Okay, good afternoon. Thank you all for coming to ShotSpotter's first Investor Day. I'm actually thinking it might be more appropriately called a Analyst Investor Day based on the mix that we have here. Thank you all for taking the time to be here. We hope that this is informative for you and you leave here maybe with some of your questions answered, and also a good picture of what we're doing, how we're doing it, and with a greater introduction to some of the other team members of our executive team. Before we get started, though, this is the contribution from our attorneys in terms of the forward-looking statement and the safe harbor. I'm gonna put this up there for a minute or so you can all read that. I know you're gonna memorize it. Okay.
Before we get started, and I apologize, there will still be some people coming in. We are hoping to have a couple of more participants in the audience as well. If they show up, they'll just take their seats. We'll keep going, though. Agenda today. Ralph Clark's gonna give a quick executive summary. Our Senior VP of Technology, Paul Ames, will give an overview of the technology. Senior VP of Operations, Joe Hawkins, is gonna go over a little bit of what a day in the life of a project manager looks like. Senior VP of Marketing and Product Development, Sam Klepper, is gonna give you an overview of our marketing programs. Senior VP of Sales, Gary Bunyard, is gonna talk about our sales a bit.
I will wrap it all up with a brief summary of our financials Q3 and a little bit of the forward-looking area into 2019. We will have a question- and- answer session at the very end. For those of you that are not participating in person here but are participating via webcast, thank you also for participating. Sorry you can't be here, but if you do have questions, go ahead and email JoAnne Horne questions during the presentation, and we'll do our best to make sure we get those answered as we can. Her email is on the webcast, but it's also jhorne, J-H-O-R-N-E, @marketstreetpartners.com. We'll go ahead and get started. I'm gonna turn it over to Ralph.
Great. Thanks, Alan. Yeah. Are you gonna advance the slides for me? Great. Good afternoon and welcome, everyone. Thank you again for joining us for our very first Analyst/Investor Day. My name's Ralph Clark, and I've been with ShotSpotter now for a little over eight years. What was so attractive to me about joining this amazing technology franchise eight years ago and keeps me still very excited and engaged today, is really the opportunity to lead an organization that is having a significant positive impact on the lives of so many people that use our technology, live in our coverage areas, work and play in our coverage areas.
We know that as a company, working with our employees, our vendors, and partners like yourself, we're making a very big difference in the lives of so many people in helping law enforcement agencies drive more positive public safety outcomes, leveraging a very unique technology that we're privileged to be a part of from the founding of Dr. Robert Showen over 20 years ago. You're gonna hear a lot today about the what and how we do what we do. We've been coached not to say too much about the when we're gonna do it, as not to commit inappropriately. But as we kinda go through these presentations, I wanna challenge all of us to kinda keep in the forefront of our mind, really the purpose or the why. Why does this company exist?
This is why we exist. We exist to really work with police departments, earning a trusted relationship with police departments, and help them provide equal protection to all communities, and as a result of doing that, reduce and prevent gun violence. That's really the moral purpose of this company, and we believe it's really important that we keep that moral purpose or that why front and center in terms of our activities, 'cause it really is quite a motivator and a compelling thing to lean in on this in the work that we do. Okay. Many of you already know, this is a category or space that ShotSpotter has created: wide-area acoustic gunshot detection technology. We created this market. I think it's fair to say that we pretty much exclusively own the market.
Our technology is basically transforming the way police departments go about the business of being alerted to, responding to, and investigating and ultimately preventing and reducing gun violence. We're doing it at a really interesting scale. If you look over to the right of the slide, you'll see that we're deployed, have been effectively deployed, and are effectively being used in over 90 municipalities and agencies, mostly in the U.S., although we do have a deployment in Cape Town, South Africa.
As impressive as the scale is in terms of the number of sensors deployed, the square miles deployed, the alerts that we've produced, at least for 2017, all the database information that we have to continually perfect and improve the system, as impressive as that is, we know that there's so much more work to be done. We're in largely a untapped marketplace, and there are so many more communities and cities that could really benefit from police departments using our service. You know, we're proud of the progress we've made as a company, but we also recognize there's a lot more work that needs to be done, and we're excited about getting at it.
Now I'm gonna walk you through the financials, hopefully, we're gonna spend a little time on the technology piece, and you'll understand how it is that over, you know, 20 years, $75 million of invested capital before going public, that we, even as a very small company, have really built this incredible technology and have been issued 32 issued patents. I mean, think about that for one second. A company our size with 32 issued patents, really kind of speaks to the solution that we've created. Okay, let's go to the next slide. To the problem of gun violence. A lot of us think we know about gun violence. We certainly read a lot about it in the newspapers and the like.
Typically, if we're watching local news, you know, oftentimes the first news segment that leads out, they call it if it bleeds, it leads type of thing. What I want you all to appreciate is that doesn't tell the true problem of gun violence. Gun violence is more persistent than I think it typically gets reported, because what gets reported is only typically blood shootings, things that result in homicide. Maybe there's an occasional gunshot wound victim that gets reported. By and large, and we know this from our own anecdotal experience, and it's since been validated from an independent organization, the Brookings Institution, that gun violence happens much more frequently than that which results in a blood shooting. To that degree, the significant amount of it, over 80%, goes unreported.
Think about what it means to not have 80% of a shooting, shootings in your neighborhood not reported. The first thing it means is that if it's not reported, probably police aren't gonna show up, right? As a result of that happening day after day after day, gun violence becomes normalized, and also the trust between the community and that police department is diluted significantly, right? If you have that kind of environment, you can't produce public safety outcomes. Flat period, full stop, exclamation point. You cannot deliver public safety outcomes and not show up to 80%-90% of the shootings in neighborhoods that have ongoing persistent gun violence. That's the thing fundamentally that our technology addresses. That's our traditional, I'll call it, you know, persistent gun violence law enforcement-oriented use case, which we're kinda known for.
There's also another interesting use case that, sadly, is increasingly becoming more prevalent, and that's the use case of the active shooter scenario. In those situations, you're always going to get the call. The problem is you get the call typically maybe three to five minutes after the active shooter has initiated his or her active shooting event, and that complicates the response point of view. There, it's not so much about, you know, kind of responding to the arrest the individual or do an investigation. Really, what you're trying to do is get there to stop that shooting event and be able to provide EMS services to likely wounded victims. Typically, when you have an active shooter situation, there's going to be blood involved.
Getting EMS support there very quickly and allowing law enforcement personnel to get cops to the dot very quickly and precisely, they know exactly where the shooting takes place, is a significant game changer. We're digitally transforming the way police departments really go about their business. That's really the what of what we do and the nature of the problem. Let's go to the next slide. That's an amazing business for us. Again, I think I've hopefully made the point that we're significantly under-penetrated into a very large TAM market. We've got just over 90 cities deployed.
On our traditional TAM analysis, you've seen and heard from Al and I several times, is that we think there's well over, I mean, well over 1,000 cities just in the U.S. that have the need for our type of capability, and we're less than 10% penetrated in that market. Again, we have no direct competition that we've seen or can really speak of, and that gives us a clean, open playing field to continue to execute on that, I'll call it, domestic core public safety business. In addition to that, if that isn't good enough for you, I mean, we've got some really incredible growth drivers that go beyond traditional domestic public safety solutions, and I'm just gonna review a few of those for you.
Internationally, gun violence is equally a significant problem and probably even more so a problem in certain parts of Latin America. That's a wide open field for us. We recently hired a senior excuse me, a VP of sales focused on Latin America to go get at the Latin American opportunity, which includes South America, countries in Central America, and also Mexico, which is technically in North America, that have very significant gun violence problems. We don't have any deployments in the Western Hemisphere outside of outside of the U.S. technically. When you combine the Latin American opportunity with what we think is the opportunity to expand in South Africa, plus there's a couple of cities in Europe that you could put into the mix, we think there's about 200 cities internationally.
We've talked about the fact that we have a little bit of pricing leverage internationally. Our price per square kilometer or price per square mile is a little bit more internationally than it is domestically. This represents a fairly significant growth driver for us as a business, and we're organizing ourselves to go get at that opportunity. Secondly, we think we have a pretty interesting security business opportunity as well. We've talked a lot about the progress we've made on selling to college campuses. We're mostly focused on higher ed from a direct selling point of view. We don't have an interest in selling individual systems to K through 12 school systems.
We have a bit of a high-minded moral purpose here, and the idea of selling to, you know, really economically strapped elementary schools and middle schools just doesn't make sense. Selling to a university that has more resources and frankly, has more of a gun violence problem, just even beyond the technical active shooter scenario, a lot of our universities where we've deployed, we're up to about nine or so to date. A lot of these universities, you find them located in parts of the city that are, I would say are transforming is probably the best way to put it. Transitional neighborhoods maybe, where there's not ongoing persistent gun violence, but there is occasional gun violence that happens. We know this. We're privileged to know this because of our nine or so deployments.
We're alerting on places that you might not think should get gunfire alerts. There happens to be a very large university that I won't name by name, but they happen to be in the Bay Area. They've actually had some shootings just off campus. Fortunately, because of our system, we've been able to, they've been able to dispatch officers there, not only make the on-scene arrests, but again, be able to dispatch EMS services there that could critically save someone's life there. You have that two, three-minute time window where getting immediate EMS care can save a life. We're pretty excited about this opportunity that has also expanded to other pretty interesting use cases, frankly, just because of the reputation we have of delivering technology in a very effective way.
We've been approached by a consortium of folks that said, "Hey, look, we've had a problem with some sniper fire on a freeway. Do you guys think you can provide linear protection in a really noisy freeway environment?" One thing about our engineering team and as our company is just a part of our ethos is we love those technical challenges to show off what we can do. We talked about the deployment we have in a number of square or excuse me, linear miles of a freeway that we're protecting, and we're really excited about watching the use case and operational effectiveness of that. Because as you might imagine, there's a lot of linear freeways available here in the U.S.
Obviously, not all of them have the same issue that this one has, but you know, you don't have to go very far outside of California, Texas, and maybe some other places where there's similar problems that we could address effectively. That's all very exciting for us on the security front. Intelligent city solutions, that's more, I would call it a, we like having a seat at the table. We like the optionality of it. There's a lot of talk about intelligent cities, smart communities, and the like. To date, there has not been a significant amount of spend, it's predicted to be a very large opportunity. What I like about that is we have partnered with 3 of the biggest companies that are focused in those areas.
If you go down Verizon, AT&T, and GE, they're all partnering with us as a part of providing a gunshot detection solution as a part of a broader intelligent cities solution. That could be something that significantly expands our TAM effectively out of the traditional urban gunfire policing solutions that we've been known to be so effective at selling and executing. We're excited about that. I say our excitement is more of a timing issue. We're not expecting big things early on, but we certainly like to have the optionality to be there if and when that market does take off. I'll just combine four and three for a moment.
We've often talked about the idea, you know, again, we love this market that we created, that we pretty much own exclusively, but because we're so customer-centered and customer-focused, we're in the conversation a lot when folks are talking about the problems that they're having in driving public safety outcomes. From time to time, there's been opportunities for us to basically leverage our core platform of gunshot detection and move it into adjacent areas that not only from your purposes expands the TAM and expands the market opportunity, and that's pretty important. What's even more important, frankly, is being more relevant and solution-oriented to the customers that we deal with, that we have to live with and want to live with over 15, 20 years.
When an agency comes to us, and we have a level of intimacy with them to say, you know, "Hey, look, while we're producing this amazing ShotSpotter data, we'd like to figure out how we can be more predictive and implement more precision-oriented policing, and we think the ShotSpotter data is a piece of that." Light bulbs go off and says, "Well, maybe being in the predictive analytics space is something very complementary to what we do as our core solution. Is that something that makes sense to you as a customer?" Literally, when you have customers kind of pulling you into adjacent markets, that's pretty exciting. That's what our HunchLab acquisition was about.
We'll probably have the opportunity to do things organically, as well as kind of tuck-in acquisitions, but those will be all additive to our core wide area acoustic gunshot detection service. I'm just gonna keep on moving. Let's hit the next slide if we could. Where do we go here? Where do we go from here? From an overall strategy point of view, we believe that we just gotta take care of business in the core market that we've created, and that means more deeply driving ourselves into this wide area TAM, wide open area TAM that is virgin. We're going to continue to move in that direction, add more cities to the platform, making sure that they're successful.
Think in terms of land and expand. We're gonna do that as a part of our strategy. The other idea that we have is like it's just not about selling day one. It's also about how agencies are successful in using the solution and are willing to talk about it. You're gonna hear a lot today about NPS, Net Promoter Score. We believe having some leading, well-respected agencies that are driving fantastic outcomes by using our system will help accelerate the time in which we get to a tipping point, which is when agencies look at what we do as a standard of care issue. Think in terms of body-worn cameras. You know, 10 years ago, that was kind of a nice-to-have. Now it's kind of a have-to-have, right?
We wanna get ShotSpotter to the point where, hey, it's a have-to-have, not a nice-to-have. A lot of that has to do with continuing to do, one, more deeply penetrating our TAM, but at the same time, making sure that the folks that are using our technology are using it the right way and can talk about the outcomes that they're driving. That's really critically important to us. We've been investing in building not only the technology platform, but I'll call it the organizational resiliency to drive growth. Internally, we've kicked off a couple of collaborative cross-functional teams called growth strategy teams, and it's all about You know, friction points to growth, how we can add accelerants to growth and continue to have impact over many more agencies than where we are today. We're being very intentional about that.
Let's talk about ShotSpotter revenue. I want to call this something different, but was advised by outside counsel that I have to choose my words carefully. We're calling this ShotSpotter revenue characteristics. I think the subtitle really is we want to generate high quality revenues. High quality revenues. What are high quality revenues? Well, high quality revenues are revenues that grow. They have to grow. You like them to be somewhat sticky. I think we've got some good stats on the stickiness of our solution. Again, it's intentional. We invest in stickiness because we invest in NPS, in client account reviews, in customer success organizations, et cetera. Our revenue grows, it's sticky, it's built on strong customer partnerships, right? That's important because they almost become the virtual sales team for our solution.
If you know anything about our vertical, for those of you who have had an opportunity to spend any time at IACP, the chief of police of Metro A does not compete with the chief of police of Metro B. They're collaborative. They call it fraternal order of police for a reason. When a solution works for some person, and it works very well, they're more likely to share it with their colleague, as opposed to if they were in banking insurance, and they're competing with one another. If they have a great secret sauce thing, they don't wanna share it with their competitor. We don't have that problem. It's not a problem for us, it's an opportunity for us to create a viral effect for our solution. Partnerships are key.
Alan will probably walk you through the income statement, I hope, and be able to convince you that our revenue is high contribution revenue. Because we don't have any direct competition, we've got leverage on the revenue line. Our cost scales very, very nicely in terms of absorption. As we grow revenues, we don't have to grow costs to the same rate. We get, you know, great gross margin and gross profit expansion. Below the line, because of our vertical focus, right, our cost to serve, our cost to acquire is extremely low. The technology is invented. We use our fifth column sales force of very highly satisfied customers to help us sell the solution. We don't have to boil the ocean from a customer discovery, customer conversion point of view.
