Covering tech services here for Citi Research. Matt McRae here from Arlo Technologies. Great to have you here, Matt. Welcome to the conference. This is a name that's a little bit new to me as well, so for investors that I think are new to the story, maybe let's dig into what you believe the market opportunity is today for Arlo, and where does Arlo fit in the broader connected home ecosystem?
Yeah. Arlo, for those that are very unfamiliar with the story, we are a spin about eight years ago from a company called Netgear, who was in the home networking business, and they were looking to develop products that would utilize a lot of the wireless networks at home to try and trigger upgrades for people to buy the latest wireless router standard, and hit upon video as a use case that would eat up a lot of bandwidth and drive people to a new router.
Decided that, Roku had just started, and there were some streaming set-top boxes and figured that wasn't the right area, and hit upon the idea of DIY, do-it-yourself security, and built the first Arlo camera nearly, I think it's over 11 years ago now, maybe 12 years ago. It took off like a rocket.
The market since then has really been going through a transition from what we call DIFM or do it for me, meaning installers coming and building home security systems in your home, to DIY, which is do it yourself. The technology that Arlo created and the market segment that we actually created was really around simplicity, powerful visual-based security, and driven from the idea that you could take two or three cameras, install them in 20 minutes. That market that was 100% DIFM 10, 12 years ago, is now about 60-plus, sometimes as high as 70% DIY, and the rest of the market is now about a third of the market.
Oof.
That's the shift. The other thing I would tell you about the size of the market is independent studies and surveys show that about 20%, it's anywhere from 18% to 22%, of broadband households in the U.S. now have a camera or some kind of visual-based security that they can access through the internet.
I've been in the consumer space a long time across multiple companies, and right around the 15% to 20% is when you move from early adopter, maybe tech-forward type customers and market segment, into mass market. We think we entered that transition about a year and a half ago. We're seeing that in many ways. One is we're seeing a new group of customers coming and buying the product.
But we are also seeing the transition of, for instance, we are seeing Walmart be a bigger part of the market than was historical over the last three years, and that is usually a signal that it is going mass market. So our view is the next 30 million households, in the United States alone, will probably come on and become connected with visual-based security in the next three to five years.
Services make up roughly 60% of the business right now. What are some of the important strategic decisions that enabled that transition, and I guess, how do you think about expanding that base going forward?
Yeah, it is a great question, and again, it goes to the history of the company. So when we spun, we were really a hardware company at that time. It was all about the hardware margin and just trying to sell as many widgets as you could in any quarter. But it was clear at that moment that home security is really a services business, right?
You are providing an ongoing value proposition to that end user. And so after we spun, we set about basically transforming the company from a hardware-first to a services-first organization, and that was a change in the culture, it was a change in the roadmap, it was change in finance and accounting practices and which metrics we are chasing. Changed everything in the company. So I would say the genesis of that transition happened probably right at our IPO, which was in the middle of 2018.
Took about two and a half, almost three years to actually really transform the company at its DNA into a services company. And if you look, we crossed over, not only service revenue being a bigger portion of our revenue than hardware, but from a profitability perspective, probably about two years ago, and it has just been a rocket ride ever since. So you are going to see that continue.
Certain metrics you look at, we are looking at our service revenue growth year-over-year, approaching, we want to stay around 20%, so still be one of the fastest-growing companies in the space. We have also seen gross margin, so blended gross margin across the company, has gone from 30% to 40% to 50%, and we think we are on our way to 60%, is where we want to target over the next couple of years.
Operating margin as a company, too, was negative when we were a hardware company, broke through into positivity, and we hit almost 20% in the last quarter. I would say we're in that 15%-20% positive operating margin, headed towards 25%-plus, over the next few years.
Should we think of hardware, and we've seen other companies that have transitioned into more of a service-based model, is hardware more of a customer acquisition tool?
Yeah. The way we look at it was a dial we turned slowly as we transitioned the company to services. But once we hit about 2 million, 2 million to 3 million paid subscribers, I think it was around 2.5 million, the economics and the profitability of the company transformed to the point where we could really look at hardware as CAC.
