No longer science fiction: Software is helping to detect crises before they happen

But platforms can’t tell the future without a human touch.

Photo credit: Getty Images
Photo credit: Getty Images

The sheer accessibility of information through news sources, as well as word of mouth and social media, means that the risk of a communications crisis looms larger than ever before.

Something like an altercation between a flight attendant and a passenger on a plane, which at one time would not have seen the light of day, can now go viral on Facebook or Twitter, creating a major headache for an airline.

How can crisis communications pros know if an incident will bubble into a full-blown crisis? Can software detect a crisis before it happens? The answer is a resounding, but unsatisfying, “yes, but...”

First, it is helpful to understand what these tools are doing and what they’re capable of doing. Leigh Fatzinger, founder and CEO of Turbine Labs, explains that there are two overarching types of technologies that can make sense of when a “seemingly innocuous” event could become a crisis.

The first category is predictive analytics, “which looks at historical data across a wide variety of events and makes statistical inferences on the likelihood of similar events occurring in the future,” says Fatzinger.

The second is advanced machine learning, “which looks for the formation of small clusters of conversation or coverage within vast sets of data, looking for patterns in the data that forecast the potential growth or amplification of a topic or event,” he adds.

Turbine Labs falls largely into the latter category, focusing on cluster detection using advanced machine learning. Another software provider, Alva, falls into the first. It aggregates millions of pieces of publicly available content every day. The software generates metatags for each piece of content and then uses an algorithm to assign a reputational value to each piece of content within a tag, generating a sentiment analysis. Other tools, such as NewsWhip, monitor engagement with social and web content.

There are countless intelligence, social media engagement tracking and media monitoring tools that compile and make sense of tens of thousands of pieces of data in real-time in a way that wasn’t possible before. However, even the founders and CEOs of these companies will tell you that software requires the complement of human intelligence to provide the necessary context to make sense of the data.

Traditionally, communicators hadn’t been able to predict crises because the wealth of data wasn’t available. “In the past, we just knew that there was a story published in The New York Times, but we had no idea if it created a stir; we just knew what the readership of the publication was,” explains Paul Quigley, founder of NewsWhip. Today, communicators have access to readership density and can see how much a story is being shared, which Facebook and Twitter accounts are sharing it and how popular the story is. That gives users a better idea of how big of a deal it is, Quigley says, and creates a large digital signal that can help inform a response.

Humans can still beat software to the punch. Quigley notes that “crises rarely come out of the blue. Usually, the relevant party knows something is coming up.” This is often the case with an incident like a data breach, which comes into public knowledge via disclosure, or an earnings report, which is made public at a scheduled date and time.

Where software becomes more relevant is helping brands or companies stay ahead of “slower, water-level-rising” social or environmental changes that could lead to a crisis, Quigley argues. In other words, software can be used to monitor for slowly changing public attitudes that could lead to a part of a particular business becoming problematic.

red plastic straw
red plastic straw

“You don’t want to be the last orange juice maker with a plastic straw on the back of the carton,” Quigley says. “It was completely innocuous a few years ago, but now it has become taboo. Monitoring tools can help stay ahead of that.”

Alva cofounder Alastair Pickering notes that even a technologically advanced piece of software needs human interaction with data to tell a story — to provide the “so what?” In other words, the tech piece is perhaps necessary, but not sufficient.

“Technology won’t replace function,” Pickering says. “Functional expertise is invaluable to make sense of shifts and patterns. It’s the partnership between the two that’s vital.”

Crisis communications specialists at some of the world’s largest PR agencies agree. Matt Groch, global lead of data analytics and innovation for FleishmanHillard’s TRUE Global Intelligence, argues that “what is key is less about the technology, per se, and more about having access to the right data and being able to analyze that data to produce actionable intelligence.” It is this “contextual interpretation” that is crucial.

What the data can do is tell you when a known issue is “gaining traction and likely to hit a flash point, or when a new and relatively unknown issue is just emerging but has a high degree of potential to emerge as a prominent reputational threat,” Groch explains.

Edelman’s global chair of crisis and reputation risk, Harlan Loeb, says that while software is “part of the solution,” human intelligence should be used in concert with data streams to “predict the multiple paths a crisis might take.” This is because off-the-shelf software might involuntarily miss critical factors, particularly if it over-emphasizes quantitative data. A human is able to give the data “qualitative depth that a software can’t replicate.”

Weber Shandwick uses software platforms, “but ultimately software is a tool that informs human analysis,” says Peter Duda, EVP and co-head of crisis communications and issues. Relying on software alone is risky, Duda argues, because crises are driven by a number of factors, at a level of complexity that only human analysis is capable of grasping.

Ultimately software is a tool that informs human analysis.

Software can compile and sift through vast troves of data in a way that might not be possible by a human. It can also identify changes in trends and patterns and call attention to anything moving abnormally, serving as an early warning system. Yet there is no magical solution that can predict a crisis. Existing technology requires a human touch to provide the necessary context, helping distinguish between positive and negative information and marking the critical difference between what could be a good news story or a potential crisis.

Have you registered with us yet?

Register now to enjoy more articles and free email bulletins

Already registered?
Sign in