How AI Is Transforming Short-Term Rental Bookings

Welcome to the STR data lab. Hello and welcome to the STR data lab. I'm Jamie Lane, chief economist at AirDNA, and we've got a really fun topic today and a great guest. We have uh Jason Sinata uh the co-founder of Kismmet uh on the show. Jason, welcome.

Thank you, Jamie. Happy to be here. This is exciting.

Yeah. So, we're gonna dive into the exciting realm of uh AI visibility, this changing landscape of how do I get booked through whether it's chatgbt claude, how AI models, how and these new tools are changing discoverability uh for short-term rentals for hotels, how the OTAAS sort of impact that and Booking.com, Airbnb Expedia, but I want to start because and it is it has been a fun topic on Good Morning Hospitality. Uh we're obviously on a a co-host and we talk about it a lot, but just last week and booking held their um and quarterly results call uh and they disclosed that AI chat bots drive under 1% of their room nights and booking is I know partner with anthropic with open AAI they're directly integrated with these systems like they've got I mean so in my mind I would think that and they would be on the cutting edge of being able to drive bookings through their platform. So if they're only able to do 1% and why do I even care about this?

I I think that's the best question to ask. I love it. Um and so we have very different numbers. We'll break in and like I I refreshed them this morning.

We'll break them down. We'll tell you what we're seeing. I think ultimately, and we're gonna try to make this interesting as two people with econ, this is a question of measurement and choosing what to measure. Um, and I think there is a big build from less than 1% of room nights to what I consider a pretty major impact already on the revenue that's being driven from these channels. Uh and let's be specific, I think it's mostly chat GBT in the chat and chatbot arena. Uh so AI search um and then the other half of it is the fact that Google search is now mostly AI mode. Um but it also matters for paid and we can get into that too but let's start with the the room night's question because I think this is a question of measurement. Uh what we've heard from Skiff and Focus right is in excess of 60% of people are planning their travel using chat bots. And so if 60% of people are planning with chat bots, it's got to be more than 1% of room nights.

Uh and it's got to be more than 1% of revenue. So let's let's talk about what we're measuring when we say 1% of room nights. I think they're looking at UTMs and referers because right now you can't check out in chat GPT and AI mode. That's about to change and I think the numbers will move very quickly once you can do that. But right now you have to go check out on bookings website, on a supplier's website, on a hotel's website, right? But if you're only looking at UTMs and referers, you're going to miss the vast majority of that traffic. and we have the data to back that up. So what we're looking at is actual server fetches from chat GBT uh which is a different metric than looking at refer or UTM. And when we do uh you know some complex modeled analysis on what's going on we're finding and the headline number is that on the most seasoned portfolios upwards of 15 16% of overall revenue is coming from chat GPT alone and that if you make any sort of assumption on what portion of you know Google search is now AI mode you might be 30% of your direct traffic is is getting touched and impacted.

from a revenue perspective uh by by AI at this point.

So I definitely want to get into it on how sort of can make the link leap to those assumptions because so let's just play out like what's happening when I go to chat GPT and start planning a trip, right? because and the incredible thing and I and I just before I went on my trip to Iceland, let's say, I built my entire itinerary and using the LLMs, uh figuring out what to do, where to go, where to stay. And when I'm kicking off that search, I mean, it's going out and maybe doing tens, 20s, hundreds of different searches of like what's out there. And what's incredible is like, and maybe it can do it on booking, maybe it can go and do a Google search and see who what are all the direct ones. And just because it's maybe hitting that direct site and collecting data on that direct site doesn't mean that it's actually referring me to actually book on that direct site. Right? So just because chat GBT is going and collecting info and maybe it's doing it thousands tens of thousands of times like I think people are seeing a ton of traffic on their site sort of attributed to uh these LLMs.

How do I make the leap between and uh chatbt I'm collecting data from my site and that someone's actually going to be booking because I would think it would be a magnitude less than the volume of traffic that I'm seeing from these sites uh as opposed to a magnitude more in terms of conversion.

Well, the there's a question of methodology there and I'll touch that in a second. But can I ask you a question? How did you book after you did all your research in Chat GPT?

Uh, I went to Airbnb and booked uh well, my wife went to Airbnb and booked our our homes. Yes.

Did she do it on the same device that you looked on chat GPT uh and in the same browser window and within

Honestly, none of the recommendations that they had on lodging we ended up taking. So we took the advice on where to stay in terms of cities and areas and then went and did our own separate search and our normal search on like where we'd go and book lodging. Uh so high level build the itinerary.

Yeah.

And then go through our normal channel of how we would go about finding lodging given the OTAAS that and how we'd normally book.

