By Sam Ruchlewicz—One of the most interesting and under-studied applications of communications data is during crises. Perhaps because these events occur relatively infrequently. Perhaps because most senior communications professionals have absolutely no idea what data is, or how it can be used. Perhaps because when a crisis situation does arise, the last thing most communications professionals are thinking about is what the data says about the situation. Or perhaps it’s a combination of the three.
Whatever the reason, the importance of data for successfully handling a crisis cannot be overstated. Today’s marketing communications professionals have access to previously unfathomable amounts of information—from social media sentiment to the relative influence of individuals to the clustering of digital activity—all of which can be brought to bear in creating, refining, and executing crisis communications strategies.
3 Phases of a Crisis
Crisis situations evolve in three phases:
- The build-up phase begins with the “crisis trigger” and ends when the situation has become a full-fledged event;
- The impact phase encapsulates the “crisis event” itself; and
- The recovery phase is the aftermath.
The Build-Up Phase
Due in large part to our pervasive use of social media, crisis situations today evolve more rapidly than ever before. Several years ago, a customer complaint or product defect (or other problematic situation) would require several days or even weeks (along with multiple media reports, newspaper articles, etc.) to balloon into a full-fledged crisis situation. Today, a similar spark in our social media-driven environment can explode into a full-scale crisis in mere hours or minutes.
In essence, the hyper-connected, always-on nature of today’s digital environment has compressed the build-up phase dramatically, reducing the efficacy of the traditional safeguards (editorial review, sourcing requirements, etc.) that have protected brands for much of the past century. Today, anyone with a mobile device is a journalist, and anything posted online can go viral in a matter of minutes, if the right conditions are in place.
Regardless of the time lapse between a triggering event and when a crisis situation emerges, a pair of ingredients still must be present:
- A critical mass of awareness about the triggering event, and
- A sufficiently high level of negative sentiment among key target audiences.
Absent the triggering event and the two aforementioned factors, a crisis situation will not emerge; the spark will simply fail to ignite a crisis.
I find one of the most interesting (and tragic) examples of this phenomenon to be a tweet speculating about the untimely passing of star singer Whitney Houston, posted at 8:02 pm by @BarBeeBritt to her 799 followers. 55 minutes later, at 8:57 pm, the AP picked up the story, and in the 60 minutes immediately following that tweet, there were nearly 2.5M tweets about Whitney Houston’s death.
While this isn’t a corporate crisis example, it is illustrative of two things:
- The importance of connections to crisis situations, and
- How information spreads in today’s environment.
Twitter user @BarBeeBritt’s connections were limited (she only had 799 after all) and, as a result, her pertinent message fell on deaf ears. It wasn’t until an hour later that the message was amplified by a sufficiently influential source (the AP) to ignite an online firestorm.
Develop an early warning system to predict triggers and plan responses
If this was an example about a company or a product, that delta (55 minutes, in this case) between the first mention of a triggering event and the amplification of it by a sufficiently influential source represents the opportunity to mitigate or altogether avoid a crisis situation – if can be detected early enough.
Fortunately, there are a number of innovative tools and approaches that can be used to do exactly that:
- Social listening tools (such as Brandwatch, Sprinklr, Crimson Hexagon, Hootsuite, etc.) constantly scour social networks for brand mentions. They enable you to create custom alerts to notify your team when the volume, frequency, or negative sentiment of online discussions spike.
- Even better, combine rules-based social listening tools with online influence metrics (like a follower count, or metrics from an influence tracking platform like Traackr, Kred, or Klout) using APIs in real time. In this way you can quantify the potential impact each participant in a potentially-triggering event can have and adjust your response accordingly.
- Combine those risk models with a detailed understanding of your brand’s perception in the marketplace (which is a proxy for the flammability of the overall environment) and you can build models that reasonably predict how probable a given trigger is to result in a crisis situation, and respond accordingly.
While in today’s environment a single individual with sufficiently high influence can manufacture a crisis (see Trump, Donald J.), it is important to have an early warning system in place. Having those few precious minutes provides sufficient time for your team to formulate a response and prepare for the coming onslaught.
The Impact Phase
Despite your best intentions and diligent preparation, a crisis situation may emerge. When it does, your team needs to be prepared to move through it as quickly and efficiently as possible.
Having an early warning system like those detailed above will provide your team with precious time to assemble your crisis team, formulate appropriate messaging, halt outgoing advertising campaigns (if necessary), and begin to get ahead of the situation.
A crisis in today’s digital ecosystem is likely to be accompanied by a massive volume, variety, and velocity of communications taking place across multiple platforms, both online and offline.
The primary challenge communications professionals face in these situations is exponential growth in communications activity. A company in the midst of a crisis may see a 25,000% increase in mentions overnight. Monitoring that volume of conversations—let alone responding to them—is effectively impossible for many companies and agencies.
The below chart, compiled by Zignal Labs, provides a visual representation of the social media maelstrom faced by automaker Volkswagen after news broke that the firm cheated on emissions tests:
The correlation between VW’s brand index score (a composite of the overall sentiment around the brand at the time) and the company’s stock price is strongly positive, suggesting that the situation had a materially adverse effect not only on VW’s reputation, but on the company’s bottom-line as well.
