Critical messaging at scale
TETRA Today: How does the Everbridge platform help people generate messages and disseminate them during emergencies?
Jaime Ellertson: We started with emergency notification and then we moved into incident communications – automating the whole process in regulated industries. Twenty-four of the top 25 airports in the US use our system, and when any one of 38 different incidents happen, from a lost child to a guy with a gun to an aeroplane crash, the FAA requires you take many different steps. So, rather than the 20 operators in the control room having to get out a manual during a crisis, our system streamlines the process and sends it out to 200 different authorities. There’s also an audit trail so that after the incident, the authority can come in and say ‘Okay, did you send the messages to X, did you communicate to Z, did you follow this process?’, and it’s all automated.
Our system has a very simple interface for creating messages. During an emergency, the typical education level you’re operating at is about 5th/6th grade – because emergency situations are so stressful, you need to keep it simple, and if you’re sending a message to three million people, there’s a need to avoid embarrassing mistakes. So, our user interface uses templates and drop-downs, simplifying message creation, and then the platform delivers it anywhere in the world through up to a hundred ways: phone calls, text, email, digital signage, sirens in cities, even internal speakers and TVs. We can lock people’s desktop computers down in a fire drill, so it says ‘You must exit the building’ and you can’t type any more.
Multiple modularity makes critical messaging work. You never ever want to send a critical message over one path, because the chances of any one individual seeing that is one in ten unless they’re using the device and seeing it right then; you want to do that serially, going through seven different paths, and then start again.
Today we’re purely a software-as-a-service (SAAS) business that automates and accelerates the process of communicating to people to keep a business running, or communicating to people to give them the right information, at the right place at the right time, so they can remain safe and be safe in a critical event.
We’re not a mobile network operator. Our messages go over all the different types of communication lines, the internet, more than 200 major carriers around the world and secure networks and private networks. We have 99.99 per cent uptime and we strive to limit downtime to planned updates.
Our system has a rich API interface and we can take that and convert it with the metadata to go over virtually any output. Its infrastructure is mainly SIP-based. Increasingly our messages are going to smartphones over HTTP and the internet or through data with the carrier, not as an SMS or a call. People like that it’s push, they can attach pictures, videos, do whatever they want, and there’s no cost to that – it’s a push notification. Push notifications are the fastest form of communication we have. Apple or Google will send out a push message roughly 30 per cent faster than virtually any SMS in volume.
TETRA Today: How is your platform used to handle disasters and large-scale emergencies where the public have to be notified?
Ellertson: We have rich profiles on members of the public, which allow them to opt into location-specific alerts, choose their preferred means of notification depending on the time of day, and our platform can deliver a message to tens of millions of profiles in minutes and allows them to respond. While we provide the infrastructure, the system is run by our customers.
I’ll use Florida as an example. Hurricanes occur every few years, and when one happens, they not only need to send messages saying ‘Here’s where we’re evacuating’, they need to receive responses from those people who decide to remain in their homes, especially those with critical needs. Once the hurricane hits and they’re in the affected area, those are the people you need to rescue with rowboats. With our system, those handling the response to the hurricane can receive confirmations that people have received the messages and ask them ‘Are you leaving your property now that the evacuation has started? Yes/No’ and then plan their response accordingly.
TETRA Today: How does it help dispatchers in a public safety environment?
Ellertson: A good example is the work we’ve done with the South Western Ambulance Service NHS Foundation Trust in the UK. They’ve used it to automate their processes and cut down on the number of steps, so they can dispatch their ambulances faster and more economically. [The time it takes for the trust to complete a communications cascade has dropped from more than 60 minutes to less than four minutes, due to Everbridge’s text-to-speech functionality, according to a case study – Ed].
TETRA Today: To what extent is your alerting system integrated with devices and radio systems?
Ellertson: In the US Midwest, a tornado will rumble through a town at 4am on Sunday; there’s no town manager, there’s not even one policeman and a control car. The national weather service puts out an automated report, we capture that in real time and they set up a profile for the town that says if a storm gets within 25 miles of the town, we start sending the notices out; as it closes, we send them out with increasing frequency, turn on the sirens and call on all the devices. That’s done all automatically, there’s no human intervention at all.
The same thing happens in a big manufacturing plant. Some of the big automotive companies that use our system are communicating across plants that are often miles in size, and you have to convert the messages to analogue radios as well as multiple languages – Spanish, English, Portuguese – depending on where the plant is, and so the system will do that conversion for you and put out that message.
TETRA Today: What are you doing to help companies protect their employees?
Ellertson: Ten years ago, most business continuity people [tended] to focus on facilities because, if you were a financial institution, you had your facility and your 100 mile away backup facility and you could recover your operations. And for most corporations, you had all your staff in your building during the day, and if you had guards out front and used passkeys, your staff were safe.
