All right, thanks everybody for joining. We're going to go ahead and get started. I'm Adam Tyndall, and this is part of my connected devices coverage here at Raymond James. Very happy to have Matt McRae, CEO from Arlo Technologies. Many of us are probably familiar with the brands. Hopefully, everybody here has the cameras in their house. I know I do. So we'd love to keep it as interactive as possible in terms of format. I do have just a list of fireside chat. We don't have any slides or anything like that. But if you do have questions along the way, please feel free to raise your hands and ask those. So Matt, thanks again for being here. Maybe we'll start just high level for those not as familiar: the background of the company, market opportunity, and key value proposition.
Yeah, so we're in the consumer security space. In fact, we've really, if you go back far enough, we're a spin from NETGEAR. So if you're familiar with NETGEAR, it's a home networking company.
They'll be here later today.
They're here? Oh, good. Good. We'll say hi to them.
CJ and Bryan.
Our brother. They were really, if you go way back, they were looking for something that was going to utilize the home network a lot more and get people to trigger an upgrade cycle for home networks, and they hit upon the idea of video and driving video utilization in the consumer space, and actually made a couple of acquisitions, developed a team, and created the first wireless and what we call wire-free security cameras, so they're battery-operated mostly, we have powered as well and Wi-Fi-operated, and you can put them anywhere you want. You can screw them in a tree and face your house. They take five minutes to set up, and actually created a market segment in the DIY consumer security space, so before Arlo launched, nearly 100% of the market was dealer-installed. They roll a truck, drill holes in your house, run cables to install security.
Arlo pioneered the DIY market segment. When we were part of NETGEAR, we were really a hardware company. And looking at making 30-40 points margin on the hardware and trying to sell as many widgets as we could. When we spun, we set about changing the business model completely, totally flipping it upside down into what we call a services-first business model. That was around 2018, middle of 2018. The spin finished in 2019. At that time, we had about 50,000 subscribers on a very rudimentary AI service. So we were one of the first companies to actually charge consumers for AI. And that was in 2018. And again, like I said, most of our business, most of our gross margin was coming from hardware. We had about 50,000 subscribers on just the beginning of an AI services. Fast forward to today, we have 4.2 million subscribers now.
We're fast approaching about $250 million in ARR. I think we guided to the year to finish around $240 million in service revenue at nearly 80% gross margin. So last quarter was about 77% gross margin on that service. If you look at the subscription revenue, which is a bulk of that service revenue, the margin's closer to 90%. So it's about 88%, 89% on the service margin line. So we've taken a hardware company that created a market segment for consumer security and turned it into one of the fastest growing and profitable consumer services in the world. The service, just to give you an idea, you have these cameras that are creating the visual-based security. The service does things like video storage, obviously, but also uses AI for object detection.
So we can see if there's a person or a vehicle or a package, for instance, at your front door and give you a different alert and filter those. And most recently, we've launched Arlo Secure 5 that I'm sure we'll talk a little bit more about. That includes a recognition engine and the ability for users to actually create their own AI models so that the cameras can actually detect and notify on very specific things in the home. Then the service goes all the way up through professional monitoring. So we've got multiple call centers that we partner with for actual emergency dispatch, fire, medical, and safety from a police perspective. So again, if you look at the market, I mentioned before pre-Arlo launching and actually creating this market segment, it was almost 100% install, dealer install market. Now it's roughly 60%-70% DIY.
So we've shifted the market two-thirds to where it's DIY and people are actually doing this themselves instead of signing up for a dealer.
Perfect. And this is probably more of a Kurt question, so you'll have to do your best to Kurt impression. Maybe just double-click on the business model and operational transition that carved a path to profitability and talk about the importance of household formation, opportunity to build on services, and how the teams use that product to grow subscription.
Yeah. So as part of the transformation from a hardware company to a services company, we've gone through a couple of inflection points in the business model. When we hit about one or two million, it was about one and a half million paid accounts, we started seeing obviously a transformation in the P&L. So we have started obviously generating profit. We've started generating cash. That's great. It's building up our balance sheet. Our switch in cash flow has been dramatic, especially over the last 12-18 months. And what we're focused on from a model perspective is driving unit volume through our channels. We can talk a little bit about those channels later, but we have what you would consider normal consumer direct channels like Best Buy, Walmart, retailers, plus e-com, Amazon.com, our own Arlo.com, and channels like that.
