Ep. #814 - Reinventing the Home Security Industry
In this Startup Hustle episode, we’ll learn about the latest developments in home security. Join David Selinger, Co-Founder & CEO of Deep Sentinel, and Matt Watson as they discuss how Deep Sentinel is reinventing the home security industry.
Covered In This Episode
Are the current home and business security systems enough? What does it take to reinvent the home security industry?
Deep Sentinel created home security systems with machine learning for smarter and more accurate crime prevention. David shares how Deep Sentinel is better than any current home and business security systems that rely only on security cameras.
David and Matt also talked about the sense of adventure and excitement in the early stages of building a company. And how the later stages become more exhausting.
Join their conversation by tuning into this Startup Hustle episode now!
- Hiring problems (1:01)
- David Selinger and Deep Sentinel’s brief history (3:50)
- How Deep Sentinel works (6:48)
- Dealing with barriers (10:32)
- 24/7 security (12:55)
- Employing outside the US – the Philippines (13:12)
- Why Deep Sentinel built its own camera (13:42)
- Building the camera (16:50)
- Phases of getting your product to the market (17:18)
- Deep Sentinel replacing your security guards (20:07)
- Improving machine learning for security (25:01)
- Raising capitals and VCs (27:39)
- Stopping crimes (28:46)
- The effects of the pandemic and shift to WFH (31:01)
- Deep Sentinel’s go-to-market strategy (32:18)
- Working with local installers (36:13)
- Business and Home insurance (37:45)
- Finding the right partners (38:38)
- The camera design (40:54)
- Early stages vs. the later stages (42:15)
- Batteries (44:18)
- Where to find Deep Sentinel (45:07)
- Wrapping up (45:55)
if you’re just building services on top of third-party hardware, while you may have a really interesting service, you have a really hard time telling people what it is you do, meaning that just human beings are so visual, and so narrative based in terms of their storytelling. I can’t show your listeners what the cameras are that we built. But let me say this, our cameras were designed to be our brand. Whether you buy our cameras or use your own cameras. When you think of Deep Sentinal as a customer, you think of our cameras.David Selinger
Anybody I’ve ever talked to that’s, that’s been in the hardware business. It’s been a total nightmare and a disaster to get the product to market. And you just highlighted the issue. It’s like we spent a year building prototypes and throwing them away. Like people need to really understand that if they want to go down this way.Matt Watson
We realize that people like you just said that the crime of offices went up. And what we found was that where those people go is typically to local installers. And so the number one thing that we changed was we changed from a mostly direct model to a mostly channel model. So most of our distribution now goes through installation.David Selinger
Stop losing sleep when it comes to payroll and other HR needs with Gusto on your side. Gusto is a modern all-in-one HR platform to help you manage your team as easily as 1-2-3. You can easily take care of payroll, onboarding, training, and tax filings with a few clicks. Get your free three months now.
In addition, find the solutions for your business through our Startup Hustle partners. Our partners make it their goal to help the startup community grow.
Following is an auto-generated text transcript of this episode. Apologies for any errors!
Matt Watson 0:00
Hi, and welcome back. This is Matt Watson, your host today for Startup Hustle. Today I am very excited to talk to Mr. David Selinger. And he is with a company called Deep Sentinel and is reinventing how the security systems work and excited to get all to talk all about that. Before we get started, I want to take a second to tell you today’s episode of Startup Hustle is sponsored by Gusto. Gusto has modern solutions for modern HR problems, which we all have too many of, whether it’s talent management, payroll, or onboarding tools, Gusto HR platform has it all for you. Be smarter than your competitors. Try it. Three free. Oh my god, I can’t even talk to you today. Try a three-month free subscription. Now just sign up at gusto.com/startuphustle to get started. That’s gusto.com/startup Hustle. Well, welcome to the show, man.
David Selinger 0:59
Glad to be here. Thanks for having me.
Matt Watson 1:01
You know, I guess going on the Gusto theme for a second. How was hiring for you? Is that? Is that a disaster?
David Selinger 1:07
Oh, man. So, you know, it’s been weird is probably the right answer. I mean, like, there’s kind of stages of growth, the companies that are normal and stages that are hard, and ones that are easier, the pandemics just been weird. And it’s an interesting combination of people that are really scared and like want to work hard and are willing to kind of really help with their elbows and the shoulders into the work. And then a bunch of people that are just ghosting interviews, we’ve had, you know, the typical story of somebody will, will sign up for an interview, show up and say, I’m just verifying that I’m here, and then hang up. And you can kind of figure out what that’s for. Yeah, it’s, it’s, it’s been ridiculous. And I mean, you hear that everywhere. But to actually have it happened time and time again, you know, in our hiring report, the ghosting, not showing up ratio. Wicked weighted IRA. Now, you know how to get talked about a little bit. I mean, we’re an interesting business in that we have a bunch of different types of employees. So we’ve got technology guys based out of are based in California. So we got most of our employees. Now through the pandemic, we’ve really gone nationwide and global worldwide. We have everything from software engineers to DevOps, which are technical positions, AI engineers, and typical kinds of back office finance. And then we also have a number of hourly positions, like our security guards. And it’s been different across the board. I would say, at a very high level, the number one thing that we’re seeing is, like, it’s weird. But the number one thing that I think is actually maybe good is I think the pandemic forced a lot of people to rethink what they want to do. Whether you’re in a job that you like, and you enjoy with, you like the people that you’re working with.