So we wouldn't have counted you as as an AI delivered guest. um you would have been a direct guest uh under our methodology, you know, a direct organic, you know, showed up at the website and and did did what you did. Um, if you had booked one of the ones that you visited on Airbnb, Airbnb using the same methodology that we would, and I'm sure they are because they have more data scientists than we have and they're more sophisticated than we are, they would count you as as as direct booking that came out of an AI channel because it's very easy to measure. What they would do is they would look at the specific properties that were shown to you in that chat search and then look for differences in the booking rate of that property using some pretty gory math at the end of the day, but a very wellestablished methodology of modeled conversions. That is the exact same way that that Meta and uh and Google count their modeled conversions for their ad performance platform.

Um and when we do that, we see that much higher number. Now the AI traffic question is a really important one too. Um and that's because we see on our portfolios anywhere from on average 4:1 ratio of fetches uh so visiting a property page within chat GPT or on an AI mode 4:1 ratio of fetches of those properties to guests actually landing on those pages from an AI channel. Um, and so that's equivalent in the olden days to a page view, right? That's a guest looking at a property. And so actually a lot of organic traffic or searchbased traffic has fallen off across the industry. Um, and this has been happening since the midpoint of last year. And that's because people are completing most of their research before they ever visit a site. So just a couple weeks ago, New York Times had a really interesting article. 75% of people who see an AI mode overview in retail never click out. So for every one clickout you get, three people viewed your content without ever leaving AI mode.

And Google doesn't tell you about that. Um in the same way G4 traditional Google Analytics would not catch chat GPT reading those pages. Right now they're giving you a little bit more than they used to. They're saying, "Yeah, there's some AI fetches." But because we're in the server logs, able to understand exactly when those fetches happen because Chad GBT and Perplexity and Claw tell you who they are and what they're using a page for when they go and get it. And our method says that about anywhere from 75 to to as much, you know, as like 85% of your traffic from AI never reaches your website. And that's a huge number. That's that's that's thousands of guests reading your properties at our most seasons portfolios uh reading multiple properties at a time in bursts uh that that that don't appear in your Google Analytics uh that are that are really researching what you have to offer without ever visiting your website.

So I feel like we we dove into the data and we we like I want to maybe step back a little bit in terms of and maybe some of your background and what Kismmet is trying to do. Uh because and so average listener, we've got property managers, we've got individual hosts, we've got investors, and there's going to be a mix of people that have their own direct booking website, and there's going to be a mix of people that entirely rely on the OTAAS, but ultimately, I mean, we're trying to drive bookings uh into our properties, and we want our properties to show up top of whatever search and and wherever the search is happening, whether it's showing up on the OTAA, whether it's showing up on the top of Google, whether it's showing up in a chat on on chatbt.

Let me answer the question on the history of a company with a personal story. And then I want to actually touch on that because I think it's it's really interesting. We've seen data that people using kismmet will impact their OTAA channels as well and we can measure that because we're integrated with the PMS. Um but uh so story of Kismmet, my background and it's really come full circle. I wrote my undergrad thesis on modeling out recovery and for sale housing markets back in the great financial crisis uh and actually went straight into economic consulting where I was running recovery models for uh different municipalities so homebuilders and investors could uh invest in these markets um and understand just how uh uh structurally you know this is a family show I won't swear uh just how structurally problematic certain markets in the country were in 2009. So I came to it with, you know, an econometrics background like, you know, we're probably contemporaries.

You know, I was doing STA back in the day. I since learned how

I still have STA,

okay,

on my computer.

No, I did not learn in R unfortunately. It would have been nice. Uh I I was STA, but you know, and now I don't have to do anything because Claude writes amazing SQL. I haven't written a query in, you know, two years. But the the the that was my background. um went into the hospitality industry, which is through a family business, which is a really interesting route, and always had kind of a very analytic bent on what we were doing, but built hotels, um invested in rental real estate, um built it ground up. And for a decade, uh, I I would be on in an asset management or an owner perspective report outs where the commercial team would tell me everything that was going on and, you know, different revenue strategies. And I would say, okay, but where's the bottom line impact of everything that's going on and I was so frustrated by the distribution world and how opaque it was and how hard it was to get in front of a guest.

So about three years ago almost to the day I made the decision to start Kismmit on the belief that guests searching via AI was such an important platform shift that all of your technology for how you read meet met a guest where they were and helped them find your property or whatever it was that you were offering someone to book would be completely transformed and that you needed a brand new set of technology to do it. So, what we started with was just organizing suppliers data for AI legibility because we didn't know whether it was going to be their AI or a guest's AI or whether you're going to have a chatbot. We didn't know how you were going to transmit it. MCP wasn't even an idea, but we knew we had to organize um long live data, commercial data, images, copy into something that was legible and operable by AI and then a lot of dynamic data. And that was a lot of the tricky part too. How do you in a second on a machine that can burst and ask 500 properties for a price at once?