As part of our customer acquisition cost, we measure ourselves on hardware unit sales because that brings new households into the funnel. We measure ourselves on blended gross margin. Even if the hardware goes negative margin, what we're always doing is looking at the blended gross margin expanding. It's a good way to be disciplined around the overall profitability of the company growing.
We measure ourselves on things like LTV to CAC ratios to make sure that our CAC is still at a very healthy level against the value of the customers that we're actually bringing to the table from a shareholder value perspective. Just give you a couple of numbers. Our LTV is approaching $1,000. I think it was $976, $967 last quarter. Typically our CAC is anywhere from $200 to $300. Our LTV to CAC ratio at the beginning of the year was about a 4x, which is considered world-class, and we'd like to keep it between a three and five. It moves up and down depending on promotions and in the-
And $300, I'm sorry, $300 CAC?
The $200 to $300 CAC on a $1,000 LTV. You can see every time we lean in and produce additional households, the shareholder return is actually very, very high.
Oh, and how do you think about pricing on the services? Is that you have a tailwind there? Is there a sensitivity? Could you just walk through some of those?
A couple things. When we made our transition, one of the things that we tested early on was the pricing elasticity on hardware versus service. What we quickly found out is that consumers are much more sensitive to the upfront cost of the hardware, so making that initial purchase. So how much value are they getting? How many cameras can they buy for a certain budget? How hard is it to install as well? Do they have to hire somebody to install it or not is part of the cost equation.
But once they're a subscriber, they're relatively insensitive to price because of the amount of value we're providing. So our ARPU right now on average on our retail and direct is just over $15. A couple of years ago, that was closer to $11, so we've had ARPU increasing substantially over time.
Most of our competitors in the broader security space are typically anywhere from $40-$60 or even $70 a month. We feel we are nowhere near the ceiling of where we can go. Our typical roadmap is we will produce a lot of new functionality, new services, new features into our service tiers, then look to earn a price increase in that following year. It is a little bit of a tick-tock-
type timing. This year you will see us introduce a lot of new features in the coming weeks in something we call Arlo Secure 7, which will be a new version of our subscription services, setting us up for potential ARPU enhancement and increases next year.
I definitely want to get into the platform in a bit here, but I guess before we get to that point, I want to position Arlo against maybe some very large brands out there that are very platform-centric, have lots of bells and whistles, DIFM, even DIY. How should we think about Arlo's wedge in the market here? What are some of the differentiating factors that Arlo has?
Yeah. If you look at the competitive landscape, it is really, I would say, in two major buckets, and one of the buckets has a little sub-bucket maybe. If you look in the retail and direct space-
You've heard of names like Ring or Nest, then there's a lot of smaller competitors in that retail space. We're a solid number two there. I would say Ring is number one, mostly because they own a channel, so they own Amazon, they spend a ton on marketing. But we're happy being in that number two and having a more innovative service, a much more profitable business, and we think we're getting the best customers out of that retail space. We even see Google kind of pulling back, and some of the other brands are starting to consolidate a bit in the retail channel.
As the services business has become a larger component of the overall business and specifically our P&L, a lot of the entries that come into the market, whether they're overseas Chinese, Taiwanese, coming in and building just hardware and trying to sell that, it's very difficult to stay competitive because we can go negative easily on the hardware because all of our economics come on the service. If you don't have that service component, it's extremely hard to be competitive long term. So that's on the retail side. I would say it's really Ring and Arlo battling it out in the retail space, and we're seeing some consolidation that we think we can take advantage of over time.
Our work in the space has also shown that integration is a major factor here.
Yeah.
Everybody can't use multiple apps to manage their home.
Yeah.
Having one integrated control center, no other way of saying it. I do want to talk about the platform a little bit, but in particular, the Secure 7 release.
Yes
Which there's been a lot of investment ahead of that. Can you just walk us through what that enables? What sits underneath the platform? It seems like it's such a strategic asset.
Yeah, it really is. Especially as we became very cash flow positive and started building up cash reserves, it's really moved to a more formal capital allocation plan as a company. And that includes stock buybacks, it includes a small acquisition we made, we may talk about in a few minutes, but it also deals with our organic investment in the platform, especially over the last three to maybe four years.