Matt Watson 3:04
There are more opportunities, right? Yeah, I mean, more opportunities out there. So like, you know, I don’t really like working here, and everybody’s hiring. So I might as well go do something.
David Selinger 3:14
Exactly. So I think, you know, though, I would say I think that’s good, right? Like a lot of people are afraid of employee churn. I would way rather have someone quit and find somebody who’s like, super stoked on their job than surviving and kind of, like, hanging on with somebody that doesn’t really like it. So I mean, it’s, it’s been interesting. We went through a rash. We have almost no employee churn at all in general. We went through a rash right at the beginning of the pandemic as people kind of rethought their priorities. And then we went back to normal again, where we’ve had very, very little employee churn in the last year.
Matt Watson 3:45
So so, tell us, you said you’re primarily based in California. How old is your company?
David Selinger 3:50
Yeah. So I’ll tell you a little about Deep Sentinel. We’re like you said, we’re a security company we started about four years ago. And our basic mantra is to stop crimes, actually, stop crimes in the backdrop of that, my background is in artificial intelligence. I got to go to Stanford. I did a bit before then. And then my last three gigs had been in AI. And so deep sentinels a little bit of a leap to the left of the rest of my career. I was at Amazon really early and ran an AI team there, the r&d Division of the personalization team. I started Redfin, which is an AI and data company in the real estate space. And then my last company was relevance, which is another AI company. So just ai ai ai, with deep Sentinel, we, our entire company is started differently. I wanted it to be taking advantage of AI in the most interesting possible way. But what we found was that there’s this humongous gap in the security space where all the solutions you might buy, whether they’re cameras or alarms, they don’t actually stop crimes. And that was the big to me like the big aha moment. My neighbor had a home invasion, she has an alarm, she has cameras. And yet this worst possible thing happen and none of them took any steps to prevent it. And that’s really the problem that we wanted to solve, we found that was an AI is really, really good about taking cameras, and transforming them from a recording device where you post on next door, here’s this dude breaking into my car and stealing my stuff into a device that actually stops driving.
Matt Watson 5:32
Well, and so I have my house when I bought it had home security cameras, and they were the old coax cable kind. Sure. And the problem was, the video quality on him was so awful that they were sort of useless anyways. And so I bought some wireless cameras, like probably a lot of people. And then the wireless cameras suck because the batteries die like every three months, or however, they’re super cool that they use some sort of machine learning AI magic to tell me, you know, that there’s a package at my front door, or, you know, can tell the difference between a car driving by or a deer, you know, or whatever, right, like different stuff that they can do, which is super cool. And it sounds like what you’re doing is trying to take that to the next level to go beyond recognizing deer to go to recognizing crime. Right. I mean, is that kind of the genesis
David Selinger 6:24
it’s even more than that. Lay it and that, Yes. And right. So, so yeah, I mean, what I observed about four years ago when I started the company was I built a state-of-the-art AI system on my cameras. They’re kind of similar to what you initially described. They’re wired cameras. They’re Poe, and they were state of the art. They’re aliens. It was 1000s and 1000s of dollars to have them put on my house.
Matt Watson 6:47
Now, we got 4k.
David Selinger 6:48
Yeah, exactly. And so the thing that I learned was that whether it’s 4k 10, ADP 720 P, or like 320, P, coaxial q1, they all do the same thing, which is just record stuff. And if you turn on the alerts, even if you put AI behind alerts, like you can’t respond, how am I gonna
Matt Watson 7:09
get bugs flying around?
David Selinger 7:14
Yeah. Even if you just filter it out to the things that might be suspicious, you’re still getting dozens of alerts a day. And if you have a life, which I hope all of your listeners do, if you have a job that demands 100% of your attention, which I would guess 100% of your podcast serve as like, how to do better at pouring my energy into crushing it with my business. You don’t have time to be sitting there looking at your phone, I gotta, I gotta give you 100% of my attention. While I’m on this podcast. I fly on a plane, I can’t be giving attention to my my phone. And those are all those missed opportunities where that’s how crime happens, right? Go on next door and you look on next door and you’re gonna see dozens and dozens of videos of while I was asleep. This person broke into my car while I was asleep. This person did this while I was away. Somebody did this. And they think the verbiage on there is my favorite. Right? It’s just so ironic. I caught this guy stealing with no you didn’t. You got a recording of a guy stealing your stuff. Congratulations, right? Like, do you wanna you want a cookie? I’ll give you a cookie.
Matt Watson 8:13
Yeah. You just, you just have evidence of it at this point.
David Selinger 8:17
And so what we do is we take the AI, we filter all that out, and then we use it to send that video within five seconds to a lifeguard fully annotated with what’s going on. So So by way of example, you come up to my house, and the second you step on my property, our cameras trigger and start sending the video to artificial intelligence that AI is looking for you to do anything suspicious. Now you?
Matt Watson 8:41
Yeah, how do you filter out the Amazon guy who comes by every day and the Instacart guy?
