or, better still, to allow the anticipation and avoidance of that by giving better insight into the overall mood in terms of key subject areas
  • allowing real-time assessment of how a response is fairing in a crisis management situation and enabling necessary adjustments.

Sentiment analysis is a measurement of media coverage and PR value that dates vanity tools and measures such as advertising value equivalency (AVE) and other tactics. It is a measure that drives continually improving performance and shows the value of PR to the C-suite by adding meaning within reports.

Sentiment analysis is an area in which AI is revolutionizing PR.

Daniel Batchelor, Global Head of Corporate Communications at travel technology firm Amadeus, said, “To inform our overall business strategy, we needed a solution to benchmark, sense-check, and understand how stakeholders and customers are reacting globally to the brand and its products.

“We can look at the quantity and quality of our coverage across various topics and visibly compare that with our competitors.”

What are the limitations of sentiment analysis?

To be truly effective, sentiment analysis software needs to understand slang, sarcasm, and nuance.

Words such as ‘sick’ and ‘bad’ change meanings with time and can challenge less advanced sentiment analysis tools. Similarly, certain words used in different ways can convey opposite meanings – such as ‘the call wait times are killing me’ as opposed to ‘this product is really killing it’.

Emojis are now frequently used by people to express emotion, which could challenge some sentiment analysis tools.

Sentiment analysis tools also need to be set up and trained to cope across multiple languages where required.

Sentiment analysis in practice – demonstrating the outcome of an advertising campaign

Signal AI’s analysis of advertising campaigns by three different challenger banks on the London Underground provided interesting insight into their comparable success.

The team behind the campaign for one bank – Revolut – achieved the second-highest volume of coverage, but only 16.6% were classified as positive.

Monzo’s campaign, meanwhile, generated the least amount of coverage, but 30.2% was positive.

Starling, meanwhile, got the most coverage and the most positive coverage. A total of 67.4% of analysed mentions regarding the campaign were positive.

A chart demonstrating sentiment towards campaigns from different challenger banks over time

This is a great example of how a look at sentiment analysis may influence how the success of a campaign is perceived.

Sentiment analysis is a powerful measure, but is only one measure – and other factors are imperative in terms of a rounded assessment of the outcomes of campaigns. If mass coverage is achieved and the majority of it is neutral, or even negative, it may still raise the profile of a product or service to the point of achieving increased sales, for example, which may have been the primary objective.

How to apply sentiment analysis to your media monitoring

Sentiment analysis should be an integral part of any media monitoring software.

Top tips for sentiment analysis usage include:

  • Ensure you are up-to-date with the feelings being expressed around key topics such as your brand name, a specific initiative or PR campaign, and the general mood in your industry to help shape future campaigns, assess the effectiveness of previous or running campaigns, and inform responses.
  • Highlighting negative feedback about your brand, PR campaign, products, services, or practices to help the business and make decisions about future campaigns.
  • Helping to uncover influencers, reporters, and publications that have an audience you could look to talk to with your messages.

Best sentiment analysis tools and software

The best sentiment analysis tools need to be trained to recognize a high level of nuance, offer flexibility for adjustment to meet your specific requirements, and react to and deal with the complex nature of language and communications.

Sarcasm and words that have dual meanings can throw less sophisticated sentiment analysis tools off. That is especially true of those words that can mean the opposite of their original intended definition, such as ‘sick’ for ‘great’ and ‘extra’ as a negative for being a bit fake or over-the-top. Signal AI takes great care to get this right.

Signal AI’s Configurable Sentiment Analysis feature gives communicators the power to choose how to measure positive, negative, or neutral sentiment, with five new ways of aggregating sentiment. 

Sentiment analysis in practice – case study

Effective sentiment analysis tools and reporting are part of the Signal AI platform.

Benjamin Thiele-Long is Director at Cognito – a PR, marketing and communications agency – and he told us how our software enables his business to create meaningful reports to demonstrate the value of its work to clients.

He said: “Hits for the sake of it is never going to impress a board or a C suite but actually understanding the value of why a particular media programme was useful, how it leads to sales generation, how it leads to awareness, how it leads to brand value, is all important.

“Being able to do that by understanding ‘share of voice’, how it works with competitors, what it meant in the landscape as a whole, is really useful for that.

“It also is useful to understand where things didn’t work well because you can put that into context and realise how you might do something differently and that informed process is really what KPIs for PR programmes and media should really be about.

“To provide meaningful insight but also staying as alert in real time as much as possible to what’s happening. That push and pull of understanding the data and information flow that’s key to the job and that’s something that I rely on every single day.”

Signal AI for PR and communications professionals

Signal AI technology offers real-time insight and alerts to global print, online, and broadcast media to support PR and communications professionals in crisis and reputation management, media planning, and reporting.  Sentiment analysis is part of that offer.

Find out more: solutions for PR and communications

You may also like

View More
Article

Sentiment Analysis Measured Your Way: Introducing Configurable Sentiment

For Corporate Communications leaders, sentiment analysis has been the key to understanding public perception. But sentiment can be subjective—from company to company, what your internal team decides is ‘positive,’ ‘negative,’ and ‘neutral’ sentiment can be a subjective reporting decision. Now, Signal AI’s new Configurable Sentiment Analysis feature gives communicators the power to choose how to […]

Read more
Article

Brands Go Quiet this Pride Month, Fearing Rainbow Washing

Every June during Pride Month, the corporate landscape transforms into a kaleidoscope of rainbows and pro-LGBTQIA+ messages. Ideally, it’s an opportunity for corporations to show support for the queer community and advocate for social change. However, it can often come across as “pride-for-profit.” Rainbow washing (a.k.a. pinkwashing or rainbow capitalism) is the marketing practice of […]

Read more
Article

How Messaging is Shaping the UK Election: Insights from Signal AI’s Election Index

The United Kingdom is set for its first election in five years. The Conservative Party, which has ruled since 2010, faces a formidable challenge in retaining power. Running behind in the polls, years of problems and controversies have made the Conservatives a prime target for critics. In this article, we used the live U.K. Elections […]

Read more
Article

Sentiment Analysis Measured Your Way: Introducing Configurable Sentiment

For Corporate Communications leaders, sentiment analysis has been the key to understanding public perception. But sentiment can be subjective—from company to company, what your internal team decides is ‘positive,’ ‘negative,’ and ‘neutral’ sentiment can be a subjective reporting decision. Now, Signal AI’s new Configurable Sentiment Analysis feature gives communicators the power to choose how to […]

Read more
Article

Brands Go Quiet this Pride Month, Fearing Rainbow Washing

Every June during Pride Month, the corporate landscape transforms into a kaleidoscope of rainbows and pro-LGBTQIA+ messages. Ideally, it’s an opportunity for corporations to show support for the queer community and advocate for social change. However, it can often come across as “pride-for-profit.” Rainbow washing (a.k.a. pinkwashing or rainbow capitalism) is the marketing practice of […]

Read more
Article

How Messaging is Shaping the UK Election: Insights from Signal AI’s Election Index

The United Kingdom is set for its first election in five years. The Conservative Party, which has ruled since 2010, faces a formidable challenge in retaining power. Running behind in the polls, years of problems and controversies have made the Conservatives a prime target for critics. In this article, we used the live U.K. Elections […]

Read more
View More