We've got some very attractive customer acquisition costs and R&D costs relative to what we're doing. We also have high quality revenues, I think there's an element of diversification there as well. We know that in order to pursue our long-term broad ambitions, we just don't wanna limit ourselves to domestic ShotSpotter. We wanna do some of these other tangential things that have impact and relevancy. International security are also very big parts of what we plan to do as a company. All right, I think one more, I think. Yeah. You're saying, "Ralph, you built an amazing franchise." We believe it. It's like, what's going to protect you from someone trying to come in and basically take this thing, this incredible jewel that you built? Well, here's what I'm gonna tell you.
We've got a sustainable, we believe, I'm told I should say, we believe we have a sustainable competitive moat. It's really kind of three elements of our competitive moat that I wanna talk about. The first and most obvious one is the core technology. 32 issued patents. 20 years of perfecting, fine-tuning the technology. $75 million invested capital, excuse me, that goes into reliably, at scale, detecting, locating, alerting gunshots, right? That's not an easy thing to do, right? I can tell you, 32 issued patents, that's something that's very real. Someone that needed to get into this business and be able to do wide area acoustic gunshot detection, they have to figure out how to work around 32 issued patents. That's not a small feat.
After they do that, think they're really smart, and they can get someone to give them $35 million and two years to do that, and they work around the patents, then they have to figure out how to get around the experience curve and tribal knowledge that we've developed over 20 years. It's just not the technology, it's the application of the technology. What does it mean to deploy in Boston? What does it mean to deploy in Oakland? Those are very different deployments, I can tell you. How do you perfect a machine learning capability and you don't have data to run against the algorithms? We've got 20 years of gunfire data. We've got 20 years of non-gunfire data, that enables us to have a great deal of precision more that someone can't get from day one, even if they invent the technology.
What are they gonna run the algorithms against, right? All that means is, let's say they're still bold enough to try to do that, then they have to have their first deployment. Well, guess what? The most critical barrier to entry we have is the trusted brand we have with customers that's evidenced by our NPS. They have to convince a chief that despite the fact that 100+ other chiefs, 'cause now I'm projecting if it's two years from now, someone decided to go today, they've got to say, "I don't believe what you, you, you, you and you did. I'm gonna go out there on a ice sheet on my own with this new thing and try to do it." What's gonna happen with that thing, right?
I don't think it has a good chance of being successful. It's gotta be rocky because the nature of what we do, it's always rocky for your first one. I mean, our first deployment was rocky, the second was a little bit less rocky. Even in our deployments today, I think it's fair to say, and hopefully Joe will share some of this with you, we're always learning, and part of that learning gets into tribal knowledge. Because we're so committed to the customer experience, we've got a reputation for delivery. We feel pretty good. We don't take anything for granted. I mean, we're both humble and confident at the same time. We're always paying attention. But I think those are pretty large barriers for someone to overcome. I think that's my last slide. Is it not?
That is. We will be taking short, like Q&A between each section. There is a more thorough Q&A session at the end. If you have a burning question right now, you can ask. If not, just hold it to the end.
It's gotta be hot burning for me. All right. Well, we'll just keep the show going then. Thank you very much. Really appreciate it. Good to see you all here. Who's next?
Paul.
Paul. Okay.
Do you want to just give a brief background please?
Good afternoon. My name is Paul Ames, and as Alan said, I head up the technology team at ShotSpotter. So I've been with the company for around about four years. My career in technology is leading technology teams, building out products, building out infrastructure. I spent quite a lot of years working within the real-time financial information industry, delivering systems to trading floors. My guess is some of you have some inkling as to what that environment is like. It's very challenging and good training for what we do here at ShotSpotter. I just wanna echo what Ralph said about this being a very unusual company. It's the first company that I've worked for that I can honestly say that I'm proud to work in.
A lot of the people in my team and across the company can say the same thing. Next slide.
I would answer.
Oh, yeah. All right. I want to spend a little bit of time talking about how ShotSpotter wide area acoustic gunshot detection actually works. Spend a bit of time introducing you to what the user experience is for the police officers out on patrol. We'll take a little look at the back-end infrastructure and take a deep dive on a couple of projects that we've had running in 2018. Next slide. For those that have sat through earnings call, we're a very coverage area centric company. We talk about square miles of coverage. The coverage areas are typically composed of between 20 and 25 acoustic sensors deployed per square mile.
The actual sensors themselves are typically deployed on top of buildings, but they can also quite happily exist on a streetlight not too far above street level. If you look at the sensor itself, the current generation of ShotSpotter sensor is the thing that we call internally [Sepis 2]. Physically, it's about the size of a lunchbox. It's a pretty boring looking device, and it's deliberately designed that way so that people don't really know what it is, and they're not gonna interfere with it. If you crack the thing open and have a look at what's inside, we have an Arm CPU, so that's the CPU that typically lives in your mobile device. We have a couple of microphones.
We have a cellular radio for communication back to the SoundThinking cloud. Perhaps most importantly, we have a GPS chip. The GPS chip is used for two purposes. First, it's used to locate the sensor when it's initially installed. Once it's installed, the sensor doesn't actually move, so it's kind of boring from that point. More importantly, the GPS provides a synchronized clock, a highly accurate synchronized clock from the GPS satellite network. We use that synchronized clock to precisely timestamp the audio that the sensor hears. This picture shows four sensors deployed in a city. You can see here that when somebody pulls a trigger on a gun, gunshots are very loud, as you know.
The sound propagates out around the city and arrives at different sensors in the sensor array at different times. It's this information, the precisely timestamped arrival of the gunshot sound at the individual sensors that allows us to triangulate the location of the origin of the sound. That's, you know, a fairly standard technique that's been around for a number of years. Conceptually, this is pretty straightforward. In practice, it turns out that it's kind of difficult. For starters, cities are pretty noisy, so there's a lot of background noise. We're deployed in New York City, we're deployed in Chicago, and these are not quiet cities. The first challenge is to deal with the background noise. Another challenge is the weather.
The weather, it could be wind, it could be rain, it could be the temperature. The temperature in the environment changes the speed of propagation of the sound. There are issues with trees. Trees filter out gunshot sounds and change the characteristics of the sounds. You start adding up all these things, including buildings, hills, et cetera, and it becomes a rather complicated proposition. As Ralph was saying, we have 32 patents. We hold 32 patents. A lot of these patents are around managing the difficulty of deploying to a city. I'll just call out three patents that are of particular interest. One is the notion of using GPS timestamping to timestamp the audio.
We have a patent on that. We have a patent on the way that we manage the audio within the within the sensor. We also have a patent on encrypting the information that comes from the sensor. There's a lot of intellectual property, a lot of know-how, a lot of knowledge that goes into making this thing work. Next slide. End-to-end, this is roughly how it works. The first three bullets I've already covered. Once the software in the cloud has figured out the location of the sound, we run various data through artificial intelligence classification with the goal of coming up with a machine classification of the audio event. Is it a gunshot? Is it a backfire?
Is it a helicopter? Is it, you know, all these other sounds that occur in the environment? This is important. This information is captured and is then given to human reviewers, along with the information about where the sound originates, an audio snippet of the sound itself, as well as the machine classification. What the human is doing is they are making an assessment as to whether or not this truly is a gunshot, as the machine classifier has assessed. The end-to-end process, if the acoustic event is in fact a gunshot, takes less than 60 seconds. Our service level agreement with our agencies, with our customers, is to publish this data within 60 seconds. Generally, the time is between 30 and 45 seconds. High level, that's what happens.
Next slide. I want to take a look at the application. How is it that the officer actually experiences gunshot data? Next slide. Really high level. We have two buckets of applications within SoundThinking. We have the customer-facing applications, and we have the internal applications that we use to manage the service. On the agency side, we have, first and foremost, an application that we developed and released roughly two years ago. It's an application called Respond, and this application is all about delivering real-time gunshot alerts to the officer in the field so that they can go to the precise location of where the shooter, where we report the shooter to be. This year, at the beginning of this year, we released an application called Dispatch.
This is an application that lives within the communication center, or more commonly known as the 911 dispatch center. Here what's happening is we're pushing the notification to the dispatcher. The dispatcher is gathering the data that we give them, radiating out to a car that's in the field and dispatching that officer to the location of the gunshot that we report. We also have an application called Investigator Portal. This is more for post incident analysis, often used by the crime analysts within the agency, but also used by the detectives as they're gathering information about the event. Ralph, I think, mentioned ShotSpotter Missions and the acquisition that we made just recently.
From a officer perspective, what they experience is that we deliver to them recommendations about where and when they should be patrolling to have the maximum possible effect of driving down crime through the deterrence effect of having an officer that's on patrol. Effectively what they're doing is, or what they're receiving is you should be here at this particular time of day. We also have an administration application, the super user, beg your pardon, within the agency, uses. This application is all about managing the user community within their particular agency. They can do typical things like add officers, take them away, monitor how they're using the application. Internally, one of the key applications is a thing that we call Review.
I have a slide on that in a minute. This is used by the human classifiers to figure out or make an assessment as to what the acoustic event is. An application called Dashboard. Dashboard is all about providing visibility into exactly what's happening end to end. You can see when the acoustic event is arriving from the sensor, you can see what the machine classifier says, the human classifier says, and which police officers are receiving the notification of the gunshot. Next slide. Two years ago, we released our first next generation application. This application I refer to as Respond. I think this is, for us, a core competency of the technology organization, but the organization as a whole.
I think we've done a very good job at delivering a very simple, unambiguous application to police officers in the field. It's critical that this application is easy to use. It's critical that it's unambiguous because they are responding to a situation that involves guns, so they need to know what they're doing. All of our applications are built as responsive applications, meaning that regardless of the form factor of the device that the police officer is using, there's a good user experience. From a small mobile device through to a, to an iPad, to a web browser, to a native Windows application, it's the same experience. From my perspective, and from the company's perspective, one of the key attributes of the applications that we're delivering are that they're built using the same source code.
We're developing an application once. The sustaining engineering load is low. We're able to make significant strides using a very small staff of application engineers. Next. Here's a picture of the Respond application. On the left-hand side is a screenshot of Respond from taken from a iPhone X. This is an application that an officer downloads from the App Store or the Google Play Store if it was an Android application or an Android device. On the right-hand side, this is the same application running within a web browser on a Mac. You can see that they look kind of similar.
When you actually use them, if you're familiar with the mobile application, it's very easy to use either the browser version or the native Windows version. On the right-hand side, I'll just point out a couple of features of the application. On the map, there's a red outline area. That's the ShotSpotter coverage area for Oakland. The yellow dots or the yellow pins are the precise locations of gunshots that have occurred over the last 24 hours. The number that's on the pin equates to the number of shots in a particular incident. We'll report an incident with 15 shots, for example, and there'll be a 15 on number 15 on the pin. It's simple. You'll notice it's a very dark color.
One of the things that became apparent when we were developing this is police officers do not want to have a very bright application on their device because it will light up their face in a situation where you really don't want to be a target. Next slide. Here is a chart that was taken a few days back from our Dashboard application that shows the adoption rate of the Respond application. Respond, as I said, was released a couple of years ago, so we've gone from zero users two years ago to roughly 12,000 monthly unique users. What that means is if I use the application once, I get counted once. If I use it 1,000 times, I'm counted once.
Bearing in mind that we have 90+ agencies on board, you're talking somewhere around about 120, 130 users on average. In actual fact, it's a little less average than that. There are some agencies, for example, Chicago, where they have a lot of users. There are some agencies that are very small where they only have a handful of users. In addition to these users, we also have almost 3,000 users that are using the prior version of the application. Over the next 12 months, we expect to see this number go, the number at the bottom go down dramatically, and the number at the top continue to drive up. This is something that we take a real keen interest on.
This is a, I think, a good proxy for officer engagement with the service. One of the recent releases of the Respond app was to provide an additional level of intelligence about how the officer is actually using the application. A traditional click-stream analysis. Next slide. I want to just introduce very high level what the architecture of the back end is that makes this stuff work. Next slide. The way to read this slide is from the bottom to the top. At the bottom level are the sensors that are deployed to the coverage area. At the top are the applications that I just ran through. Sensors are listening for acoustic events. They're sending pulses to our cloud, their database. We figure out the location.
We do perform machine classification on the incident. We perform human classification. If the incident is deemed to be a gunshot, the button is pressed, it's published to the officers in the field. On the right-hand side, with all infrastructures that are all back-end infrastructures, you have to worry a lot about security, you have to worry a lot about user authentication, monitoring, et cetera. All of that stuff is of course present. Next slide. Here is a deployment model. Here's the same stuff, here's where the software lives. On the left-hand side are the applications, on the right-hand side are the sensors.
My primary reason for showing you this is to point out that we have what's referred to as a hybrid cloud infrastructure. What that means is, two years ago, we had a private cloud infrastructure, so we were hosting all the back-end computers within our own colo facility. Since then, two years ago, we've migrated to about 50/50. At the end of next year, we're gonna be entirely within the public cloud, specifically, AWS. Next slide. I wanna run quickly through a couple of 2018 projects. This effort that I was just referring to, the migration of our back-end components from our private cloud into the public cloud, internally, we refer to it as ShotSpotter 2.0.
The, the drivers for this are all about staying ahead of Gary and the sales team's success. It's critical that we have the back-end capacity to scale and add new customers or add new square miles for existing customers. Now, one of the interesting things, perhaps a little like online shopping, is that the usage of the service effectively is very, it's not at all flat. There are times when there's a lot of activity, and there are times where it's pretty quiet. Generally, people shoot guns when it's when it's the night. They also shoot guns over the weekend, interestingly.
Another strange thing about doing acoustic gunshot detection, wide area gunshot detection, is there are two times in the year in the U.S. where there are acoustic events that cause us some amount of angst. Those days are July 4th and New Year's Eve, where there are a lot of fireworks go off. If you look at the back-end compute requirements, it's really important that we introduce elasticity into our back end. That's another major driver for why it is that we're investing in this move to AWS. There are ancillary benefits relating to security, disaster recovery, performance, et cetera.
This effort is taking a fair amount of resources within my team, and as I think I mentioned earlier, will be done at the end of 2019. Our colo facility with our existing hardware can be repurposed for internal uses, such as training machine learning. Next. One of the services that we migrated into the AWS public cloud a little before July 4th, was the classification service. This is an incredibly important part of our back-end infrastructure. Up until fairly recently, we were using a single machine learning technique, a thing called Naive Bayes. Typically, Naive Bayes is used for spam detection or maybe classifying documents. We've historically used it for classifying acoustic events.