David Selinger 8:47
Well, so here’s what we do is if they’re bringing a package on we the AI triggers, and we’ll still elevate it to a person, we don’t alert immediately, we still believe that you’ve got to have an effective, well trained, contextual human being in the middle. And that person’s job is if it’s the Amazon guy, I’m going to watch him make sure that he leaves the property. If it’s the neighbor coming by to bring cookies, I’m going to watch make sure they knock on the door and somebody answers the door. And then they respond appropriately. If that neighbor that brings cookies all of a sudden pulls out a crowbar and starts banging on the door. Within three seconds, they’re going to hear this is deep self-security. I need you to stop what you’re doing. The police are on their way. If they just sit there at your door stop. Then after a minute of kind of loitering they’ll they’ll say Hey, this is deep sounds secure. Just want to make sure that you know you’re supposed to be here. And in all of those scenarios. What you have is you have an intervention before damage is done to the property. And what that means is 99% of the time if that is a criminal they pulled out their crowbar, bam, they’re gone. Because you know what’s the neatest thing about this too, is when our guards tell someone the police are on their way. That’s not a bluff. When we When we say that we call the police instantly. And instead of let’s say you have ADT, and you get a, you get an alarm, okay, hey, I’ve got a motion in the living room, or I’ve got a garage door, the guy probably just left his garage door open. Can you send a cop? What cop wants to respond to that? None of it? Is there. 99% false alarms. Yeah, we call police, we get an instant response. I’ve got a white male and a hoodie with a crowbar banging on the front door, I need an instant response. So every cop across the nation responds to that.
Matt Watson 10:32
So when you went to start this business was an in it’s very clear the benefits that it can provide, right? But what were the daunting things you’re like, Man, I have to like, also invent, like, my own video cameras, or like, what were those barriers? I think, I mean, I can apply machine learning. And that part of it makes sense. And I can provide, you know, remote surveillance and whatever. But wasn’t the like, developing having to deal with the hardware side of this, like a daunting thing that
David Selinger 11:01
you did everything about this business was daunting. I mean, yeah, this isn’t my first business I started. It’s just by far the hardest, right? I hope to think of this one as my swan song, right? Like, the culmination of all of my past successes, and, more importantly, the culmination of all my past failures, like bringing all that knowledge to bear, I would say there are two things that were really hard about this business. One was what you said about hardware, I’d never build hardware. I did research on robotics in college, and build robots. And so that was kind of my experience with hardware. I’d never done mass production, you know, design my own printed circuit board, and assemble it, and then manage that through quality and delivery. So that was definitely a new experience for me. The second thing, though, which I think is just as rewarding, and by the way, we like, we built our own cameras, and they’re sick, right? Like they’ve got, they’ve got huge batteries, they have solar power, so you never have to change the batteries. We have a built-in charger and all this stuff, right? So that was a great and rewarding experience. The part that actually has been the hardest was gluing it all together. Once we had the cameras, and we had the AI, and we had guards sitting at a console, it’s figuring out, how do you make sure guards do exactly what you just said? They respond right to the Amazon guy. Okay, now Amazon’s using these white vans and people without uniforms to deliver, which happened. Yeah. How do you deal with that? How do you use AI to make sure that not only are they doing certain to not only say that we know what we’re supposed to do, but we use AI and data to tell that our guards are doing the right things and oversee that and ensure that we’re delivering a great experience for our customers? And that really has been the hardest thing of building this from an operational perspective to scale. So I’ve got about 55, 60 employees right now. And 40 of them are guards, and to make sure that they have what do you want to.
Matt Watson 12:55
You must have employee all around the world right to do this? 24/7?
David Selinger 12:59
Yep, we do. We’ve got a corpus of them here in California. We’ve got a bunch of them all over the US. And then we’ve got two stations outside of the US to make sure we’re covering 24/7.
Matt Watson 13:12
Where are your employees outside the US?
David Selinger 13:15
Most of them are in the Philippines.
Matt Watson 13:16
Okay, so we can relate. We can relate at Full Scale. We’ve got over 200 employees in the Philippines.
David Selinger 13:21
So yeah, I mean, I, I found the culture there to be fantastic. They’re really good at this job. They’re very attentive. It’s a great light, you know, you talked about hiring right at the beginning, labor pool, there’s wonderful, you know, variable provide a high quality of life to these people. And they’re, they’re able to work really hard for us and the night shift, which is pretty tough.
Matt Watson 13:42
Yeah, absolutely, we can definitely relate it at Full Scale and that side of it. So, to me, I guess I’m still wondering, in my head, why did you have to create your own video camera is Wasn’t there something like off the shelf? You could use it for that? Because that seems like
David Selinger 13:57
That is such a good question. And you know, looking at the business where right now, over 80% of our sales are for third-party cameras. So we have a whole camera integration layer where we support third-party hardware that’s not manufactured by us. But we found two really key things. Number one, if you’re just building services on top of third-party hardware, while you may have a really interesting service, you have a really hard time telling people what it is you do, meaning that just human beings are so visual, and so narrative based in terms of their storytelling. I can’t show your listeners what the cameras are that we built. But let me say this, our cameras were designed to be our brand. Whether you buy our cameras or use your own cameras. When you think of Deep Sentinal as a customer, you think of our cameras. Our cameras are they’re uniquely designed they’re more aggressive looking than like if you bought ones from nest or ring out up and running
Matt Watson 14:59
Yeah, looking that online here. I see flashing red lights and a dude running away with a crowbar.
