--- title: "2024: A Slow Goodbye to Third-Party Cookies" description: "Advances in first-party data collection, the new advertising techniques emerging to replace third-party targeting, and where the deprecation timeline actually landed." series: "Blind Data" episode: "4" year: "2024" speakers: "Kelly Nyland, Ryan Hornberger" words: "15339" author: "Kelly Nyland" url: "https://kellyny.land/transcripts/blind-data-4-a-slow-goodbye" --- # 2024: A Slow Goodbye to Third-Party Cookies _Blind Data · Episode 4 · 2024_ > Advances in first-party data collection, the new advertising techniques emerging to replace third-party targeting, and where the deprecation timeline actually landed. **Speakers:** Kelly Nyland, Ryan Hornberger --- **Kelly:** Welcome back listeners to the Blind Data Podcast series. I'm your host, Kelly Nyland, founder and CEO of Whym. Again, I am joined by Ryan Hornberger, VP of Engineering. And today we're discussing the current landscape of web browsers and recent changes, specifically as it relates to cookies and data collection online. Just a quick recap that the last episode explored motivations behind key players in the landscape, including Apple, Google, Shopify. Amazon and Facebook. Today we're going to focus on advances in first party data collection this year, development of new advertising techniques and strategies to combat this deprecation of third-party cookies, and the evolution of privacy compliant tracking methods online. So welcome Ryan, here we go. **Ryan Hornberger:** Awesome, let's get going. **Kelly:** So it's been a couple minutes since we had our last episode. How have you been doing, Ryan? Any latest insights on what's going on in the landscape? **Ryan Hornberger:** Well, I have been doing a lot of reading and podcast listening and that kind of stuff, just watching how the industry is responding to everything. Google is 100% launching their cookie death scenario starting in January. So we're going to be very exciting and interesting time here as this finally rolls out, but it's definitely not going to be a fun time for a lot of people. **Kelly:** Yeah, January 2024 is when we're expected to hear the results of the antitrust trials. There's actually quite a, there's a series of trials going on right now and we may end up doing a special episode. just on those antitrust trials in the new year, depending on what the results are and if we have any insights or maybe, you know, if we have something special to share. Today we're going to dive right in. We're actually going to start with Google and Facebook's approach on the advertising side and we're going to talk specifically about the new privacy practices. So what's interesting about... how we're going to be moving forward here. And in terms of how Facebook and Google are going to continue to make money has a lot to do with the fact that they've already collected all of this data about us for years and years, basically two decades or more. And they're using that data to now create these prediction models of how to advertise us. So the bottom line is they have so much data that they believe that their advertising businesses can survive knowing what we're going to do before we even do it, essentially. Right, Ryan? It's pretty fascinating. So we'll talk a little bit about maybe sentiments around that, but specifically the new Google practices have to do with a couple key factors here. Number one, **Ryan Hornberger:** Yeah. **Kelly:** is in the browser as cookies go away. We know that what they've said is that the data will now stay on. your device. So we're going to talk about what do each of these things mean. So the data is going to stay on your device. The browser is now, meaning Chrome, Safari, you know, specifically we're going to talk about Google, but the browser is going to sort you into topics. So topics could be something like, you know, Kelly loves athleisure, I'm a yoga fan, or I'm a young conservative, something like this into these topic categories. So they'll know which categories you are in based on your search behavior, but they will not necessarily have any first party data or know exactly who you are. This is this is really the new mantra. And then, and so we're gonna talk through each of these things, like what does this really mean? What are these changes? So let's start first, Ryan, with what does it mean when we say that data stays on your device and how is that different than how the data was treated, you know, previously? **Ryan Hornberger:** Yeah, the articles that kind of kicked off this plan are like announced this plan around what Google calls the privacy sandbox. If you read the initial article, it basically says we're launching this technology because we want to be tolerant to possible incoming regulations and things like that. That's like the middle of the article basically talks about how other strategies are not very good at solving regulatory problems. kind of bulletproof. So this is what they have been going with. Now the technology that Google is launching are not launching in any of the browsers. So Firefox, Safari, et cetera, et cetera. They are still all in a world of their own making around this kind of protections. But Google's technology in particular is putting this thing called the Google Privacy Sandbox into the world. And when you're on the Google browser, a number of the technologies that originally ran on Google servers as Google built indexes **Kelly:** Mm-hmm. **Ryan Hornberger:** what they thought your interests were and those kind of things. Those technologies have been migrated into the browser. And some really smart engineers spent a lot of time trying to figure out how to make this as, um, Defensible as possible. In other words, uh, protect your privacy as much as possible without dropping Google's ability to make, to, to deliver ads. The big shift here though, is that a lot of these key technologies used to rely on third party cookies now because it's. those have gone away, only your browser can know this information. So only your browser is going to be able to develop strategies and thoughts and technologies on around how that data might be used or put together. And that means anybody who's outside of the browser maker is now cut out of the process of innovation, cut out of the process of ideas and working on figuring out new models for all this. Only the browser will get to innovate in this space now. And so it's... the two-edged sword. On one hand, Google's saying you're going to be more private, and on the other hand, everyone's saying, but that means Google's the only company that can hang on to your data now. **Kelly:** So I'm just curious for like a non-technical person, Ryan, what does it mean for the data to go from a server basically into the browser? How, what type of technology can we use to actually basically create this wall, right? To say, okay, we're not even gonna be able to see beyond, we're gonna intentionally make a wall, we're not gonna be able to see beyond the wall. That sounds like a little bit of like... like a little voodoo-y, right? Like how do we actually make things so that we can't actually look at them? So can you explain a little bit more why this is significant in terms of privacy and how we've intentionally actually built walls that we can't see around? **Ryan Hornberger:** Sure. So historically the way that advertising technologies, including Google's worked is that your browser just had a number or an identifier of some kind that followed you around the internet. And every time you visit a different website, you just say, hey, here's my number, here's my number, here's my number. You're just constantly giving out a number to these advertisers. And you didn't create that number. That number wasn't shown to you. We called that number a cookie, all these hidden terms. And it was generated by the third party, for the third parties. And so All that number did was say, hey, this person is the same person over and over and over and over again. That way, as you went around the internet, they could track you. This is where you were and this is where you were and this is where you were and this is where you are. Then at any moment, these companies could then gather other information about you. This is your name, this is your phone number, this is your email address, and then something that can develop this portfolio. This number represents this person and they carry it around. So the key difference here is that number is going away. But now that they can't identify you, when you're visiting site to site to site to site, they can no longer come up with strategies on how to present to you the ads that they think you'll be most interested in. And so what they're doing instead is they're saying, hey, imagine that I was an advertiser who wanted to talk to people who are interested in cat toys. Can I just tell the browser, hey, I've got ads in... **Kelly:** Mm-hmm. **Ryan Hornberger:** Cat toys, would you like to take one? And then the browser can say, I'm a cat toy, interested browser, yes. And they can kind of deliver that ad that way. They're doing what's called putting the auction into the browser, which basically means that a number of ads will appear to the browser and say, hey, any one of these could be the right ad. And then the browser can make a determination, this is the ad for the person. And then it will make that decision inside the browser instead. **Kelly:** Got it. **Kelly:** So I'm really curious and then we'll move on from this topic, Ryan, but is there a way for me to try to trick the browser right now? Like, could I open a tab? Could I open a new window? Is it based on an IP address? Could I go from like my desktop to mobile? How could I, you know, once these changes in this rollout happens, is there a way for me to test whether I get the same response for cat toys? it how would I what would I do in order to see how this works? **Ryan Hornberger:** So part of this, part of the problem presented by third party cookies has already hit the market because you're on multiple devices already. So you might find that depending on how you're using your browser, that if you are doing certain behaviors on your laptop, sometimes then you move to your desktop and it already shows you an ad from the same product you're looking on your laptop just a minute ago. Sometimes depending on the scenario, you might, they can't make that connection. If you want to really prove they can't make that connection or actually enforce it, you can use a browser other than Google immediately. So if you use Safari, you'll see that connection kind of breaks between device to device. Or nowadays in Google Chrome, there is a setting that says disable third