Up until midway through this year, the quality of those classifications was pretty low. After we took advantage of the database of gunshot acoustic events for training, we were able to drive that number up pretty significantly. I'll give you one number that I think is kind of interesting. Historically, the Naive Bayes classifier was around about 40% accurate. We were able to drive that up to 70%. That's one, that's one classifier that we use. The other one that was introduced fairly recently is the ResNet Convolutional Neural Net classifier. This is technology that I think everyone touches pretty regularly. It's the technology that photo service providers use to classify pictures and recognize your mom, for example.
That's the technology that is used there. What we're doing is we're repurposing that or reusing ResNet to improve or up our game for machine classification. This has been enormously successful and looking forward is gonna be a continued focus on within the technology team. Looking to the future, a couple of things I'll just point out. As our classification accuracy goes up, we're able to do interesting things. For example, the ResNet classifier is very good at classifying helicopters for whatever reason. If we feel confident that the ResNet classifier gets it pretty much 100% accurately correct, then we can not present those acoustic incidents to the human reviewers for review.
The net effect there is we're able to add more coverage area or more miles of coverage area without adding more reviewers because they're having to classify less acoustic events. That's one example for helicopters. On the flip side, we can do things like, say if we're supremely confident that this is a gunshot, maybe we don't need to have a human review in the process. The benefit there from an agency perspective is that we can push that alert out quicker than we would if a human was involved in the review process. Next. Here is the Review app. We are releasing this app to the IRC users next month. The goal here is kind of what I just said. It's all about doing more with the same number of people.
What we don't wanna do is, for every square mile that we add, we have to add, you know, a part of a human reviewer. The goal here is to make the classification process more efficient and make it more accurate. One thing I'll just point out is the concept of workflows. This application is going out with the ability to create workflows for particular customers. One example is, in our security, for our security customers, if a gunshot is detected, then the reviewer physically picks up a phone and makes a call to the agency.
We're able to build into the reviewer console, that workflow and actually remind them that they need to pick up their phone and call a police officer. Next. That's it. Okay. Again, if there are any burning questions now, then great. If not, then we can catch them at the end. Okay.
Thanks, Paul.
My pleasure.
Thank you, Paul. Good afternoon, everybody. My name is Joe Hawkins, and I've got the privilege of heading up the operations organization for ShotSpotter. Our operations organization is made up of two divisions. First, there's a project management team that is responsible for leading the design and deployment of new and expanded services for our customers. Second, there's a network and field operations organization that has the responsibility for installing, remotely monitoring, managing, and servicing the thousands of sensors that we've got out in the field. I've been with the company for about 6.5 years.
Before ShotSpotter spent a long time in software development and network and telecommunications companies way back in the day, and the last 15 years or so, running similar sorts of organizations, managing distributed networks of computing devices, what today we might call the Internet of Things or IoT. This has been a great convergence for me of bringing sort of that network and technology background into, as Paul said and Ralph said, a business that's really uniquely rewarding in terms of what we do for our customers and the communities they serve.
What I'd like to do today is just give you a brief glimpse into the sort of the day in the life of a typical project manager, who, as I said, is responsible for driving the deployment of our service for our customers after the sales team delivers a contract. A little bit about the project management team. First of all, these, I've got a team of seven project managers that in aggregate have about 35 years of ShotSpotter experience. That's on top of the years of experience they may have in project management in general. One of the things that Ralph noted in his introductory comments in terms of the competitive moat, was this experience curve.
What we do at ShotSpotter is very different than what I've done in past network operations companies, and other technology companies. The project managers do have a traditional project management role, which you see sort of here in this first big bullet. They lead interdepartmental cross-functional teams to get things done, and there's a lot of work involved in that. Secondly, they have an individual contributor role, if you will, as the architect and designer of the sensor arrays that are the backbone of the service, and they also go out and secure permissions from literally hundreds or thousands of building owners and property owners to allow us to put those sensors in place. Then they do a couple of other things that are unique to our business as well.
At the end of a project, we will actually validate and verify the performance of our service by doing a live fire test with law enforcement. Then they finally have an internal review with the entire project team, basically the entire cross-section of the company, to make sure we're ready to flip the switch, start delivering service, make sure our service is ready, and then secure a sign-off customer acceptance from the customer. Typically, these project managers are running multiple projects concurrently. They're distributed geographically all across the country, so we've got people in the South and people in the East and people in the Midwest. That it's easy pretty easy to get everywhere where they need to be. They are road warriors.
Let's take a look at the next slide, and we'll get a little bit into a little bit more what the project looks like. When I say this is a day in the life, that's a little bit of taking a little bit of license there because what they're doing on any given day depends on where we are in the project life cycle. What you see here, I apologize for it's a bit of an eye chart, something that looks a little bit like a traditional project schedule. This is very simplified, but it's a good model of what we see in terms of a repeatable pattern of things that have to be done, what things have to be done sort of linear sequentially, which things can be done in parallel across the business to bring service to life.
There's sort of a pie chart view I'll come back to later that describes the relative of scope of effort involved and the major pieces of work that get done. The overall things I'd like to have you take away in terms of key themes are, this is a cross-functional, multidisciplinary project effort that covers the whole business, and much of it happens in parallel. Second is that the long pole in the tent, the thing that defines how long it's gonna take to get service delivered in every case is the permissions acquisition process.
When you think about the nature of our business, and this is the first time I'd run across it in a, in similar kinds of endeavors, is where we're putting our sensors are almost always on properties that neither we own or control, nor does our customer. They're independent third parties that live in the community, that work in the community, that probably don't think of themselves as having an opportunity to contribute to solving the problem of gun violence. We give them that opportunity by asking them to voluntarily allow us to put a sensor on their property. We typically don't compensate them for that. The reason they'll do it is because they'll recognize there is a problem to be solved, and they've got an opportunity to help make a difference because they do care about reducing gun violence.
They would like things to be better, and now we give them a chance to do that. That said, it's a pretty long, exhausting, you know, pounding the pavement, knocking on doors job of trying to identify, once I know where I wanna put sensors, who owns that building? Who lives there? How do I find them? How do I contact them? Then it's a one-on-one sales effort to basically get them to allow us to say yes, and there's a piece of paper they sign. Sometimes it goes very fast. We certainly love it when we've got relationships with property owners who have lots of properties, but sometimes it's gonna be a one-off in lots of cases. That's the thing that takes the longest amount of time, and as I said, it defines the critical path.
We also have a series of control gates. These are internal review sessions, and you'll see them highlighted in yellow on this project schedule to give us an opportunity to circle the wagon, so to speak, internally and say, "Where are we? What have we learned so far? What have we learned about reducing risk and improving our confidence in a delivery date that we can ultimately commit to the customer as uncertainty in the beginning is replaced by certainty and knowledge, and confidence as you progress through?" When to commit money, when to ship and start installing sensors, where your variable costs go up, and so forth, as opposed to just sunk salary dollars. There's a couple of them at the very end, right before we do turn on service.
I won't walk through all those things, but you may be wondering about the color codes on these activities. There's some green ones, yellow, red ones, blue ones, and so forth. Those basically indicate which part of the organization is responsible for doing those tasks, while the project manager is responsible for making sure they all get done. Looking at the pie chart up there, again, the major things, it's hard to read, kind of at the 12 o'clock going around the horn, you'll see there's a planning effort, there's a design effort. They're relatively small compared to the other parts of the project as a whole. The very, very large blue on the right is the permissions effort. Again, it is by far the thing that takes the most amount of time and work.
Then we've got materials management, sensor installation, provisioning of the back-end hosted services, live fire tests, the latter two of which are generally very short. The hosted service, in fact, in the back end, the tools we've got now are very, very well known, very much an automated, a near automated process. It's very, very easy to manage. Then the next large part of the effort that you'll be hearing more about later today is the work that goes into onboarding the customer. Pre-preparing their minds, from strategy and tactics, all the way to training, making sure the customer's ready so that when we do turn on service, they're ready to hit the ground running. Next slide. A little bit closer look at the project team.
This is in the big table here, you'll see roles that are present on all of our projects, whether it is a first-time deployment for a new customer or an expanded deployment for a current customer who said they want more miles under coverage. Project Manager, as already mentioned, is both a team leader and individual contributor, architecting and designing our sensor arrays, securing permissions, running our live fire test. The Customer Success team, again, is charged with all aspects of onboarding the customer and getting them ready and making sure they're healthy before we go out. We've got our Network Operations teams that deals with provisioning and configuring the back-end services in the cloud and also managing the sensor networks as they're installed, deployed, and working. The Field Services team actually does the sensor installation.
We do this through contracted third parties. We've got some very large partners that are typically set up regionally, but are moving more towards a national model as well, where we've got redundancy in terms of our service providers there. Supply chain has to do with inventory management, purchasing, materials management, supply, freight and shipping, logistics, making sure that the troops out there are fed, so to speak. We involve our customer support team, who of course, their major role is once a customer is live, early in the planning and deployment process as well. They'll be involved in things like integrating with third-party systems using our API, whether it's a video management system or a CAD system or something similar, and just making sure they're ready to deliver service, they've got some knowledge of the customer before we turn the switch as well.
Depending on different projects' needs, we'll bring in other folks from within the organization for specific piece work on an as-needed basis. Next slide, please. Let's look a little bit at sensor array design, because that again is something that's very, very unique to our business. Again, Ralph mentioned earlier, he was talking about the experience curve and how it's one thing to have a great technology that we know works, and it does, but it doesn't really work in the real world unless it's deployed and used properly, and that's where this great experience of our project management team comes into play. We have some design principles that do drive all of our work.
The first is we design and build these services to make sure they are going to perform, if not over perform, all the time. You may know, you've probably heard in the last couple of years, we've increased our commitment to our customers in terms of the service commitments we have around detecting gunfire. Raising the bar from 80% of outdoor gunfire to 90% or better. To do that, in some cases, we have to go out and deploy more aggressively with more sensors than we used to in the past to ensure that we get that. The second principle is cost.
What I mean specifically there is, we invest, we overinvest upfront with a capitalizable part of our investments, put more sensors out there we need, quite a bit more sensors than we need, so that we don't have to spend as much or as unpredictably on maintenance in the out years going forward. We like to see a nice, smooth, predictable spend on maintenance. One of the interesting things, if you think about the environments we're deploying in, often underserved communities, a lot of turnover in buildings sometimes, especially in the residential side, but even businesses will close. Suddenly, where you used to have power to the sensor, it's not there anymore.
we can anticipate that at any given point in time, we're not gonna have 100% of our sensors up, and that cannot matter to us. We overinvest upfront to make sure that we've got a lot more sensor technology out there than we'll ever need at any point in time. That all comes together with this concept of overbuilding. We do overbuild intentionally for redundancy and reliability. That's the R&R part. If I lose power in part of the town or a building goes down or it gets closed or boarded up, it will not affect our service. If, you know, we've got flakiness with one cellular coverage for whatever reason, we tend to have cellular carriers out there, we're gonna be resilient to those sorts of vagaries that are beyond our control.
We also wanna make sure that, if I can't get into a building to replace something if I ever need to, getting building access may not be as expedient as I want. I don't rely on needing to send a truck out because one thing is offline for a few hours. It gives us a lot of latitude on the back end, basically allowing us to shift the cost to the upfront. In terms of how we translate these principles to practices, this is definitely a mix of art and science. Certainly the science that Paul told you about, speed of sound, triangulation, is well known to us.
What our project managers take in applying that is their experience, the expertise they've built out of experience in knowing how to look at the environment that they're deploying in and knowing how best to execute with the technology we've got. They start off by looking at basically remotely, an aerial survey of a coverage area that's under contract using tools like Google Earth. They're going to be looking for features in the environment that work to our advantage and are gonna be things we need to be able to overcome in terms of being acoustic challenges. We'll look at, is it a hilly or flat terrain? Is it a densely built up urban center with narrow streets or very wide boulevards that don't have a lot of traffic?
Do you have industrial noisemakers, you know, industrial plants, railroads, very big, busy freeways? What's the weather like seasonally in the worst times of weather and when the weather is the calmest? We take all these factors into consideration to decide how many sensors and where would I like to put them so that I get the appropriate density, the appropriate spacing, a nice design on paper that looks good in a two-dimensional space. Then we augment that by going out there and conducting physical surveys of the environment, driving around and getting to know the city very closely, because only when you hear it and see it in three dimensions can you really get a handle on how I'm gonna make adjustments to that design.
That's very much heuristic driven. I don't wanna say quite intuitive, but it requires expertise and experience having done this a few times. When we get a new project manager on board, I guarantee you they don't go out and fly solo right away. They're somebody's wingman until they've done it a couple times and learn the tricks of the trade from the wise sages that preceded them, if you will. It's that site survey that allows us to dial in our design, maybe make some adjustments. We're gonna find out a building that we thought was there is no longer there, or a building is now there, and it's 15 stories high, that didn't used to be there. That allows us to change and fine-tune what we call the emplacements. These are the sites where we're gonna put a sensor.
When we look at those things, we're looking for, is the site suitable for me? Meaning, is it in the right location? Is it a good height? Does it have a clear view to the horizon? In an ideal world, we'd like our sensors to be able to hear things at great distance where the sound's not being impeded or attenuated on its path to the sensors. That's rarely gonna be possible in most cases. When we can find it, we love those things. They've got to be math friendly, which is a weird way of saying, kind of sort of counterintuitively, we don't like our sensors lined up in a perfect symmetrical grid pattern.
We like to have asymmetry because it's very good in terms of drawing nice hyperbolae that intersect very cleanly on the back end because of the different spacing, the likelihood of time of arrival being quite diverse, and things like that. We also have a number of tools that we use. There's a picture of one here on the lower right. This is a two-dimensional density map. It's a heat map that shows relative density or spacing of sensors in different parts of the area where red and even purple means we've got more compacted, denser sensors. As you cool off on the color spectrum, it means it's less dense. That's intentionally done based on what we find in the geography out there.
For example, where you see a lot of deep red and purple, we're probably dealing with a downtown urban, highly built, lots of tall buildings, lots of compressed space, where we expect sound to bounce off a lot of surfaces before it hits sensors. We've got more sensors there to rely less on distant sensors that may not pick things up. As you get to the cooler areas, that might be more the flatter residential areas where buildings are not tall, where sound's gonna travel a great distance, and we're simply not gonna need the number of sensors. I mentioned surveys. That's a big tool using project managers, eyes and ears. There are other things that we use as well, but probably run out of time if I had to talk about all those things.
Lastly, and most importantly, our project managers need to deal with adapting to what the real world throws at them, because as Mike Tyson once said, "Everybody's got a plan till I punch them in the face." What we design in two dimensions on paper and pick buildings that we like to get on to make a perfect service, we've got to rely on the cooperation of building owners who can grant us permissions. We've got to be able to find them and get them to say yes. They don't always say yes. They usually do. We also have to find buildings that have suitable power, reliable power that we can get to, and we've got to be able to get in and out of the building, both to install it, but then later on downstream to service it too.
