Keeping a U.S. University Secure via Real-Time Social Media Monitoring and Threat Assessment
The client is an American higher education institution and member of its home state’s university system, a postsecondary organization encompassing a dozen schools and serving more than 400,000 students.
At a glance:
In April 2023, a student posted a threatening tweet toward the campus community. The student was arrested, but in the incident’s wake the institution became very concerned about future threats and negative social media conversations about the university.
The university asked Fullintel to implement an enhanced social media monitoring system to gauge sentiment and help ensure a secure campus environment.
The client communication’s team’s goal was to track anyone discussing the university and its campus community on all relevant social platforms, including Discord, Twitter, TikTok, Instagram, YouTube, Facebook, and LinkedIn.
The team needed to be alerted in real-time to any potential threats to keep the community safe – not an easy proposition when monitoring several thousand posts per day on popular social and discussion platforms.
The project included the following objectives:
- Develop and implement a robust monitoring system that scans social media platforms for posts and comments containing threats of violence against any university community member, as well as content that may be detrimental to the university
- Proactively detect and categorize threats based on severity, while establishing clear response and escalation protocols for each threat severity level
- Provide real-time notifications and alerts to university management and local police for immediate action
- Continuously monitor for accuracy and effectiveness and evaluate outcomes to make necessary adjustments
Fullintel’s social media monitoring and crisis management services were well-positioned to meet this challenge through real-time alerts, daily analysis reports, and human-curated monitoring of all relevant social channels.
Fullintel proposed a comprehensive, human-curated, real-time social media threat detection and alert system including real-time alerts and weekly reporting. This proposed system needed to operate around the clock, every day of the week – even holidays.
Fullintel’s combined human social media curation, real-time alerts, and weekly reporting powered by a series of qualitative and quantitative metrics and insights including sentiment and reach.
To avoid excess noise or missed posts, monitoring is conducted manually with some automated assistance. This hybrid approach is how we provide deep intelligence from LinkedIn and other platforms unavailable in automated SaaS tools, and provides a more accurate, relevant, and reliable dataset than fully automated monitoring alone.
This approach also allows for much faster and more accurate threat assessment in situations where seconds can make a huge difference.
Fullintel’s strategy included the following elements:
Automated monitoring with human curation: Fullintel continually receives automated content feeds from social channels, 24 hours a day, seven days a week. This content is ingested and validated by an expert Fullintel analyst to assess the level of potential risk.
Manual monitoring: Fullintel analysts also manually comb through social media channels to fill any gaps that may have been missed. Fullintel manually monitors the university’s Discord channels to identify any conversations that may pose a potential threat.
Owned media monitoring: All private and direct messages and comments received by or posted on the social media pages handled by the university are monitored and analyzed.
Real-time alerts: Fullintel promptly alerts the university via phone call or email if any posts or comments contain threats of violence against any member of the university community, or if the content negatively impacts the university.
Weekly curated reporting: Weekly curated wrap-ups of the week’s social content is provided in Excel file format. This facilitates convenient tracking and monitoring for university executives, informing them about weekly occurrences that have a negative tone or may pose a threat.
Fullintel analysts now track social media posts and comments using multiple monitoring platforms to assess sentiment, with a particular focus on negative sentiment.
Comments on posts are monitored to identify any changes in attitude exhibited by individuals towards the university.
Fullintel collaborated closely with the institution’s emergency management services, campus police, and communications staff to develop the system and its threat classifications. All social media posts and comments containing a threat or that could have a negative impact on the university are identified and classified as one of the following three threat levels:
- Red: Posts/comments with threats to public safety, sexual violence, or threats to property. Fullintel promptly notifies the communications team by phone call, followed by a detailed email informing stakeholders.
- Orange: Posts/comments that involve hate speech, suicidal ideation, or targeted harassment. Fullintel immediately notifies the communications team by phone call, followed by a detailed email to stakeholders.
- Yellow: Posts/comments that involve negative commentary of a general nature, including any mentions of hazing. Fullintel notifies the communications team via a detailed email to stakeholders. A phone call is not required for Yellow Alerts.
Fullintel uses a combination of quantitative and qualitative metrics for media measurement:
- Quantitative metrics: Number of posts, audience drivers
- Qualitative metrics: Post/comment sentiment, sentiment over time, alert/threat categories
Comprehensive sentiment analysis, which combines automated algorithms and analyst-generated sentiment data to enable a more accurate assessment, is also used. Analysts manually verify and classify automated sentiment data into one of three threat level categories.
Fullintel’s human-driven content collection, curation, and analysis – combined with a unique reporting approach, including classifying threat levels by type and contacting the university with real-time threat alerts as required – now provides a clear view of the nature of social conversations about the university, potential threats, and potential bad actors.
Not only does human curation and analysis allow the inclusion of important LinkedIn content (a platform not readily available on SaaS media monitoring platforms) but it helps better separate relevant signals from noise.
This approach also helps the university detect threats and keep relevant stakeholders informed in a timely manner while dismissing innocent content that could erroneously be deemed a threat via a fully automated approach.
The university now relies on Fullintel’s real-time alerts in the event of any potential threat or excessively negative commentary, which are communicated to stakeholders who then initiate appropriate response procedures. The FGCU communications team now relies on weekly, human-curated media analysis reports to track and analyze any potential issues.
Thanks to Fullintel’s diligent reporting, the university met its objectives of:
- Implementing a robust social media monitoring system
- Proactively detecting and categorizing threats based on severity, along with establishing clear response and escalation protocols
- Establishing real-time alerts to spur immediate action in case of a threat
- Continuously monitoring for accuracy and effectiveness, and making necessary adjustments
Most importantly, thanks to Fullintel, the university met its goal of being able to track anyone discussing the university and its campus community on relevant social platforms, being alerted in real-time to any potential threats, and having better tools to keep the community safe.
Fullintel’s categorization of alerts into red, orange, and yellow categories enables university officials to navigate and respond to posts or comments on social media in a structured manner. This approach aligns with the severity and urgency of each alert, ensuring appropriate actions are taken.