party cookies for all sites. And if you turn that setting on, you'll start to see exactly what this world looks like for everyone after January, or you wait till January and it's just going to happen to you. When it happens, you the ads in your space. on the websites you're visiting, you're going to see something different about them. It's going to be hard for you to articulate or understand exactly what has changed because while you are observing those changes, there's thousands and thousands of people around the world trying to find alternative ways to give you valid ads. And every company is coming up with different answers and different strategies and different partnerships to make that happen. And so you'll see some stuff work less and some stuff work more, but you will notice difference if you turn off your cookies. **Kelly:** Got it. **Ryan Hornberger:** It just depends on how well you'll detect what those differences are. **Kelly:** And I definitely noticed that right away because it will require me to sign in or any cookie settings that I've already toggled on are no longer valid or everything from like little things like my carts are disappearing. So let's talk about interest categories. Ryan, is there any more danger in the way that we're now getting sorted into these interest categories versus how we were sorted before? **Ryan Hornberger:** Yeah. **Ryan Hornberger:** Well, so before and after, by this point, pretty much every major advertiser was using some form of AI, plus algorithmic strategies to kind of come up with categories for you as an individual and historically those buckets were managed by advertisers, depending on the platforms you're visiting. So if you're visiting, let's say the Wall Street journals and the CNNs and the other publishers of the world. they might partner with one advertiser and that advertiser may or may not be Google or they might partner with Google or they might partner with whoever. So depending on the sites you're visiting and where you're going, the platforms that are secretly behind the covers powering everything that you do or everything you see with relation to ads, they all had very unique solutions. Now, as this change rolls forward, we're quite confident that Google has solved **Kelly:** Mm-hmm. **Ryan Hornberger:** their situation. So anybody partnered with Google is gonna be happy, they're gonna continue to work. They might have some rocky roads, but they're not gonna have nearly as rocky a road as these other advertisers who are not the owners of the browser and who do not have control of how Google's technology empowers their business. And so anybody who's working on those companies are going to need new strategies. The most... **Kelly:** Mm-hmm. **Kelly:** Mm-hmm **Ryan Hornberger:** detectable strategy you're going to find and you're already seeing is that everywhere you go is demanding an email address from you as soon as you get there. And the reason they're doing that is because the email address has become the final thing they can use to figure out who you are and the, a big percentage of the internet is publications, putting out free content that is powered by ads. And those ads used to deliver a certain amount of value for advertisers. **Kelly:** Mm-hmm. **Ryan Hornberger:** Now those ads are not going to deliver nearly as much value unless they know who you are. So you're getting asked to show your, to identify yourself immediately and early just to use their sites now. **Kelly:** Mm-hmm. Got it. You talked a little bit about they're using a combination of algorithms and AI to create these predictive models. Are these predictive models new, Ryan? Are they just being used more? Can you talk a little bit about what is different? **Ryan Hornberger:** 100% not new. Yeah, they're not new. And for anybody who is not Google, their technology is gonna continue to evolve exactly how it was. Although key technologies that were really dependent on incoming data coming from cookies have probably been deleted. In fact, you'd probably notice if you're an advertising company who's paying for ads on Facebook or Google or other places, when you go to your platform and say, hey, I wanna put an ad out, the number of features that are available to you. to do that, how to target and those kinds of things. A big chunk of those have disappeared in major platforms primarily because the technologies, whether they were algorithmic or AI driven, a big percentage of those needed third-party cookie data to make them effective. And now they don't have it. So these platforms are coming up with new strategies to deliver similar results without those same technologies in place. If you're a Google, and you really thought this particular technology was working really well, then you just move it into the browser now. And the browser does the exact same work it did before, but the AI is actually running on the person's local computer. And then you just have to figure out how to communicate with it in a way where it doesn't reveal the user's identity as you're trying to make to leverage that AI benefit. **Kelly:** Got it. So not necessarily tons of new things. It's really just sort of where the information originally was either being pulled from or hosted in the server component. And now. We've created this intentional wall that we can't see around when it comes to the user's privacy or first party data identifying them through using this number. And we've brought all of that bidding process, silent bidding really, into the browser. And... Okay. **Ryan Hornberger:** That's what's marketed by the way, whether or not it will succeed. That's, that's still a, yeah, huge question. **Kelly:** So last quick question, Ryan, what do you think are the different ways to think about how all of our data has been used to analyze and basically create these predictive models? What do you imagine are the different opinions here on how people feel about this? **Ryan Hornberger:** Actually, yeah, I'm torn opinionated wise about this. So from the one hand as a consumer, I'm not a huge fan of all the tracking that happens about me on one hand, right? And in particular, it is kind of weird that I'll go look at something on one, like I'll look at a product on my laptop and then I'll walk to my phone and send an ad for that product is appearing for me. That does come off a bit strange. How did the publisher know merchant that was just showing me that product was just showing me that product. When you actually dig into the technologies that power it, it seems a little bit less scary, but it is just, it does, it does come off weird. And I, and personally, I've, as a consumer, I've felt concerned about it. But as a business builder, as a business owner, as a person who's trying to get my product out in the world, the last thing I want to do is spend $50 per person just to bring them on to sell them a $50 product, right? And so the last thing I want to do is be sending out ads for, let's say, nerdy technology toys to my daughter's soccer team, right? Like they don't care about my nerdy stuff. Why am I showing them those products? So I do on one hand, feel concerned about the privacy. And on the other hand, understand that it is better in a world where I see ads that actually matter to me. And then the final point of view, this is that the big part of the internet is all the free content, right? If you had to pay for every article you read, if you had to pay for every YouTube video you watched, if you had to pay for all those things, at some point you'd stop paying and the amount of information you consume would go down. That may or may not be good for your health, but definitely from my point of view, like if I need a piece of information right now, I want to go find it, read it, and move on with my life. **Kelly:** Mm-hmm. **Ryan Hornberger:** and not stress about having to pay 25 cents every time I read an article or 99 cents every time I read an article. And so I do personally at this point feel that the advertising industry needs to be protected because it is what powers major parts of the internet that I use every day. **Kelly:** Yeah, and in some ways we have been paying with our own personal data, but we just haven't been aware of the exchange or maybe the value of that exchange. Okay, well, great. **Ryan Hornberger:** Yeah, honestly though, the information they're collecting about me is not terribly crazy, right? Like, this is Ryan is basically the majority of what they need to know. This is Ryan. He looked at this product and he's now looking at this website, show him this product on this website. That's basically the connection that they're storing, but they're not like recording my phone calls. They're not looking into my medical history. I'm sure some companies out there are wanting to do all that kind of stuff and maybe doing that kind of stuff. **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Ryan Hornberger:** But at the surface, that's not what third-party cookies power. Third-party cookies simply just power that this is Ryan and now this is Ryan. And that's happening outside of that is probably where the real creepy stuff happens. **Kelly:** Got it. So in some ways we're about to go to a place where we may not have as much personalization because now there's going to be limiting factors on how many connection points or cross domain data. Like as we go from domain to domain or device to device, um, it's the, the picture of us is going to be much more blurry. Um, and that probably means at least for the foreseeable future, potential little like step back here in terms of personalization, unless we find a way to share a more holistic picture with these third party advertisers or content creators, content publishers basically online. So we're gonna talk a little bit about how these new features within Facebook and Google are, there's some storytelling around how these new features work and we're just gonna maybe... **Ryan Hornberger:** Definitely. **Kelly:** on a few of these new features within the advertising platforms so that we can understand a little bit more. Recently, Ryan, I met with some of these different agency leaders at these agencies that focus specifically on, let's say, Google or specifically on Facebook. What I understand about a lot of these new changes is that... we're kind of, these platforms are now basically saying, give us your creative, give us your headlines, give us your products, and we will essentially go out and find your audience, right? So it's that kind of saying through these, the different way that we have for identifying. what's in your creative and for, you know, the contextual analysis of your titles, your description, you're basically