We design not only for making sure service works on day one, but that we're going to be able to reasonably keep the service working for years to come. That's sort of where we run into the real world considerations. If I can't get this building right here that I want, as much as I might like it, you know what? I'm going to compensate that by going to the two buildings over here and a streetlight pole over there. Between those three, they're going to be better than the one building, even if it takes me a little bit longer to get those three. Next slide is just an example of two similarly sized coverage areas to give you a sense of what I'm talking about, just in the very, very simplest terms, around sensor density. These are two different cities.
Both of them are 3 sq mi. You can tell the red outline that Paul showed you earlier defines the coverage area boundary. The customer's paid to have service inside that perimeter and not outside. The one on the left is a typical flat residential one-story house, 3 sq mi of coverage. I'm able to cover that very, very well with 17 sensors per sq mi. You can see the asymmetry that's in there. That's taking advantage of what features there are to install on, mostly, along with the other principles we have around density, spacing, and so forth. The image on the right is a large East Coast city, very dense. You might even be able to see from the how tight the street grid is there.
These are a lot of very, very tall buildings, lot of very densely packed narrow streets. In some of these areas of that coverage area, you've got public housing projects that tend to be quite tall buildings, where sound is not gonna propagate to great distance like it will in this beautiful flat area over here. We hit that with a lot more sensors. That's got 25 sensors per square mile on average over those 3 square miles, and there are gonna be parts that you see, like on the lower right-hand side, that's probably closer to 35 sensors per square mile, and as you move up to the left, it's gonna be closer to 20. Again, that's just to give you an example. This is typical of what we'll see across our 100 cities.
It varies kind of greatly, in practice, in application, despite the fact that they are the same principles being applied over and over and over again. I hope you will find that to be somewhat interesting in terms of how we bring our service to our customers. Again, once we get done with this, the sensors are installed. We do our live fire test. We have our infamous go live date. That marks the end of the project and when service starts for the customer. It's a major milestone of great importance for us, and it's what we're here to do. If you have any questions, I'd be happy to answer them. Otherwise, I will introduce our next presenter, Sam Klepper.
Thank you, Joe. Good afternoon. I'm Sam Klepper. I'm responsible for the marketing team and product strategy. I came on board at the end of Q1 this year, and I have a team that was focused at that time on content and PR and events. I thought to myself when I arrived, wow, the company has been doing really well, and there really isn't a lot of lead generation going on at the in any systematic format. What a great opportunity to introduce that kind of sub-function of marketing into the company as well as more formal product planning, product management. That's really been my focus, is to expand the really the scope of marketing and make a greater direct impact on the business.
Focus has been lead generation, sales enablement, and product strategy in particular. My background is a mix of corporate with Microsoft and with Bain and startups in Silicon Valley, a lot in the energy space, IoT, subscription SaaS, that kind of thing. I'm very happy to bring that skill set to this company at this time. I'm gonna focus on some of our major initiatives that we're working on today, give you an idea of how we're making that greater impact on the company that you'll see translate into results in, you know, late part of 2019, early part of 2020. Folks, focus first on lead generation. Lead generation in this vertical is pretty different from most verticals.
There is definitely an opportunity for the classic digital marketing 101 email campaigns. I believe that is in front of us and an opportunity. I see it almost like there's an art and science to the whole deployment strategy that Joe talked about. There's another aspect that I think is very important to be effective in this market, and that is almost a grassroots community level engagement with different parties. I'm gonna focus on the first two initiatives are more on that grassroots side of things, and then I'll get to more of a classic digital marketing program that we're doing. The first initiative I want to talk about is called the Healthcare Initiative.
What I noticed, shortly after I arrived is that there were a couple deals that the company had made that were partially funded by hospitals that were in the area. As we look further into this, I realized that there might be a program, turn this from one-off into a program, and maybe change the standard operating procedure in which we approach a city. We now have a, we're in execution of some of the materials for this program and expect to launch next quarter. We have a goal in this initiative to convert healthcare institutions into both advocates and funders of ShotSpotter in cities.
The value proposition is quite interesting to these healthcare systems as we interview them, those that help fund, as well as potential hospitals in other areas. The value prop is multipronged. There's a piece to it which is ShotSpotter improves patient outcomes because we're able to get the cops there faster, first responders there faster. Those who are hit by a bullet can get to the hospital faster, to the trauma care center faster, and their outcome is more likely to be positive. There is a monetary piece to this for the hospitals. Half of the gunshot wound patients in the U.S. are not covered by insurance, the hospital's on the hook for that. The hospital has a patient outcome plus an ROI benefit. There's pieces around employee safety.
People who get shot, often, they are associated with other individuals who might put their own hospital employees at risk, as well as other patients. There's a whole piece around just the hospital wanting to be seen as a positive community partner. We have pretty clear value proposition. We believe the entry point into this marketplace, ultimately, we want to get to the hospital CEO and the foundation, the hospital foundation. But the entry point, we have very passionate trauma surgeons, who are dealing with these gunshot wound victims, and they want to see a reduction in gun violence. One of those trauma centers on the right picture there, Cooper in Camden, New Jersey, where we have a deployment, did their own research.
They weren't intending to do anything about ShotSpotter, they were doing some research. They found that in ShotSpotter coverage areas, the time it took the patient to get from the scene of the crime to their trauma center was much reduced relative to non-ShotSpotter coverage areas. What they are working on now is actual patient outcome data to understand to what degree ShotSpotter actually contributes to saving lives because of that reduction in the response time. They published some data that was at a trauma surgeon conference in September in San Diego. What I did was I hired someone who has healthcare background to then work with other hospitals, gather other data. There's multicenter collaboration going on for additional research. We're very excited about the opportunity to work with these folks.
We believe the trauma surgeons will give us access to the CEOs. That's how it's happened in the past with such a couple of these deals. We are currently working on asset creation, standard things like a pitch deck to the hospital. It's a little bit different than our pitch deck to a police department or a mayor. Testimonials, we're shooting a video in December with Cooper University Health Care and other items that are being created to give credibility to what we're doing and educate others. We'll be launching this in three cities, approximately in Q1. We're looking at a large city, medium-sized city, and a small city to experiment with the impact of a healthcare organization on those size cities.
One of the thoughts is maybe in a medium and smaller sized city, the hospital might have greater influence over a decision, possibly working with the mayor and the police department to, again, advocate in some cases actually fund a ShotSpotter deployment in full or in part. We're very excited about this initiative. Let's go on to the next initiative at the community level. The idea here is to work with the actual community members in gun violent areas in target cities to have them pull in ShotSpotter. This hasn't been done before. We're obviously working top-down, executives at the police department, mayors, city council members. This would be in addition to that, working with community members in two different ways.
In one sense, my belief is that we have an opportunity with prospects that are educated on ShotSpotter, but they're just not moving forward at a quick enough pace. That something has slowed it down. To create a sense of urgency by having community members, in essence, put pressure on the mayor, on the police departments to move forward. The way we're gonna do it is there's a couple different ways. The first piece is we wanna educate the community on the public safety gap, a term that we coined that describes how some parts of the community receive a certain level of, a certain standard of care in response to gunshots, whereas other parts of the community don't receive that same level of service.
We wanna highlight that contrast, that gap, and inform them how they can take action to help provide more equal protection for all parts of the community. What we've done is we've created an animated video, showing two different 11-year-old boys from different towns and how they experience gun violence in their areas. I'm gonna show that to you in just a minute.
We're gonna be launching this video in Q1, and it will be accompanied by a landing page and materials so that folks who get hooked emotionally, hooked by this video can learn how they can get involved, how they can do more, who they can talk to, how they talk to that person in the city to help create some pressure, some advocacy around ShotSpotter, as well as other technologies and techniques. The other piece to this campaign will be surveys. We believe there's an opportunity to create some urgency by actually understanding on a one quantitative level what community members, what their perception of the level of gun violence is in their community and what the level of the police response is.
That may be at a different level than the mayor and city council and police department believes. If there's a gap, that creates an opportunity to educate those influencers in the sales process. If we go to the next, we're gonna play about a two-minute video here. Before you do it, actually, I apologize to the folks who are watching on the webinar. I understand you won't be able to see this, but you will be able to hear it, and you will, in the replay of the webinar, you'll be able to see the full video.
I heard some loud noises. It was like a pop. I thought maybe it was firefighters. Then suddenly there were police cars and sirens all over the place. There were police everywhere. They told us to stay back. Then one came over to talk to us. They asked us a lot of questions. Did I see or hear anything? I didn't. I was really far away. I was outside my house, and I saw a car. There was a lot of popping noise. The sound of bullets, not firecrackers. Firecrackers sound high, like a squeal. Bullets, they're lower. I crawled under a car. That's when I saw my friend's brother on the ground. He was bleeding. He was dead. My mother told me to stay away from the north side of the park. We think he lives there.
I'm worried that he knows I saw him. I stayed home from school for a few days after that. We put up flyers in the neighborhood to ask people to call if they had any clues or tips. When I went back to school, so many kids were asking, "What happened? You saw a shooting? It was so crazy." Since that day, things have been pretty normal, I guess. No one really talks about it anymore. Most of my friends have seen a gunfire or have had to hide when they heard shots. How many times have I heard the guns fired since that day? All the time. I get scared, and I get anxious. Like, maybe I'll hear a noise from outside, and it scares me, and I hate it. I get angry at the noises because they make me afraid.
This media, this video will be delivered via social media in Q1. You noticed probably, we didn't put ShotSpotter branding in the video itself. There'll be ShotSpotter branding surrounding it in the social media. There'll be free social media. There'll be paid social media. Some healthcare partners have already volunteered to help spread the word with this video. Two examples of some kind of grassroots level, very targeted types of programs that we're gonna be experimenting with in the beginning of 2019. Now I want to take you to the next slide to talk more of traditional digital marketing calling campaign. This is with our campus product, SecureCampus. What we've done here is we've looked at, there's no such a list of universities and colleges.
They're very long. There's lots of colleges in the U.S. Where do you start? Do you start in the As, the Bs, the largest colleges, the smallest, private, public? What we did is we took our existing customers, and we developed a scoring tool based on a number of factors, including proximity to crime, proximity to ShotSpotter, if ShotSpotter was somewhere in the area, the size of the institution in terms of number of students, the tuition. The higher the number of students and tuition was a proxy for the institution's ability to pay for the service. We put those all together into a single index score. From that scoring tool, we're able to focus and prioritize where to email and where to call initially.
Our messaging, we in the past have talked a lot about active shooter. It's a very unfortunate incident, obviously. It's also a very rare incident, and a lot of colleges just figure, you know, "It's not gonna happen to us. Do we really have to worry about that?" We have come up with the idea of messaging around a proactive mindset, and don't wait until it's too late. Be prepared for gunfire incidents and use existing customers who have that attitude, customers who have even never had a gunshot in their college or surrounding area, who are expanding their deployment because they think that's the right thing to do. They've got to be ready. They don't wanna be unprepared, caught unprepared. What we've been doing is we've been enhancing the website.
Articles from customers have been published. We have a webinar we're gonna be working with one of our customers on. She's very enthusiastic. One of our latest customers experienced very soon after deployment the two gunshot incidents, both led to arrests, and is very high on the product being valuable to their community, keeping folks safe. We are producing a new video. The campaign will be focused on calls and emails, more traditional. I'll just note some of the images from our upgraded website for this area. I will note that we say that we protect 50 campuses today. As you know, we have around nine SecureCampus direct customers, but we do have over 40 campuses that are covered through our ShotSpotter Flex product.
We really are protecting over 50 campuses today. I'm gonna transition into Missions, which is the HunchLab product that we've renamed, the acquisition, and tell you what we're doing there and why we're excited about it. Background, we acquired this product just before IACP, the big police show that was in Orlando. This is a category that was a little bit sleepy. Three primary players, all small. Limited resources to develop and sell their products. We found this product to be from a technical standpoint and a user interface standpoint, as well as a reputational standpoint, the best product in the category.
We acquired the product, and what we see as an opportunity is to take this great platform and to add our significant brand equity we have with law enforcement officers, to add our sales and marketing infrastructure, product development resources, to launch this into the marketplace and really, grow this business. We feel there's a natural product tie that we have with gunshot detection. On one level, some of our customers tell us that ShotSpotter, not HunchLab, ShotSpotter helps predict where crime is going to be by looking at historical. That has merit. That is the traditional way that police departments use to deploy their resources. They look at historical.
With a product like this, there are many other variables that you can add to that to have a better risk assessment or risk forecast, using machine learning to decide where to deploy your patrol resources and when and for how long. We see a natural tie with our the way that the model that we use to work with patrol officers, which today is, before ShotSpotter, it's they react to a call from 911, and they go to a specific location. We come in with a digital approach that gives them a notification digitally, and they go to a dot on a map. ShotSpotter gunshot detection has been very effective because the model it uses is very similar. It's just more efficient, and right in the hand of the patrol officer.
We feel that this product can use the same model. Now for crime that hasn't happened, that is likely to happen, they can get a notification there's a high risk of a certain type of crime in a specific area and go to that spot. It really does fit together, and we also have seen, and strongly believe that the addition of ShotSpotter gunfire data to the ShotSpotter Missions product will improve its accuracy around violent crime. We feel there's a natural integration and tie with this product. In terms of the total available market, domestically, we see the TAM as about $100 million, focused on cities with populations of 25,000 or more.
While we're not announcing pricing today, we will approach it on a subscription basis like we do the gunfire detection service. This will be on a tiered basis based on how much the service population is for that particular region. What we've been doing and what we're currently working on, we did launch this at IACP in Orlando in October. Top right is from the show floor. There's former police commissioner from New York and Los Angeles, Bill Bratton, demoing the product, bringing folks to the booth around the product. We had tremendous interest. We had interest from both our current customers and customers we haven't talked with, as well as ones we're that are considered prospects at this time.
There was a general I would say there was two kind of segments of prospects for this product at the show, at least. Those that had heard of it, but really didn't know much about it and saw us as trusted advisors to explain what the category was about, what the product does and how they could potentially use it in their agency. There was another segment which really was completely unaware, and again, an opportunity to educate others. That's where I think because of our reach, our ability to reach current customers and other customers, through our sales and marketing channels, we have a great opportunity to educate the market and increase adoption. We do believe our best opportunity initially is with our installed base. They already trust us.
The reaction of several of the customers of HunchLab before we purchased the product or as we purchased the product, and they learned that that was happening, was a very positive reaction because they know that we will invest in the product and make it work the best it can for their particular agency. We, in fact, are working with a couple customers we're calling innovation customers who have offered this, do focus groups with their command staff and patrol officers, to help inform us on what it really needs to have going forward. Even though it's a great base, there's a lot of different directions the product can go. There's been a lot of great ideas we've heard already.