Plus we have what we call strategic partners, where it's a little bit B2B to C. So we have partners that we deploy to, and then they actually do the sales and marketing and actually position that out to consumers in the marketplace. So those are our channels. Unit volume is what we look for that drives household formation, that then drives service attach rates. Again, give you a couple of metrics. Our service attach rate is roughly 50%-60% on services. So again, we used to see a 1%-2%, maybe 3% attach rate on services. With this transformation, now we're seeing over half subscribe to the service and drive the service revenue to the business. Our long-term value of a customer is roughly $700. So if you look at LTV, it's roughly $700. An average customer stays on about seven years.
Our CAC, our cost of acquisition of that customer, is just around $100. It's anywhere from $100 to maybe $150 when we do promos. So you're seeing an LTV to CAC ratio of almost seven, which again is a very high metric for any kind of service that you've deployed. So we're now generating gross margin dollars just off the service business. Ignore the hardware business. That's obviously significantly higher than even our OPEX. And so we're generating free cash flow and profits as a company, which is exciting. From an operational model, as we've grown, we're gaining more and more leverage on the business. So as we're adding another million subscribers, you're seeing our OPEX only raise slightly. So we've built a cloud platform that can take millions, if not tens of millions of customers. And as we're onboarding customers, we're generating more service revenue.
We're generating more gross margin dollars. But you're seeing our OpEx line actually grow at a very slow rate.
Yeah. Yeah, it's been a great story. Maybe just zooming out to high-level demand. We're here at a tech and consumer conference. So Arlo represents a great cross-section for that investor group. Maybe just talk about what you're seeing from consumer behavior through Black Friday/Cyber Monday and into the holiday season. Where's the addressable market opportunity at these days?
Yeah. Yeah, actually, we were just talking about that before we started. The holiday period, it's actually very similar to last year. So the consumer is, you can tell they're stretched a little bit. What they're doing is really looking for deals. So the consumers are still there. They're still buying. When we gave our guide to Q4, we actually mentioned that we thought we'd be up 20% or so year over year on units. And again, units is what drives our business model. So units becomes households becomes subscribers. So a 20% lift year over year on units is obviously a great goal. But we are seeing consumer behavior where they're looking for a deal. They're being a little bit more selective. They're waiting things to be on ad. And we're seeing a little bit of diversity between channel partners.
We're seeing, and this is something we had talked about earlier in the year that was one of our suppositions of what would happen in the holiday period. We're seeing retailers that have more foot traffic. Retailers that have grocery or think of a club doing better on a relative basis than kind of pure play electronics type customers. I think you're seeing that in some of the public commentary from the retail channel as well. Overall, consumers are there. They're buying. We're doing the household formation that we hope to see. Our goal is still to hit roughly 20% year-over-year lift on units.
How about the competitive environment and consumer tech? What are you seeing most out of the major players? Obviously, anytime we talk about Arlo, oftentimes investors ask about Amazon Ring, Google Nest. What are you seeing out of them? How does it compare to previous years?
Yeah, it's the same. So we're not seeing anything different this year. Coming into Q3, Q4, we thought the promotional activity would be slightly higher than what we saw last year. And we actually, in the last earnings call, mentioned that we were going to turn the dial just a little bit more. Because again, we don't need to make a certain gross margin on the hardware. The service gross margin dwarfs anything that we return on the hardware business. So hardware is really that device that creates that relationship with the end user and drives future service revenue. So from a large competitor, even small competitor, it's very consistent to what we've seen in the past. It's slightly more promotional than last year, but that's also what we had anticipated coming into the year.
Maybe even double-click on the higher level of competing with Amazon Ring and Google Nest. How does Arlo differentiate from them?
Yeah, I think the biggest differentiator is we're focused. So when you look at both the retail channel and when you're talking to consumers and building solutions for consumers or in the B2B to C, what we call strategic accounts, our focus on the space is a very large differentiator. So Arlo has always been on the forefront from a technology perspective. So our hardware is considered the best in the marketplace from just RF perspective, video quality perspective, audio quality perspective, battery life perspective, and power management. And we always strive to continue to stay on that forefront. Increasingly, over the last three or four years, it's been our leadership in the services element that's driven us a lot farther than some of the big guys. So our AI is actually more accurate and more diverse from a feature set perspective.
Again, we just came out with custom AI detection where users can create their own AI models. We're the only one in the world with that feature set. If you look at some of the larger companies just in the tech space, what they're really focused on is developing a very large general purpose AI. And so they're trying to solve a trillion-dollar bet and get that done to serve multiple industries. What we're focused on is we have a dedicated AI team building models specifically for consumer security. And again, it's created a much more feature-rich capability from a service perspective, but also a much more performant and accurate output that we get credit for. So if you look at CNET or you look at other reviews of products, Arlo typically is number one.