David Selinger 15:05
That’s right, it’s got two functions, it has a camera. And actually, the speaker is separate. It’s the largest speaker of any camera that you can buy out there. It’s got red LED lights that spin around that tell you somebody’s watching. It’s got a huge battery in it because it’s designed to always be on it has built-in solar panels if you want them, and, and that all those things that I just said, are indicators of trust, we take your security seriously, if you hang up a cute looking white camera that’s designed to disappear in the background, how much do you really trust it? And it turns out, like, that matters, right? And so I love having our own cameras, because it allows us to tell our story in a way that’s fundamentally different than anybody else. And, and it’s you in all of your startup listeners know, your story is almost as important as the way you tell it.
Matt Watson 16:00
Well, for sure these cameras look different and scary as I’m looking at your website here. They look pretty cool. But I still wonder like there wasn’t a middle ground where you could have taken like a off the shelf, like some of the hardware and then wrapped it in your own like stuff like did were you able to reuse some of that like, or do you have to start from scratch on like, how to build a video camera 101?
David Selinger 16:22
No, no, we did a little bit of that. Right. So the chips on the inside of this are the same ones as the ones that you see in like an Arlo camera. Okay, electronics design is pretty similar to that. Yeah, so we didn’t go all the way down. But we did go down to the kind of best-of-breed components that you would look at in other cameras. And but still, even doing that is pretty ridiculously well, managed supply chain and everything like that.
Matt Watson 16:50
And that that that’s got me I want to ask the next question is like, how many iterations of this hardware design did you have to go through where like, you order 1000 of them, and then you get them and you’re like, oh, shit, this didn’t work at all. And you’re like, throw them all away, like did you have to,
David Selinger 17:05
In three phases. The first phase is the one you just described, where you’re like ordering stuff, testing it, and then throwing it away. And that lasted for probably a year when we didn’t have it.
Matt Watson 17:18
And this is a really key point to listeners out there that are thinking about this kind of business. And that’s why I bring it up. Because anybody I’ve ever talked to that’s, that’s been in the hardware business. It’s been a total nightmare and a disaster to get the product to market. And you just highlighted the issue. It’s like we spent a year building prototypes and throwing them away. Like people need to really understand that if they want to go down this way.
David Selinger 17:38
Especially for a software guy, understanding that’s very different than anything you’ve ever done.
Matt Watson 17:45
Oh, yeah, just fix the bug and ship a new version?
David Selinger 17:48
That’s right. Yeah, I mean, there’s, there’s a little bit you can do with existing hardware. But when there’s a bug you got to deal with, you got to learn as much as you can with what you got. And then you got to do a turn. And it if you think about software, usually takes five or six turns to get that done. Well, there you go, there are six months right there. And, and that’s in kind of the best of times. So then that’s the first phase, then you get to something that’s working, it’s sellable, I can manufacture it, and I can manufacture it and guarantee the quality. And by the way, those last few things are what you spend the whole last half of that year. So the first half of the year is getting something that really worked and function. And the last half of the year was making sure that we build something that when we manufactured it on a manufacturing line, it would be consistent, we could test it on the manufacturing, we do a good job of it. Then the second phase is when you’re in production, and for the first six months, like we had a cable that one out of 100 of them to cable melted under just under power. And then, it would cause the camera to stop functioning. But it took a long time to find that because you had to use it for six months. And yet, it only happened very, very, very infrequently. And that second phase you have, you know, hopefully, your goal is to have like 5% returns, and then move on with life. And then you do like pretty significant iterations of your hardware. And then the third phase is I’m locked in the hardware. I haven’t changed it. I haven’t changed our hardware for a year and a half. And I’m just updating the software. And we launch software updates all the time. And so it what’s neat about it is that once you get that platform, what you can do in software is typically made you just got to make sure that you’ve thought about the right things and enabled those things in the hardware. So then your software updates can be super powerful. And think about it like the Tesla. They launched self-driving capabilities after they launched the car because they knew what they wanted. But then they had to come up with a new version of the car in order to support that absolutely. of self-driving. And it’s an iterative process. You know, I mean, it’s, it is a lot of work. It definitely was harder. If I had to do it over again, I would say absolutely do it because our brand is so incredibly strong. If you want to solve this problem, if you own a small business that has assets, you want to solve this problem. And we are the only company to go to to do that. And we own that space.
Matt Watson 20:07
So speaking of that, if you’ve got HR problems, managing your team can be as easy as 123 with Gusto, no more late nights for processing payroll, or dealing with business tax filings, no more painful spreadsheets for attendance, tracking, say hello to our new smart HR platform, check out gusto.com/startuphustle to get a free three-months subscription now that’s gusto.com/startuphustle. And so I was looking on your website, and it looks like you also have a service for businesses. Right? And so in some, in some way, do you guys replace like paying like a real physical, you know, security guy that would just like walk around all night and do nothing?
David Selinger 20:52
Matt Watson 20:52
I mean, you put him out of business.