saying like, now we are smart enough to know if you just give us your creative and give us what you're trying to sell, we can actually go out and find those audiences for you. And this has been being done under this privacy umbrella so that you can't go in and say, I want to target this particular age group and this household income in this demographic. Some of these levers are still available, but what these advertising platforms are saying from a big umbrella perspective is that we know how to find your customers. Just give us your stuff and we'll do it for you versus you having to go in and program **Ryan Hornberger:** Yeah. **Kelly:** and sort of pick. So we're gonna talk a little bit about customer match and pair, which is this new, basically some of this new information for Google. And so can you walk us through a little bit on both of these technologies, right? And customer match and this pair, which is a capital P-A-I-R. **Ryan Hornberger:** shirt. **Ryan Hornberger:** Yeah. So Facebook, they, uh, they, they and Google are coming at this from different approaches, but both of them are employing AI to power their solutions before. Now that what you were talking about at the earlier part of this, uh, section was, um, a technology that Facebook is pioneering where you no longer give targeting configurations. You know, for those of you who have never done advertising before, what people usually do is they go into an advertising platform and say, I want to target this audience and they target a slightly different audience. the exact same ad, but a different configuration. They'll target, they'll do like four or five different targets like that. They'll run the ads and then they'll see which one of these campaigns did better. They'll delete the four that didn't work, keep the one that did work, and then try some new variations and they'll keep on making those changes over and over again. And that's the job of your agency, is to basically keep on tuning the advertising until it gets exactly the outcome you're looking for. And the platforms supported these agency operators or just even individuals who are doing this on their own with a number of tools. But those tools also required a heavy interaction with the advertising employee or agency. Then with all these changes to cookies, a lot of those targeting capabilities had to be adapted to not have third party cookie data and the innovation that Facebook came up with was, well, AI is probably better at this than all of us. So why don't we just let it do the caption, the, the version of Try this, nope, that didn't work. In fact, if your agency was doing those changes, let's say once or twice a day, three times a day, maybe even once every couple of days or a week, this advertising AI engine theoretically could do it every second. And so that they'll try the ad in this space and then it won't work and the AI will just quickly bubble the ad to where it believes the most accurate targeting is. The challenge and the problem with that **Kelly:** Mm-hmm. **Ryan Hornberger:** is now you've lost complete control of who's seeing this ad and all those kinds of things. You actually have no idea who showed the ad to, why it chose to show those ads. In fact, Facebook may not even know the why for years to come because AIs kind of do things that we don't quite understand until we learn to understand them. And so it's going to be very fascinating to see how Facebook's approach works or doesn't work. Will I still see ads that are relevant to me? **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Ryan Hornberger:** Will advertisers get the effectiveness they need? Can the AI handle a bunch of advertisers all competing for the same eyeballs? Those are very fascinating things that it'll be fun to watch, see how Google solution comes out. Google is also employing AI to kind of do those kinds of automations as well. They're not advertising it as directly as Facebook is, just saying like, we're turning off all of this and you're just gonna give us the ad and we're gonna figure it out. Google's being a little more specific about how you're still going to advertise on their platforms the two technologies you mentioned customer match and pair Really focus on accomplishing the same thing but in different parts of Google's ecosystem so for YouTube versus for web advertising but the Key feature here is something that not just Google is doing but all advertisers are doing which they're saying Hey, we can have when we don't know people who people are we can advertise so effectively But if you can tell us who people are, if we can figure out a way to figure out who people are and what their interests are, then we can give them more accurate advertising. In the case of Google, let's say you're on the YouTube platform watching a lot of content as a user, they know everything you're looking at, they know who you are, they know all that kind of stuff. What they don't know is that you recently visited this merchant anymore, right? Because the cookies are gone. And so what they're asking the merchants to do now is to upload. **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Ryan Hornberger:** basically spreadsheets of email addresses to their platform so that they can say, oh, here are the people who visited this merchant and I know who they are matching it to my platform. So now I know who to advertise to. The problem with that technology, if you haven't picked up on already, is uploading spreadsheets of emails. There are ways to automate it, but it's still uploading emails. And so there's... things that Google is doing in their customer match and pair programs. They basically say, you're not actually uploading the email, you're uploading a version of it that we can't actually turn back into the email. There's a way to do some math on any piece of text and turn it into something you can't read anymore. But it's used for matching. And so there is some protection there, but it's still kind of weird that now a human has to manually go and upload these spreadsheets every day. **Kelly:** Got it. **Ryan Hornberger:** And if you think really about what those spreadsheets contain, there are lists of people who have given you their email address, which likely means they are lists of people who have recently purchased from you. They are not lists of people who have not yet purchased from you, which is what advertising used to be very focused on. And so really this technology does not solve the part of the space where, you know, 95% of people who visit merchant sites aren't logged in. **Kelly:** Mm-hmm **Kelly:** Great. **Kelly:** Right. **Ryan Hornberger:** and they're just browsing and then they're leaving, they become a little bit harder and impossible to track unless you can find fancy ways to identify them. And right now, the most aggressive thing you're seeing is you're getting a strong demand to give an email very quickly so that they can do these pairing and matching technologies. Pair the technology from Google, add some additional protections to the user, by the way, on this, that use levels of encryption and match publishers with advertising companies like the merchants that you buy from in a way that no third party in the middle can listen to. But it's still very, it's a lot of work to get done what we used to be able to do with just a cookie and without email addresses at all. **Kelly:** That's in, yeah, and in the name of privacy, right? But there's a lot more manual work. Ha ha ha. Too shy, okay. So if I'm a brand or a merchant, in order to be able to help Google recognize the traffic that's coming or. **Ryan Hornberger:** In the name of privacy, yeah. In the name of privacy, we're now sending lists of emails to people. Ha ha ha. Ha ha ha. **Kelly:** Which to your point Ryan is, is advertising, especially for search has, has really been about where are your future customers, right? Where is the top of your funnel? How are you getting to discovery? And a lot of this behavior is, you know, **Ryan Hornberger:** Exactly. Yeah. **Kelly:** And I would say different agencies, organizations, leaders would probably argue that by the time you go to a search bar and punch something in, there's a pretty high level of intent of something that you're looking for. Now it's just really about finding the thing, discovering the thing that you're looking for and how much relevance does that have and how closely does that. **Ryan Hornberger:** Definitely. **Kelly:** How quickly do you find kind of what you're looking for? Once you go from this once you go from that search in the browser to one of these websites Your point is that at that time the customer if this is a new cut if this is a brand new customer That's just discovering you they don't have any they don't have any data on you So this customer match system wouldn't necessarily work for that That would be the Google scenario the Facebook scenario is a little bit different right because they have They're using they're not using search, right? you don't get a Facebook to search, you know, just to search for things or get results. You're, they're just, yeah, on certain platforms, but they're more optimized for discovery based on like things that they think you're gonna like, right? And that has, obviously that interest graph is very different. It's based on **Ryan Hornberger:** In small ways you do, in certain platforms, yeah. But not web search, yeah. **Ryan Hornberger:** Exactly. **Kelly:** you know, the type of content that you can consume. And Facebook, we've already talked through this a little bit, but they lost this cross-domain data ability a while ago because they don't have a browser. And they're not going to make one probably. I mean, maybe they will. And so their way of, yeah. **Ryan Hornberger:** Yep. **Ryan Hornberger:** If they haven't been playing around with the idea that they should have been. But it's tough to launch a browser that says, hey, we want to track you in a crazier, even more ridiculous way. So yeah, I guess that's why they don't launch browsers. **Kelly:** How of a serve you better at? Okay, so then the last question is, you know, and we talked, I think we touched on this a little bit, but you know, there's now a difference between visitors' uniques and confirmed uniques. So if you are... **Ryan Hornberger:** Yeah **Kelly:** a user of the Google Analytics platform, you will likely notice a difference between Google's ability to identify a total number of users, a total number of uniques, and then basically confirmed uniques. And is there anything else that we could say about those three metrics changing and shifting a bit? **Ryan Hornberger:** So as we're talking about Unix also, I think this might be a