Now we’re increasingly becoming a mobile world. I think International Data Corp has predicted that 70 per cent of workers in North America will be mobile by the year 2020. If they’re spending their time outside of your building, how do you recover, how do you keep them safe or even know where they are during a normal day? There’s also the growth of corporate campuses. Because employees could be in any one of a dozen buildings during the day, knowing which ones they’re in during an emergency can be challenging.
We saw a clear need for businesses to be able to track their employee’s locations, but not by requesting access to location services on their devices 24/7, as that’s an extreme invasion of privacy. We realised that most corporations have a lot of data on their employees’ locations. We have a mobile app that ties into location variables, including travel itineraries, scheduling systems [such as Outlook calendar] and corporate Wi-Fi – routers now get us down to within nine feet on a specific floor. And if the employee has badged in, we’re connected to the big four security companies. We’re adding ride share platforms, Uber and so on. We aggregate all that information together and use a formula to work out where employees are likely to be. So, when a building catches fire, we can put out an instant muster list that might say ‘Here are the people who are supposed to be in the building… but you also have another 100 occupants and six of them have issues and they’re on the third, the fourth and the 17th floor’.
Specific locations matter when there’s an active shooter event, which is a huge driver for corporate safety in the US right now – it’s the number-one reason we see people put in these systems on a company-wide basis versus just a location basis. These events typically last 12 minutes and it takes 18 minutes on average for the police to arrive. There’s a clear need to alert people immediately and give them location-specific advice – either stay in place or evacuate in the right direction, whichever is best for their safety.
During the Brussels attack, [one of our customers] sent out a notification to all their employees to stay where they were for the next 24 hours because it was an uncertain environment. One of their employees sent them a note two days later, saying ‘I was supposed to get on a train that afternoon. Had I gotten on the train, I would have been dead as that was the train that was blown up. I didn’t because I got the notification and was told to shelter and stay where I was.’
TETRA Today: What differentiates Everbridge from its competition?
Ellertson: Our scope and size. We have traditionally focused on a global market, so you could send a message here, but it could be received by people everywhere from Paris, Moscow, Dubai, Capetown and Korea, all in different languages and interfacing with the system.
This year, we’ll expect to send nearly two billion messages. The economies of scale we have and can pass onto our customers is dramatic. In many of our contracts, the messaging is included so you’re paying [for the messages], because we want people to use the system and be encouraged to use it as much as they need during an emergency. We’ve moved away from a charge per call or message, because that creates the misconception that it’s going to cost you in an emergency to communicate to people.
When we bought Vocal here in the UK, we changed the pricing model because they had a lot of customers who said: ‘I’ve got the emergency communications service, but it’s an insurance policy. We only want to use it when things are bad.’ And the problem with that is if you only use something when it’s bad, you won’t know how to use it and everyone will freak out and it will be a disaster and it will compound the problem. If they use it regularly, then in an emergency it becomes a seamless operation and you’re that much more successful in keeping people safe and businesses running.
We charge on the number of people, places and things you’re communicating to and from and then the application purchase price. Our customers can also pay for premium services such as automated threat or weather feeds.
TETRA Today: How did you grow your business so quickly?
Ellertson: We took on all the verticals from day one. That was risky and stretched us a little thin but it was great experience and it forced us to build a system that could support enormous scale. That’s why we were the first in the cloud. When a big emergency happens, you could swamp any single carrier or data centre, so we invested in the cloud, we keep the data about the people locally resident because of privacy laws but we can use cloud resources to process the calls faster because we may have to send a million phone calls through 28 public exchanges in two minutes. To do that, we migrated to a big data platform, a Mongo database for profile management and in-memory next-generation architecture.
This is a space where the proven ability to do it time and time again, flawless execution at enormous scale at the biggest events, makes a difference. Our scale and financial strength are helpful in our conversations with potential customers, and that has created a bit of a snowball effect. A lot of the work we do is around data. Today, we have almost 130 million people in our database. It went from 30 million in 2015 to more than 100 million in 2016; 2017 will probably achieve the same magnitude of growth.
TETRA Today: What’s next?
Ellertson: We’re currently building an analyse function for critical event management. It looks at the way [our customers] processed data in the past and then makes suggestions on the way it might be done better in the future. That can be done in real time when they’re managing an incident, it can be when they’re creating a message – ‘Don’t use that template or that term, if you’re sending to a group, send it to this group of managers’ and so on.
We’re managing about a million critical events a year. If we could mine that on a no-name basis and say ‘The last time you tried to respond to that flooding you took these three steps. We recommend you take these four steps in this order this time, because last time it didn’t work’, we could help our customers solve their problems faster. That’s the final piece of the puzzle.