David Selinger 20:54
There’s my sales pitch, right? You know, I mean, the yes, that is exactly what we do. If you look at the market, you can spend 10 bucks a month or 20 bucks a month, and yet cameras, maybe 40 bucks a month, and yet ADT and then your next upgrade, if you’re still having crime problems, which by the way, none of those things deal with auto burglary, none of those things deal with package theft, none of those things actually prevent most actual break-ins, none of them protect your fleet of cars, your next step up is a guard. So you’re going from 50 bucks a month, and then your next step up is 10,000 bucks a month, maybe 20 If you want to do 24/7. And so we effectively own that entire price band for anybody that is having security issues. And you know, if you’re paying attention, this last two years, especially for businesses that have hacked, you got an electrician fleet, and you got copper, I mean that there you go, you own an RV sales site, you own a truck fleet, and that it has catalytic converters, you have to be protecting these assets. And so that’s, that’s what we do. And we, you know, you mentioned the guard not doing anything, one of the things you have to also learn is not only are we cheaper, and we’re able to like do that, effectively, we’re way better than a guard. Because if you have a job, let’s just put you in the seat of this guard, hey, I want you to sit at this chair for four hours. And look at this screen where nothing’s happening.
Matt Watson 22:24
And instead play on your phone and watch Tic Tok videos.
David Selinger 22:27
Amen. Right. Like, that’s why tick tock is doing so well. Because of all these security guards out there. They suck. They’re expensive, and they suck. And so I mean, the number one thing that you see all the time, as you see you actually one of the inspirations for starting Deep Sentinel was, I started researching this space and talking to like Mark Zuckerberg, security staff. What do they do between 10pm and 4am, jumping jacks, man, you’re paying these guys a full-time job. There may be former secret, sort of special forces, not secret service. Most of them are former special forces or military, and you’re paying them to do jumping jacks for eight hours a day to do nothing. It doesn’t matter if you’re the best-trained person in the world. You’re gonna start screwing up because your job sucks. And one of the needs is boring. It’s horrible. So then, go ahead.
Matt Watson 23:18
So how do you, how does your team manage that where I have to feel like 99% of their job is watching Amazon delivery guys walk up the front doors?
David Selinger 23:27
Yeah, about 95% of it is. But there’s two things that we do. Number one, is we make sure to give them the information, hey, this is a low priority, high priority event, here’s what I want you to look at. And then we also measure the pace at which we feed videos to our dogs. So unlike you’re used to kind of seeing a camera system where you go to your hotel lobby or your, you know, whatever commercial property lobby and you see the person that just has a CCTV, and there’s a bunch of stuff not happening, we flip that over, where instead of the guard being in control of what they look at, we actually feed videos to the guard, and we require them to respond. And so we’re able to do two things with that. Number one, we measure the pacing of what we send to the guard. Number two, which is actually the most important we then measure the guards response. So we look at did this guard do this for five seconds or 25 seconds and they do it for five minutes did they call the police and then three we close the loop we optimize those feeds based on what produces productivity from the guards. In fact, we even use that for training machine learning models and what what is actually working and what’s not working. You know you can you can have AI that like you said on your cameras identifies a package. What we are measuring is not just Is this a package we’re measuring, does a package plus a gun plus a person driving by result in a guard action? What’s the action? What’s the outcome?
Matt Watson 24:58
And are you guys working?
David Selinger 25:00
Isn’t that the outcome?
Matt Watson 25:01
I would imagine, then you’re working nonstop to figure out how to improve the machine learning and computer vision to identify like, that’s, that’s the weekly UPS guy like, is that something you guys are continuing to work on? So it’s like you kind of like filtering out that kind of.
David Selinger 25:15
That’s exactly right. Like, again, let me use Tesla as an example, we’ve been able to look at a camera and say, is that a car? And is it going? And is this a person? What Tesla focuses on with their autopilot? Is not? Is this a person they’re focusing on? Did my driver hit the brakes? Did they veer the car slightly to the left of the rear of the car to the right, did they speed up? Did they slow down, all of those things are actually what you care about. And with the most recent advances in machine learning, it’s, it’s called, what we do is called self supervised deep learning. And with with those types of data inputs, you can start predicting what the what the actual physical outcome is, instead of just what are these things and draw a bunch of boxes on a screen? Well, because again, boxes on the screen are getting a bunch of alerts, that package alerts, it’s just as noisy and just as meaningless. For the purposes of actually stopping crime.
Matt Watson 26:11
Well, to re-, probably rephrase what you just said in a different way is, you’re basically your guards, look at these videos for five or 10 seconds or whatever. And then they basically say it’s an issue or not an issue, right? And then you basically have the history of all of that over 1000 times that then you can feed into some kind of machine learning algorithm that says, here’s all the things that we’ve watched that we said were not a problem, and here’s how we reacted to them, and go and then try and break the signal from the noise. Right?
David Selinger 26:39
That’s exactly what we are basically taking. One of the problems in machine learning it all express technically, is that if you just have a video, and you didn’t label it with this is a person, this is a car, it’s actually not that useful for making an AI better, we have this really unique closed-loop data set that includes the actions of the guards. And so we’re actually training the AI based on the actions of guards, not just based on as a tree or not just That’s right, that’s one of this really unique kind of breakthroughs. From a technology perspective, that means we took what is considered the state of the art a solved problem. And then we opened up an entirely new realm of problem that isn’t solved. The space of analyzing the response to videos is completely open. It’s bleeding-edge technology. Tesla’s at the forefront of that as it relates to cars. And we’re at the forefront as it relates to security.
Matt Watson 27:39
So to do what you guys are doing, did you have to raise a lot of capital and VC capital and stuff to get this off the ground?
