Beyond the Pixel: Jer Tippets on AI, Customer Data, Privacy, and the Future of Digital Measurement
- 14 hours ago
- 54 min read
Every AI breakthrough begins with something far less glamorous: clean, trustworthy data.
In this episode of Signal & Noise, Brett House and Rio Longacre sit down with Jer Tippets, Director of Digital Tagging & Implementation at Hyatt Hotels, to explore the invisible infrastructure behind modern digital marketing. While everyone is focused on generative AI, personalization, and autonomous agents, Jer explains why none of it works without a strong data foundation—and after more than 15 years building customer data systems for global brands, he's uniquely qualified to explain why.
Rio first met Jer while speaking together at Tealium's Digital Velocity Conference, where Hyatt shared its customer data strategy and philosophy of treating customers like humans. Hyatt has long been a Tealium customer, making this a conversation we've wanted to have ever since. Together, they unpack why poor data quality quietly derails AI initiatives, why companies continue collecting mountains of unusable data, and why "collect everything" has become one of marketing's biggest mistakes. Jer explains why first-party data is now a critical business asset—and why governance, identity resolution, and privacy have become strategic priorities.
The conversation also explores Hyatt's approach to digital experience, balancing personalization with customer trust across millions of guest interactions. From consent management and CDPs to synthetic audiences and AI-driven personalization, Jer offers a practical perspective shaped by years of real-world implementation. Jer also shares why AI should amplify—not replace—human expertise, why data governance is becoming a competitive advantage, and how AI will reshape digital measurement as agents increasingly generate and consume marketing data.
If you've ever wondered why some brands deliver seamless, personalized experiences while others struggle with fragmented customer journeys, this episode reveals what happens behind the scenes—and why the future of AI may depend more on better data than better models.
We also discuss:
Why AI depends on quality data
Hyatt's approach to customer identity
First-party data as a business asset
The cost of poor data quality
Privacy, consent, and customer trust
Why "collect everything" fails
CDPs, identity, and synthetic audiences
AI's impact on implementation and analytics
Why governance is a competitive edge
Preparing for an AI-first future
Whether you're a CMO, MarTech leader, data engineer, product owner, analytics professional, or simply interested in the future of AI and customer experience, this conversation offers an inside look at the systems powering every click, recommendation, and personalized interaction.
Watch the full episode and join the conversation.
#SignalAndNoise #AI #Data #MediaQuality #Attention #Advertising #Privacy
🔑 What We Cover💡 Key Takeaways🎯 Why This Episode Matters
Read the full transcript below.
Brett House (00:01.109)
Everybody, welcome back to Signal and Noise. This is Brett House with my co-host Rio Longanker. today's guest is Jar Tippets, the director of digital tagging, excuse me, and implementation at Hyatt Hotels. Welcome to the show, Jared. Great to have you. And and I'm glad I didn't I'm glad I asked about your name because it's J E R as in Jeremy, but it's it's Yeah, we're like, it's Jerr. It's it's like a Viking name. Right?
Rio (00:16.045)
Welcome.
Jer Tippets (00:17.181)
Thank you guys.
Rio (00:21.944)
We thought there was some Nordic pronunciation like year or something. We we yeah right We completely screwed it up.
Jer Tippets (00:22.355)
Yep.
Nope, nothing cool like that. I've had my DNA done. am British and Italian and I have nothing exciting, nothing Nordic in me. So it's just Jer, just nice and boring.
Brett House (00:36.897)
Yeah. Yeah, yeah. And you know, so you've spent more than fifteen years helping organizations collect, manage
Jer Tippets (00:43.955)
Have.
Brett House (00:44.213)
and activate customer data across pretty complex and increasingly complex digital ecosystems, right? You've worked across hospitality, retail, enterprise commerce, really with a with a hands-on perspective into the the evolving role of d AI and data collection, advertising and content personalization, governance, and of course measurement in terms of how does all all this work and and data cleanliness and quality starts at the collection stage, I think, right?
Jer Tippets (00:49.843)
Yep. Yep.
Jer Tippets (01:02.899)
Yep. Yep.
Jer Tippets (01:09.639)
Yes.
Jer Tippets (01:13.427)
Yes. Yep.
Brett House (01:14.107)
and you know the goal here I think is to to ensure that organizations have accurate trusted data to deliver these types of experiences and outputs in a trustworthy, credible way, right? I mean I think that probably captures it. and so so so clearly AI is reshaping digital measurement, it's re it's reshaping content production and engagement in in data management. And so I think you play a a very critical role.
Jer Tippets (01:22.334)
Yes.
100%.
Jer Tippets (01:34.099)
so much.
Brett House (01:41.286)
you know in in this ecosystem and in you know being at Hyatt I'm sure there's a lot of data from a lot of of locations, what fourteen hundred and fifty hotels across the world in 79 countries. Right? So and I and I've I've seen how hard no knowing knowing Marriott I've seen how hard and how poorly their data collection and unification is because they continue
Jer Tippets (01:48.937)
Yes.