One of the things that's happening is the AI level layers in the service is getting more and more sophisticated, and we've always been at the forefront of developing those new technologies and rolling them out. To give you an idea, we rolled out probably what is one of the world's first consumer subscription AI services in the world. In 2018, we rolled out a subscription service around object detection and computer vision using AI capabilities.
Ooh.
And I remember at the time we called it AI, and it actually freaked users out, so we changed it to computer vision because AI wasn't even being talked about, and people didn't really understand what it is, and it's kind of gone through, I guess, a bell curve where people are getting freaked out about it again. So that's all.
Come along then.
Yeah. That's how long we've been doing this, right? We created the first object detection. Is it a person? Is it a package? All that almost eight years ago, and have been building on that. The overall roadmap of where we see AI going in the space is one is moving from detection, and now what we do is recognition. I can see facial recognition, but we also do vehicle recognition.
Is that my wife's car in the driveway? Is it an unknown car in the driveway? We've started doing custom micro models and all these things in that area. Where it's moving next is interpretation or assessment. One of the things you'll see from Arlo Secure 7 is the beginning of what we call threat assessment. That's not just trying to detect objects in a scene, it's actually looking at the overall event.
Do I see a person? What are they holding? What are they wearing? What time of day is it? Are people home? Are people not home? Do I see a person at the front door and the back door at the same time? Are they approaching a window versus a door? We take all of this information in, and the AI interprets a threat level to that event.
Ooh.
Then can trigger certain responses based on that. Most security systems, actually all security systems today, are binary, right? They either do nothing or they call the call center, the police, and set off the siren. It's either on or off and there's no in between. What threat assessment allows us to do is actually score things on a much more granular level. If it's nothing, maybe we don't even give you a notification.
Maybe it's something interesting, we give you a normal notification. Maybe it's at night and something looks suspicious, we can give you an escalated notification that punches through any do not disturb. Then if it's really high on the threat level, we can actually dispatch first responders without even contacting you, and then letting you know that it's already happening. It's a much more powerful area.
Its interpretation and assessment of what is happening, not just, is there a package at the front door? I think that is going to open the next maybe two to three years of innovation in that area. Then you brought up platform, kind of broader platform. The other thesis we are operating under, and you will see a little bit of this in Arlo Secure 7, but a lot more in Arlo Secure 8, not to get ahead of ourselves.
Boy.
The smart home and home security is really becoming smart home security. We see those worlds combining. What is interesting is there is a lot of incremental benefit from the smart home capabilities. For instance, we could envelop a lot of those devices and capabilities into our threat assessment. Maybe if somebody does approach your house late at night, it could turn on all the lights, or flash the lights, or do something to actually deter. It is actually using your smart home to actually become part of your security layer, right?
What is interesting also is almost all of the economics, especially on a recurring revenue model from smart homes, are coming from security. People do not pay for a smart lock on a monthly basis. They do not subscribe to their smart light bulbs. What they do is they subscribe to security.
The economic comes from the security element of the relationship and the control and the end user. Really what is happening is it is going to become smart home security combined, and that you will see start to ratchet up over the next couple of years as well.
The reasoning functionality, that is fascinating. You could think about all of the applications that you could tie to that.
Yes.
And personalization involved in that as well. Wow. That's fascinating. The revenue opportunity here, is it a significant up-step, or do you think it's going to be more of a market share play?
Yeah, I think it's both in a way. Because again, I think we'll be at the forefront. It'll be great for the brand. It'll drive people who are interested in the latest capabilities to come to Arlo for that from a marketing perspective. But also, I think you'll see it shift our ARPU higher. As we mix people into higher tiered services, have the opportunity to maybe increase price as that gets into the marketplace.
And that'll be paired with a new platform launch that we'll do next year from a hardware perspective as well. So, a lot of that organic investment, you'll see some of it in Arlo Secure 7 in the next couple of weeks, and a big part of it about probably next summer, Q3 of next year, and that's part of our capital allocation plan.