How do you have the technology to answer that machine with accurate pricing and accurate availability for length of stay and all the rules that you have to compute? Um it was a really interesting, you know, uh uh a data science problem uh and and and really compsside problem uh that we had to solve to make that work. And and it's really only the last couple, you know, really since the start of the year that there's been a commercial application for what we're doing, which is how do you help guests find the right home? And then we work on behalf of the property management companies, the owners to help them distribute their homes and their hotel rooms uh through AI channels and traditional ones. uh uh uh we do we do ad buying, we do email marketing, but really using the technology to help them be in front of the guests who will love them most and and and build that relationship over time. So that's really helpful context.

Can you help me frame the state of the market today? So like we've got the OTAAS and my understanding is that Booking and Expedia are leaning really heavily into partnering with the like let's say Chat GBT Open AI whether it's through MCP or who however they're connecting on the back end so that their inventory is infinitely searchable and show up and that they can connect. My understanding is that Airbnb is still kind of like a walled garden. they don't really want their inventory showing up. They don't want it bookable. They want to I maintain their inventory on their platform and that I and they're not let's say playing as nice as the other OTAAS and then I want to break direct and into maybe the big hotel companies because like the Marriott, the Hiltons, they have the scale that they can partner directly. they have their own apps, they have their own infrastructure that they can maybe partner directly with the OTAAS and then maybe independents uh independent hotels and I would broadly categorize every single property manager as like equivalent to like an independent hotel operator.

That's how we think about it.

Yeah. So can you just walk through how you view the current state of all this and what is the and how do you see and LLM visibility for these different buckets?

Yeah. So I think it's really interesting just as a business nerd um that we have three OTAAS that are playing very different strategies right now and they're all competing at once. Uh, okay. You know, Airbnb has basically said, "We're going to have all the same inventory that everyone else has and we'll get there quickly, right?" Um, and you know, I I don't want to speculate, and they haven't said this publicly, but the idea that you're going to book a Marriott on Airbnb a couple years from now does not seem fanciful. Uh, and a rental car and an airline and everything like that, right? But Airbnb, I agree from the outside that they seem to be playing the walled garden game and they see AI as something that you will show up and use their AI and you will build a relationship with Airbnb's AI to plan your travel. Um, booking um seems to be dipping toes in a lot of places. Um, they are absolutely along with Expedia launch partners with chatbt and AI mode.

um the big brands are launch partners with Google for AI mode agentic booking via UCP. So yeah, the big guys are going to show up and they're going to play in these channels and especially when it's a paid channel, which is where they've always excelled, they're going to be really good at it and they're also come in with a huge advantage from lifetime of like SEO and technical SEO understanding how to make their offerings machine readable because that's actually what SEO was. I mean like we just we didn't we didn't call it machine learning. We called it machine learning. We didn't call it AI. But like you know Google invented AI for like you know basically how you search the internet. Uh and and and and incubated all of this stuff by the way that we're we're dealing with today. Um and then Expedia obviously has leaned heavily into B2B via MCP. Um and booking seems to be leaning heavily into B2B too with some of their reorganization.

So you have some overlap in strategy. You have two that are very clearly wanting to work with big top offunnel partners. Airbnb sees themselves as top offunnel technology that can earn the right for someone to log in directly and they've had that advantage with their their mobile application basically since they were invented. And then you have everyone else. So you got the big brands that are probably going to do it themselves and you have us and some other software providers that are that are thinking about this from a small and medium. And when I say medium, you know, we we work with companies that do hundreds of millions of dollars of booking value, you know, annually upwards of a billion. And so I think if you're less than a billion dollars, you're not building this yourself, like this is pretty sophisticated technology.

And the long tale of of of of you know, people out there that offer lodging, most of them are under a billion in in topline booked value. So, uh, we want to be the technology provider to help those guys be as good as the bookings, the Expedias, the Airbnbs from a technology perspective, and we want to give them the tools that they really haven't had over the last 20 years to compete in this emerging emerging distribution channel on on equal footing. Um, because if they're not if they don't have these tools, the middleman will capture all of aentic commerce. That's our belief. But if we can be out there and I'm long run bullish on the little guy, I think they can end up with a very different world in terms of percentage that is third party versus percentage that is direct uh in the out in the aftermath of kind of this two to three years of transition that we're about to go through. That's going to be pretty rocky.

Okay. So that that makes sense to me. So like if you think about what the advantage that booking has, it's like they they've got large structured data that they can feed to the LLMs and they can get and itineraries, they can get prices, they can get and everything that they would need in a clean structured way to and feed in the context and provide great and bookable properties that and are going to meet someone's need and then they have the ability to pay for that traffic, right? That that like, all right, you've got a search coming through. It meets this criteria. Like, I'm willing to pay $2 to make sure that I show top and that the Booking.com offering is what shows up top that someone is then pushed to book through. So, what what is it? And and maybe I'm missing something else, so you tell me. But what so is that what Kismmet is ultimately offering a independent uh uh property manager or hotel the ability to get great structured data into the hands of LLM?

Yes.