We've collected a lot product requirements in just the past month. Those customers see the value in being able to shape it in the way that they wanna shape it. In terms of go-to-market materials, Gary and I will be doing a sales training at the end of this month with his team so that they're prepared to go out and really hit the pavement and start selling. It involves content, pitch decks, data sheets, videos, et cetera. We'll be having that training. Then we'll be announcing pricing at that point. What you see on the bottom right, I was visiting Chicago.
They're the largest customer of HunchLab, and this is one of their SDSCs, Strategic Decision Support Centers, and there it was, right on the screen, prominently displayed in 1 of the districts, with the red boxes indicating, they had it tuned only for gun crime at this point. There's 7 different crime types that the product does predict or forecast for a risk assessment. They're focused primarily on the on gun violence. You can see those red boxes, and that's where their patrol missions go. Okay. I'm gonna transition into a very different topic. Ralph talked about our purpose, building trust with police departments. Net Promoter is a very important concept at the company, really one of the key tenets for future growth.
I remember when I worked at Intuit earlier in my career. Scott Cook brought me into the company, before there was ever the words Net Promoter, Net Promoter survey scores. He had this concept with Quicken that you invest in customers to make them apostles. It's so much more efficient to take a customer who is somewhat satisfied to very satisfied and will give word of mouth at no cost, relative to customers who are, he called them terrorists, who would, you know, negative word of mouth. You know, that has stuck with me, and I saw an organization grow very quickly to hundreds of millions of dollars in revenue with that kind of strategy. I see really that DNA here. I see that part of this company's culture.
You know, a direct impact that I just really a quick story. I'll disguise the names of the police agencies, but this is the kind of impact that we want, and this is happening. One of our largest customers was at a conference with a police agency that was not a customer, that was not a prospect, that was not in the pipeline, not on our radar screen. And that non-customer said, "We have a serious gun violence problem. It's getting worse. What do you think we should do?" Customer of ShotSpotter. And that customer said, "ShotSpotter and cameras." They had a discussion about it. We were not involved in it. We were not aware of that conversation.
When we became aware that this happened was when we read on the internet that this city, this non-customer city, had gotten city approval and budget for ShotSpotter and cameras. We looked at each other on the executive team and said, "Who's out there? Which sales director has been out there talking to this city?" We hadn't. We hadn't made the pitch. That's the kind of power that positive promoters of a product can do for us. There's a saying that many of you have heard, "Blue respects blue." Police trust other police. As a technology marketer, they're not gonna listen to me. They're not gonna listen to my team. There are materials that will help support gain some awareness and interest, but they're gonna make a decision based on what another police agency says.
The recommendations they provide are key for our growth, and we're already seeing that happen. I'm gonna tell you about the Net Promoter survey we recently conducted through the back half of September and early October, and we have the preliminary results. How we use Net Promoter. It's a quantitative measurement. It's a pulse of how are we doing. We know, regardless, we're in touch with customers all the time, but this is a formal opportunity to have them say, "It's going well. Maybe it's not going so well," and why. We get that pulse, and as we know, their sentiment can impact sales in a positive or a negative way. We identify the accounts. Over 70% of our respondents offer to be a reference for testimonials, references for other potential customers.
We identify issues, and we rectify issues. We have a whole process for getting the feedback and not just sitting on it, but doing something about it. It's so important to us, we've tied it to company-wide bonuses, what our score is each year. Because we feel everyone has either a direct or indirect impact on Net Promoter Score within the company. Just for those who haven't heard about this particular type of study, I refer to it as a customer satisfaction study on steroids. It's, it's kind of a grueling type of grading system. What you do is you ask a key question. There's multiple questions in the study, but this is the key question. How likely are you to recommend ShotSpotter to another police agency that has a gun violence problem?
It's a zero to 10 scale, 10 being I'm gonna definitely recommend it. What you do is you take those who score a nine or a 10, and you subtract all the people who said zero through six, and seven and eights are considered neutral. Even though most people say eight, that's a pretty good score, they don't count. They're not, they don't add positively to the score. You take your promoters minus your detractors to get your Net Promoter Score. The scores range from negative 100% to 100%. We go to the next slide. The grading curve, so to speak, is you're doing good if you get between 20% and 35% Net Promoter Score. Excellent between 35% and 50%. You can see world class is above that.
ShotSpotter scored, if you can hit the button. ShotSpotter scored a 50% NPS. We're considered excellent on the border of world class. We're very pleased with that. We don't wanna sit on that. We wanna get better and better each time. Just for context, you know, I tried to look up and see what other companies have scored. I found evidence that Microsoft was at 45%. I found Google at 35%. I found Amazon, everyone loves Amazon, 62%. Very good score, something to strive for. Comcast was a negative 2%. Something we take very seriously. Now I wanna share some of the specific results. The thing we're very pleased about is our response rate. 91% of the agencies that were asked to participate responded.
That's an all-time high for us. We've been in the 70s and 80s in the past, we really put on kind of a full court, full court press to get those responses. We wanna hear from everyone if possible. Our 50% score was slightly down from last year, which was at 55%, overall, we're up 18 percentage points from 2015. What we looked at to see what happened this year is our detractors decreased. We went from 14%- 13% detractors, the passives were up a little bit, they stole from the promoters. When we look further into this, what we found was, first of all, terminology. There's decision makers, this is internal term, decision makers at the police agency.
Those are the executives who would literally be asked: Would you recommend ShotSpotter? Those are typically the chief of police and possibly assistant chief. Then there's kind of everyone else. The everyone else, sergeants, lieutenants, captains who are involved in some way with ShotSpotter, patrol officers, they are called influencers. What we found is if you look at the decision maker scores on a scale of zero to 10, they were actually a little bit higher than last year. There was a phenomenon this year that more influencers participated relative to the prior year, and the influencer scores were lower. We think in terms of the sales process, we focus on the executives. Those are the decision makers, and I think we do a really good job of explaining the benefits and getting them to buy in.
I think this indicates we have an opportunity with the rank and file, with our customer success team to do a better job of understanding their needs and getting them to buy in that this is a great product for them too. Don't get me wrong, these are 8.6 is a fantastic score, we still have an opportunity to do better in that area. Here's some selected comments from the folks who gave us a nine or a 10. You can see an integral part of their crime prevention strategy, making the community safe again, a must-have tool, a tremendous officer safety tool, it's deterring gun violence, improving the community perception. These are the benefits that we tout in our marketing, and we do that because we hear it on surveys like this. This really happens.
I'll finish this section with a piece of our research that asks customers: To what extent would you agree with the following statement that does ShotSpotter contribute to. There's a number of things we ask them about. Does ShotSpotter contribute to an increase in shell casings recovered because we provide a precise location where the gun was shot? Percentage of gunfire shots identified, guns recovered, et cetera. The thing I wanted to point out overall, high scores overall. The blue is strong. The dark blue is strongly agree, the light blue is agree. Relative to last year, everything has increased, and in some cases, very significant increases. Guns recovered, plus 15 percentage points. Arrests, plus 7%. The perception of our contribution to these items is on the increase.
One of the biggest ones is in reduction in gun violence. Now we did revise the question a little bit. It used to say reduction in casualties, but to reduction in gun violence. That jumped significantly. One more agree strongly agree is on some other things around keeping officers safer, more efficient use of resources, all improved relative to last year. We introduced two new questions. We keep hearing about police community relations being enhanced, and Ralph talked about that. You have to have the community involvement to solve crimes. If a community doesn't trust the police because they're not responding to gun violence, then you don't have that support of the community. 63% agreed or strongly agreed that we contributed to improving police community relations.
On the NPS, we did our first ever NPS with our campus customers. There's a limited number of those customers, not statistically significant, but directionally significant. NPS off the charts, 0.80, +83%, no detractors, 100% likely to renew, you can see some of the great quotes. This is a great basis for me as a marketer to go forward with a campus, lead gen campaign with this kind of customer support. I'm gonna finish with product strategy. When I came into the role. Let me go to the next slide. When I came into the role, my first blush was that this is not about adding functionality, because you don't wanna overfill the product with extra features. This is about doing a core thing really well.
That's a killer app is very simple to use. Paul has gone through that. What I thought about was, when is this product used? It's used. The gunshot detection is while that incident is happening, gets the cop to the dot. The red items that you'll see in this slide are what groups within the police department are getting value from this, are using it. I've got patrol officers, special units sometimes respond in different agencies. PSAPs are essentially 911 centers. They're all involved. Device integration, for example, we send through our API a pulse to a camera and have it turn in the direction of the gunfire, helps with at the PSAPs and real-time crime centers. I looked at it from a perspective of gunshot went off, crime's happening, where are we adding value?
Who are we adding value to? Which group? Which department? I looked at Next. We have some functionality that helps post crime. We've talked about evidence collection, identifying witnesses. You're getting there faster. The witnesses are not, are still there in many cases. This helps a number of other folks by getting the shell casings for those agencies that have a Crime Gun Intelligence Center, they can feed these into the NIBIN system to see connections that which what gun it was fired from and maybe connect to other crimes. There's forensic reporting that is court admissible. We work with PIOs on the media side to communicate results and resource allocation with the chief and command staff. We're very strong on the real-time crime side.
We have some functionality available on the post-crime, now with the acquisition of HunchLab, next, we're in the pre-crime area. From a temporal standpoint, we kind of span the globe, in a sense, from before the crime, during the crime, and after the crime, we're involving lots of groups within the police agency. What I decided to do was get a kind of a standard org chart for police department and see how many departments we were impacting. That's the next slide. The ones in green are departments that we are positively impacting. This isn't the org chart for every single agency. This is kind of a standard medium-sized agency. We're really involved in many of the divisions and partially involved in others. We are really used across an agency.
This gives us opportunities, given our product is successful, given we are respected and trusted, to identify other product opportunities, other issues that they're having, where we can improve how processes are conducted and possibly automate those processes. We saw that on the post-crime area, investigation, criminal, the analysis of the crimes, the whole NIBIN system, ways to automate. Those are areas that we're looking at as potential areas where we can add additional value. That leads to what are our goals and key investments on the product side for 2019. Three key goals: increasing our Net Promoter Score, increasing revenue per customer, and supporting customer growth overall. Big focus. We have high accuracy. We wanna improve accuracy. 100%'s the goal. We may never get there.
We're close now, but we gotta improve accuracy, increase the moat, increase customer satisfaction. That is an opportunity for us. We wanna improve investigative capabilities, as I mentioned, and we think there's some automated ways to increase app adoption. You saw the curve from Paul's slides of Respond usage. There's opportunities to make that go up even more. In terms of revenue per customer, having a second product to sell into the same customer, like ShotSpotter Missions, and then researching additional product areas that we could build or we could buy in the future. Finally, on supporting customer growth, Paul talked about making sure that our system has the capacity and the performance when we add those next 100 customers, and then supporting international requirements as we go into new geographies, and there are different requirements for those geographies.
I think that's it.
That's it. Any questions for Sam?
Yes.
Yeah. Do you see the Pleasantville, New Jersey, bond measure that just got passed as a potential marketing opportunity or something you could use with other cities down the road?
Yes. just for those who don't know-
Can you repeat the question for those?
Yeah. There was a question about a recent ballot measure in Pleasantville, N.J., that had ShotSpotter on the ballot. This is a city that has gun violence, that has researched and wanted ShotSpotter for years, filled out multiple grants, was not able to get those grants, probably because of the size of the city, small nature of the city. What they did is they ultimately put it on a ballot measure for citizens to decide were they going to pay for it themselves directly. The measure did pass in a landslide victory, 70% in favor. Yes.
Initially, when that got on the ballot and got approved, they thought it wasn't gonna pass. It initially was 60/40 against.
Okay. For those on the webinar, this gentleman has said initially the polls indicated that it was not going to pass. It did pass very significantly. We're getting a lot of press interest too, 'cause this could be It's a first of its kind for ShotSpotter. This could be a way for future cities to be able to fund ShotSpotter. They don't have the funds today and are unable to get grants. And we've been asked to be interviewed for an article that's happening this week about that. Good question. Anything else? Okay.
We're gonna be kind and give everybody a 10-minute break to stretch the legs, use the restroom. Those on the webcast, we'll be back in 10 minutes.
Will they just pause the webcast when you pause the webcast, but when they do the replay?
The restrooms are down the hall to the left and then another left.
[Break]
Table eight. We ready to get started again? Well, thank you for coming back after the break. I'm Gary Bunyard. I have been in public safety technology for about 27 years now. Had the honor and the pleasure of coming to work for Ralph almost two years ago January. I, like so many others who have presented before me here this afternoon, was really moved by the value that we offer the law enforcement agencies that we partner with and seeing the impact our services have had on the communities we serve. It's meant a lot to me and consider it my most important assignment of my 27 years in public safety. I'm gonna take the next 25, 30 minutes and talk to you a little bit about sales.
Gonna talk to you a little bit about our sales model, how we approach the marketplace organizationally. Talk to you a little bit about our regional orientation. Speak to you briefly about our playbook for how we try to manage these regions and our territories, something we refer to as a territory business plan. I'm gonna shift our attention a little bit to a life cyclical of a typical sales opportunity for ShotSpotter and talk about some of the more common gates and steps that we often take to turn a lead into an established ShotSpotter customer. I'm gonna finish with a little bit on the role customer success plays.
You've heard several references here this afternoon to our customer success organization, and I want to take just a few minutes to make sure you understand the important role they play in preserving our client base and helping us expand that client base. With that, start off with the sales model. As Ralph so eloquently described early on, our traditional market has focused on this box to the right, neighborhoods with persistent gun violence. That has really been the bread and butter of our operation over the years in assisting law enforcement agency, law enforcement in taking steps to respond to gun crime in these troubled neighborhoods.
On the left of this chart is the other end of that spectrum, it's more of our ShotSpotter SecureCampus focus. In helping campuses be prepared if they should ever find themselves in a situation where they face an active shooter situation. More recently, we've been focusing on and addressing some opportunities and more of a blend between the two. That's focusing on the reality that a lot of the violence in and around these campuses are not actually on campus, but it's in the community surrounding the campus. We refer to this as more of a campus community focus, where we're looking at steps we can help law enforcement take to protect the students, the faculty, not only where they work, but where they live, where they play, and on their way to and from campus.
Our sales organization is really geared towards those three value propositions. The most important role in that is the role of the regional sales director. We'll talk more about the regional aspect of that, but we have sales directors that own a given geography and everything we do within that geography. We've been building up a support organization to really back them with a lot of skills and expertise to do things on their behalf. On the right side, Director of Public Safety Solutions, Public Safety Sales Support. These are individuals who have a law enforcement background, are there to assist with presentations, analysis of crime data, the development of proposals and quotes, things that the regional sales directors need to really respond and move quickly in the marketplace.
We also have several individuals who have more of a campus security, campus law enforcement background, again, supporting those regional sales directors in providing the expertise, the consultation, and backing in terms of proposals and demos and what have you. We also very recently added another position referred to Director of Sales Operations. These boxes across the bottom are individuals that are there to make the regional sales director as efficient, as effective as possible, whether it's taking steps to support our existing customers or taking steps to expand our footprint. We also have the role of the customer success team, which plays a vital role in this, and we'll come back to that here in a bit. In terms of regional sales focus, we've taken the country, we divided it up into five distinct geographic regions.