The other element that's kind of on the side, but it's becoming more important, is as a focused player in the space, we tend to win when people or partners especially are thinking about privacy and data security. So we have a very strict privacy pledge. We are very much focused on data security. When you're a customer of ours, it's your data. We're just processing it and doing what you've asked us to as part of the service element. That makes us not only, I think, a better position from a brand perspective in the retail space, and we know we're winning customers because of that. But when you look at the strategic accounts and the business partnerships, we stand out and are considered probably the best partner in this home security vertical.
Any questions? Yeah.
Yeah, I wanted to follow up. So into the holiday period, you're seeing maybe about 20% hardware uplift. I'm wondering if you're seeing any more interest or activity in the channel. It's not a brand new trend, but it seems like it's getting worse. So we're all getting more things delivered while we're away. And we've all seen it on the news and.
Yeah, package theft. Yes.
So it might require multiple angles, things like that. Is that maybe not right now in the holidays, but are you seeing any kind of lift in the channel?
Actually, we do see that. So it's interesting. The first time we saw a big shift was actually during COVID because the thinking and consumer behavior pre-COVID was very much about security, security. And I mean perimeter security, somebody breaking in the home. COVID hits and everyone's staying home. And it's a little bit less of people breaking in your home because everyone's home all the time. It was much more about package delivery, what we call front door use cases, meal delivery, even so much as people, we actually could tell that people were watching the video as the food was being delivered, a meal was being delivered. And they would wait for the person to leave. So you're not face to face during COVID with somebody delivering you something. So we saw a big shift.
You could see it on just the background platform level where you would see cameras behave a certain way, then suddenly there's all this activity on a single camera or a doorbell. Every year we see the same thing in Q4. I think it's heightened a little bit this year. In fact, it's gone up year over year a little bit. It starts with, funny enough, Halloween. Halloween is actually our biggest day of the year from a traffic perspective because you have everybody coming to your front door all day long, right? People like to keep those recordings. Yeah, unless you hide in the back of the room. Yeah, I guess that's true.
Not that I would ever do that.
But it's funny when you look at it from a platform utilization perspective. We see huge spikes. Actually, our platform technology team actually anticipates that spike. We actually adjust kind of platform resources to make sure that we can deal with the spike in volume. So it usually starts there. Then we see kind of higher utilization on front door use cases, package theft, gifts being delivered, and things like that through pretty much the end of the year. It's even the week after Christmas we'll see a little bit of that. It does generate a lot of awareness. So there has been a lot of package theft last year. It's continuing this year. In fact, there's new types of package theft, if you're following this, where people are actually either dressing up as gardeners and waiting for something to land.
So if you have a gardener on the wrong day of the week, that's something maybe you want the AI to be able to tell you. There's people dressing up as delivery people. There's people standing in front of the yard and showing a fake ID and walking away with a laptop. So the sophistication has gotten a little bit higher. And what that does is it triggers a lot of news stories about it. You have a lot of neighbors that maybe tell you, "Hey, yeah, I had a package theft last week." That generates a tailwind of demand. And we're continuing to see that. So the more online shopping, just in general, and the delivery, to your point, the higher there's been from a need for a front door solution that can give you an idea of what's happening there. Other questions?
Maybe a double-click on the AI point. Arlo has been doing AI way before all of us here have been kind of inundated with it on a daily basis. So talk about what AI means to Arlo and some of the exciting new things that you're previewing here.
So, AI, on a very high level, what we look to do with AI is something we call cognitive offload and threat assessment. So, what we're really trying to do is allow the homeowner, if it's a homeowner, we have small businesses that use our product as well, allow them to worry less. We think everybody has a right to feel safe in this world. So, the AI is really there to help filter and start to make some determination of what's happening on site, and again, it started with very simple object detection in 2018, 2019.
So being able to tell there's a person walking towards your front door instead of tree leaves blowing in the wind or a car driving by or something like that, or to be able to tell, did a package hit your front door or not, and give you a specific alert for that because maybe you want your neighbor to go pick it up per the discussion we just had. So it started with object detection. Where it's gone already with Arlo Secure 5, which we launched about two months ago, and where it's going is something more and more sophisticated as AI continues to grow. So we're now in what we would call recognition mode. So we can do facial recognition. So we can tell who's at the front door. And you can make decisions based off of that.