This type of information can be leveraged in three ways to successfully navigate a crisis situation:
1. Provide insight into “hubs” of crisis conversation to prioritize efforts and resources.
As mentioned above, most crisis situations take place across a number of different platforms, both on- and off-line. Your goal shouldn’t be to monitor, engage or respond everywhere these conversations are taking place—doing so is virtually impossible—but to focus resources in the right places.
What are the “right places?”
Generally speaking, they are the nexuses or “hubs” driving and amplifying the crisis conversation. In most situations, there are a (relatively) few number of digital properties that fit this bill. Identifying and prioritizing efforts on those platforms can have a disproportionately large impact on mitigating the situation as a whole.
Pivoting to another crisis example, here is a spider chart (again courtesy of Zignal Labs) that represents the activity surrounding the 2016 Apple iPhone crisis that emerged after the company refused to unlock the phone of the San Bernadino shooter:
In the above chart, the blue dots represent domains where the conversation is happening (with the size of the dot representative of the level of activity); the orange dots represent stories, and the fuchsia/turquoise dots represent tweets and retweets, respectively.
This single chart provides an incredible amount of actionable information in an easy-to-understand format. As you can see, the vast majority of crisis activity is taking place around a single story posted on a major domain (in this case, the site is TechCrunch). There are several other stories on major platforms (NY Times, Wired, etc. as denoted by the other large orange circles), but none are generating the volume of activity of the Techcrunch piece.
This type of chart allows communications professionals to prioritize activities and focus responses exactly where they are needed. Responses could focus on a number of different tactics, from asking the publisher to correct any inaccuracies to offering an exclusive interview in order to present the company’s side of the story. Depending on the sophistication of your marketing and communications teams, it is possible to concentrate advertisements on specific, high-value properties that take readers to a specific landing page with the facts of the situation.
Here is similar example from a Danish bank crisis, in which a visual mapping tool was developed based on Facebook data. In this case, network visualizations allow identification of key social media players and political stances.
Regardless of the approach you choose, the reality is this: in a crisis situation, there are thousands of skirmishes occurring in every corner of the internet being fueled by a handful of major battles. The best approach is to focus your attention on turning the tide of those major battles — after all, those conversations are driving everything else.
2. Understand what’s working and what isn’t.
As most communications professionals can attest, turning the tide of major battles is significantly more difficult than it sounds. There are dozens (if not hundreds or thousands) of different messages that can be used in crisis situations, each with varying degrees of efficacy.
Take, for instance, the chart below that breaks down average sentiment by age and gender across several different platforms, as well as the trend for each. Having this information at your fingertips while crafting a message allows you to identify what group(s) are driving the crisis and tailor messaging to them.
The applications of data to crisis situations don’t stop with audience identification. The same data can provide brands with real-time updates on the impact of various messages on the overall sentiment of the conversation.
One way this can be represented is a semantic analysis map with a sentiment overlay, as shown below. Various messages around “vaccines” are shown, with sentiment color-coded in the dots accompanying each.
A chart such as this can be updated in real-time during a crisis, providing your response team with a clear picture of how the situation is evolving, which messages are shifting the conversation, what new issues are arising (if any), and how the overall sentiment around your company/brand/product is changing.
Data, when deployed in this manner, can provide your team with incredible insight into tangentially-related topics (like SB 277, Rand Paul, and Chris Christie from the above chart) that may merit additional investigation or risk-management issues. It can go one step farther, providing your team with an analysis of the centrality of each tangent to the over-arching conversation, as well as the net sentiment of the activity surrounding each topic.
This data can be invaluable in anticipating how a crisis situation will unfold, identifying potential risk factors and crafting the best possible responses.
3. Respond at scale to diffuse the crisis drivers.
Once you’ve located the nexuses of crisis activity and refined your messaging, the final step is to deploy these messages at scale to begin to contain the situation.
Depending on the crisis situation, you may not want to launch a full-scale outreach program. In fact, your data may show that the crisis situation is contained to one or two regions, as the below chart from Zignal shows:
In this case, the most effective response might be to prioritize the two regions most impacted by the crisis (western U.S. and eastern Canada in the above map) with a full-scale outreach effort.
Again, the smarter your team can play, the more quickly and efficiently a crisis can be addressed and normalcy returned.
The recovery phase
Once a crisis has subsided, the work of the communications team is far from over. There are debriefs to hold, brand reputation analyses to conduct, and on-going monitoring to put in place. For each of these activities data can play a central role, helping you and your team understand why the crisis happened and what can be done moving forward to facilitate better outcomes.
Most importantly, data can be used to quantify the success your team had in resolving the crisis. While there are dozens of approaches (all with their merits) that can be used, one of the best is counterfactual economic modeling, which asks, “If we hadn’t responded to the situation as we had, what would the outcome have been?”
Assuming your organization has a robust data collection process in place, modeling the impact of likely crisis scenarios isn’t particularly complex (as far as modeling goes), but it can be insightful.
What if your team hadn’t quickly identified the crisis and been prepared with comments and responses the second a story was posted? Well, odds are that the story would’ve mushroomed, reaching millions of people before a response was posted. By that point, your brand index would’ve dropped by X, resulting in a net loss of $Y.
Assembling this type of business case—complete with actual dollar values—is how communications professionals demonstrate their value to the rest of the C-Suite.