David Selinger 27:48
I did, you know, I raised about a little over $20 million to get this going. And, you know, I’ll be frank, at the beginning, that was easier. And then once the pandemic hit, the pandemic was not good to our business. So it’s been tough for the last two years. And really, just in the last six months since October of last year, we broke out. And, and that’s been awesome to see I mean, a lot of hard work. As you can tell, we’re a super mission-driven organization like we. We believe strongly that stopping crimes can change security. And you know, it’s just funny to say that because you’d be blown away how other security companies don’t think that so what we’re seeing really should be table stakes. But ADT doesn’t actually stop that many crimes on it. We probably stop as many crimes in our as a small company as ADT does across their entire business. Just because that’s what we do, we actually stop crimes.
Matt Watson 28:46
Is that something that you guys can actually measure to say, this is how many crimes that happened. And this is how many, like you actually know that?
David Selinger 28:53
Absolutely. Every morning I wake up to a list of here are all the crimes that we stopped. Here’s where we intervene. Here’s where this trespasser stopped. The one thing that we do have that a little bit of a bummer, which is it’s ironic and funny, is we stopped 99% plus of our interventions, and so are a lot of our videos. If you go to our YouTube channel, we have a YouTube channel and deep set on YouTube. We take the ones that are more dramatic. We’re like you see the person pull out the crowbar. But honestly, most of our videos are like person is hanging out just like just loitering and we say Hey, this is deep central security, and then they walk away instantly. And so we stopped the crime so early in the cycle that most of our interactions are really not that dramatic.
Matt Watson 29:35
You detain people hanging around being being creepy.
David Selinger 29:40
I mean, if you look at the video again, go on your next door and look at this. You look at the people that actually get broken into a lot of the times the burglars are outside their house for three 510 15 minutes getting ready, because nobody’s watching the videos. In fact, let me let me actually even take that a step further. If you have A video doorbell, and somebody is about to break into your house. You know what they do? Well, a couple of years ago, they used to avoid it or maybe put tape over it. What they do now is they ring it, they make sure they triggered as many times as possible. Why would they do that? That doorbell is giving them information about you. It’s supposed to be your way of seeing your front door. It also tells them, hey, nobody’s home. Nobody’s paying attention to their camera. We got the house for at least 20 minutes. We’re good. And so the act of intervening at that moment, it doesn’t take a lot. It just takes them knowing that there’s a human being Washington has this entire psychological shift. Even when we have people that are fully prepared to break in. There’s two or three of them. They’ve got masks. They’ve got hoods on. They’ve got crowbars, they’ve got bolt cutters, they’ve got a battering ram, some sort of like jerry-rigged battering ram if they know someone’s watching, and the police are on their way, crimes done, no damage.
Matt Watson 31:01
They’re looking for the low-hanging fruit. Yeah. So the effect of the pandemic on your business is definitely interesting to think about. Because I could see home security maybe being a little less important. If everybody’s working from home, or like I’m home all day, I’m not as worried about security anymore. But I would have thought that Office security would have been like through the roof, like there’s nobody in any of these office buildings. What are we doing for security? So, so, is that what you saw?
David Selinger 31:29
Absolutely. We only launched our business line of business. I think it was in March 2020. So it was right at the beginning of the pandemic. And so we had to launch a new product in the midst of a pandemic. And what was particularly what was hard was, was changing our brand, changing our mode of operation, changing all of our operations. While going to the pandemic, you’re right, though, when I look at the inflection point we hit six months ago is because we launched that product, we got enough momentum behind it throughout 2020, then we found out how to sell it, how to market it, how to price it in the early part of 2021. And then by Q3 2021. We figured all those things out and the market was literally just, you know, elbow curve, in terms of every one of the metrics and
Matt Watson 32:18
I’m intrigued, you know, what is your guys’s go to market strategy? I honestly never heard of your product before, right. Like I don’t see it at Best Buy or Target. You know, what is your go to market strategy? How to how do you reach your your audience?
David Selinger 32:30
Yeah. So we have three key thrusts to our go-to market strategy. Number one is we realize that people like you just said that the crime of offices went up. And what we found was that where those people go is typically to local installers. And so the number one thing that we changed was we changed from a mostly direct model to a mostly channel model. So most of our distribution now goes through installation. The second thing we learned was as we launched our business line, residents wanted installers maybe 20% of them wanted an installer, 30% on it, pretty much 100% of the businesses wanted installer to just didn’t want to do it themselves. If they’re hanging up, you know Arlo cameras or whatever, then they’ll do that themselves. But they’re not really trying to solve a security problems. They’re doing that the people that want to solve a security problem one and professional and they want to professional on site. So we had a launch an entire network, we now have almost 380 installers all around the US almost every major metro areas we ever know.
Matt Watson 33:29
I just talked to one the other day, like literally yesterday.
David Selinger 33:33
There you go.