Every region is owned by one of these regional sales directors. These regional sales directors are very seasoned, experienced sales professionals who have a great passion for what we do and the impact we have on the communities we serve. They own everything we do within that geography when it comes to expanding our business base, whether that expansion is through the securing an additional customer or whether that's through the expansion of an existing customer. They also have responsibility, with a lot of help from customer success and other members of the company, to ensure that every customer renews and that we retain every customer.
Through retaining every customer, we have a very established network of net promoters that are not only willing, but eager to help us close that next deal and help that prospective law enforcement agency understand the success that they are having in their organization. Each of the regional sales directors have a number of metrics that they work under. Common core to all of them is a focus on growing our public safety business. There's two metrics that they are working under. One based upon booking dollars associated with our public safety business and the other based upon just new miles of coverage.
They are also all focused on annual subscription renewals and making sure that 100% of the customers in their region renew and we're expanding that subscription revenue and that customer base either through price increases or through the expansion of our footprint in those established customers. Last but not least is security bookings, a focus on dollar annual recurring revenue brought in the door as a result of incremental security or SecureCampus bookings. Each of the regional sales directors are asked to manage their region like a business. Through that, we've established and are refining a playbook for how we want to manage these territories. There's a variety of elements that make that up.
I'm not gonna walk you through every one of them, but I'm gonna walk through here and hit just a few of the highlights. First, you've heard reference to our addressable market. We've taken our TAM analysis, and we've broken it down to the regional level. Within each region, we're looking at at least four established tiers in the marketplace and our level of penetration in each of those tiers. Let's see here. Looks like we're missing a slide there or something. Within that TAM focus, we're using that as a guide for the sales directors in looking at what do we need to do to expand our footprint, our market share in each one of these tiers.
By doing that, we're recognizing the fact that going into a Tier 0 opportunity is a little different than going into a tier three opportunity with a service population of 50,000 people. The approach that we take, the approach that may involve more of a regional play, we often have to approach those opportunities a bit differently. We're looking at each segment of the market and what we need to be doing and what the regional sales directors need in the way of support from the company to expand our market share in each one of those tiers. We also look at a variety of forces that may be impacting our ability to expand in that geography. This is just a quick sample I've illuminated here in terms of recent ShotSpotter activations.
As you've already heard, this marketplace is very, very focused on what my buddy, the police chief down the road just did. Word of mouth is so key. We're often looking at gun violence trends. We're looking at the activities of our net promoters. We're looking at what we refer to as anchor cities. There is typically, in these geographies, some key core agencies that are seen as a leader in that geography, and a lot of other agencies are watching what they do. Obviously, economics are economic barometers, factors that we're often looking at, recent press, and any regional national events that may be going on in that area that may bring a group of customers or prospects together. I've already touched upon business objectives. Existing customers. We focus on making sure we retain every customer.
One way we ensure that we're focusing and every customer remains front and center is twice a year, we ask the regional sales director to rank every existing customer in terms of most likely to renew to least likely to renew. We may not have a problem with the agency at the bottom of that list, but we wanna make sure any agency that falls in that lower 50th percentile is front and center in our thoughts and our actions. That if they, for whatever reason, they end up at the bottom of that list, we need to have some action items, some things that will ensure that the next time we go through that ranking, they move higher up on that list. Account reviews.
We're gonna talk about those in a bit, but part of that territory planning is mapping out a game plan to ensure we do an annual account review with every established customer in that region. We have a very robust tool, Salesforce, that we use to manage our sales funnel, our sales pipeline. We break that sales funnel down into five different stages. We track all opportunities through that funnel, opportunities for new business, as well as opportunities to expand our footprint with existing customers, and we do that down to the regional level. We track the movement of opportunities through that funnel at the regional level. We take an extract of that, and we roll up every month and an update every week, a new forecast. A new forecast that looks at those opportunities in the funnel that the regional sales directors are comfortable committing to.
Looking at current quarter, next quarter, and looking at a variety of elements that we track along with each of those opportunities. We always maintain what we refer to as the top five opportunities. It's a dynamic list. Things move all on and off that list, but there are a variety of data elements that we track with the top five opportunities that are front and center in what those regional sales directors are trying to do to drive our business in the current and next quarter. We're always looking back at any challenges. Are there anything you're dealing with that you need our help on? It may be assistance from our marketing organization. It may be assistance from customer support. Are there any challenges you face that you need the company to support you on, whether they may be funding issues, support issues.
I've got the reference here to competitive. Are there any competitive activities going on that may be getting in the way? In referencing this, in planning for this presentation, it dawned on us we didn't really have a slide up here to talk about competition. And we've already touched upon the fact that we really don't have any competitors. There are companies that pop up from time to time representing themselves as offering outdoor gunshot detection capabilities. As you can imagine, selling specifically to city, county government agencies, everything of any significance goes through a competitive bidding process.
However, we enjoy the reality that over 90% of our contracts are secured through a sole source justification, where agencies purchasing city legal have determined, "Okay, in this case, there are no viable competitors, so we're gonna allow you as the using agency to move forward and avoid the competitive bidding process." Last but not least is in every refresh of our territory business plan is the business outlook, looking at how are we doing against our annual objectives, how are we doing on a quarterly basis with projections and what does that suggest we're gonna do in terms of % of our annual plan. This is our playbook for how we ask each of the regional sales directors to manage their territory, and they have an opportunity several times a year to present to their peers and to the executive management team how they're doing against this plan.
Now I want to turn your attention for a couple of minutes to a mythical opportunity, an opportunity that represents many of the steps that we often take in taking a lead and turning it into a new ShotSpotter customer. What I would say up front is, very few opportunities are gonna include everything on this chart. In many cases, we may follow these steps in a different order. There are some cases we may have to repeat several of these steps. This is just to give you a feel for some of the more common steps that are required as we go through the sales process. I'll hit on a few of the highlights. Starting with lead generation. Our leads may be a contact we made at a trade show. Leads may come in through our website.
Leads may be, I've been knocking on their door for two years, they finally invited us in and want to talk. The most common source of our leads today is, yeah, we've been in there. We've met with them several times over the last four or five years, they just never were really ready to move forward. We activated that system down the road last year, there's been a lot of great press and a lot of good things happening. Now they want to sit down. Now they want to talk about what we might be able to do within several troubled neighborhoods that they've been struggling with. Most of the leads are coming from a success that we're seeing and the impact it has regionally as these agencies are always looking at what their partners are doing down the road.
Early on, we go through a qualification process. Before we get too deep, we do a lot of research. We understand the history, the background of the key players, stakeholders, their affection for technology, the kind of gun crime issues they're dealing with. We look at a variety of factors that help us navigate through that organization. We're always looking for who could be our best advocate, who could be our inside coach, who's gonna help prepare us when we go in for that briefing to say, "Make sure you've got an answer to this question because you're gonna get hit with that question." We often go in and go through a series of briefings where we go into the present to the police chief, command staff, sometimes some of the elected officials. Sometimes it's a series of briefings, sometimes it's one.
This is typically the pinnacle event in our sales process where we get everybody, all the decision together in a room, and we ensure they understand what we do, how we do it, and the impact it has had on other communities who have taken this step before them. As a part of that, we're often bringing in some net promoters, putting them in touch with a peer down the road, the other side of the state, who is going to sell the system for us. It's going to help them understand the impact that the technology and the services had on that community. Somewhere in there, we're going to start the process of analyzing the customer's crime data and start recommending coverage areas and where we can provide the highest return on investment, preparing to present a proposal.
As I've already talked about, we often assist them in helping them justify the sole source procurement and helping them understand why there are really no competitors out there. As I mentioned earlier, have been successful at that in over 90% of the procurements we're involved in. We often will assist the agency in looking at funding options. Very often, going into a new organization, there's not a line item for gunshot detection technology in their budget. They've got to get that line item added. We provide a lot of assistance in helping them understand what they can do locally within the police budget, city budget, asset forfeiture dollars, public-private partnerships involving maybe healthcare, as Sam talked a little bit about earlier. We also have assistance in Washington in helping them explore federal grants and external funding.
A variety of funding options available for these agencies, and we do a lot to help them navigate through that. We may be brought in for community meetings. The community may be very interested in what's going on and what law enforcement is gonna do about gun crime, and they're hearing about this company that's coming in to analyze gunshots. We may be asked to come in and present to those communities, typically with very positive results. If we're successful, then we move into the process where we're working with purchasing, we're working with city legal, and we're going through the process of hammering out a contract, typically using our paper, typically our paper supplemented with some unique terms unique to that particular agency.
As we go through that contract process, culminating with this thing has to go to city council or to some board, some group for final approval. That whole process can go as quickly as three or four months. That's very rare. It can be as long as three or four years. On average, it's typically 12- 18 months. Customer Success. We've referred to the Customer Success organization a number of points this afternoon. Wanna make sure you all understand the important role that this group plays in our path moving forward, growing our business. We have a team of people. Their mission is ensure that every ShotSpotter customer maximizes the value, the success of the ShotSpotter gunshot detection location service.
Very simply put, their role is to do whatever they need to do out in the field, leaning in, engaging in the weeds with these agencies, helping them figure out how to maximize the value of what we're doing. The team of people we have assembled, and it is growing, today made up of seven people that have almost 200 years of law enforcement expertise amongst them. Some of that is local law enforcement, as in city county law enforcement. Some of that is federal, specifically ATF. These are very senior, very experienced law enforcement professionals that are highly regarded, highly respected in our field, and are very comfortable sitting down with the officer in the field, sitting there in their patrol vehicle, talking about how they can use our mobile app, or sitting with the police commissioner and helping the commissioner plan his or her strategy moving forward.
The first role they play is in the onboarding process that Joe touched upon very briefly. They may be involved in the sales effort, just helping the sales team articulate the role they're gonna play in the deployment and ongoing support. The onboarding process is where they typically now come in, they're officially introduced and engaged working with the customer. What does that mean? When Joe's team, he talked about the whole sensor deployment process, when they start the process of deploying the sensors, the backroom technology, we're out in front working hand in hand with the program manager and the users, the key stakeholders in the organization. First, we're looking at best practices. Helping them establish the best practices for how they're gonna use the products and services.
How they're gonna use them more effectively, meaning how are they as an organization gonna operate with this new information that's becoming available to them, and what is their response protocol going to be when that first alert comes in. Spend a lot of time, obviously, on user training. A combination of computer-based training, train the trainer, user-based on-site training, working hand in hand with that agency, scheduling issues, shifts, doing what we need to do to make sure everybody can be an effective user day one. We also spend a lot of time helping that agency looking at analytics in terms of what metrics, what KPIs do you want to establish to track the ongoing efficacy of this program. Helping them not only establish those metrics and KPIs, but institutionalize routines to maintain these statistics on an ongoing basis.
A lot of support in terms of go live, before, during, and after go live, followed with a number of post go-live checkpoints to check back in with each of the key users in the organization and find, okay, how are we doing? How's it working out? What do we need to be doing within your organization? Their role as an ongoing consultant to that agency, to stop in from time to time and check in on how the system is performing. Which leads me to this last topic, and that is annual account reviews. We, as a part of our customer support playbook, want to sit down with a customer at least once a year. When we do, we're focusing on asking three key questions. How are you doing in achieving your goals with respect to gun violence reduction?
How is ShotSpotter doing in helping you achieve those goals? What can we be doing to better serve you in achieving those goals? These annual account reviews are where we sit down with typically police chief, police commissioner, command staff, key stakeholders in the organization, sales director, customer success director, often several members of the customer success team, and a member of our executive management team, sits down in that customer's conference room and over a course of an hour or two hours, explore these three questions. In every case, we come away learning more about what we can do to better serve that agency. In terms of 2018, we're wrapping up our account reviews. I share this just to let you know this is real. This is not something we talk about.
These are all annual account reviews that have been completed so far in 2018, where again, three or four of us set out with that police chief, that police commissioner, members of their command staff and rolled up our sleeves to talk about how they're doing and how we're doing supporting them. As a result of that, we've identified now, like, over 160 action items that came out of there. Specific, defined, documented action items that we walked away with, things we wanted to do on behalf of that agency. We did this year, kick off district-level reviews in Chicago. Chicago operates in a very decentralized fashion. We made the decision that each district where we're supporting the Chicago Police Department deserves its own account review.
That has turned out to be a very good decision based upon the feedback we've received so far. We have 12 more reviews to complete by year-end, and our goal is 100% of our customers see us, sit down with them and ask those questions every year. Any burning questions?
With the sales cycle of 12- 18 months, when you go into the year, you pretty much have good visibility about what you're gonna be doing that year, right? I mean, you're pretty much your sales force is focused on next year at that point.
For those on the webcast, the question is steering on that sales cycle and the visibility we have.
Right.
Looking over the horizon based upon a 12- 18-month sales cycle.
Correct.
We, as an example, through our territory planning process, we've already identified the top five opportunities in each region for 2019. Now, that list could change, but we know based upon where we are in the sales cycle, i.e., where these opportunities are in the sales funnel, when we think we should be able to close those.
We've mapped out deals that we think we can close in Q1, Q2, Q3, et cetera. Yes, we, because of the length of the sales cycle, have a lot of visibility into the deals we're working. There's a lot of factors there that we don't control regarding the timing and the gates that the agency has to go through to get final approval. In some cases, the people we're working with are not all that clear on the gates they're gonna have to go through until they get to them. Visibility, yes. A great deal of confidence in knowing exactly what month we're gonna close that, sitting out there, you know, two quarters out.
Chicago example was double over 100. When did you roughly know that Chicago was going to be doing that?
The Chicago deployment happened in several phases.
Right.
Those phases were preceded by some pronouncements from the mayor's office regarding steps they were taking to combat crime. There was a fairly public announcement regarding the steps they were taking. It does make them unique, not only in size, but the visibility that came with the way those were rolled out.
Last question. If you had what would you like to see in terms of miles per year in an ideal situation that each of your regional managers is executing in terms of expansion and new cities? What should be a mix that they should be trying to target? Is there something? Did you look at it that way? Like, we should be getting 50% from expansion, 50% from new customers.
Two thoughts. One, looking macroscopically, we're looking between 150- 200 sq mi per year in terms of new miles. The mix can vary considerably by region.
Right.
When you have an opportunity like Chicago represented for expansion. The variation between our expectations, new business versus expansion varies tremendously year by year as well as by territory. I can't really give you a hard number there.
Right
because it only takes one major expansion opportunity to shift that weighting pretty dramatically.
If we're doing 200 a year, we're doing all right.
All right, Ralph.
Are you up next, Alan?
Yeah.
Alan's really enjoying this.
