PR measurement can feel like a thankless task—all of that sifting through content and analysis—but it is vitally important to assessing the effectiveness of PR work.
Ongoing analysis can surface useful trends in stories covered, audience response, and can even help to predict crises. One of the most promising features of artificial intelligence is its potential to help public relations professionals better use the data and information collected by media monitoring tools.
Assessing large data sets to predict the response of target audiences and forecasting outcomes of messaging campaigns are just two examples of how predictive analytics can be deployed for better PR results.
Media Trends
The “predictive” part of predictive analytics can help in the development of strategy and planning.
Media monitoring, based on keywords and key phrases, provides an ample data set. Analysing content collected over time can reveal information about journalistic priorities, for example.
It can also uncover topical trends. For example, intuitively it makes sense to assume that journalists will cover environmental topics around Earth Day. However, it takes a more detailed analysis to discover effects like which types of environmental stories are more likely to be covered, which journalists cover the topics appropriate to your specific industry, when these stories start to be published before Earth Day and when they tail off, and how stories are framed.
This more detailed analysis can help to predict the types of pitches that will be more successful, when to pitch them, and who is best to pitch.
Audience Response
Reactions of target audiences are important to both capture and understand—but they can be challenging to predict.
Human responses are going to depend on a variety of variables, up to and including things such as current events, or social and cultural mores. Responses can even be impacted by fatigue with a particular fad or trend, and it can be difficult to predict when people will tire of a particular craze, whether it is eating an entire cucumber in one sitting or Georgina Rodriguez’s potato omelette video.
In the event of a natural disaster, opinions shared online can even be affected by the weather.
With so much potential variability, paying attention to how audiences are reacting to events in near-real time can be a huge benefit. Predictive analytics can make sense of large amounts of data, tracking sentiment and providing rapid feedback so that a brand can adjust responses or messaging if needed.
Growing use of AI
AI can help to alleviate one of measurement’s biggest hurdles: the fact that it is time-consuming to measure the things that really matter, and PR professionals struggle to be compensated for the additional effort. Budgets are tight, analysis takes time, and time is money—that is the challenge for a lot of measurement professionals.
Artificial intelligence continues to make strides, and with the proper training models it can produce highly useful results. Ideally, AI will tackle the routine tasks associated with monitoring, assessing large amounts of information—tasks that used to be done manually—revealing relevant data points. This can then be analysed and the resulting strategic recommendations are based on data, rather than guesswork.
The Human Factor
Recently, the New York Times ran an article about a small study that used ChatGPT to assess medical case histories. The study randomly assigned medical files to three groups: doctors with access to standard resources but not ChatGPT, doctors provided ChatGPT along with standard resources, and ChatGPT alone.
Of these three groups, success rates in correctly diagnosing illnesses were as follows: doctors with access to ChatGPT achieved an average accuracy score of 76 percent; those without access scored 74 percent; and ChatGPT alone scored…90 percent.
Why did the AI alone score better than doctors who had access to the AI? Although this was a small study, researchers suggest that doctors prioritised their experience over the AI and ignored its conclusions.
How do we as humans know when to trust our instincts versus putting our trust in conclusions reached by artificial intelligence? This is likely to become the key question for PR professionals as we move forward in incorporating AI into our everyday work lives.
Trusting results generated by AI, particularly when they may run counter to long-held assumptions, will take time. One way to build trust in results is to start small, using AI for lower-risk decision-making.
Prediciting Crises
PR professionals know that a crisis can happen without warning. That said, the more you can do to prepare for a potential crisis ahead of time the more you will be able to manage one if (or when) it happens.
Paying attention to rising negative sentiment can be an early warning sign of a crisis in its early stages. Although this is the most direct indicator, it is not the only way to use predictive analytics to detect a crisis.
Keyword analysis can also help to point to trouble brewing. Depending on the industry, using technology to detect words that might be an indication that problems are increasing could help to pinpoint issues that could lead to a crisis—before things hit crisis levels.
Conclusion
Public relations professionals have long struggled to get the most out of monitoring, particularly when trying to convert media results into actionable data. AI can be a valuable new partner in measurement.
The use of algorithms and predictive analytics can help to improve strategic recommendations when designing PR campaigns and communications programs. With better messaging, better targeting, and an improved understanding of possible outcomes, the use of predictive analytics is the future of PR measurement.