I ultimately being able to maybe pay for some of that traffic and and and [clears throat] what else is there?

Yeah. No. Um that's a perfect summary of exactly what we do. Um and you know we when I say our ambition is to give a uh a booking I mean I give an independent operator equal footing to a booking that's exactly what I mean. I think we take you to parody on the technology. Uh you know and I say that humbly because they have awesome big teams of both data scientists and computer you know scientists who are doing great work and have been doing this for a long time. But these are solved problems how you communicate to an LLM and we will give you that ability. We'll give you an MCP server. We'll build you that technology. Um and and we make it accessible because you need the platform to run it as an independent operator and it needs to be smart and opinionated in a way that allows you to just keep doing what you're doing because you are not a data scientist and nor do you want to be. No one got into hospitality because they love uh uh uh math um and technology.

They love they did it because they they want to be in front of guests and give an awesome give an awesome experience to the guests that they can find. But this is part of the business now. And so our job is to get you to equal footing and then help you explain why it makes sense to book directly with you or at the very least why it makes sense to book your properties. And a supplier will always have better information, a better reason to book direct. Uh it will have better videos, it'll have better content. It's just a matter of choosing how to expose it. And I think that's why I'm very bullish on the ability of individual property owners all the way up through large management companies to develop brands that can be in front of guests and and really create a reason for guests to book directly from them through their AI channels.

So, I want to push back on something and

have you seen the websites from most property managers? [laughter]

Guess what? Uh, you know, with with a with a with a $20 a month lovable subscription now, you can code up something that looks better than Booking's website.

I I I totally agree.

That's why we're No, but that's why we're publishing an SDK, right? That's why we have a builder and a template system and we will give you a website that looks better than 99% of websites out there for free because I really think that the ability to execute on website design has commoditized. Good design ideas, designing a brand that is appropriate for both your guests and your homes or your boutique hotel. That's your special sauce. That's your knowhow, how you want to talk about it. But you can use technology now to create gorgeous, awesome, immersive experience that's more true to your brand than booking will ever be because they have to be every brand at once. And that commoditization of that layer, the technology to build it and express yourself is why I'm so bullish on the ability for them to produce direct booking experiences that convert.

Okay. So what what I'm hearing you say is you agree and because of that like you're like oh we we not only need to make your website discoverable, we need to probably rebuild your website. And so that that

that's why one of our first products was a vacation rental booking engine that was just completely different that was mobile optimized that felt good. We we would turn on somebody's te tracking telemetry uh uh you know recognition tools and we'd look at devices and we'd be like listen no one converts on mobile and a majority of Airbnb converts in mobile why do you think that is and it's like all right we we can build this experience right and so we need to help give the tools to the agencies that larger brands work with to the individual little guy who wants to use Lovable to express why their homes are We we need to help those people create an experience that produces the trust level. And there are I mean, by the way, there are great firms out there, you know, and I won't name names, but there are a lot of ones that are out there and accessible if you get to 10 units to help you have a better looking website and a better booking experience.

I think that the opportunity to really do something special that expresses brand is going to be a big part of why direct starts to take off in the vacation rental space. independent hotels have always been able to leverage that, right? Um, and the best ones do excellently on direct because they have a much better feel and experience. And these are aspirational purchases. And, you know, I look at some of them out there and I'm like, wow, you know, these Aspen homes that I go for, you know, $30,000 a week on what looks like something that I would have built in 1996 on WordPress. Actually, it's a little early for WordPress, but you know, something that I would have built on Angel Fire in 1996, like, you know, in the copy serve era, right? This is this is this is not good.

Yeah. But you but just to be clear, you and you work with people that have their own site that you sort of hook into or like you know what like you're going to be much better if we just start from scratch and build net new and but it ultimately it doesn't and you might convert better if you've got a a modern better site but yeah

we're we're we're a tech company. Our focus is on the technology. We have several templates that can be incredibly modified very quickly and cost-effectively because of the changes in technology to express your brand. You give us like design direction. You give us a a brand document and we will put something out that like will knock your socks off. And and and we really believe we don't believe that form sites or or or template sites like and just lightly themed is the future. We think that everybody's going to run their UI not just on their website but in these MCP channels because you can do UI in MCP channels with their own brands feel, their promos, their landing pages, their uh you know their their their membership cards and what it feels like like you're going to express you know down to the individual small property manager what's special about what you do. Uh and I'm excited about that future.

So I when doing prep for the interview, I was I was looking through some of past posts and things like that and correct me if I'm wrong that you guys sort of started at least initially like looking at hotels whether it's pivoted or or not towards short-term rentals. So maybe explain to me like what's the difference here between hotels and short-term rentals and like what why do you think the opportunity is better for short-term rentals and and and the current focus?