I was telling an investor at one of the meetings this morning that it almost felt like we were getting to a plateau of capability in the space when you could pretty much detect any object and be able to recognize people and cars and everything. And I would say over the last 12 months to maybe 18 months, there was a punch through to this more interpretive, intelligent reasoning capabilities that have risen to the level that I think we're going to see another three to five years of intense innovation and ability for us to extract additional value from customers that are getting something real, something beneficial from our subscription services.
That and I guess we're also starting to hear more about Wi-Fi sensing as another-
Yes
potential technology, which is amazing the potential that we can do with that. That is fascinating. I do want to dig into that capability, that computer vision layer, the intelligence layer that Arlo has. From your perspective, what is unique about it? What is unique about Arlo's AI strategy, and where do you believe the company has an advantage?
Yeah. First, we were one of the first to build most of these models and actually get them deployed. So we have a very deep team, a bunch of PhDs that work on this quite a bit. We often get the question of why are not you using off-the-shelf models or why are you developing this internally?
I would tell you our experience is that focused technology, and specifically if we talk about AI, models that have been trained, built to solve very specific use cases, perform dramatically better than a big generic model that you are trying to ask questions of specific things. I will give you a couple examples of that. So, even when you do person detection or package detection, a big generic model, if you ask it, is this a package or not, or is this a person or not?
They are trained on pictures of people that they are pulling from stock photos and newspaper articles and all those kinds of things. When in reality, when you actually deploy a security system in the field, most cameras are mounted up in the top corner and they are looking at people from a downward angle in most cases, right? It turns out those generic models can only get to a certain accuracy level.
We train all of our models on real security footage stuff that is taken from real security environments, real deployments around the world. So our models are dramatically more accurate. That is one example. Another example is if you use these large models, they can be very expensive. They can be compute-intensive.
When you build a model that is specifically trying to determine what is at the front door, is it a threat or not, is it a package, is it doing certain things, and you can boil the model down, not only is it potentially more accurate because of your training, it is also faster, it is cheaper, right? Much easier to scale across the business. In fact, we can scale our models so well that we actually do something called micro models. So if we have one person type in, "Did I leave my trash cans on the curb?
Right.
We can build a little micro model and actually put that in that user's account. Somebody else could type in, "Did I leave the back gate open?" And it can put that model in that account, and those are user-specific models that are not even pulled into our main model, and so they are private, and they are not shared from a training perspective or actual execution perspective. Those are some of the benefits. It is scale, speed. One of the things in security is latency matters, right? Very much. You want to be able to interpret a scene as it is unfolding within tenths of a second, not multiple seconds. The smaller models actually perform better from a consumer perspective as well.
What we typically do is we will start with bigger models, figure out what we are trying to hone in on from a great user experience, start building our own internal models side by side, and relatively quickly, our models perform better, are less expensive, are faster, and can be deployed a lot quicker, relatively quickly, and that is typically what we go to market with.
In some of our other work on the security business, we do focus a little bit on how the use case for security is where false positives and misdetections are major super important kind of elements of the quality of a service. Where do you think Arlo needs to do better than its competitors for AI to become more of a true differentiator rather than just a feature?
Well, I think there are multiple vectors. One, it has to be faster, better, all the things we already talked about, just from a raw performance perspective, and we are definitely there. From an accuracy perspective, it is two things. It is one is that training I talked about. Training on real-world data instead of just stock photos and kind of general training is really important.
What is interesting too is when you tune for security, false negatives are actually much more important than they are in the normal world of just trying to determine is it a person, is it something else, for other purposes, right? Because if missing, having a false negative, meaning there is actually somebody at your door and you totally missed it, is a big event for the end user.
False positive, where maybe it thought there was a person, but it turned out it was a tree that looks like a person, is less important. Some of the tuning of the models and what you allow from a false negative or tune for on false negative versus positive is very different than what most of the general models are tuned for as well. That's important. Then just like we talked about before, I think it's extremely important that Arlo stays at the forefront.
We were the first with person detection, first with package detection, first with vehicle recognition, still the only one. The only one in the world with micro models. We're going to be the first in the world with any kind of interpretive threat assessment. We will stay at that forefront.
