At 12:58 AM on July 27, 1996, a pay phone near Centennial Olympic Park in Atlanta rang through to a 911 operator. The caller was Eric Rudolph, though that would not be known for nearly two years. His message was eleven words: “There is a bomb in Centennial Park. You have 30 minutes.”
The 911 operator took the call and immediately tried to notify Atlanta Police. What followed was a ten-minute cascade of small failures that, taken individually, might seem forgivable. Taken together, they consumed more than a third of the time between the warning and the detonation. Bad phone connections to the command center. Jammed lines. A frantic search for the address in the CAD system. The word “Centennial” would not enter. The operator asked how to spell it. Then tried removing “Park.” Then wondered aloud if this was Zone 3, where Olympic Stadium was.
Meanwhile, at roughly the same moment, Richard Jewell — a 33-year-old private security guard working the AT&T pavilion — noticed an unattended olive-green knapsack under a bench near the sound tower. He alerted Georgia Bureau of Investigation officer Tom Davis. Davis called in a bomb squad. They began moving people back. The crowd of beer-cup-waving concertgoers did not move quickly.
The bomb detonated two to three minutes into the evacuation, before all spectators could leave the area. A 40-pound pipe bomb packed with six pounds of nails. Alice Hawthorne, 44, of Albany, Georgia, died in the blast. She had driven to Atlanta with her daughter just to be in the atmosphere of the Games. Turkish cameraman Melih Uzunyol died of a heart attack while running to film the scene. One hundred and eleven others were injured.
The comm center story is specific and documented: there was a 911 call, there was a ten-minute delay in getting officers notified, and there was a private citizen who saw the bag and started an evacuation before any of that notification reached the park. The question this exercise asks is not whether what happened was anyone’s fault. The question is what your center’s process looks like when an anonymous bomb threat comes in, and whether that process is faster than what Atlanta had in 1996.
The following excerpts are from the actual Atlanta Police Department transcript of the 911 bomb threat call, released publicly after the bombing. Times are from the official record.
CALLER: “There is a bomb in Centennial Park. You have 30 minutes.” [caller hangs up immediately]
[911 Operator attempts to reach Atlanta Police Command Center — bad connection, jammed lines]
911 OPERATOR: “I tried to call ACC but ain’t nobody answering the phone… but I just got this man called talking about there’s a bomb set to go off in 30 minutes in Centennial Park.”
DISPATCHER: “Oh Lord, child. One minute, one minute. I copy Code 17. OK, all DUI units are Code 8 and will not be able to assist on the freeway. Oh Lord, child. Uh, OK, wait a minute, Centennial Park, you put it in and it won’t go in?”
911 OPERATOR: “No, unless I’m spelling Centennial wrong. How are we spelling Centennial?”
DISPATCHER: “C-E-N-T-E-N-N-I… how do you spell Centennial?”
911 OPERATOR: “Centennial Park is not going. Maybe if I take ‘park’ out, maybe that will take. Let me try that.”
DISPATCHER: “Wait a minute, that’s the regular Olympic Stadium right?”
DISPATCHER (to units): “1594, that’s affirmative, got a Signal 73 at 145 International Boulevard. It came from the pay phone at the Days Inn. The caller is advising that he has one set to go off in 30 minutes at Centennial Park. Sounded like a white male.”
[~7 minutes later — units directed to locate the caller, not to the park. State police not notified despite instructions to do so.]
UNIT 1546: “1546. I copy. Advise the state police, they police that park. I’ll go to the Days Inn and see if I can locate the caller.”
[1:20 AM — 22 minutes after the call]
UNIT 2924: “2924 to Radio, be advised that something just blew up at Olympic Park.”
The transcript was released publicly after the bombing. The Washington Post reported that the ten-minute delay resulted from “bad phone connections, jammed lines at the police command center, and a frantic search for the address.” State police, who were responsible for policing the park, were directed to be notified by a field unit — but the transcript contains no indication they actually were before the bomb detonated.
The 911 call and the security guard’s discovery of the bag happened independently and nearly simultaneously. Jewell spotted the knapsack and alerted GBI officer Tom Davis at roughly the same time the 911 call was being relayed through the dispatch system. The evacuation that saved lives was initiated by a private citizen who had no idea a 911 call had already been made. The comm center and the park were not in communication. This exercise asks: in your jurisdiction, for a large pre-planned special event, are those two tracks connected?
No right or wrong answers. Click to expand.
The Atlanta transcript shows the gap between receiving the call and dispatching to the threat location rather than the pay phone. The first instinct was to find the caller rather than to move on the threat. That decision cost time the dispatcher did not have.
Most centers have a bomb threat protocol that includes: notifying the supervisor, contacting the threat location, initiating law enforcement response, and contacting the bomb squad. The question is the sequence and who owns each step. In Atlanta, the confusion about whether the dispatcher or a field unit was responsible for notifying state police — and the absence of confirmation that it happened — is the single most correctable failure in the transcript.
For your center: pull your bomb threat protocol. Does it specify who notifies the venue directly? Who calls the bomb squad? Who confirms state or federal partner notification? Are those steps checked off by the dispatcher, or does the protocol assume someone else is handling it?
This is not primarily a technology failure — it is a training failure. The dispatcher and operator did not know how Centennial Olympic Park was stored in their system. For a venue that had been the centerpiece of the Atlanta Olympics for two weeks and that was drawing 15,000 people a night, that gap is a pre-incident planning failure.
CAD address matching for major venues, event locations, and non-standard sites is something that should be validated before events begin. How is your county fairground stored? Your sports stadium? Your downtown concert venue? If a caller says “the park by the river” or the name of a new development that was just built, does your dispatcher know how to find it?