Okay, so um this is a fun story. We for years had been getting inbound from larger rollups of property management companies, usually PE backed, that would call us and say, "Hey, we found you guys. There's no one doing this for vacation rentals. Would you be willing to do this uh for us?" And, you know, I'm getting, you know, top down pressure. We need to put AI into our systems. I don't know where to go. This seems interesting. And I'd say like, yeah, yeah, yeah. Look, look, I'm a hotel guy. I came from the industry. Um, this is who I know. This is my network. You know, I feel comfortable operating here. I've been looking at, you know, these report outs for the last decade. This is this is good for me. I'll build like a vacation rental object in the back of the uh in the back of the database and like, you know, we'll play with it at some point. And really at the beginning of this year, I got a a whole burst of inbound from vacation rental managers, including one who I think might have been shopping me and seeing whether he could uh he could build it himself.

And uh and I've been talking to investors about it, and he came back to me and was like, "Oh man, you guys are on something. You're so far ahead, but you got to do vacation rentals. If I bring you a manager who will give you their guesty API keys, you know, would you would you try it?" Right. And I was like, "Sure, I I'll spend a weekend. I'm going to build a toy." It was not using rates. It was just like semantic search, the ability to look across vacation rental inventory. And this is a small portfolio, about 130 units, right? Not a small portfolio, a very midsize, but a good portfolio. And the question was whether it was a much better experience to look across homes if you semantically embedded and did some vision models against what they were doing. And it was a light bulb moment, right? When we were working with hotels, even big resort hotels, you know, 30 rooms, all somewhat different, you know, 30 room types is is is the highest end of whatever we worked with.

Usually in the teens, the ability to search across an individual properties inventory wasn't that helpful. But in a brand level in multiple locations, it was very helpful. But the brands tended to have incredibly large fragmented tech stacks that would be very hard to do kind of the fullfunnel telemetry that we do. that makes this magical. Um whereas in the vacation rental space, there wasn't that tech stack that we would have to come in and you know deal with because they were running far fewer pieces of software. Some of the hotels were like 15 to 20 pieces of software just in the commercial space. But the inventory was amazing to AI. The moment that AI was interacting with it, it was making much better recommendations. We were we we had this like toy model where it would where it would flip through the carousel of 130 images and pick the exact one of the hot tub, you know, if I was searching for a hot tub.

And you know, the average hotel room had like eight images, right? Like this was such a rich fun data problem that it was a Eureka moment. The more differentiated the product it is, the harder it was for AI to use the open internet to find out about it. But the richer the product was, the more impact making it legible was. And so we immediately like wrote up a deck, went to 10 managers, were like, "Do you want to be in the our initial cohort?" Seven signed and I was like, "Okay, this is this is this is this is Kismmet to, you know, use our our company's name like this feels right." So it's been an explosive growth. We'd only launched our first vacation rental manager um in April, but we have dozens signed and tens of thousands of units now in the pipeline that we'll be releasing. It's a very exciting time for us because there's real need and impact that we're able to demonstrate with a catalog like a vacation rental manager has.

And the ability to take something that wasn't in an official catalog format that was legible at all to a machine has downstream impacts on AI visibility, but it also means that we're able to impact, you know, Instagram and help them run catalog based ads and none of the vacation rental managers were doing that, but a lot of hotels were. So, so now we're able to do so much more just by having good structured commercial data and really organizing things in a different way that uh the the opportunity to really impact these guys. We're seeing you know massive five figures of impact even on a small small portfolio, five figures on a monthly basis. So, it's it's been really great.

Yeah. So hearing you say that I sort of in channel channeling the excitement I hear from the Airbnb team of like how AI is going to change search for this entire category, right? Like obviously they think they can keep it in and that that like the AI mode of search is going to fundamentally change how someone finds their lodging on Airbnb and that ultimately the AI whether it's search and and bringing it forward the visual aspect of it and will fundamentally increase conversion. So I think I hear the same conviction on your end, but that a lot of that conversion is now going to happen through chat GBT. But ultimately, I mean, you're we've got to make the leap between the LLM itinerary planning and finding the inventory to then converting on the direct booking website. So, can you walk me through one, how do you think search is going to change for short-term rentals on with AI? And then two, how does the leap happen from in chatbt to to direct booking in this new form factor of of searching and finding short terminals?

Okay, so there's 100% alignment between me and the Airbnb team that this is a transformational way to search for things that were really really hard for a machine to understand before,

right? Like it all the information about a short-term rental trying to push it into like 10 filters and and amenities and like I it's it's it's a fundamentally really tough question of like how to find the right short-term rental for the type of trip that you're doing, right?

Yeah. So there are two things though that are downstream of that as you've correctly hinted at, right? The first is um Airbnb has what 7 million homes is what they quote. We do not and we're helping the direct channel. So if you're talking about even an individual manager, you know, even if you have a thousand homes on your portfolio, searching across your portfolio alone, it's a better experience once they're at your portfolio to use AI search, but you also have to be able to be searched top of funnel too, right? And Airbnb has aggregated all the inventory. So our bet is simply that the AI models which are actively out there ingesting individual websites training against them using their search indexers and using them right now to answer questions and we measure that and we see all that activity. We think they are in of themselves an aggregator of what is available. And that means that you can take advantage of technology to be aggregated in a direct way the exact same way that Google search or or paid Google aggregated you in a way that could end up with a direct conversion.