good time to talk about attribution. Is this a good time? Where are we moving on to that? Okay, so Unix, in particular, this is most obviously seen in Google Analytics, the number one, just standard analytics platform people use on their websites. When you used to go to Google Analytics, it would show you a whole bunch of information about your users, what countries are coming from, even what states are coming from, zip codes, those kinds of things. It would tell you... **Kelly:** Sure, yeah. **Ryan Hornberger:** gender and related topics, demographic type topics that could tell you types of devices people are using and those kinds of things. But one of the most important things it would tell you is that, hey, you had, let's say a million visitors, or sorry, a million visits to your site, but of those million visits, let's say only 100,000 visitors. Right, so 100,000 visitors saw a million pages on your site. So, you know, anyway, so that's what it used to be able to say. Now if you go read Google Analytics, it says, hey, we're going to have a heck of a time now trying to tell you if these visits are the same person or not the same person as they move from page to page. It's going to be a little more complicated than it was before because we're a third party compared to the website. The website is themselves the first party, but as your analytics platform, we're not considered first party here. So we're going to lose visibility into who is unique. And as we lose that, we're going to start making guesses. And the way we're going to make guesses is we're going to use AI, quote unquote. And, um, AI is probably going to be effective without any third party data for a little while, but you can imagine that if it has no input data coming into teaching it new information all the time, then it's going to become less and less accurate over time. So the question is, where is Google's guessing AI getting its new information from? And that's where some creepy questions happen about what's Google actually doing. Because nobody knows. **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Ryan Hornberger:** But how are they going to make guesses about how many uniques you're having if they aren't feeding the AI with new stuff all the time? Now, so you've got your uniques. It's important to know your uniques because if you don't know how many visitors you have, you might have one visitor making a lot of traffic for you and not very many unique visitors, but the unique visitors are the ones that matter because one user can only pay you so much. And so knowing your uniques is really important. platforms like Google are now coming up with strategies for you to communicate to them this person is unique. They're the same person from page to page to page, teaching advertising companies how to bypass the third party cookie limitations, at least in a small way, to bridge that unique gap. So there's a part of Google Analytics that's gonna be guessing the uniques, and as many as possible that you can get confirmed. **Kelly:** Mm-hmm. **Ryan Hornberger:** the better and you confirm them by changing the way you use their analytics technology to send them, Hey, this person's the same person as it was before. The most interesting nerve wracking part of that whole thing is that oftentimes what's going to be used to feed that unique detection system is going to be some sort of email address. Right. **Kelly:** Yeah, these lists, this list of the emails. Yeah. **Ryan Hornberger:** Yeah, not just list, but like the moment you enter an email into a form, these advertising companies and Google are all telling their partners, the moment you get an email, hurry up, package that up in a way that we can't reverse it and actually read the email. But just give me that number that the email generates and hurry and fire that off to me so that I can identify this as the same person over and over again, as soon as you possibly have it. And so what's really going on is that in order for advertising to be effective, they now have to collude. a little more aggressively with the merchant or the publisher. Whereas before the merchant publishers kind of turned this thing on and forgot about it. Now the merchant publisher have to put a bunch of effort into kind of share information with their advertiser behind the scenes, intentionally. **Kelly:** Mm-hmm. **Kelly:** Yeah, it really impacts attribution, which is old. And attribution has been impacted for years, right? There's this whole conversation around attribution. At the end of the day, it just means who gets credit for the sale, right? That's a simple way of understanding attribution. And then the question becomes, well, how do you know? that entity, whether it's an advertising platform, an influencer, a referral from a friend, how do you know who gets credit for the sale? And marketers, brand, business leaders, they also have to ask themselves the question of like, well, when did Kelly actually discover me, right, for the very first time? How long was her consideration phase? How long did it take her to think about my product? How many times did I have to serve her a retargeting ad? **Ryan Hornberger:** Yep. **Ryan Hornberger:** Yep. **Ryan Hornberger:** Yeah. **Kelly:** finally make the decision and which platforms were responsible for bringing her back to me at the lowest cost. Right? So these are all the dynamics, but really attribution is a simple thing to understand. It's just saying, likely what I'm doing is I'm going to go to the platform and I'm going to go to the piece of content as a brand or as a retailer that is generating the most sales for me. And so that's the other part. **Ryan Hornberger:** Yeah. **Kelly:** of how this process becomes very inefficient quickly is because it's harder and harder now that you don't know that this is the same user that just came from Facebook, advertising, Instagram, TikTok, search, referral from a friend, all of those, all of those. touch points, you can't tell if that's one person or many people. And so the question of who gets credit for the sale becomes much harder and then therefore it makes less efficiency within the model. And ultimately, it likely drives up costs and drives up experimentation on behalf of brand, retailer, agency, like whoever is spending money to find new customers, right? So that's the biggest challenge with. **Ryan Hornberger:** Yeah. **Kelly:** you know, I would say on the business side, on the consumer side, we've already touched on the fact that because this mess is going on, you just might not see as much relevancy or personalization around how these systems serve you ads, contents help you discover new products, you know, save your cards, all these different types of things. So just a little bit of, yeah, go ahead. Yeah. **Ryan Hornberger:** For anyone, did I interrupt? So for anyone who's listening, who's like, what is attribution? So just wanted to throw in there a couple of thoughts here, which is basically the foundation's attribution allow an advertising platform to stop selling an advertiser just impressions, which means the simplest form of selling ads is like, hey, I've got a billboard on the subway. You buy that billboard, 10 million people see that billboard every week, right? And... And they could be at any age, they could be of any gender, they could be of any persuasion, they could be of any political party. Doesn't matter, like this is just that space. Then attribution advertising saying, hey, I don't wanna pay you advertiser unless you get me a sale. So you better show ads to people who matter and not to people who don't care about my product because I don't wanna pay you unless I know that you close that deal. The problem then gets... **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Ryan Hornberger:** super complicated when Facebook identifies that user and Google identifies that user and Twitter identifies that user and then they all say, I get credit. You just got a sale. I get credit. And you're like, well, who actually deserves the credit? That's where attribution gets really complicated. Well, attribution was primarily built on third party cookies. And like in order for us to know that, you know, this person was just looking at Twitter and then moved over. **Kelly:** Mm-hmm. **Ryan Hornberger:** to this merchant and purchased in 20 minute timeframe, but to know that that's the same person. And that was really hard before. So attribution lately has been successful, or its success has been tanking. And new models are coming together to try and solve that primarily based upon email addresses matching on both sides. But let's say you use a junk email address to log into Facebook and you use your email address and you make purchases, that link can't be made anymore. **Kelly:** Mm-hmm. **Ryan Hornberger:** And so it just becomes this giant machine problem they're all working to try to solve and everyone's coming up with different answers. But at the end of the day, it's a very complicated, difficult, hard, and right now not a very effective technology that they're working on. Google, by the way, in their privacy sandbox has launched the Attribution API, which allows, let's say you're browsing on Facebook's website, see a product, click through to the merchant. Facebook's website can say, hey, heads up, I just sent this user to this website with an ad, advertiser browser, remember that. And then the merchant's website can say, hey browser, I just had a closed sale, remember that. And if the browser sees a connection, the browser can notify both parties, hey, this sale is because of this platform. What happens though, when you're in Instagram's app? Or what happens when you are in Facebook's app, or you're in Twitter's app, they can't make that link. and the browser can't be used for that anymore. And so you get into another bridging problem that just is very, very difficult to solve. **Kelly:** Yeah. And having used a lot of these platforms, sometimes those numbers aren't very high when you're initially testing or you only have a little bit of money to test. So it's a difference between something that I made two sales on versus three sales and that it's not very conclusive and you're spending money and you're, you know, it's, it's this lack of efficiency, um, that I think is the biggest, uh, chat, you know, that's really the biggest challenge. So, and I like what you said, Ryan, about billboards. It was just something that I've also heard from. **Ryan Hornberger:** Yeah. **Kelly:** marketers, which is we're kind of shifting back to almost this idea of brand marketing, almost questioning whether, you know, whether billboard advertising, which seems to have a very broad reach, would potentially