We actually are already doing vehicle recognition as well. So if you want to know if your wife's car or husband's car pulls into the driveway, or you want to know only if an unknown car pulls into your driveway, we can actually determine which vehicle has actually pulled in via vehicle recognition. And like I said, we've rolled out a beta of custom detection. And what that is, it allows the user to use and actually create their own private encrypted AI model to ask certain questions of the platform and get a response back. So things like, did I leave the garbage cans on the street? And every Tuesday night, an AI can run, take a quick snap, and determine if that's happened and give you an alert, yes or no. Did I leave the garage door open? Did I leave the back gate open?
Did I leave the lights on? And so we have the ability now for a user to type in a question, right? Did I leave the garbage cans on the street? Our platform takes the text of that question and starts to build a model. And it'll know garbage cans, street. So what does a street look like? What do garbage cans look like? And it starts to build a model to actually determine what's happening. You as a user then can add images. So you can add an image of your cans on the street, remove them, and do an image in the other state where they're not on the street. And then the model will start to learn over time as well. So if there's some indeterminate states, the app will actually ask you: "It looks like your garbage cans on the street.
Is this true, yes or no? And if you say yes or no, it's training and refining your model. So it gets smarter over time. So again, we're the only company in the world doing that. But the idea is to allow you to worry less and be able to kind of cognitively offload, let the platform start to think about what's happening around, and you have the confidence that you're actually getting notified when it matters. So that's one. The other area that I think you're going to see additional AI capabilities and just linkages to other platforms like emergency response is in that threat assessment and kind of emergency response area. So we do a lot of work on what we call time to response. So if something's going wrong, how many seconds does it take for you to react or get help?
And so there's a lot of work we do there to the point where you'll see an announcement, some rollout of some new services where we're going to have the capability in certain municipalities to actually push metadata and sometimes video all the way up to the police or fire crew that are responding to maybe an emergency, right? So as a user, you have the ability to turn on specific cameras to share with emergency response. And that information can get all the way out to a laptop that's sitting in the first responder.
If you have a grandmother who's got diabetes or a bad hip and they're on the second floor, or you have a certain gate code for the neighborhood that you need to give to the 911 operator immediately, and you're not home or can't do it, all that information can be passed securely through the 911 center and all the way to first responders. Those are the two areas that we spend a lot of time thinking about and where I think we have some of the biggest advantage over the competitors in the market.
And what does that mean for the financials? It's obviously a compelling qualitative story, but I think it's also bleeding through into things like ARPU and metrics towards that maybe could double-click on.
Yeah. So we've been experiencing a lift in ARPU over the last two or three years. So if you rewind, three years ago, our ARPU, I think, was $7 or $8. It then raised to $11. The most recent quarter was just over $12. And I mentioned we launched Arlo Secure 5, which is our new security platform, is at the end of September, really started rolling out first week of October, so about a couple of months ago. And we've seen a dramatic shift in user behavior on signups. And this shows, I think, there's still a lot of untapped potential to provide features and get people to sign up for a higher tier plan. So I'll give you a couple of data points. One, on new subscribers, now that Arlo Secure 5 is rolled out, ARPU's already jumped above $14.
We're seeing a substantial 10%-20% lift in ARPU with our new service pack that we rolled out in October. The other thing is we have four plans. We have two basic plans, kind of a basic single and a basic multi-cam plan. And then we have two premium plans. There's four plans total, two premium, two basic. Historically, before Arlo Secure 5, about 18% of users would sign up for one of the premium plans. So most of the users were signing up for the basic plans. And then our job is to try and get them to mix up over time with features. With Arlo Secure 5, we're seeing 42% sign up for a premium plan. Now, some of that is the new features. Some of that is plan construction. Some of that is we're always A/B testing how we present the plans.
So there's a lot of work that goes into every service launch. But on a blended basis, the changes that we've made and the new features that have rolled out, including the new AI, has had a dramatic impact on ARPU and also what we call plan mix, which portends to future ARPU increase as well.
Great. Maybe we'll tackle go-to-market. You had alluded to it a little bit before, but there's been a substantial expansion in your go-to-market, a lot of different partnerships. If you could just touch on Verisure and what that has meant, and then maybe expand it into Allstate and some of the more recent partnerships.
Yeah. So again, if you look back at spin, something like 84%-86% of the revenue coming into Arlo was from retail. So it was the typical channels, Best Buy, Amazon, Costco, and the like. Part of our transformation of a company to a services-first company also had a track of diversification of revenue. So now, given the quarter, anywhere from roughly 35%-45% of our revenue actually comes from partnerships, strategic partners. And those could be an ISP. Sometimes they're mobile carriers because we have cameras that have cellular, but they also sell our Wi-Fi products. We've got Calix as a partner. Adam mentioned Allstate. So Allstate is an insurance carrier. And they found that the pairing of smart home technologies and security technologies is actually beneficial to their business for engagement, retention of customers, but also damage limitation.