But again, part of that is because we're so focused on what we do. We have an entire team that wakes up every morning and tries to build the best security experience in the world. We're not waking up every morning and trying to build this massive generic model-
Right
that is supposed to pass the bar exam, teach you to cook a recipe, and maybe also figure out if there's an intruder at your front door. So it's a very different business model, and it's proven, I think, extremely effective, and it's a big differentiator for Arlo.
Oh, it's a huge tech benefit for purpose-built. Absolutely. I do want to touch upon the acquisition of Aloe Care Health, expanding the company in the Age in Place and the wellness monitoring kind of business. Why don't you give us an overview of that opportunity-
Yeah
What drove you to acquire the company?
Yeah. I mentioned our capital allocation plan around organic investment, which we just talked a lot about. We're doing stock buybacks. Historically, you should expect us to probably do more at the current stock price. We made our first real acquisition earlier this year. We've always told investors that if it's going to be in an adjacency or kind of a new market, it'll be a relatively small acquisition. If it's in our core market, it could be bigger.
This is an example of a small acquisition that moves us into the smart aging or the Age in Place market. Small company, huge potential. To give you a little bit of background, the market for Age in Place in the United States alone is around $26 billion, $27 billion right now. Very antiquated technology.
If you're familiar with Life Alert and some of these really old pieces of tech that haven't moved forward. No AI to speak of in any way, terrible user experiences, if there is a user experience at all from an app perspective. That market's going to grow from about $26 billion, $27 billion to nearly $300 billion in the next 5 to 10 years. So you're looking at a tenfold increase in that market.
That's driven from technology and other things, but also just demographic changes of the aging population. One in five to almost one in four Americans will be over 60 years old in about 5 to 10 years. So it's huge movement, and there's a big push across all the industry, including the government, insurance, and everything, to have people stay home longer.
That means they need to be monitored and make sure that it's a safe place, and safety and security is what we do. Aloe Care is a really interesting small acquisition, and we really purchased them for two reasons. One, they had the most interesting Tech forward, innovation forward pipeline from a roadmap perspective we had ever seen. So they have a very innovative hub. It can check monitoring conditions, it checks movement, it can do two-way phone calls.
They have an AI calling service that can check in and have a discussion with the user. It actually transcribes all that, can flag concerns to caregivers. It's amazing technology, and that's just starting to roll out. They're also moving towards fall prediction. So you'll see some solutions out there where somebody wears something and it detects a fall. That's good.
What's much better is if you can actually predict a fall before it happens, right?
You can do that through environmental measurement. You can do that through gait analysis and radar. You can do that through the discussions and transcribing. Did they sleep well? Do they not? Are they dehydrated? All that. Movement through the house. If there's a lot of movement at night, that can mean there's an infection or something. Actually being able to predict fall is considered the holy grail of being able to really help people at home. They're pretty far down building an AI model that can detect falls.
Right.
That's one. The other reason we're really excited about the company is they have a pipeline of potential partners and deals. We announced one recently, expect another couple announcements in the next couple of quarters. But we believe that the revenue from this piece will grow dramatically in 2027 and then continue from there. The other, there's a lot of synergies long-term in the roadmap.
The ability to sense things in the home and call first responders, it's very similar in a lot of ways. I think some of the AI models will continue to be able to benefit from each other. We're finding a lot of homes where you want to check and make sure your parents are safe at home, also means you want to make sure they're secure at home.
We believe a lot of these households will both use an Age in Place solution inside to monitor the safety of the person while they're home, but also the perimeter security.
Right.
You will probably see over the course of a few years us bundling subscriptions together that nobody in the market has done before.
That is fascinating. As a user and of having elderly parents in that situation, that is fascinating. You recently discussed additional growth factors beyond core consumer, including service provider, insurance partnership opportunities in the SMB market. How should we think about the next phase of expansion, and where do you see the most compelling opportunity?
Yeah. Post-spin, one of the things we did is we worked really hard to diversify the revenue for Arlo as we were transitioning
Right
simultaneously to a services business. What that meant is when we spun, we were primarily, and actually almost solely, a business selling through retail partners. So the Walmarts, the Best Buys, Costcos of the world. We specifically built our platform to be partner aware in a lot of ways. It is built for partnerships. So Verisure is a very large partner, if you are familiar with their story. They are similar to an ADT. They are in Europe.