reach just as many eyeballs or generate just as many sales. **Ryan Hornberger:** Yep. **Ryan Hornberger:** Yep. **Kelly:** it, I wouldn't say it's that bad, but I would say it's like kind of starting to feel like we don't know, you know, it, we're painting with much broader brushes here. And so it does actually elevate this idea of brand marketing in the mix of marketing mixes a bit. And because there's a, there's broader broader targeting and basically less, we're lacking the ability to sort of pinpoint. The last thing I'll just say about the advertising, like Facebook advantage shopping campaigns, which you can get some insights from, but the biggest changes over there are because Facebook is saying, just give me your creative and I'll go find your people. This is also gonna create challenges for... brand marketers that are looking to create what they call their ideal customer profiles, right? They're constantly looking for new niches where there's an affinity, right? Or there's a correlation for growth. And this is typically like someone in someone brand or in someone growth marketing, you say like, what are these additional audiences that we're going to test? **Ryan Hornberger:** Oh yeah. **Kelly:** And so we're, so they're going to be losing a little bit of visibility that way. And also, you know, these types of campaigns don't work for specific types of products or industries, which would be like, you know, niche or specialized products. meaning like if you're an audiophile or if you are looking specifically for sailboat parts and things like that. So there's a bunch of different industries where this won't work for it as well, meaning. **Ryan Hornberger:** Well, yeah, this only works for the big guys who want broad reach. It doesn't work for anybody who's trying to get very narrow because now you're competing with all the broad guys. And if Coca-Cola can pay X number of dollars per person and you, a small guy trying to target a niche category can only pay this much, you're gonna lose Coca-Cola every time. And, you know, so these movements serve to make bigger companies bigger. **Kelly:** Mm-hmm. **Kelly:** Mm-hmm. **Ryan Hornberger:** and smaller company is smaller. That's the outcome of what's gonna happen here if the technologies don't improve. **Kelly:** which we'd hoped for the last couple of years, right? That we would be championing the small business. Like there's been so many initiatives, right? Around buy from local, buy from small business. And this is just, this is a challenge. If you can't target. **Ryan Hornberger:** Yeah. **Ryan Hornberger:** Yeah, we're seeing small companies moving off of their websites and into Amazon and into Walmart and things like that. That's where they have to go to get eyeballs now because they have a very hard time finding them anywhere else. **Kelly:** Yeah. **Kelly:** Yeah. So next we're going to talk a little bit about some of the, you know, industry sounding board more recently. Um, we got our hands on an article from Digiday, which they had just, um, it was kind of coming out of the performance marketers summit, essentially. And There were there were some basically they did almost like a anonymous like over her You know, we have these meme accounts like overheard LA right where it's like we heard or the thing but we're not gonna Necessarily talk about who the source of that information was so they kind of did this forum and there were some interesting things That came out of that. I would say the overall sentiment is that this is Sometimes we read things that say like, oh, this is like we're solving these problems, right? Like these problems are going to be solved for third party cookies. And what was interesting about this article is like, nope, these problems are still on the horizon. We're going to have a lot of work to do. 2024 is going to be gnarly. If you're not testing, you're dead. And I think the, what's going to happen is finally like the rubber is going to hit the road or meet the road. Right. And we're really going to understand whether these solutions. which are promised to be solutions are truly going to be solutions. So we touched a little bit on attribution, and I can link this article in the show notes as well, but it talked a little bit about why Google and DSPs are pointing the finger at each other for invalid attribution. So from a technical perspective, Ryan, can you walk us through a little bit about why Google and DSPs would now be sort of for this like return of invalid. What, what does that actually mean? **Ryan Hornberger:** Yeah, so DSP, just for those that don't know it, Demandside platform, they are the... So supply side of advertising is any publisher who's publishing and showing you ads. Demandside is anybody who's buying advertising to be published into those spaces and the companies in between are who make those links happen. And so DSPs are the companies that advertising companies buy from to put their ads out and they... you know, sell ads and they need attribution to have these kinds of ads where you can pay for when a deal closes, et cetera. That's with attribution. **Kelly:** Can you give us some brand names, Ryan, of DSPs just so that we can... **Ryan Hornberger:** The ones that are most known would be Google, Facebook, TikTok, Snapchat, things like that. Those companies sell ads on their platforms, et cetera. And then you have to think about all the places where there are advertising happening where you don't know who's actually powering it, such as any publishing websites, so CNN, sorry, Wall Street Journal, et cetera, et cetera, et cetera. Any website who's creating content, blogs, mommy bloggers, those kinds of places. **Kelly:** Lost your journal. **Kelly:** Mm-hmm. **Ryan Hornberger:** Like small niche environments all need to be powered by ads. Who's powering those? That is not usually them on their own, selling ads one-to-one, making phone calls and trying to close deals by themselves. They usually import some sort of platform that's selling ads at bulk. Some names that might be known out there are like Traydesk is the second largest one in the world in that space, drawing some blanks on a few others at the moment. But... **Kelly:** Mm-hmm. **Ryan Hornberger:** But there's a number of companies who are kind of under the covers powering all this advertising and selling it in bulk environments so that you can buy from them and then they make it work on 10,000 websites all at one time. **Kelly:** Got it. And so one of the observations is that Google and DSPs are pointing fingers at each other for these invalid links or attribution. Can you talk a little bit about that. **Ryan Hornberger:** Yeah. So attribution is, the attribution is typically powered by historically cookies, right? So they know that you were on, so let's say your trade desk and you work with Hurley, the merchant, and you work with wall street journal and you, and someone in the user visits Hurley and trade desk puts a cookie on that user and that cookie now can follow the user everywhere they go on the internet. By follow, I mean like the next time you visit another website, that cookie will present itself as well. And all it is a number. This is the number of this person. So. You then leave Hurley and you go visit Wall Street Journal, not having purchased that surfboard you were looking at. And suddenly you're seeing a surfboard ad from Hurley on Wall Street Journal's website that was connected by that cookie that was just a number. Hey, this person is here and now they're here. Didn't have an email address, didn't have any of that kind of information in it. It was just an ID. And now they know that you're the same person and they can create that ad. And then if you make a purchase, they can go, oh, Ryan saw the surfboard. He went to Wall Street Journal. He then came back and bought the surfboard, Wall Street Journal and the technology that connected us all deserve credit for that sale and a little bit of kickback for it making that possible. That's how that's how attribution used to work. Well, then you kill cookies. And now, you know, Ryan visits, hurriedly looks at a surfboard, goes to Wall Street Journal, does not see the surfboard ad. Maybe somehow they figure out how to get me to see the surfboard ad. When I go back to buy it, they don't know that I'm the guy who looked at the ad. And now, **Kelly:** Mm-hmm. **Ryan Hornberger:** They can't confirm that this was the reason why I made the purchase. And so attribution based advertising to the, if you're, if you're Hurley, the company used to be able to tell the advertiser, Hey, I don't want to pay you unless you come to close a deal. Now you can't say that anymore. Now it's just, Hey, show my ad, please just show it. Please show it. I'll pay you for whatever you need to show. There's no confirmation coming back. Google invented a technology that could theoretically solve it only works in their browser. It's powered by the browser that thing I was telling you guys earlier where like the Hurley would put a little notice in the browser. Hey, I just had a customer look at this product then Wall Street Journal say Hey, I had a customer look at this product and then the customer would buy it and Hurley would say this purchase happened The browser can make those links and then tell everybody hey this worked out So Google's saying hey DSPs just use our technology and you won't have any more problems and DSP are saying hey We don't want Google being the intermediary between who gets to decide who gets credit for the purchase **Kelly:** Mm-hmm **Kelly:** Mm-hmm. **Ryan Hornberger:** And if you're the only ones who can make the technology, why would we want to do that? So DSPs are now pointing their fingers at Google, Google's pointing fingers at them. It's really a fight for power. It's not a fight for anything else. It's just who gets to control the outcome and the knowledge around this information. **Kelly:** Got it. **Kelly:** But ultimately, anyone needing the positive attribution information is the one that's suffering, basically, at this point. **Ryan Hornberger:** Yes, yeah, this is a fight between providers that they work with, but it's not their fault, and they're in a lot of ways helpless to solve it. **Kelly:** Got it. We're going to just touch on two of the quick things, Ryan. So IP addresses are apparently still on the table in terms of first party data that is allowed to be collected. Can you talk a little bit about the IP address and also as it's related to sort of these, how it's being used with what we call clean rooms? And if you want to discuss what a clean room is really quickly, that would be great too. **Ryan Hornberger:** So IP addresses and a number of other features