If you have a smoke sensor or a water leak sensor, the financials change quite a bit if you have a household that's considered smart and has these sensing capabilities. So we think there's a big play in insurance over time as well. Verisure is one of our largest partners. We signed with them actually in 2019, and we did a five-year deal with them that just got renewed for another five years, where we partner with them on basically the entire European region, so we partner in retail with them, but they are also considered. They're very similar to like an ADT here in the United States, fast-growing, very innovative security company that is what we would call a traditional security company where they knock on doors and actually do the installations. In Europe, there hasn't been quite as big a shift.
We're actually their exclusive provider for cloud platform, all the AI that they use, video storage. And we sell them all of the cameras that they use for both the retail channel, but more importantly, the direct channel where they're doing these installs. So again, the strategic partner area has grown from basically zero when we spun to nearly 40% of the business. And we believe, as we look at our long-range targets as a company, kind of our five-year goals, we think about 60% of that incremental growth will come from strategic accounts because we just see so much opportunity.
There's so many potential partners, so much activity recently, especially in the last six months or so, of companies now coming and saying, "I want to deploy security as part of their service pack or to their population of users." One of the macro trends we've seen that I think is driving some of this is DIY security, smart security, has just hit about 18% penetration in the U.S. market, and we're seeing it transition to mass market, so Walmart now is becoming a much bigger player in the space, and that's a signal that the awareness has hit a certain level and the market is starting to open up, and we think another wave of growth is coming as we hit mass market, that's often what triggers potential partners to want to deploy it as well.
When you look at the consumer home, there's only a certain number of services that people are willing to pay a monthly fee for, especially when you look at technology and smart home, and security is definitely one of them, and there's not many else, so it's a great opportunity for recurring revenue. It's high attach, low churn, and there's these huge populations. If you're an ISP, as an example, a broadband service provider, you may have millions of customers, and this is a great area to actually increase your engagement, potentially lower your churn, and drive incremental recurring revenue, and as Arlo, we're OK either way. We look at the market as how many households can we address, and if that's through a Best Buy channel, that's great. If that's through a broadband service provider, that's great as well.
We're just trying to address the maximum number of people as we can.
Great. Any final questions?
Talk about improvement in AI and efficiencies of the platform. Can you talk about hardware improvement? I think one of them can be high definition security camera, sensors at night, but I'm sure there are many more.
Yeah. So it's a little early. But I will tell you, we see innovation come in waves. And I think I mentioned as we led the initial hardware innovation wave and still, I think, are at the forefront of what's currently out there. I think then there is and still continue to be a big innovation wave around AI and services. And again, we're very focused on leading that. To your point, though, I believe over the next 18 to 24 months, we're going to see another wave of hardware innovation. And it's going to come from alternative sensor technologies that can go into the products. We have an interesting most of our product is battery operated. So often, the technologies we choose in our product are balanced between performance and battery life because we want all of our products to last months and months on battery life.
So often, a given technology has to have a certain shrink on die size, power utilization, and everything for it to really become viable. I think there's several technologies, three I can think of right now, that are going to move past that sweet spot of performance and actually power utilization being low enough that we could deploy it en masse in the consumer space. So think of longer range, higher accuracy, more informative from a data perspective. Motion detection technologies, I think, are going to start penetrating a lot more. The other thing I would tell you is there has not been a lot of what I would say ecosystem and consumer experience from a camera perspective, not from a service perspective, but from a camera perspective, since Arlo basically created this market segment seven, eight years ago.
As we're moving to mass market, we're spending a lot of time thinking about what is the next user experience for consumer security, and it may not be the form factors we're seeing today, and we have some ideas. We've actually kicked off some interesting programs that I think, again, will land within the next 24 months that'll drive, I think, another way for somebody to actually install, interact, and actually drive a very sophisticated consumer security standpoint with form factors that look very different than what we're seeing today, and so we're very excited about that, so I do see, to your question, the point of your question, I do see another innovative cycle coming on hardware, and it'll really hit probably next year and continue for a few years.
As every industry goes through this, as you kind of rethink, you start from a clean sheet of paper when you hit the mass market. You start to ask very different questions, and I think that'll layer on top of the more traditional format that we pioneered when Arlo actually launched seven or eight years ago.
With the teaser, I like it.
Yeah.
We're going to have to leave it there. Thanks, Matt.
Thank you.