And we think there are other secular trends so to speak, you know, good trend lines for why direct ends up being something that is more formidable. But again as I said earlier we can measure the impact across channel because we are integrated with the PMS and we believe that the behaviors that lead to good AI visibility actually impact your transactions overall on other channels as well even paid channels you know even third party channels because some guests look at your website or read your your properties in chat GPT and just check out on Airbnb and that's good too like you want the revenue that's what's most important at the end of the day. So, I don't think you have to believe that all or none of traffic will eventually go through a direct channel. Like, I think that's a bit of a a red herring here.

You just have to believe that you want some direct business in the future. So, set that up. And the things that you do that impact your direct business are also going to impact your performance in thirdparty channels, too. So, it's it's it's worth it to make your homes really visible. So, that's how I'd answer that question. I got one more question on high level and then we we'll go back into the data. Uh because I I I definitely wanna and and bring it back to where we started. Just the the whole OTAA direct debate. I want to get a sense of where you're at like the OTAAS it and and and Airbnb booking Verbbo. Um it uh they've seen enormous growth over the past decade, over the past year, over the past five years. Like any way that you want to measure it and they've been doing really well, uh their stocks and support it uh they're and you look at their growth of bookings like in any which way like I can measure the data like it appears that overall industry is growing.

the overall share of bookings that the OTAAS are are getting um continues to grow and it's generally at the expense of direct and whether it's and people that used to call and book their vacation rental are just going onto the on Airbnb or Verbbo or whe whether it's convenience of and and traffic's moving to mobile and most people's direct booking website on mobile are are terrible or whatever reason it is and when I think about Yeah. Yeah. [laughter]

And and then thinking about and and AI and it it feels like that's only going to accelerate that adoption or push just because and I I see the integrations with booking and expedia and that they're going to and drive and whether it's and they're in bed with Google and Google pushes to booking and open AI has an incentive. They've got to make money and they're going to push track to booking too. So am I wrong on that thinking or or what is your approach?

So like there's a somewhat flippant answer which is that you know if you monetize through ads that are based on an auction you want as many biders as possible. And so if you push all the traffic to booking like you're going to hurt yourself in the long run uh if you're an ad platform uh that works on an auction. So the little guys is going to have to win every once in a while and we're the technology that helps them manage their ad spend to to do that. So there's a place for us in that world. But as a second answer that I think is a little bit more inspiring, a little bit more exciting. The problems that the OTAA solved, Booking is a web one technology Airbnb is a web two technology. Bookings embraced web 2 technology as well, but their website was where booking became booking and Airbnb was amazing at mobile where these were awesome platforms that changed how people wanted to book. And as you've pointed out, the trend line has been away from phone, which really relate which which 100% helped the little guy, right?

Because Airbnb could never be as good on a phone call as an individual property manager, an individual property owner. It's moved toward areas of convenience that meet the guest where they like to plan, right? And mobile is an awesome planning experience and has been for a decade. Mobile is not going to be the interface that I plan travel on five years from now. I'm certain and and to the extent that you believe me on that is the extent of how quickly you should adopt our technology. But I am

so so explain that like we're both millennials.

Yeah. Like I mean I still pull out my laptop when I go to book a trip. Like I haven't I haven't adopted mobile then and I see my younger brother that's and 10 years younger and and he books everything on his phone. So like and where is travel booking going?

It's going into AI and and what that means is not necessarily chat bots in chat GPT or claude on your mobile application.

It's going to be everywhere. AI is going to diffuse. It is a platform in of itself and it will diffuse through devices throughout our life. And the same AI that's in your laptop is also going to be in your phone. It's going to be in your smart speakers. It's going to be in your television. It's going to be in your car radio. And your ability to say to AI as you're driving along, hey, what was that place that I booked at like six months ago that I really liked? Is it available next weekend? You're talking to your wife and your kids and you're like, we really want to get away. like your phone is going to be able to connect your your your phone or your car speakers are going to be able to connect with your AI look that up answer the question and tell you whether it's available and so that interface that change in platform means there's a whole new way that guests are going to interact with what it is that people sell the cataloges and the options that can be booked and and if you believe that that's where this all is going that AI that's connected to your calendar and your phone and your iMessage and ability to talk to everyone in your life is going to be what helps you plan and execute travel, then the ability for an individual supplier, an individual property management company or small hotel to put what they offer into that interface is an existential question for their direct booking channel.