of your browser that are hard to cut out of the knowledge tree because deep down the internet works in a very certain way and IP addresses is how it's powered. And so at the end of the day, both parties get to know who's talking to who in a lot of ways, at least on an IP address level. And so they can use that information to kind of still learn enough about if this person's the same person or not. It's really complicated though, if you're ever let's say on a corporate network or in a hotel at the moment, or on a public Wi-Fi unit at the airport, you might be with three million people sharing one IP address as you're talking to the internet. And so it becomes more and more difficult to track somebody across places. If you're traveling and moving from place to place to place, they can't track you anymore because you're constantly on the move. So there's a number of reasons why IPs don't work, but they're better than most because a lot of people spend their time doing a lot of browsing in home. And at least when you're at home. We have a little bit less noise there. We can kind of figure out what's going on and potentially create an idea of who you are. That said, IP addresses are still way less useful than, say, an email address. So... **Kelly:** And would this mean that any device on this IP, I mean, let's just call it, like, let's say not your house, right. But like a standard, um, home, every is, is device also identifiable now just IP addresses, does that mean all devices on the IP are now treated basically the same, which means you, your wife and your three kids, which have all very different interests in demographics are all sort of one. thing now. **Ryan Hornberger:** So the question has actually two answers. For a percentage of the world, even at my house, if I've got 15, 20, 30 devices, 60 devices in my house, all talking to the internet, from the perspective of the internet, they're all the one IP address, from one side of the internet. Now there's an upgrade to the internet that some people have and some people don't. And that upgrade makes it so that every distinct device has a unique IP address while on your personal network. **Kelly:** Great. **Kelly:** Mm-hmm. **Kelly:** Okay. **Ryan Hornberger:** Um, and that's, uh, that's a little more trackable. However, the moment that device moves from your house to the local coffee shop, its number changes, at least the IP number changes. And, um, and so the only thing that really protects you in that world is moving around a lot, if you don't want to IP track, you got to move around or VPNs can kind of obfuscate that and bundle your IP with millions of other people. **Kelly:** Mm-hmm. Okay. **Ryan Hornberger:** coming from Finland, even though you live in the United States kind of thing. But, but yeah, that's the, that's how IP is going to work is they, they're, there can be many of them in your house, but the moment you move your devices, numbers change temporarily until you come back. In addition, your number might change in your house from time to time, let's say every day, every seven days, whatever. It'll just change up all by itself in a lot of people's houses. **Kelly:** So foundationally, and we could talk a little bit more about this, but for now I think we'll move on. So IP addresses are probably not going away for the foresee, meaning the ability to track an IP address just because of, okay. And so as it relates to clean rooms, what's sort of the theory here that IP addresses might be clues for us? How are those used inside of clean rooms? **Ryan Hornberger:** Yeah, it's pretty hard to get rid of that. **Ryan Hornberger:** Yeah, so clean rooms is a technology that has been around for a very long time, not just in advertising, but the technology or the business model makes sense. Two parties want to know sometimes what each other knows about a customer, but that's not the case in this case. All these two parties want to know if they're using a clean room is, is this person the same person as this person? Is Ryan visiting Hurley and Wall Street Journal? Are they, is that the same guy? Well, the question is, what does Hurley know? And what does Wall Street Journal know? They can both create sets of information about what they know about this person, send them to a third party intermediary clean room company. That company, without revealing what either party knows to the other, can at least say, this is a match. So the word clean intends to imply safe, or whatever this person, whatever this company knows is not gonna bridge over to this company, and what this company knows is not gonna bridge over to that company. All they're gonna know is connected. **Kelly:** Mm-hmm. **Ryan Hornberger:** Yes, yes, no, yes, these are the same pieces of the same customer. **Kelly:** So they're not sharing additional information. **Ryan Hornberger:** They can't, the information's not supposed to be able to get from party A to party B, but party C in the middle has to have enough of that information to see if there's a match. And so let's say you know a phone number, they know an email address, the cleanroom company might know that both, they might know both your phone number and your email address. So you say, hey, is this the same person as this person? I've got a phone number, I've got an email, cleanroom knows that's actually the same person. Yes, they're a match. Um **Kelly:** And so this is a strategy that may help these different parties still deliver relevant content if they can get a positive result. **Ryan Hornberger:** Yeah, so if you get a match, then you at least have some form of attribution. Even if you don't have all the information you need to really match that customer. Some companies have a lot of information. Let's say Google, Google uses clean rooms in their technology to match customers to advertisers. Um, and so, um, Google knows everything about everybody. And so, but they're not going to give up that information. It's very protected. Our information, we want to sell it to you over and over again. We don't want to sell you the information we want to sell you access to knowledge around the information. And so they'll serve us a clean room and just constantly help you be a matching company without actually revealing to anybody outside of themselves who you are. **Kelly:** Interesting. So there was a quote, Ryan, through this article, and then we'll just kind of wrap up this, this part of the discussion, but it basically said that like, giant brands have never really had to use clean rooms before, whereas medium sized and small brands just haven't had the budget to do it, right? Or maybe that's hasn't been as high in their priority list of like, getting to this level of sophistication or needing this level of data. **Ryan Hornberger:** Exactly. Yeah. **Kelly:** And what was interesting is that this person basically said, this is really just like leveled the playing field, right? Because now everyone needs, it's, what's your perspective on, okay, this is now leveled the playing field. What does that mean to you? **Ryan Hornberger:** Well, I think anyone who is saying it's leveled the playing field must be very, uh, must be coming from a lot of back. They must have deep pockets they can access when they say that, uh, because, um, if you can afford a clean room, uh, at least some of these more advanced ones, um, and you have the time and attention necessary to be able to send information in and get information out, match everything up and do all the work, you know, you, uh, **Kelly:** Ahem. **Ryan Hornberger:** You now feel like you're on equal footing with your competitors. But if you're a small guy who doesn't even know what this is and doesn't understand half of these words, yeah, the playing field's been equalized and it's cut you out, you know? So that's going to be, it's been equalized for these players and not for anybody else. Companies who sell clean rooms are going to try to find ways to make it more accessible, cheaper, that kind of stuff. And in a lot of ways they're becoming effective at it, but it still means a lot of effort. **Kelly:** Got it. **Ryan Hornberger:** on your side to help get the information linked up. And so it used to be for everyone, big or small, just install this cookie thing on your website and anytime someone visits, we'll do the rest. That's what cookies did. We'll take care of all the complex stuff, the messy stuff, the hard stuff. We'll figure it out because we've already got the number. We have the ID, we know who this person is. These ID wise, we don't need to know an email address, we don't need to know anything about them. We just need to... **Kelly:** Mm-hmm. **Ryan Hornberger:** know that this number matches this number when they visit different websites. And if we learn more of that information, it doesn't matter, but the point is that we didn't need anything other than that to make it work. Without cookies, now you've got to go to these big knowledge brokers and buy and sell your information as much as you can to continue to make those connections. Kelly (01:00:13.09) Got it, so now we're gonna go quickly through two more, two more big, now we've kind of laid this foundation of like how all this stuff works together. We're gonna talk about Universal IDs and then this idea of Whym's individual profiles, which we're calling Persona API. So talk a little bit about one of the solutions in market is Universal ID, another solution that we've, Ryan Hornberger (01:00:28.949) Yeah. Kelly (01:00:38.722) written this API documentation around is called Persona API. And we'll talk about that second. Talk first about, OK, so some of these big companies that are competitors to Google, or maybe they're not completely competitive, but their business is really third party advertising. They've had to come up with these solutions. And one of these ideas is a universal ID, which does not Ryan Hornberger (01:00:58.082) Yep. Kelly (01:01:06.93) Involve the end user at all right, so can you talk a little bit about what is a universal ID? Why do these parties think that this is a solution like an acceptable solution in what ways is this going to succeed and fail? Ryan Hornberger (01:01:09.326) Yep, correct. Ryan Hornberger (01:01:22.625) Yeah, so Universal ID is a brand name for trade desks technology, Universal ID 2.0. And there's a bunch of other companies coming up with IDs that have very similar names, not quite Universal ID, but I think we've read about at least 10 now. Other providers all marketing something similar. And the gist of what these IDs are, are vehicles to power clean rooms, which we've talked about a little bit here, and if you're still not following what a clean room is, we can... going into