Normally, I'm the one in the hot seat with Ralph. All right. Thank you. We will have a chance for some additional questions to wrap up. We are kind of getting close to the end of the time, though. We want to make sure we get out of here so everybody that wants to can go see the Incident Review Center as well. That'll be following my presentation, some closing remarks by Ralph, and then a short Q&A session as well. For the benefit of those that are on the webcast, most of you that cover us or covering analysts probably know this already, but some highlights about ShotSpotter I want to cover is It's a SaaS subscription model that has very low customer acquisition costs.
You heard from Gary about how we only have five direct sales reps, how they're geographically dispersed. In 2017, we only spent $0.34 per dollar of our annual contract revenue. That's very low compared to most SaaS companies. We have a very rapid payback of our installation costs. We charge $65,000 per square mile. There are some legacy customers that have a slightly lower price point than that. In general, we've been at that price point for about two years now. We don't say exactly what those, a sq mi costs for us, but what we do say is that if a customer signed up for a one-year contract and never renewed, we would still make money in that one-year contract.
We do break even in less than a year on a project basis. It is high margin, low variable costs, and we have significant operating leverage. When you think about the model right now, in our cost of goods sold, we have a couple components. We have components that are related to our Incident Review Center. We've got components that are related to our direct contribution to the projects that roll up from the operations side into COGS. When we actually deploy a project, we roll up all the costs related to the materials, the sensors, and then we depreciate that from a GAAP basis over five years. Those are the three components of the cost of goods sold for our model.
As we grow, we have the ability to leverage that, so that when we grow from $30 million- $60 million in revenue, we don't necessarily need to increase our, let's say, our Incident Review Center costs, from, let's say, 20 people to 40 people, it might only go from 20 people- 28 people because it ends up being a peak load problem. We'll look at it a little bit on the income statement to show the progress on that and why that's important for our gross margins, and they're improving. We have a very low customer attrition return, which leads to high revenue retention rate. 2017's revenue retention rate was 141%. You can think about that sort of on a same-store sales basis.
For example, customers that we had in 2016, net of any losses, plus the expansions that those same customers had were 141%. 40% footprint expansion net of any losses, which translates to a very high customer retention and a very low churn. Our goal, and this is very important for us, and everything that you heard today talks about that customer intimacy, building that trust and that relationship. Our goal is not to have a high churn on our customers, it's to have a very long customer relationship. Think 10, 15 years. We have one customer that's been with us about 20 years, one of our very first customers in Redwood City.
From a value of a customer, if you think about our average footprint, it's just slightly more than 6 sq mi, and we charge $65,000 per square mile, that's about $400,000 per year on a subscription basis for an average customer. If you have that customer for 15 years, that customer value, $6 million. It's important for us to get the customers, but even more important for us to keep them. This shows our recurring revenue model. You can see that the arrow goes up and to the left. That's always good. We wanna make 1 highlight here. For those on the webcast, I'm highlighting the Q3 2017 last year, which by the way, is our comparative quarter for Q3 of 2018.
In Q3 of 2017, we had about $900,000 in revenue, was an acceleration of the setup costs related to when Hurricane Maria wiped out Puerto Rico. Otherwise, if you were to take that out, that would've been about $5.9 million, and you'd see a significant and very nice upslope to the right. Nearly 100% of the revenue subscription base. We say nearly because there are those setup fees that we charge. In addition to the $65,000 per year, we do charge a one-time $10,000 setup or activation fee per sq mi. There are occasions when we might not charge that, generally it's related to a multi-year deal, a three-year or more contract, then we might waive that setup fee. Generally, we do charge that.
We talked about the very low and attractive customer acquisition economics, only $0.34 to generate each dollar of new annualized revenue while keeping that revenue retention rate at 141%. Here we have a snapshot of the last couple years. The CAGR growth rate from revenue from 2015- 2017 was over 42%. You can see here highlighted in blue the actual gross margins. I'm sorry. Yeah, gross margins. 2015 was 30%, 38% in 2016, and 49% in 2017. Year to date, just the last two years, for on the far right, 49% gross margin, up to 54% year to date through Q3 of 2018.
The revenue growth rate at 45% for the year to date 2017- 2018. We are accelerating. All the things that we're doing are adding to our growth rate, and they're also adding to our improved margins as we progress. One thing that helps us from a visibility standpoint is our strong deferred revenue. You could take a look at the deferred revenue that we had at the end of 2016 of $13.9 million, the end of 2017 of $18.5 million. After third quarter, already over $20 million in terms of deferred revenue. It's great for us in terms of visibility because our short-term deferred revenue is generally recognized over the next 12 months.
For us, 12/31, we know kind of what we're looking at when we look into 2019 based on our short-term deferred revenue already. We do a strong balance sheet. As of today, we have over $12 million in cash, and we have an untapped line of credit, $10 million. We have not used that at all. That has the capability or potential to grow to $25 million as well. You may note that the cash balance today is slightly lower than we had at the end of our third quarter results, and that reflects the HunchLab acquisition as well as the settlement of the litigation and the lawsuit. I would also mention, though, that for Q3, we generated cash flow from operations of $3.4 million.
Things are going well in that direction as well. Some highlights for Q3. Revenues were up 35% to $9.2. Again, that is compared to the quarter last year of the actual revenues with the inclusion of the acceleration of the Puerto Rico revenues. 36 net new miles for the quarter. That's up from 17 in third quarter of last year. Three new cities and two new campuses/sites. Gross profit of 55%. I would note, as we mentioned yesterday, that had we not written off the remaining indoor sensor inventory that we had, that write-off went through COGS, was a little over $270,000, our gross margin percentage would've been 58%. Gross margins continue to march up nicely.
In a couple slides, I'll talk a little bit about our short term, or rather midterm and long term, models for gross margin. Cash flow from operations provided $3.4 million in the quarter. Talked about current cash balance. Our net income was a little greater than a loss of $1.4 million. It was actually - $1.441 million. We already talked about some of those one-time costs. About 270,000 of that came through the COGS with the indoor sensor write-off. Approximately $1.1 million was in G&A related to costs related to the HunchLab acquisition and the settlement of the litigation. Taking those out, we would've been very close to a profitability. Project Patrol was successfully completed.
That is what we called HunchLab, ShotSpotter Missions, now renamed. We completed the bank line of credit. 2018, we again raised our revenue guidance to a range of $34.4 million-$34.6 million, and we are on track for fourth quarter GAAP profitability. Those of you who listened to our call yesterday know that we gave our initial revenue guidance for 2019 of between $45 million and $47 million, mid-range of which would represent a 33% increase from our now increased 2018 guidance. Just some highlights here. For those on the webcast, I'm focusing more primarily on the far right side for the year to date, Q3 2017, looking at the gross margin %.
The gross margin percentage here was for the operating loss was 12%, but for the adjusted EBITDA in Q3 2017 was a -21.6%. Compare that to the year to date for 2018, where we had an operating loss of 12%, but an adjusted EBITDA of almost 6%. We are making great strides as the revenue increases. You can see the revenue increased through Q3 of 2017 from $17 million up to $25 million, with a corresponding increase of the gross margin percentages and the adjusted EBITDA as well. This is an example of how we have leverage in the operating model itself, both in the cost of goods sold side, but also in the operating expense side. I do wanna talk a little bit about the changes in the operating expense.
For example, Sam talked about a lot of opportunities and things that we're now doing in the marketing side that we weren't doing before. Gary talked a lot about the build up on the customer success organization from where we were and the importance of those and the account reviews that we do. Those are the drivers that are increasing our sales and marketing costs, as well as our international operations, which we really just kicked off this year as well. From an R&D perspective, you heard Paul talk about some of the things that we're not only doing to improve the ShotSpotter Missions side, but also in terms of scalability, some AI techniques, some things that we're doing on the core technology side that we continue to invest in in an R&D side, and we'll do that into 2019.
From a G&A perspective, the numbers here look a little odd because they include in the Q3 2018 number, $6.764, that does include some of those unusual or one-time costs that flowed through G&A. That said, we are continuing to increase and invest in our admin staff into 2019, so G&A will continue to go up just a little bit as well. Here we talk down at the bottom a little bit about the increases and what the stats were from a revenue retention rate, 2016 versus 2017, and the spend per dollar of 2016 versus 2017, all still very low. We haven't yet calculated exactly what the marketing spend per dollar is gonna be for 2018, but we expect it to be similar to what we've shown here in 2017.
This is a slide a lot of you have been waiting for. In terms of the longer term model that we are trying to achieve and I believe we can achieve. If you think about the midterm here in the center, we know we can achieve gross profit margins. Midterm, medium term, think of that as like 2020 of 61%-65%. Sales and marketing percentages around 22%. R&D around 14%, and G&A around 19%. With that, total operating expenses as a percentage of revenue around 55%, or now having an operating margin of around 8%+. Longer term, this is longer term, we can think about that as we get closer to the $80 million-$90 million, $100 million revenue range.
Gross profit margin 65% or more, corresponding reductions in each of the areas of the operating expenses, you can see the percentages there. Sales and marketing, 13%-15%. R&D, somewhere between 8% and 10%. G&A, somewhere between 10% and 13%. Giving an operating margin of about 25% on a relatively low revenue base versus other SaaS companies. Our EBITDA margin would be above that, you can think about adding another maybe 10 percentage points into an EBITDA, above the 25%.
I think that's my last slide. Turn it back over to Ralph before we open up the Q&A.
Thank you, Alan. I'll be really brief because I know we all want to get over to the headquarters location and view the IRC. I first just wanna say thank you so much for investing this time to get to know the company better. For those of you who've been online, hopefully the whole time going through almost 80+ slides, we're incredibly grateful for your patience and your willingness to invest in getting to know the company a little bit better. I want to close on this slide because I think you hear us talk a lot about the purpose of this company and our commitment to really serving customers and how much alignment we have with our customers in terms of helping them become successful.
We think it's really important that our investors that choose to partner with us are a part of that purpose equation too. I think the bigger and we can make that surface area of purpose be, I think the more alignment we can have between our company, and when I say company, I mean employees, I mean partners to the company. When I say customers, I mean customers, not only the buying center customers, but also the community, the broader community that our law enforcement agency customers are supporting. Investors, you the investors. We're really privileged early on in our company's history to find private equity investors that were very much aligned with the mission of the company.
Although We had to perform as a company, obviously, I think they saw a long-term vision about what this company could be, not only in terms of the impact we can make in communities, but how we can really break the mold in terms of operating a profitable company at scale in a much more efficient way than a traditional SaaS company, in addition to having kind of this moral purpose. You've probably heard the concept of double bottom line. This is very much a double bottom line here too. I hope that you appreciate that. Again, thank you all so very much.
I'm really personally grateful for the energy and time that you spent. We're now looking forward to taking any questions that you have before we run over to the headquarters location and see the Incident Review Center. Any burning questions or not even burning questions? Yeah, Will.
Yeah.
Yeah.
Yeah. A question on missions and the acquisition. I think, you know, Sam that said that Chicago's really using that product and partly for gunshot detection. I guess what I'm really curious about is what are the analytics that product is providing that Chicago wouldn't already be able to get through this for ShotSpotter, you know, product that really provides additional value, you know, for using it?
For Chicago?
Yeah, Well, yeah, because.
Sure. I'll try to answer. Dan, sort them up here too, 'cause I think you're gonna need to fill in. The question is, Will asked a question about, "Hey, Chicago's using ShotSpotter. They're also using Mission, formerly HunchLab. What is it that the HunchLab's application is giving them beyond what ShotSpotter would give them?" Is that in terms of data analysis? I think the short answer is, ShotSpotter data is really good about providing real-time alerts, gunfire alerts that can, within 30- 45 seconds of the trigger being pulled, send an alert to a both dispatch and patrol officer telling them three shots have been fired at this specific time, at this specific location. Extremely powerful.
We've seen some agencies have take the culmination of that data historically and be able to do things, kind of really simple analytics. Okay, you know, it looks like I have gunfire that happens mostly on the weekend between these hours of kinda 10:00 P.M. and 1:00 P.M., and they're generally happening in this particular area. It's kinda 2 sq mi. You know, kind of interesting. Helps me better plan my patrol resources, et cetera. It has a more of a historical view, not quite as precise.
I think the power of HunchLab is now renamed ShotSpotter Missions, is ingesting the ShotSpotter data, combining other pieces of data as well from CAD, computer-aided dispatch, also record management systems, RMS, things like weather, other data attributes, and run it through a very sophisticated machine learning algorithms that can provide a risk score that does something really interesting. We use the term prediction, but I think risk is probably a better way of looking at it. You have a heightened risk for this type of activity to happen in this particular area spatially over this very narrow time period temporarily. It's taking space and time. It is a little bit more predictive in terms of having elevated risk.
It goes further on a real-time basis to say, "Okay, given that this is your risk profile, if you have available patrol resources that aren't already dealing with a call that they're dispatched to deal with, they should go have a law enforcement intervention in this area. They should drive their patrol car to this particular area, here are three or four things that they might wanna do for the 15, 20 minutes we're asking people to do that." Instead of responding to events like right after they happen on a real-time basis, what you're trying to do is implement law enforcement interventions to prevent crime from happening.
Just by that law enforcement intervention, that officer being in that area, maybe hanging a door hanger on someone's door, just maybe going and knocking on a door introducing themselves, that creates an anti-vaccine, I guess, toward violent crime or some other crime attributes. It works really well with robberies, for property crime in particular. That's just where it is today. We believe we can take that forward in fairly significant ways, 'cause now we're really combining this real-time intelligence that's more focused on pre-crime versus at crime. We think it's pretty powerful, and that's really been part of the toolbox that Chicago has used to be so successful in driving down violent crime. Did I get that right? Could you add on to that, Sam? Thank you.
Yeah, I agree 100%. What I wanted to add on is the way that traditionally police departments predict or forecast crime is a method called hotspotting. Not ShotSpotter, hotspotting. They look at historical crime patterns only. What this product does is a couple things. It adds in all these other variables that give it a higher degree of accuracy for forecasting risk. Okay, they're getting better data. There's comparisons against the traditional hotspotting method, it's significantly higher forecasting accuracy, it's for a much narrower temporal and location base. It's going to a 250 m by 250 m box at a specific 15-minute window. They could be much more precise about where they're going. That's number 1.
Number two is, this isn't necessarily for Chicago, but for other agencies that look at all different crime types to understand robberies, property theft, other types of, They call 'em Part I offenses, in addition to gun violence, and we haven't been in that business yet, so now we are in that business. I think the third thing, this is particularly for Chicago, as well as some of the folks at IACP who I said came up and really showed a lot of interest in the product. This is from a patrol management standpoint. There's forecasting risk, and then there's where are my patrol resources? What are they doing?
They are a huge percentage of their budget, and they're out there, and many agencies don't know where they are, don't know what they're doing, and this is a product that can start to quantify that, summarize that, report on that, so they can make better decisions overall on how they deploy resources. Did that answer your question, Will?
Yeah.