But then the direct channel can give you that answer on which property you had last time. And it can offer you uh hey a returning guest benefit that's different than what it can get from somewhere else. Uh what that AI could get from somewhere else. And the AI, they're tireless. They're going to check for you. They're going to go look at booking. They're going to look at Airbnb. And they're going to look at the direct channel. And they're going to put all three in front of you and say, "Hey, I checked. you have a membership at this, you know, property manager in in Denver that you you always go visit and like, you know, why don't you go book through them and uh, you know, they they'll leave out, you know, your favorite, you know, cheese that they that they that you mentioned in a review because you bought it at a cheese shop nearby, right? Like that's where this is going. And if that's where this is going, then I think direct has a huge advantage.

And and Airbnb can play a different advantage there. They they they've seen different things than a direct supplier will ever uh uh uh uh uh see and they can offer certain things and if Brian wants to build an ad network, he can build an ad network, right? Um but the direct guy, they're going to they're hospitality is about connection and and if you if you want to play that game and you want to play in that pool, it's going to be easier than ever. Um, and I think that the way we book through AI is gonna fundamentally open an opportunity for a different type of hospitality. And that seems counterintuitive to a lot of people because they think of AI as like the Terminator, whereas I see it as like a helpful assistant that lives, you know, in your phone and in your computer. But that's what booking is going to feel like in the future. And if you believe that, that's part of the reason that I'm bullish bullish for direct.

So, so let's say you're right and that we're the form factor is going to change of how we search for travel, how we book travel. Now, let's relate it back to where we started on terms of measuring it because we all want to measure where our bookings are coming from. We want to make sure we're on the channels that and showing up in the places and if we're not tracking it, we're not uh we're not understanding and how we drive our our business growth. So maybe help me understand what what are we measuring now or what is the average property manager seeing today? Uh and what do you think they're they're missing and and how would you be tracking it going forward?

Yeah. So, this is part of the reason why I think 1% of room nights is maybe a self-motivated reporting stat where the street is nervous about what effect AI has on booking an Expedia and it behooves them to talk about 1% of nights. But here's things that we see. So, and we see that AI delivered guests convert at one and a halfx the rate of an organic search from a traditional channel. And that may be a flaw because right now Google doesn't break out a guest who arrives from AI mode versus a ga guest who arrives from traditional search. So it could be even higher, right? And that's because the guest who arrives from one of those channels is better informed, has done their research, learned about the homes across managers before they ever landed on the website. So one and a half the conver uh times the conversion rate. But then we see a 2x basket size versus organic.

So when you add it all up, we think that the average guest who arrives to your website via AI is worth $52. And this is just in our portfolio. Guest who arrives via organic is worth $18. And a guest who arrives via traditional paid search. Now, this is a mix of people using our technology to buy the ads and overall is maybe worth as little as $13. And at traditional pay-per-click, you know, you're scraping exactly where you want to be on rorowaz and what the industry has accepted historically at $13 per paid guest uh arrival.

But if organic AI and the paid AI channel is just starting to emerge, but if organic AI is is worth 52, that's a huge value difference, right? And and that's because of the trust that the AI builds of like, hey, I did all these checks. I I did all this analysis and like th this is the perfect spot for you to book and like the the propensity of someone to actually then convert is that much higher.

And then the then then the easy explanation on basket size is people spend more time researching bigger purchases, right? So they're going to spend more. They're thinking about planning a whole week and they're doing it over time and we're able to measure that. And so when we add that all up as a percentage of direct revenue on our most seasoned portfolios just looking at the the chat GPT of it all and it's like 97% chat GPT that we measure from AI search. So that's where you're playing. We think it's as much as 16% of your direct business right now in terms of revenue. And that's ignoring AI mode which we think is probably another 15 16%. So that's a long way from 1% of room nights. And it leaves you to ask two questions. There's really two possible explanations, right? which is one, the way that we're measuring and modeling conversions would not work for booking because booking already is so the default option and it's such a default behavior to search for something for booking and open your booking app on your phone or open a new tab and go directly to booking that they're not seeing the incremental benefit from AI.

That's one possible explanation. The other possible explanation, and realize it's a biased sample because it means that we've implemented a lot of technology on their website to make them more visible to AI. The other possible explanation is that AI prefers direct. And if that's the case, if organic AI does prefer direct, there's an interesting story there, too. And again, I don't know what it is that they're measuring out of the server logs over at Booking or Airbnb or Expedia for that matter. I can only report on what we see. But the 1% of room nights, if you measure via just UTMs and referers, we'd get the exact same number. But if you model it out, and remember, you know, 75% of AI traffic arrives dark, right? These are people landing on individual property pages on Safari who would have potentially typed, you know, a 90 character URL string into Safari.