The practical fix: before any major planned event in your jurisdiction, verify with your CAD administrator how the venue is addressable in your system. Dispatch it. Confirm it resolves. That exercise takes five minutes and could matter enormously. Atlanta had the Olympics in their city and did not do it.
The Centennial Park bombing produced a partial evacuation because a private security guard acted correctly without knowing about the 911 call, while a 911 call was being processed without any direct link to the security team on the ground at the park. Two parallel warning systems that were not connected to each other.
At large pre-planned events, best practice is to establish a unified command structure that integrates private event security with local law enforcement, and to have a direct radio or phone link between the event’s security command and the 911 comm center. If Jewell had been able to call a direct line to the officer in charge at the park rather than flagging down a GBI agent he happened to find, and if that officer had been in contact with the comm center, the evacuation might have started faster.
For your jurisdiction: what is your pre-planned event protocol? Does your comm center have a direct contact at the event security command post? Do they have a way to reach you other than 911? Is there a designated liaison between the event and law enforcement before the event starts?
This is a mass notification question. Verbal instructions from officers moving through a crowd of 15,000 people at a concert are the slowest possible evacuation method. The most effective tool available was the concert sound system — if it had been commandeered to announce a clear evacuation order, it would have reached all 15,000 people simultaneously rather than requiring officers to personally direct each cluster of attendees.
From the comm center perspective: who has the authority to order a venue to use its public address system? Is that call made by the dispatcher, the supervisor, the incident commander on scene, or nobody? Does your center have contact information for the event sound production operator? In 1996, this was not standard. In 2024, it still varies by jurisdiction.
Additionally: given that only 10 of the 22 minutes between the 911 call and detonation were actually available for evacuation after the delay, the practical window for moving 15,000 people was never realistic without immediate, large-scale notification tools. That math matters for how you plan large event response.
Centennial Olympic Park had no metal detectors, no bag checks, and no formal perimeter screening for a park that was open to the public 24 hours a day during the Games. The security investment went to the ticketed venues, the athlete compounds, and the official event sites. The public gathering space — described as the “spiritual heart” of the Games — was open access.
Rudolph walked in carrying a backpack full of pipe bombs. He set it down. He walked away. He called 911 from a pay phone. The “massive security effort” did not prevent any of that. What stopped the death toll from being far higher was one alert private security guard.
For dispatchers and training coordinators: visible security presence is not the same as effective threat detection. The lesson for comm centers is that the first notification of an actual threat often comes from someone who has no formal role in security — a vendor, a passerby, a security guard from an adjacent property. Your protocols should be as ready to receive and act on those calls as they are to receive calls from law enforcement.
This question sits at the edge of dispatch operations but is worth raising: the information environment after a major incident is shaped by what law enforcement communicates, and the comm center is often the first repository of information about who made what call, who was at a scene, and what was reported. Dispatch recordings, CAD notes, and call logs become investigative materials. They can also become leak material.
Jewell’s wrongful public accusation destroyed three months of his life and is now studied as a case study in media irresponsibility and law enforcement information leakage. The anonymous source who identified him to the Atlanta Journal-Constitution was never publicly named. The FBI later acknowledged it had no direct evidence implicating him.
For comm centers: your records are sensitive. Your personnel have access to information that, if disclosed without authorization, can cause real harm to real people. Post-incident information security — who can discuss a call, to whom, in what context — is a protocol question that belongs in your training alongside call handling and dispatch procedures.
This is an actionable pre-event checklist item. Before the event opens, someone should test-dispatch a call to the venue in your CAD and confirm it resolves correctly. If it does not, work with your CAD administrator to add or correct the address before the event begins. This takes minutes and is directly traceable to one of the most documented dispatch failures in modern American history.
Also worth confirming: does your center have a direct contact number for the event’s security command post, separate from 911? Is that number in your CAD notes or on a printed reference at the console?
The Atlanta transcript shows a field unit being told to notify state police rather than the dispatcher making that call directly. There is no documented confirmation in the transcript that state police were actually reached. In a protocol with multiple parallel notification requirements, the weakest link is always the one that depends on someone else to make the call without verification.
Best practice: bomb threat protocols should assign specific notification responsibilities by name or role, not by “advise the appropriate agencies.” And each notification should be confirmed back to the dispatcher and documented in the CAD record, not assumed to have happened because someone was told to do it.
Your center holds Emergency Action Plans for dams, pipelines, and hazmat facilities. This series turns those plans into dispatcher exercises โ built from real incidents, told from the console. Free to use, no login required.
See the Infrastructure Series →- Atlanta Police Department. (1996, August). Transcript of 911 bomb threat call, July 27, 1996. Released publicly following the bombing. Archived at: bton.ac.uk (risks-forum archive)
- Washington Post. (1996, August 9). Foul-ups followed 911 call in Atlanta. washingtonpost.com
- Atlanta Magazine. (2016, July). Fallout: An Oral History of the Olympic Park Bombing. atlantamagazine.com
- Atlanta Magazine. (1996, December). Presumed Guilty. atlantamagazine.com
- TIME Magazine. (2016, July 27). The Summer Olympics Bombing in Atlanta: 20th Anniversary. time.com
- Wikipedia. Centennial Olympic Park bombing. wikipedia.org
- FBI. Eric Rudolph: Famous Cases & Criminals. fbi.gov
- FBI. Phone Bank from CENTBOM Investigation. History artifact. fbi.gov
- CNN. (1996, August 9). FBI defends Olympics bomb probe; 911 transcript shows delay. cnn.com
- Columbia University Graduate School of Journalism. Richard Jewell Case Study. columbia.edu