Okay, great. Thank you. There's another question here. Yeah.
With the initial installations that you guys are doing, is there a path towards also adding on differentiated and more subs capability that go beyond what ShotSpotter Flex does?
The question is, if I might repeat it or rephrase it a little bit, given that we have a lot of sensors deployed out there that are finely attenuated or tuned or engineered to detect gunshots, are there other things that we can maybe use those sensors to detect? Is that a fair characterization?
Correct. Yeah.
The short answer is no. I mean, we're really happy having a very narrow field of view with respect to the sensors that it's not there to detect anything other than impulsive noise that have a high likelihood of being gunshots. Why is that the case? Is why that's the case because, A, from a privacy point of view, we wanna be very sensitive not to broaden our field too much and create issues for ourselves. Secondly, there's so much work to be done just in detecting gunshots. When people ask me about, say, "Well, what other things can you do?" I say, "One of the things we can do is take our penetration of, like, 8% and make it 25%." That's a lot of work.
We don't need to, you know, get particularly fancy beyond out over our ski tips doing anything else, 'cause we have a really wide open field just to continue to dominate this particular market segment that we created. That's what we want to do. Great. There's a question over here. Yes, sir.
Following on the previous question a little bit, you said, and I remember I asked you the same question, and you kinda said, "No way, we're staying focused," right? You've got your very tightly focused gunshot detection mission, but then HunchLab does seem to take you down that other path of broadening the types of things that you do. I was thinking about this in the context of, you know, this category you call pre-crime, which is actually also a term from a dystopian science fiction novel.
I'm not the marketer in that regard.
How much risk do we have here of you getting pushback from communities about being overly policed and overly surveilled as you try and expand beyond this very tightly focused service?
Yeah, that's a fantastic question. I think first do no harm has to really be a mantra of ours, and also for police departments to be effective, because if it's not about a service to being a better service to communities, then it's not a productive effort. I'm really proud to say that, and 'cause we looked at some other options, besides HunchLab, but One of the things that we liked about HunchLab was the fact that they take the whole question of civil liberties very, very importantly.
For example, one thing that you can pretty much expect us not to do is be very, I call them anti-civil liberty space with respect to using predictive capabilities, where it's unfairly characterizing or criminalizing a neighborhood, as an example, or criminalizing an individual to the individual level. That's when you start getting into some trouble, appropriately trouble and appropriate trouble, I would say. We don't need to do that, right? There's so much work that we can do around just helping police departments become more professional and precise in their law enforcement interventions.
In fact, there's some incredibly wonderful management consulting research being done and writings being done, driven a lot by Commissioner Bill Bratton, former NYPD Commissioner Bill Bratton, and also a ShotSpotter board member, where he talks about this concept of kinda CompStat 2.0 or 3.0. It's really about how do we enable a law enforcement agency to become more professional and more precise in a way that you don't do greater harm to communities. A concept I like to refer to is you don't wanna over-police communities and under-serve them at the same time.
Serving the community, not over-policing the community. We wanna have tools that help build that message. I think we're gonna have very strong allies in that regard. Great question. We have to pay attention to that. Super. Yes.
In 2017, we lost a customer in Puerto Rico because of the hurricane, which was about nearly a $2 million customer. We've grown nicely through that. Does any of the guidance in 2019 assume Puerto Rico coming back?
The question is, we lost a very large customer in Puerto Rico, and I would add USVI as well, due to the hurricanes, and we've been able to effectively grow through that. The question is: to what extent is Puerto Rico in our plans for 2019, I presume?
Yes.
I guess I would answer the question the following way. We're not talking specifically about how we're thinking about our pipeline, but we have a lot of visibility, I would say, into getting to that number that we gave guidance around for 2019. We don't wanna be too specific about what customers may or may not be in that. A lot of those customers already exist and are customers.
Right.
It's like what's in short-term deferred, plus given our track record with renewals, right, as they tail off, getting those customers renewed. There's another number of customers that we haven't talked about publicly that we've already booked and are lighting up, and if anyone was creative enough to go do a Google search on ShotSpotter customers in life, they could go see those customers. As an example, and we don't have a contract yet with this customer, but I heard there was an election somewhere in New Jersey where, you know, residents voted 70% to fund ShotSpotter. I'd like to believe that's maybe a customer that we can get in 2019.
The funding for Puerto Rico came from HUD, right? Previously.
We're talking back in 2017?
Yeah.
Yeah.
Very previously.
The originally. Yes. Mm-hmm.
Yeah.
That's correct. Yes.
That would be there's funds for Puerto Rico from HUD again, right?
You're asking me a question that I'm not gonna answer about what am I thinking about the odds of Puerto Rico coming back in 2019. I can't. I'm not gonna comment on that. Thank you.
One question for Al.
Al's not gonna answer it either.
Not that question. If we strip away the write-off on the gross profit line-
You'll be around 57% or so for the year versus 49%. Looks like the incremental gross margins were around 75% or so. Is the incremental gross margin going forward of 70%, 75% the number, or are there, like, costs that will be layered on into the cost side? Because that is where you have so much operating leverage.
Yeah. The question was in terms of gross margins, really where we are and we'll end 2018 versus what we see going forward, particularly as it relates to an incremental gross margin. What I would highlight, and part of the reason that we can have a higher gross margin as we grow, is the incremental gross margin that he's referring to. When we deploy a system, we do roll up all those costs, and we do amortize those or depreciate those over five years. However, on a cash-on-cash basis, on that year two that that contract renews, the actual costs related to that subscription fee that we charge, are very low. For example, it's telecommunications costs, might be break/fix costs.
In that second year, the incremental gross margin on a cash basis is north of 75%, just like you say. What I would expect, the punchline to all of that is ending the year somewhere between 57%-58%. We already talked about the medium-term gross margin, you know, somewhere around that 63%. I would expect the gross margins to increase, notwithstanding any of the one-time costs, in a fairly linear trend between now and then.
If we have high number of expansion or renewals of larger customers have favorable gross margins, if there are new deployments, large amount of new deployments with new cities, that has a little bit of a dampening effect to the initial-.
Just a little bit.
Not material.
Not material.
Yeah.
The renewals From a GAAP point of view, the renewals are going to include some amount of depreciation amortization.
Exactly.
Yeah. They're the same basically, unless it's a renewal of a customer that we deployed six years ago.
Yeah.
Right. In which case it will have already fully depreciated its cost. Yes, sir.
Yeah. Looking at your sales expansion, so you've got five territories in the United States. Even adding in one person, maybe two, would cause a fair bit of dislocation in the customer base and what have you. Can you walk us through what you're sort of, maybe from a high level, what you're thinking about for the next few years, how you handle that? Maybe layering in that, maybe that's just a reason to maybe look internationally more aggressively than in the U.S. How do you add salespeople in? Where do you add them in?
I'll go ahead and start, and then hand it over to Ralph or even Gary as well.
Can you just repeat the question? I'm sorry.
For the webcast, the question is, we've got five direct salespeople right now. Possibly adding additional salespeople would be beneficial. What is our plan in terms of adding salespeople as we go forward internationally or otherwise? I would say that, if we take a look back sort of historically, back to 2016, we really only had like three salespeople and one customer success rep and no one really to speak of in the marketing side of things. Fast-forward from 2016 through 2017 and 2018, we now have five direct salespeople, a senior VP ahead on top of those. We've got someone focused on the campus security side. We've got seven customer success reps, and we've added an international salesperson.
We went from, like, that four or five to about the 15 in terms of capabilities in the sales and marketing side. I think right now if we were to add and look at additional incremental headcount, it might be something that's more related to the maybe some of the HunchLab or the Missions aspect or some of the reseller coordination with Verizon aspect. That would be incremental on top of what we already have.
I think we're probably appropriately staffed right now, maybe more a question for Gary, in terms of geographics, with the sales reps and the areas that they cover, given the other support that we're building around them in terms of the marketing, the lead generation, the oversight, the customer success, and building up that sort of fifth column sales force of those 100+ chiefs of police that are out there selling without getting paid, without us paying their medical, dental, and pensions, et cetera.
Yeah. I'll just add one thing, just to give a shout-out. I think we had a, like a super critical hire in the sales organization that is just focused on renewals, right? That's been super impactful. That's why we're in fact seeing a little bit more smoothing out of our revenues because there is less kind of catch-up capability. We took a lot of work off the plates of our traditional sales directors just by bringing in to a central location someone that's focused 100% on maintenance, renewals, and that's really paid off. That was a smart, impactful hire. It's really more about how we, I think, make our current five sales directors more productive by taking things off their plate.
The marketing organization hires too are really key toward that in terms of doing the lead gen and also kind of managing those leads until they're really much more impactful prospects or potential prospects.
Very helpful. Thank you.
Yeah, great question. Ray. Oh, I'm sorry. Yes, sir. Mm-hmm.
Yeah, I know it's not an issue from a technology or scalability impact, but from an optics standpoint, would it ever make sense for you guys to open an additional IRC, whether it's on in the East Coast or internationally?
The question is, it might not make a real P&L business sense to have a separate IRC from a Incident Review Center from a capacity point of view, but from an optics point of view, could that potentially make sense? The short answer is that's something that we've kind of discussed, debated internally for a while. I think your instincts are right around it's not about really needing to add additional people to deal with capacity, 'cause we're doing so many interesting things on the technology side, where in fact, we have a lot of slack actually in the Incident Review Center. Optics-wise, I think that's an interesting point. I mean, it's to be determined, I guess.
Having an Incident Review Center in Washington, D.C., as an example, so, you know, legislators that are both involved in authorization and appropriation be able to roll down the hill and actually see ShotSpotter in action is kind of interesting. I think I would love to ask you the question, or collectively the question, after you go through the Incident Review Center and experience that, tell me, you know, what you guys think about that, right? I think our experience has been when people come and see it really kind of changes their understanding of what we do and gets them a lot more excited about it. Having that kind of capability closer to D.C., where we get a lot of funding, could potentially make sense.
I don't know that the cost would be all that great, actually, 'cause we don't have to pay a lot of money for a review center.
We don't.
Employees. Yeah. Yeah. Mm-hmm. Yeah. Yeah. Great questions. Yes, sir. Mm-hmm.
Yeah. The scaling up of large cities like Chicago and New York, given some of the smaller customers you've won recently in smaller cities, confidence to go out into bigger initial engagements than they would have in the past, or is it still more about landing in a specific area, seeing success, and then growing that over time?
Yeah, the question is there is a differentiation between trying to go for the big kahunas versus smaller cities. Do we have a preference in a way? Is that your question? Give us more confidence.
More so as those customers saying, "Okay, I've seen Chicago has scaled up significantly over time. This is a system that clearly works.
I mean, clearly Chicago's impactful 'cause so many people are looking at Chicago. They obviously looked at New York. I think that was our real coming out party, really, once we deployed in New York City, which is a really tough, demanding customer who really loves what we do, right. I think we really surprised them given our kind of relatively small size. They'll tell you, we're one of their favorite vendors 'cause we really stepped up. They expanded multiple times. That's also been the story of Chicago. People pay attention to that, both from a mayoral elected perspective as well as a professional policing perspective.
I gotta tell you, don't I mean, I know you're not suggesting this, we shouldn't discount the small cities, 'cause probably the most important thing from my point of view, and Gary, jump in here if you disagree or wanna add, the most important thing is a prepared mind, right? When you have a prepared mind of a chief that's, like, forward-leaning, that really wants to get at the problem, They're not technology-afraid. They're very sophisticated in terms of using technology to drive in gains, in outcomes with better community engagement, more higher levels of service to underserved communities. That's a game changer. When you get those folks, I mean, they could be from a place like Cincinnati that had a impact on Louisville, which, by the way, Cincinnati was influenced by Sacramento.
City of Sacramento influenced Sacramento County. Fresno, Jerry Dyer, one of our big promoters, guess what? Bakersfield, boom, like that. That's how these regional things Gary talked about, these regional aspects. You get a really prepared mind chief that wants to go get at it, we love those. It doesn't have to be a big, fancy city. I mean, we want those, you know. There's still Philadelphia out there that needs to come on board, Houston, Philadelphia, Los Angeles. They're gonna get on this platform eventually, I mean, 'cause they all have issues of gun violence in part, parts of their city. In the meantime, you know, we'll take the El Segundos, and we'll take the Monrovias and all that stuff. The Shreveport, Louisianas. I mean, those are miles.
I mean, you know, five miles in a small city versus five miles in a big city from economic point of view makes no difference to us. Yeah, we'll take them when they're ready.
We probably have time for maybe one more question then we go head over to the.
I don't want to go.
Any other Okay. Yes.
That NPS data was really interesting, and one of the things that it showed is that there's almost universal agreement around things like the efficacy of the service when it comes to You know, recovering shell casings for NIBIN or, you know, responding more quickly, right? It seems to be harder to prove that gunshots, gun violence actually is reduced. I'm just wondering, as you look at that data, does it cause you to maybe think a little bit about how you might position the service when you're trying to sell it? You know, say, "Hey, look, I'm not saying you can get in front of your city council a year from now and say gun violence has gone down, but I can absolutely say I'm picking up more shell casings, and I'm absolutely getting people to the hospital faster." I mean
That's a brilliant question, point. I mean, it's so great, I don't even know that I can repeat it for the folks dialing in. It was something about the nature of the NPS study and how we're asking specific questions about how much you strongly agree with these following value attributes. What you notice is that a lot of it had to do more with getting cops to dots, recovering shell casings, running through NIBIN, and not, maybe less so with respect to preventing homicides or reducing homicides.
Right. Exactly. Yeah.
Yeah, I think does that influence our thinking in terms of how we're marketing? I sure hope so, 'cause we've invested a lot in the process, and that's really valuable insight that's unique to us because we have this relationship with 90+ agencies that are kind of telling us how they think about it. That's a bit of a clue how the next 100 customers might think about it either. I think that's right. Sam, Gary, you guys are out there. Yeah, Gary.
I was just gonna add as well, not only on the marketing side, but when I referenced the role of customer success in helping these agencies establish metrics, KPIs.
Right.
We try to help them look beyond just our homicides or gun violence going down. There's so many other factors that are so key to evaluating the ongoing efficacy of what we're doing and what they're doing, in many cases, combined with other technologies. You're absolutely right in the conclusion you quickly pulled out of that data, and that's front and center in marketing, in the way we coach the customer, so to speak, with respect to the ongoing KPIs and metrics that they're tracking. Thank you. Sam, anything to add? Okay.
Okay. Thank you all. Appreciate it. We're gonna go ahead and head over to the do a demonstration then tour of the Incident Review Center. Rob? I think we're gonna try to break these up into two parts so we can, we're gonna try to get everybody in the review center at the same time, but No, we're gonna break it up into two. Into two. Okay, got it. If you all just head over with us, we'll break it up. One group will go in to the demo first, maybe 10 or 12 of you, and then we'll swap. If you all just head over to the demo room first.