Like I find that hard to believe, but we can measure it. We look at device. We look at, you know, whether AI previously fetched that room, you know, that that specific property or that specific room before they do that deep link. When we use traditional statistical meth methods to to to measure that impact, we're seeing massive impacts already. And we believe that there is massive amounts of AI traffic that is already making a huge difference for some of our managers in the types of revenue that they're booking direct and the type of revenue they're booking overall because it's cross channel impact. No, and I I could I I can see that leap, right, of like if you're measuring travel itinerary planning and starting the AI like and a lot of research now shows that and a lot of people are starting that way and then ultimately you get to a direct booking site and whether you open an app, whether you did a and and you you're going to get to it somehow and and you start with the AI and end up getting there just like and we're probably under indexing and what impact Google was having on overall travel as well of like, oh yeah, I'm starting my search on Google and and maybe I click on an ad, maybe I and open a new browser eventually and go to a website or go to booking.

Like it it's just the the front door of travel is changing,

but Google has always taken credit for that, right? They model that and and and

well, it's it's in their interest to take credit just like it's in bookings interest to

but it's right for them to take credit, right? you know, at uh and you know, things like that that happened in iOS like really changed the game. Safari eliminating third-party cookies, you know, Google, of course, for self-interested reasons maybe walk back how how how third-party cookies work in Chrome, right? Um and and and even first-party cookies, right? So you do need to model some of these conversions like it's not clean and these are accepted methodologies from everything from TV advertising all the way to you know instore purchases that are affected by Google search and and it's something that people have been comfortable with in e-commerce for a long long time. It's how the biggest brands measure their impact and marketing will always be a science that involves some sort of modeling and the question is what you're choosing to measure and my argument is simply this.

If we measure self-reported surveys that say that 60% plus of guests are starting their search in an AI context, then only 1% of bookings, I'm sorry, only 1% of room nights being impacted by the AI search seems awfully low.

Yep. So maybe starting to wrap up um and because I I totally agree with you and and it's probably just misaligned incentives of like and and wanting to and not scare investors of like hey like our our top funnel is changing and and verse and and and our bottle funnel of like and how we're actually what we're actually converting through and that we still control our own traffic. So what is it that property managers should be thinking? What do they should be doing today and to make sure that they can show up and better in terms of and and uh search?

Yeah, I honestly like this I don't want to just shelf for kismmet. There are other people working on this problem but we have a specific way of attacking it that is very individual brand you know oriented right individual property manager individual you know collection of vacation rental properties individual small hotel we believe that power exists at that level now and so our technology is built for helping you do that at the individual level but there are other technology providers out there that are taking a different tact on this. The point is this is all very technical. They're now competing with the bookings of the world and their team of data scientists, computer scientists. So you should probably leave this to the professionals. Don't vibe code up your ability to distribute into AI. Where you should be vibe coding and where you should be spending your time and you should be working with agencies that are really great at this is building your front end to be something that converts because you are going to have the opportunity to drive traffic to a direct channel and the quality of your website is 100% uh absolutely the most important thing that is going to is going to control that conversion rate.

I am only as good as the conversion rate that your tra your website will allow. So, it better be really good-looking. It better be really functional. And we'll give you some technology to help with that, or you can do it yourself. But like, I could send you all the AI traffic in the world, and you're limited in your capacity for your ability to tell your story and say why your brand and your hospitality is perfect for that guest. So, spend your time there. Go to a professional to help you find more traffic because there are ones out there. And it's far too technical for you to do on your own. But, focus on what you're good at. Provide good hospitality. Tell a story about why your hospitality is great. Invest in your imagery. Uh get new pictures, write good copy, have good, you know, blog posts, have a beautiful homepage. Um and have a really nice booking engine. And the opportunity is is is is there for you.

You can take it. And the best performing portfolios, the ones that have invested in that historically that we're working with are absolutely seeing results. Um so time is now. No, that's I totally agree. If like you think of the ones that are if I and you can make the argument that AI is going to like direct like you'd think all of us are equal and it would push you direct and if some number of listings are optimized towards showing up and you can make sure that you're there like and all you have to do is out and visibility your your competition, right? like and and if you can do that, then you're going to get more than your fair share of bookings and ultimately that's what we're trying to do, right?

Yep. They're all competing for for a little bit of share of revenue and to find the guests who who will love what we offer.

Awesome. Well, Jason, if people want to reach out, if they want to find you, uh what's the best way to do it?

Yeah, you can, um hit us up at kismmit.travel, uh is our URL. kisme.travel. I am at jasonkismmit.travel. I'm gonna really regret that I put that on the podcast in like three years when we're gigantic. But, you know, uh uh that is a future us problem. I'll change my email then. Uh or I'll have an AI to read it all. Right? Or you can hit me up on LinkedIn. I'm on LinkedIn. Uh I I'll I'll take connection requests. But yeah, we're we're publishing information on our website. You can schedule a time to meet with us. Like implementation, we're getting really fast at it and we we always want to talk to people who are are looking to solve this problem. Awesome, Jason. Thanks for joining,

Jamie. Thank you for the opportunity. Appreciate it.