When a crisis hits, every word in your emergency notification either helps people or hurts them. The University of Oregon spent the better part of two years rebuilding their alert system from the ground up. Here is what they figured out, and what any campus emergency manager can take away.
Where They Started
In February 2024, UO sent a campus alert describing sexual assault incidents using the word “fondling.” The community response was swift and pointed. Respondents called the term outdated, noted that its everyday definition implies positive and wanted contact, and said it “belittles what these actions actually are.” One person wrote directly that they expected better from the Emergency Management department when it comes to language that empowers survivors.
That feedback became the catalyst for a full rethink of how UO writes and sends alerts. The central insight was simple but profound: words matter. Not just legally or procedurally, but to the people receiving your message at the worst moment of their day.
Principle 1: Tell People What to Do, Not What You Think They Already Know
The biggest shift in UO’s philosophy was moving away from detailed, incident specific descriptions toward action focused common language. Old alerts tried to explain what was happening. The new framework tells people what to do about it.
During an emergency, clear direction matters more than information alone. A message that says “Law enforcement is responding to multiple reports of fondling near Barnhart Hall” gives the reader information but no clear next step. A message built around an action word like “Avoid” or “Evacuate” does the opposite. It leads with the behavior you need from your audience and lets the details follow.
Your first alert should function as a heads up. Your second should add information. Neither should try to do everything at once.
Principle 2: Keep It Short and Keep It Consistent
UO found that the I Love You Guys Foundation’s standard message length of 100 to 140 characters is a useful target. Text messages have hard limits, and even where they do not, long messages get skimmed or misread under stress. Short means scannable.
Consistent means your audience builds familiarity with the format over time, so they know what to expect even when they are scared. The same core message should go out across every channel, text, email, web page, worded the same way. Your alerts webpage should be the source of truth, with the text message functioning as a pointer to it, not a replacement for it.
One practical note: text should not be the end all be all. Email and your webpage can carry the fuller picture where Clery Act requirements kick in and more descriptive language becomes necessary.
Principle 3: Build a Framework, Not a Script
UO went through several iterations before landing on their current Alert Actions framework. They started with the I Love You Guys Foundation’s Standard Response Protocol (Hold, Secure, Lockdown, Evacuate, Shelter), then built BLARES (Be Aware, Lockdown, Avoid, Run/Hide/Fight, Evacuate, Shelter) as a customized version, and ultimately even retired BLARES when it proved less effective in practice.
Lockdown was removed because it caused campus confusion and was nearly impossible to practice given access control realities. The order of actions in BLARES did not reflect urgency, which undermined its usefulness in a real event.
The current framework:
- Be Aware
- Avoid
- Evacuate
- Shelter
- Run/Hide/Fight
- All Clear
Was built around what people can actually do, in language they can actually follow. Each action is paired with specific behaviors for students, instructors, and staff. “All Clear” was added as a formal closing action rather than an afterthought, which matters enormously for community trust.
When building your own framework, pull from multiple sources: look at what peer institutions are doing, review your own past alerts honestly, factor in Clery considerations, and think carefully about your specific audience. What works at UO may need adjustment for a community college, a health sciences campus, or a smaller residential institution.
Principle 4: Know Who Is in Charge and Make It Policy
One of the sharpest operational lessons from UO’s experience is around authorization. Typically dispatch identifies that something is happening and can take initial action, but often the admin on duty makes the call on whether an alert goes out. This distinction matters enormously under pressure, when the instinct can be to either act too fast or get mired in consultation.
UO made a deliberate decision to eliminate the consultative call before sending alerts. The reasoning is direct: it wastes time, and time is the resource you have least of. Instead, clear role ownership and written policy replace the committee dynamic. Someone owns the decision. That person is reachable 24/7 via duty phone.
For any institution building or rebuilding this process, the questions to answer in advance are: who is authorized to send, who decides, and how does the dispatcher connect to that person at 2am on a holiday weekend?
Principle 5: Use Notification Levels and Let Your Webpage Do the Heavy Lifting
Not every incident requires a full campus alert. UO structures their notifications in levels, with alerts at the top and web page updates handling ongoing or lower urgency communication. This prevents alert fatigue, which is a genuine risk. If your community gets too many messages that turn out to be minor, they start ignoring the ones that matter.
The staff message board and other internal channels receive alerts too, but something worth remembering: be wary of misinformation spreading through those channels when an incident is active. The official webpage should be clearly positioned as the only authoritative source, and your messaging should direct people there explicitly.
Principle 6: Practise the Hard Scenarios, Not Just the Comfortable Ones
Run scenario workshops with all administrators on duty — Safety and Risk Services, Police, Communications, Environmental Health and Safety, and optionally Dispatch supervisors. Many cooks in the kitchen is a feature of emergency management, not a bug, but it requires someone steering the kitchen. A sample agendas may look something like: review alert actions and templates, cover notification levels and authorized campuses, walk through the workflow, run scenarios, then debrief.
Keep this phrase in mind “correct or fast”
When those two things are in tension, your team needs to have already decided which one wins. Practice helps you make that decision before the adrenaline is flowing.
Test your weird scenarios, not just the obvious ones. Run drills after hours, because that is when your real capability gaps will show. Expect mistakes and treat them as data, not failures.
Principle 7: Anchor the Purpose
The goal of an emergency alert is not to demonstrate institutional competence, cover legal requirements, or satisfy every stakeholder’s desire for information. The goal is to help the person receiving it take the right action, as fast as possible, with as little friction as possible.
Everything else should serve that anchor. When a decision is hard, go back to it: does this help the person on the other end of this message know what to do?
A Final Note on Language
The “fondling” incident is worth sitting with for a moment. It is easy to read it as a simple mistake, a word someone picked without thinking. But the community feedback suggests it was experienced as something larger, a signal that the people sending the alert had not considered the perspective of the people receiving it.
Trauma informed language in emergency communications is not a nicety. It is a functional requirement. The words you choose either close the distance between your institution and the people you serve, or they widen it. UO’s willingness to take that feedback seriously, and to use it as the starting point for a genuine redesign, is the most instructive thing about their whole story.
Sources: University of Oregon Safety and Risk Services presentation, “Getting Ducks in a Row: Standardize Campus Emergency Messaging,” and session notes, June 2026.