Using early detection to head off a crisis

APCO Insight’s Darren Bosik on detecting and countering a potentially ugly scenario.

(Photo credit: Getty Images).
(Photo credit: Getty Images).

For PR pros, data is always an essential ingredient for a job well done. From social listening to glean what a brand's customers are saying, to media monitoring to understand how a brand is being portrayed in the press, insights gleaned from the growing availability of data means that brands can ensure they're sharing the right messages.

Having access, and paying close attention, to this data is even more critical when it comes to existing or potential crises. We spoke to APCO Insight senior director Darren Bosik about how the data and analytics arm of the agency averted a crisis for a client.

APCO works with a large U.S.-based energy company that serves more than a million customers in the Southern part of the country. The agency has created a proactive and predictive corporate communications strategy to address negative coverage related to the company's reputation. This includes weekly monitoring for issues that could threaten the brand's corporate objectives or its ability to do business.

Recently, APCO Insight's system detected a potential crisis brewing over the CEO's compensation when influential voices suggested he is overpaid. But how did the system "know" this?

"APCO Insight leverages data from multiple sources, including social listening and media software and developed an early detection crisis and issues management system," Bosik says. "Using AI, the system can predict when an issue in the marketplace is going to pose a risk to the energy company's communications business objectives and corporate reputation."

APCO Insight hooks up the API of a variety of third-party tools, such as Talkwalker, Quid, Meltwater and Cision, and feeds them into its own data lake. As necessary, the team would also include secondary data, such as lifestyle behavior, attitudinal data and proprietary APCO opinion research derived from surveys. This is a basic framework that is customizable based on a client's needs.

All of this data is regularly fed into the data lake. From there, the team can query it and apply algorithms like artificial intelligence to gain a deeper understanding.

"We apply predictive analytics, modeling and AI to detect when an issue is likely turning into a crisis, or conversely, if it is something positive for the client," Bosik says.

APCO's team can write python scripts that are able to monitor for specific issues that may need addressing, either negative or positive. Typically, it tracks metrics like issue volume, scope, silhouette, influencer involvement, audience reach and rate of change.

The changes in data trends are what will typically alert them to the possibility of an emerging crisis.

"We particularly look at momentum, or rate of change over time," Bosik explains. "The system can detect if there's a potential spike based on that rate of change. If it's a movement toward the negative or a high rate of change, it will be able to detect that."

From there, the firm can predict the mix of media and key audiences that will likely continue to engage over a period of days and can develop appropriate response scenarios.
In the case of the energy company CEO, the team detected increased chatter among influential actors about his compensation. It alerted the account team and shared its findings with the client, pointing specifically to the data about critical metrics that raised the possibility of a crisis scenario.

After providing possible scenarios and responses, along with a list of targeted media and proposed messages for different audiences, the company's comms team agreed to a course of action. The data told them to counteract the misinformation by providing factual information about how the CEO was being appropriately compensated. They also developed talking points on the positive contributions the CEO had made not just to the company and its employees, but also to the local communities it serves.

APCO's team can segment data by key audiences and media, which enables them to target a response and gives them insight on what channels to focus on and which messages would result in a highly engaged response to minimize negative impact.

This early detection system helped not just to avert a possible crisis but also enabled the company and the CEO to turn what could have been a negative story into something that was a net positive.

"The key to all of this is that it's a data-driven response. We want to optimize reach and potential engagement and minimize potential impact," Bosik adds. "We're not taking screenshots from say, Talkwalker, but taking this third-party data and creating our own customized solutions for our clients. That's our 'secret sauce' that sets us apart from the data offerings of other agencies."

Based on weekly monitoring of media, online conversations and stakeholders, APCO's staffers update its scripts and approach to ensure it is including the most relevant information about the client and their industry.

The software itself is necessary, but not sufficient. Analysts, data scientists, coders and research professionals are critical to the total effort. Bosik calls it an art and a science, with the human component key to ensuring that the scripts and code remain up-to-date based on the changing needs of the agency's clients. The APCO Insight team also works closely with the account teams across the agency to provide them with advice based on their insights.

"What we're doing is uncovering hidden truths of what's going on related to a potential crisis," Bosik adds. "We articulate those hidden truths first internally and then externally to the client, and then come up with a plan and response to minimize the possibility of a crisis arising."

Have you registered with us yet?

Register now to enjoy more articles and free email bulletins

Already registered?
Sign in