a little bit more, but the universal ID effectively is like, okay, the universal ID, like if you get rid of the branding of the word universal ID, the last time we had a universal ID was a cookie. It was just a number that you just walked around with and every time you visited a site, you'd share that same number with every tracker who needed to know it and then those trackers would use it. The universal ID is like, hey, we're gonna create something like that, but it's gonna have to be based upon information people tell us. We're not gonna make them like log in with this advertiser's log in. We're not gonna make them do things like that. What we're gonna do is that you're just gonna figure out a way to identify them. Whether you can get their phone number, whether you can get their email, whether you can get their whatever. If you can figure that out, we'll give you a tool to turn it into this thing called a universal ID. It's a way to share this information without actually sharing the information. It's a way to say, hey, I have this customer based upon this email address. I don't wanna tell you their email address. But I want to tell you that I have this customer. And then another company who finds the same customer can use this program to create another ID. Those IDs can match in a space that doesn't necessarily always need a third party central company matching them up. Kelly (01:03:01.598) meaning a third party's clean room. Ryan Hornberger (01:03:04.593) Theoretically, that said, every time I see a Universal ID, there's always a third-party cleanroom in the middle. So it's always gonna be there. It's just a branding exercise around how these cleanrooms market their technologies. So what you'll see in a lot of these advertising companies is in the code documentation for engineers integrating with the advertiser, it'll say, hey, the moment you can identify this user, send us a command that says, I've identified this user, this is who they are. And we'll turn that into a Universal ID and then we'll try to match it across the bridge. Kelly (01:03:09.597) Okay. Kelly (01:03:15.159) Got it. Ryan Hornberger (01:03:33.953) You've got to get this, that information as soon as you can, which is why you then see every website you're going to now just demanding an email address the moment you arrive. It's the very first thing they ask you, even if you're not ready to give it to them, they want it. And if you're seeing some of these content news outlets, they won't even let you read the first article now without identifying yourself. They need that email address or they need that phone number so they can start bridging the Universal ID communication channel and identifying you across the space. Kelly (01:03:41.975) Got it. Kelly (01:03:52.818) Mm-hmm. Ryan Hornberger (01:04:04.317) So that's the big takeaway of Universal ID. You won't see it happening, but under the covers, they're taking any email you give them, turning into a number basically, that can't be turned back into the email, and then just saying, I know this customer. And then somebody else saying, I know this customer. Kelly (01:04:18.326) Got it. So it kind of moved from this world of, we used to go into a browser and individuals would then have an ID put on them and that little ID would then follow them all over the internet, cross domain, wherever they did, wherever they were, consume content. Now that's going away. The browser has basically said, we're gonna just... put you in a group of interests. That's what's gonna happen. You're now just can, the Google browser is now basically said like, you're just in a group, you no longer have this identifier, we're just gonna contextually, you know, you've come into the browser, you've searched for something within that session likely, and now we've just grouped you into this interest group. These, Ryan Hornberger (01:04:45.905) That's what Google's browser said, yeah. Ryan Hornberger (01:05:05.883) It's one of the things they do, yeah. Kelly (01:05:08.766) These additional like third parties with these advertisers. They're basically saying like hey, how can we collaborate with each other? Where now it's sort of domain by domain or let's say party by party I'm an identify this person over here. You're an identify this person over here Now the user has to share some amount of information for that to be attached to versus just this more bubble of like ad interest, right? They can't attach anything to that. So I have to identify myself in two places that can get reconciled. And then those two parties, where multiple parties can kind of do their job, right? To... Ryan Hornberger (01:05:46.341) Yep, exactly. Kelly (01:05:48.69) Okay, so those are two different constructs that are now as we move forward into 2024 redefining how the internet works. Those are two things that are on the table for the solution. Now we're gonna talk about a third solution, which is This Idea of what if there was a completely separate type of cookie right that was based on something that the user could create for themselves and share with different parties, with the internet. And so we're gonna just dive right into this, Ryan. So talk a little bit about Whym's approach here and why did we feel the need to come up with this idea? Ryan Hornberger (01:06:38.049) why we needed it is by far the most important question. So as, you know, we're a company providing a service, but at the same time, we're users of the internet. And in a lot of ways, we're the generation that was first raised on the internet. Pre social media, but post like no computers at all. And... Ryan Hornberger (01:07:05.157) And we came up with an internet that was free. You know, like we could send emails, we could read a ton of articles, we could be watching YouTube videos, or not just YouTube, but videos and content in general. We had access to a ton of information that was free, and how it was free was the cookie, because advertisers could figure out that I was looking at a surfboard here, and then I went to this publisher's website, and they could show me that same surfboard while I was there, constantly dragging me back at it. And those content publisher websites, that's how they made all their money. They didn't make a single dime from me. I didn't pay them anything, but they made their money from that stuff. And what we see going on now is that publishers and merchants who are trying to sell to you all have to demand that you log in and, or at least identify yourself in some way before they're able to service you properly. Otherwise their business model is dying. And, um, so you're seeing merchants moving in to bigger. big box online retailers like Amazon, because you're already logged into Amazon, that works. And they're moving, and then you're seeing publishers going, hey, I gotta show ads, but I can't show you an ad until I know who you are. So I'm just gonna make you log in. And so all these places that used to be free and easy to access and walk into, they first started out by just saying like, allow cookies buttons, and that wasn't good enough for regulators and good enough for companies. And so now they're going, you just gotta fully log in before you can access my content. And that is an internet that we think will kill major sections of the internet. You know, if you just wanted to read that article and get out of there, now you got to reveal who you are to everybody before you can get in there. So the way you start before the analogy would be that as you walked around the internet, advertisers would put little stickers on your back and it'd just be a number. Then as you walked around, you just have a pile of stickers on your back. We called them cookies, but you could have just as easily called them Post-it notes. They just are numbers attached to you as you walked around. And every time you visited a place, the advertiser would pick up their sticky note and go, that's number 392, okay, that's the same person, and they'd bridge that gap. Now, every door you walk in, you gotta show ID. Very different world. It's like, I used to want to show ID when I walked into a bar, I used to only have to show ID when I, like we're going to government offices to pay bills and that kind of stuff. Now, I gotta show ID just to walk into the grocery store or just to walk into the clothing store. I gotta show ID just to read the newspaper article that's sitting there right in front of me. Ryan Hornberger (01:09:27.737) That's how it's going to have to work now. And that's an internet we don't want. I don't want the us as a company. We don't want. So we began thinking about how do you fix the internet so that you don't need a show ID everywhere you go. And we believe that what we call the persona solution is the way to work on it. Now, if you're a marketer, advertiser, you've used the word persona for your entire life to describe groups and categories of people that have certain like interests, we're not talking about that kind of persona. Kelly (01:09:31.982) Thanks for watching! Ryan Hornberger (01:09:56.921) We're talking about a user creating almost an avatar for themselves and sharing it intentionally instead of their ID to get around the internet. It's an intentional ID that doesn't have their name on it. It's an intentional ID that doesn't have their email address, doesn't have their phone number, doesn't have any of that kind of stuff. But it's saying, hey, I as a user am willing to at least show you this number, random number, everywhere I go so that I can read this article for free. I don't wanna pay every time I go, I just wanna show this number. And every time I show this number, I want you to let me in. And it's not gonna tell you my name, it's not gonna tell you my phone number, it's not gonna tell you my email address, not gonna tell you not that, it's gonna tell you that I'm the same person who was here and not here. And if any day I stop trusting you, I'm gonna stop showing you that number, I'm gonna show you a different one. Or I might stop showing you entirely. The big problem with the way that cookies work though, is that they turn off that number for everybody. Stop using cookies. Kelly (01:10:53.768) Mm-hmm. Ryan Hornberger (01:10:56.057) The way that Persona API works is that every time someone wants to read your number, we're gonna ask them to identify themselves to you. You're not gonna get pop-ups everywhere saying, hey, now this person's tracking you, hey, this person's tracking you, but any point you want to, you'll be able to go into the Persona platform, your Persona, and you'll go say, who'd been watching me? And you'll say, hey, Google watched you on these 36 websites, and Trade Desk watched you on these 27, and Facebook watched