Matt Watson 33:34
I had somebody at my house that provides our normal home security. And I’m actually looking for a video system like this because I have an Arlo system, the batteries are all dead. So all the thieves that want to come by, they’re all dead. They’ve been dead for months, probably like most people who have wireless systems, it’s a common problem, right? And I’ve actually been shopping for this and I might be your next customer, maybe even for
David Selinger 33:56
People once you want to be in play around with our lives all you want. Go ahead, play with Arlo’s because all they are is a toy, right? You want to watch your baby and have a baby cam suite. You want to look at your kids in the driveway suite, you want to post up on Facebook suite, you want to solve a security problem. You insert as soon as you ask yourself the question, is this actually going to make me safer? You know the answer, you know, and I was actually kind of embarrassed the way that I learned this was again my neighborhood this home invasion and we had this neighborhood watch meeting and asked the police officer the he’s now, he’s now more friendly to me now that I help him with his job. But I asked him like hey man, she has everything. How come this still happened to her? And he said to me what did you expect the cameras to do jump out and stop the crime? They did what they’re supposed to do we have a recording. But now you just have a picture of six guys with masks
Matt Watson 34:52
Recording of the guy in the mask who smashed the window. What do we do now?
David Selinger 34:56
Congratulations, right. And you know you you I developed a relationship as I started, like really digging into this. And he started showing me like, look at this, this is my drawer of thumb drives of all these videos of people that we don’t know, we don’t have a facial recognition database. We can’t do CSI Miami on this, like there’s, there’s nothing we can do with most of these things unless we happen to know the perp. The second we see the video. And so I started to realize the only way to solve this was to go to prevention, you have to move into a prevention mode in order to solve this. But But again, it did to complete your answer about, you know, selling into businesses. Once we added the installer, we added the business and we marketed it, we got it right, we figured out how to do it incredibly well. So now we stop all these crimes every day, then we had to start getting the word out. And that’s what we’ve been doing with content. So we have a ton of content, if you search for preventing crimes with cameras, and that kind of part of the world, which we think is the future. We’re owning that world from an organic perspective, we own that narrative. And we needed to make sure that we had a good foothold on what it is we do how we get it to you. And then let’s tell the story. And that’s what we’re doing now.
Matt Watson 36:13
Yeah, and I can see working with local installation partners all over the country would be at an excellent. Oh, man, they love to go to market for you. It’s just my experience, like in the past has also been very difficult, like, how do you get them to sell your product versus somebody else’s product? And how do you train them. And just all of that is got to be a nightmare. And you must have a handful of people that all day long just have to deal with trying to turn so.
David Selinger 36:26
That’s what we had to do this? Why didn’t it just take off right off the bat. And that’s why we had to wait to build that. And by the way, building that in 2020 was almost impossible, because they weren’t going on site. Like they didn’t have the demands to go on site. Because there was this weird kind of up and down of the pandemic. Yeah, 2021 was where that really turned around, businesses started to open up, people started really kind of being open to having contractors on site. And so that’s, that did take some time, we do have to get the training down. But here’s the key thing, though, for that our close rate is wicked high from proposal to close is like 80%. And so again, right? Like they’re putting a proposal in front of someone who says, hey, I want to stop crimes, hey, I’m Matt, I want to protect my business, I want to actually have something that stops, I got three bids, this bid for just cameras, and that’s 5000 bucks, this bid for just cameras, and that’s 5000 bucks, or this bid from Deep Sentinel, and it’s 5000 bucks up front and then 300 bucks a month. But it actually stops the crime. Are you able to do one, anybody who’s serious about stopping crimes is gonna get like, it’s just our people.
Matt Watson 37:45
Is there any angle here where you’re able, like people are able to get discounts on like home insurance or things like that? Because they use this type of system?
David Selinger 37:51
Yeah, so a lot of people both have business and home insurance, depending on their policy? Do you have a monitored system? Do you have live 24/7 guards, there’s industries like cannabis, where they have to have a live 24/7 guard, where not only is it discounted to insurance, it’s actually part of their accreditation, like they need it in order to be certified. And so they were filling a gap where again, like, let’s like be cannabis, their alternative is 20,000 bucks a month. And so we’re able to provide that for significantly less at higher quality. So yeah, I mean, it interfaces into that industry. For the average homeowner, though, they get a discount, that’s pretty similar to having an alarm. But again, effective.
Matt Watson 38:38
Yeah, it’s interesting. It’s just interesting to think about all those go to market strategies and partnerships. And, and, you know, it’s just for our listeners to think about right, like, you know, potentially you partner with an insurance company, all of a sudden is, you know, recommending your product, everybody, and those are all the weird, you know, channel partnerships, you have to think about over time, and some of them work. And some of them don’t. And
David Selinger 38:58
It’s, in fact, we have an insurance company, one of the top three in the nation is one of our investors. But you also have to remember, like big companies are pretty risk averse, right? So how do you how do you navigate that we’re a small company, if they recommend us is that cause problems? Data? And so we’ve got to, we’ve got to make sure that we manage both sides of that equation, like how do you find the right size partner for my business? I also know companies right, there’s a there’s a great story, I think it’s a pickle company in 1990, that got the best distribution partner for a pickle company at Walmart. And Walmart squeezed their prices so that they were one penny unprofitable, and every jar they sold, then they went out of business three months later. Wow. And so you got to pick your partners and you know, distribution is not the only name of the game. Right. And you know, I think Walmart learned a good lesson from that that made it onto the front page of the Wall Street Journal. And so what was Walmart’s procurement department is now required to make sure that their vendors are at least profitable on the This is where they have a profitable business plan through partnering with them. Because, you know, you get these big distribution deals, you got to be able to service, you got to make sure that you do them well, and it fits into your overall business.