They just went public actually last year, have about 6 million plus subscribers. We are their primary camera provider and run all of their visual AI systems in the back. We just recently signed a deal with ADT, and are doing the same for them in the DIY space. They just launched. We will see them ramp through this year and probably ramp a lot more next year as they get into more channels.
We just announced Comcast
as a partner as well. They have 31 million broadband subscribers. We are in the middle of integration with them, and hope to launch with them in the middle of next year. That will provide additional growth in the second half of 2027 and a full year impact in 2028. We have a deal with Samsung and some others.
When Kurt and I look at how we get from where we are to our long range plan, which is 10 million plus subscribers, $700 million in ARR, and I mentioned the 25% operating margin, we think about 60% of that incremental growth from where we are and how we are getting there is going to come from these strategic partners. We talked about how Arlo is I think the strong number two and the more innovative number two in the retail and direct business. In the partnership channel, we feel we are number one.
We are winning predominantly almost all of the major partnerships over there. I think it is because we are focused on the space. This is not a side hustle for us
Right
like it is with some of the large companies. This is all we do, and we also have a very deep focus on data privacy, data security. Like I said, we have built a platform specifically for partnerships. So we have robust APIs with documentation, source code. When somebody partners with us, they are seeing the most secure platform on the market, the most innovative feature sets, and roadmap, and a very robust set of APIs and documentation to get them up and running really quickly.
Fascinating IP. Very interesting distribution, which at least in my analysis of the space, I haven't seen before. I do want to talk about, though, some of the things in terms of trust which is foundational to connected home products, and there's obviously a lot of chatter about this topic. How does Arlo think about privacy, data protection, responsible use of AI? Are there red lines? How should we think about
Yes
Arlo branding itself with all these AI products?
Yeah. It's core to everything we do, and it's core to the culture of the company. We approach it, I would say, in a very unique fashion compared to all of our competitors in the space. We collect a lot of data and it's private. It's totally encrypted. But we believe, and this is even in our policies, is it's not our data. It's your data as a user, right? And we're processing it and storing it on your behalf.
We never do anything else with that data than what you expected us to do for it. We don't sell it. We don't use it for advertising. We don't use it on an e-commerce site like some of the other companies out there. We have very strict data security policies. We're the only company I'm aware of that actually has a cybersecurity and data privacy board committee.
With board guidance and governance over this entire space. It is a massive differentiator for us, and I think it's not only helping us in the consumer market, but it's one of, like I mentioned, it's one of the main reasons why some of the largest companies in the world are willing to, and comfortable partnering with Arlo over some of the other companies that are out there.
I can imagine touching healthcare in particular.
Actually, it's funny you say that. I just completed my training for HIPAA.
Oh.
That's me and Kurt's laughing.
There we go.
All the executives. We decided as part of the Aloe Care acquisition that we're not yet, but we might be starting to collect information that would be borderline health related.
Right.
We had a couple choices as a company, which is, okay, should we silo that data in a different repository in our cloud infrastructure and just treat it separately and try and kind of firewall it against other stuff, or do we actually envelop it and we turn the entire company into a company that's even another level of privacy and data security, and actually have the entire company become HIPAA compliant instead of just this one little section? Pretty easy.
We made the decision very quickly to say, you know what? It's another level of our commitment around privacy and data security. We're going to have the entire company become HIPAA compliant. We're going through that process right now, and that's ahead of future roadmap items that might be able to integrate some of this data in some interesting ways.
That's quite a stamp. Next year when you're here
Yeah
and we're talking again, what do you think is going to be the hottest topic?
Yeah. Let's see, what's that? September? I would hope we're sitting here talking about some of the partners we already named launching and ramping.
I would hope that we'd have another two to three partners that are of sizable scale, and that we'd be able to characterize the growth that we're going to see in this Aloe Care acquisition and what the year-over-year is going to look like with some very large deals that are driving material revenue on that side.
I'm looking forward to it.
Yeah.
Thanks again, Matt.
Thank you.
Great to have you.