you on these 13. And if for some reason you very much hate Facebook, or you very much hate... trade desk, you very much hate one specific advertiser, instead of killing the entire marketplace, you could say, stop playing that guy, track me, and everybody else is fine. And so when, you know, if a person, if a platform you block walks up and says, can I see your number? And you're like, who are you? Oh, nope, not you, sorry. But you're okay with these guys, totally fine. And... Kelly (01:11:44.822) Yeah, I mean, it's a way to fine tune your own data set for what you actually, who you want to give your information to and what you want to show online. So in a way, we talked about these two different ways of approaching it, right? The browser's approaching it by. Ryan Hornberger (01:11:48.43) Yeah. Kelly (01:12:02.566) you know, really just identifying you in a group of interests, right? Like an interest category. And the third parties are saying, here's an ID you will connect and just, you know, sort of like confirm what we know about you as two separate, you know, entities, but what this is, is the way for one user to actually do both of those things intentionally, right? Through their own first party. So the, um, the, the persona API, which is an individual digital identity, Ryan Hornberger (01:12:07.189) One of the strategies, yeah. Ryan Hornberger (01:12:24.139) Yes, intentionally. Kelly (01:12:32.4) It's a profile built specifically for the internet and all it is a group of ad interests that you can either collect in a really easy way through our browser extension and our iOS app and we'll kind of show you the things based on your browsing history that we that you may want to add just automatically you know. to have to add into your profile, or you can go in and create these ad interests. It's a way for you to do both things, what the browser's doing, what these third parties are doing, but in a way that you can control it, you can see it, and it will be persistent. Whereas Google's only going to be able to treat you that way for four hours. These entities have to rely on... you've somehow shared the same email addresses, right? We all have like multiple email addresses. We use junk phone numbers. Like we use all kinds of methods to basically control how and when our data is shared online. And so this is a... first party cross domain, meaning you can take it across domains, right? Like any website, user centric cookie. So this is the really the third option in the market, right? And that we've seen, we've not really seen anything else like this. And tell me how it's like special to you. Ryan Hornberger (01:13:44.335) Yep. Ryan Hornberger (01:13:55.217) I like, let me jump in real quick for. So what's exciting to me about this model is that under all the models that we've seen, cookies, and now the models are coming out after post cookies, they all still are making the same mistake. The mistake is we're gonna try to figure out how to track this person without involving this person, without telling them what we're doing, without, you know, like we want it to be so seamless and so undetectable that they don't even, yeah. Kelly (01:14:17.191) Mm-hmm. Kelly (01:14:25.546) Yeah, they don't want to be bothered with it. Ryan Hornberger (01:14:28.153) Well, like they're trying to make it, they're still trying to make it creepy. Does that make sense? They're still trying to figure out a way to do all this without telling you they're doing all this. And it's quite annoying that the solution to the problem is just making it way more complicated, way more expensive. And in many ways, it almost feels like way less, yeah, way less private. In the name of privacy, it kind of got less private because now everyone's sharing emails across. Kelly (01:14:33.123) Mm-hmm. Kelly (01:14:53.429) Mm-hmm. Ryan Hornberger (01:14:56.621) with each other over and over and over again. And then there's some conditions that might be okay, but there are a lot of conditions that just doesn't seem like that's the right answer. And so what I love about the Persona platform is we said, okay, look, what if we could invent a way for the user to be intentionally involved in the process? And so it functions very much like a third party cookie. We get to call it a first party piece of data because the user is sharing it intentionally with you everywhere they go. And the way we're packaging it up for the user is we're saying, hey, like, this is not you. This is your character. It's like the video games, movies that you see where people are in playing as characters. This is your character online. And at any moment, your character can be this character. In the next moment, you can have a completely different character and play somebody else entirely. The trackers will stop tracking you. They'll just start tracking your character. in exchange for you being willing to share your character with us, your persona, one of your personas with us. We at least now have enough information to show you some ads, but you now have control. The moment you don't like what we're doing, you can tell us to stop or you can just change characters entirely. Or maybe you can use characters in a creative way. This is my character for when I'm thinking about work, and this is my character for when I'm having fun, and this is my character for when I'm dealing with my kids. You can pick the character you want to play any given moment, and the trackers only track the character, not necessarily you. And we think that puts you in the driver's seat finally of your information. You can't decide what they're going to do with the information and how they're going to do it, but you can immediately change what they see from your point of view. And so we hope that customers will want to walk in and say, I don't want to show you my ID. I want to show you this one. It's not going to tell you who I am. Doesn't tell you enough to treat me right. Kelly (01:16:52.854) That's right. Ryan Hornberger (01:16:52.869) And that's what we hope the person on the technology does, is make the next version of the internet just as awesome as it was before with a little less creepiness. Kelly (01:17:05.664) Awesome. Well, thanks for thanks Ryan for sharing all of that. I think what I'm excited about this the most is that it really puts the user in the driver's seat and gives them a way to see. and participate, I think we're becoming more and more familiar with this idea of fine tuning data sets and it's a way to start thinking about what's available about you online and how can you use that data to help create a better experience for yourself. And really at the end of the day, as much as these companies think they know about us or as much as they're able to you know, glean from our search behavior, really we know ourselves the best. You know, we're the only ones that are the hosts of all of this information, all the products we've ever purchased, the things that we're interested in looking at, and how those behaviors or how those interests change over time, right? We're not always consuming content or shopping the same way. Sometimes I'm shopping as a mom, sometimes I'm shopping as a partner. Ryan Hornberger (01:18:18.562) Yep. Kelly (01:18:19.886) And we should have the ability to tailor the internet or discovery or results based on those different behaviors. So tell our listeners, Ryan, how they can actually learn more about the Persona API. Ryan Hornberger (01:18:38.069) Yeah, so we will be publishing websites and documents on this soon. For now, keep following the podcast and we will keep on updating you as that goes out. There are limited access right now to try it, but you'll need to reach out to us to get that access if you'd like to try it yourself. But more to come for sure. That's the right thing to say right now. It'll be coming out here soon. Kelly (01:19:01.418) Awesome. In the meantime, they can download our... Chrome extension through the Chrome store for extensions through typing in Whym shopping. And on iOS store, download the application. Our first instance of this technology will be running through those applications. And you can sign up for our beta list through those in-market consumer applications. So. Ryan Hornberger (01:19:31.953) And real quick, just to plug that product, the shopping app, just what it does, guys, is it follows you around the internet. It's connected to you as you're browsing around. And every time you see a product, it just remembers it. And its intention is that you can start closing all those tabs you have sitting opening a browser that you haven't closed because you haven't made the product purchases yet. You can delete all those images you put in your photo roll. that you never opened because you screenshotted the product but didn't actually remember to look at the image again. This software, while you're in your browser, will observe the products you're seeing and it'll just create a search engine of your personal information just for you. And so let's say you've been looking at products for six months and you're finally ready to buy some shoes, you go type in shoes and it will say, here's all the shoes you looked at in the past six months. And now you get reminded of the products that you're most interested in. without having to do any work to make that happen. It's not perfect, we're still making it better every day, but we love it, we hope you will too, and once Persona's become a public service, you will first see it there. So go get the app and start finding it now. Kelly (01:20:45.814) Thanks so much, Ryan. Thanks for that overview. I'm gonna do a quick wrap up. So thank you for tuning in to the Blind Data podcast. You can visit us on Spotify, Apple podcasts, and also on YouTube. We hope that you enjoyed today's discussion. You can reach out to us with any feedback or to sign up for our beta list at hello at whym.tech. We'd love to hear your thoughts or suggestions. You can also subscribe to the podcast newsletter by visiting wimwallet.com backslash blind data. By subscribing, you'll be the first to know when we release new episodes and receive exclusive insights and updates. And lastly, if you enjoy today's podcast, please find our... podcast on one of these platforms and we would love to hear a review from one of from one of you. Your reviews always help us reach more people and help us to continue creating relevant content for you. So once again thank you for joining us today we look forward to bringing you more engaging discussions in the future and stay connected, stay curious and until next time. Bye! Ryan Hornberger (01:22:00.689) Thanks everybody. --- [View on kellyny.land](https://kellyny.land/transcripts/blind-data-4-a-slow-goodbye)