Matt Watson 40:11
I mean, that’s the key to businesses, it’s got to be a win win win for everybody, right? Otherwise, it just, it doesn’t work. So well. Once again, today’s episode of Startup Hustle was sponsored by Gusto. If you’re looking for an all in one HR platform, it’s time to check out Gusto. You have everything you need, and just a few clicks of a button, you’ll get you’ll even get three-months free. I don’t know why I can’t say three-months free. It’s almost like it’s the first time you set out when you go to gusto.com/startuphustle. That’s gusto.com/startuphustle. Well, I really enjoyed the conversation today. And one thing I’ve been thinking this whole time, I’m like, man, maybe I need to buy one of these things. And honestly, the cameras are super cool looking, they almost look like they would shoot you with a laser
David Selinger 40:54
That was exits. So if I can, I’m gonna, before we wrap up, I have to tell a little story about the design. So most design firms, you hire a design firm, and they’re like, we’re gonna make your product look like apple. And I was like, I don’t want to really only apple, apple looks, you know, comfortable. I want this product to make you feel uncomfortable. I want you to feel like a Gatling gun. I wanted to I wanted to be able to be both safe when my wife and I come home, and then intimidating if the wrong persons there. And and so I went through this massive design cycle. In fact, we won a bunch of awards for the design for kind of standing out in that way. But it was hard to get people to do it. And so I literally went and I found designers who had worked on video games to design future esque security cameras, because in video games, you do have that they have definitely video cameras that do shoot you. Yeah, like sweet, I hired a video game artist out of the Ukraine actually, to design some examples. And then I brought those to a top tier mechanical design firm and said, I want us to get somewhere between Apple and here. It can’t be all aggression, I’m gonna kill you. But it can’t be just soft and friendly. And I’m gonna put butter on it and eat it for breakfast like that. It has to be somewhere in the middle. And that was, that was a really fun experience to get to design that that kind of soup to nuts.
Matt Watson 42:15
It’s always fun to build new things, man that that’s the the fair, the best part, the hard part is the part you’re in now or it’s the operational side of this and growing and scaling it and the the brain damage that comes along with that, right like having like, we have 50 guards that look at TVs all day, right? Like, is that part of it?
David Selinger 42:36
Is that well enough? Before he gets to scale?
Matt Watson 42:39
Yeah, it’s a whole different, whole different problem than the initial stages where you’re trying to build the thing. And honestly, there are guys like me that love and enjoy, like the early stage of it, but dread every second of the later stage. I’d rather go build something new, like I want to build cool new things, I don’t want to like, manage 50 employees that are doing that, that’s a whole different problem.
David Selinger 43:03
It is it is and it requires a completely different skill set. Absolutely. And sometimes you got to change some of your leadership, you got to be willing to change the way that you look at the business. I’m fortunate and I really enjoy both phases. And with this business, you know, we didn’t talk about the mission too much. But, you know, our our company’s mission is to make people safer. And the fact that that has just such a huge impact, you know, I get an email every week from a customer. I slept soundly for the first time last night, I cannot believe that this is a real product. I can’t believe I didn’t have this a year ago. I can’t ever imagine my life without this. And that to me, whether I’m in the positive phases of building products, and the exciting phase of that, or I’m in the doldrums of the first that first year, the pandemic. That’s what drives me. So the one thing that I would note about this company, I think it’s really, really been neat for me is that I chose really carefully I spent a year and a half choosing this business before I started it. And I wanted to make sure that when it got really demotivating when all those external factors went away, that what we were doing was meaningful enough that it would overcome that. And that’s that’s probably the single best decision I’ve made when I started this company.
Matt Watson 44:18
Well, it’s very cool. And I love what you’re doing and the mission of it. And I am seriously looking to buy a camera system right now. So So tell me how long does the batteries last for these things.
David Selinger 44:28
So on the screen, we’ve got we’ve got two systems, we’ve got our home, which is our wireless, those batteries last about three months. You can also add solar to them and then they basically last forever. And then we have our commercial offering, which is our Power over Ethernet and that integrates with a bunch of different power over Ethernet cameras out there. And those obviously lasts forever. Yep. Okay, very good. The price difference is significant, right? You’re going kind of from a consumer product to a commercial product. So it’s it’s a pretty big jump up in terms of the upfront costs.
Matt Watson 44:57
Well, very cool. I love what you’re doing and I’m sure you’ll get a few listeners that go check out your website. And we want to tell everybody where they can find more information about you.
David Selinger 45:07
Yeah, sure. So thanks for asking that. I would say, instead of directing us to them to our website, I would say the most interesting places that YouTube channel I mentioned a little bit ago, go to YouTube, search for Deep Sentinel, s-e-n-t-i-n-e-l, and we publish a video every single week of the crimes that we stopped. And that’s my, that’s my pride and joy is that video, we take the best crimes that we stopped every week, we put them together in a little Compendium, and we publish it and it’s called our stopped line. And it’s just hard hitting it’s crime after crime after crime of just stop, get off the property, stop that off the property stop, get off the property. That’s the number one way and then the second way, you know, if you’re interested in kind of following our company and my journey, I do publish stuff on LinkedIn. And so direct people to my LinkedIn and you can follow me on LinkedIn.
Matt Watson 45:55
Awesome. All right. Well, thank you so much. appreciate having you on today.
David Selinger 45:59
Matt, it was awesome. Thanks for all the great questions. All right. Thank you.