Jer Tippets (01:54.985)
So many. Yes.
Jer Tippets (02:03.977)
you
Brett House (02:07.319)
to mess my ID up. They've got some there's an ID conflict with my wife's email address and me and the and they and everything I've tried, they can't resolve our separate identities aligned to a a
Jer Tippets (02:12.978)
Huh?
Rio (02:19.095)
They don't mess mine up by the way, 'cause because I'm lifetime platinum. But sorry. We tried.
Brett House (02:21.783)
Yeah, unique yeah, and and we've tried. And yeah, yeah, Lifetime Platinum, the guy he's like I've they they they they think of me as more important. Yeah.
Rio (02:29.005)
No no but that's yours as a consultant. There really is. I that that's it is there's nothing glorious or glamorous about spending three years of your life in in a in in hotels, but hotels are great. So we're so yeah. Welcome.
Brett House (02:37.257)
In hotels. So Jared, welcome to the t welcome to the podcast. And did I get everything right about your background? What would you like to tell the tell the audience in terms of your trajectory?
Jer Tippets (02:37.353)
There's not.
Thank you. Yep, you did. Yeah. So, you know, data is so interesting and data and collection is so interesting. It's so boring. My wife tells me every day. She does not like to hear about my everyday job, right? Because it is very, very boring, but it is so important. And so I actually started my career way back in the day at a little tiny startup called Omniture on the implementation floor is where I started. And before that, I was actually...
Brett House (03:06.135)
Yep. Web analytics. Yep.
Jer Tippets (03:07.933)
data analytics, that's I'm the lifetime or so. But before that, actually start out in graphic design, which is a weird jump going from creative to analytical like this. But the reason I got into analytics in the first place is I had a creative director who did not agree with my designs. And so I got into analytics to prove him wrong. And that was the whole reason I got into analytics. I love this idea that I can use, can collect data to show the effectiveness of how something works and why it works. And I've been using that in my career ever since.
And it's kind of been my whole theory is that you can collect data, right? People collect data all day long. And so many people say collect all the data. And I love when they say all, because that means they don't know anything. And they're just hoping that somebody else will figure it out. But my whole career has been about finding ways to find actionable and relatable and useful data in all of the static that we collect. Because there's so many data hoarders out there. mean, how many many projects you've got on there? We just want to collect everything in our data lake. Data hoarders, man, they are the worst. And so
Rio (04:01.197)
Data hoarders.
Brett House (04:02.892)
De Yeah.
Rio (04:04.416)
I love that.
Jer Tippets (04:06.151)
But we need to be able to focus on actionable, decisive, realistic data. And that's what my entire career has been about is how do we collect it and act on it? And now how do we bridge that to make it actionable using AI? Because you can collect a ton of garbage and AI is going to give you more garbage. It's garbage in, garbage out, same with AI, right? And so my whole career has been about how do we find the best data to act on to really drive the business forward and make the most impact possible? So yeah.
Brett House (04:24.149)
Yep. Yep.
Jer Tippets (04:33.213)
That is my career, that is what I'm all about, and that's I hope to be able to do in until I retire, because I love it every single day.
Rio (04:34.839)
Cool.
Brett House (04:39.264)
How long have you been at Hyatt?
Rio (04:39.671)
Well Chair we're
Jer Tippets (04:41.541)
I've been high for quite some time now. Started out as consultant, kind of went back and forth a little bit like that, but it's been quite a while. In my career though, I've worked with 26 different organizations, everything from large media companies. I worked on the Olympics for a number of years. I've worked for nonprofit organizations, for retail organizations, went from consulting to full-time, back to consulting, back to full-time. I've kind of played the whole gamut, which is a lot of fun, because it gives you the perspective of what other people are doing right and what other people are doing wrong.
Brett House (05:04.055)
Yep.
Jer Tippets (05:11.505)
I loved in the intro when you talked about your relationship with Marriott, how they can't get your identifiers right and identifiers are the cornerstone of what we do. And I love being able to work with so many different organizations because even though you're a multi-billion dollar international behemoth, doesn't mean your data is right. And I love going in and learning from these different organizations. It's so...
Brett House (05:11.735)
Yeah.
Brett House (05:27.468)
Yeah. It's so frustrating. I'm like 'cause I know how this stuff works behind the scenes. And I'm like and my and I is like, why isn't this working? And I've gotta you know, like even from using the app, I have to get her to do the two factor authentication because they cannot resolve the two identifiers, persistent identifiers which are our email addresses, to one household, to one account because we've got different last names. And and I no matter what I've done, they can't fix it. And I and I'm like, This is bizarre.
Jer Tippets (05:37.837)
yeah!
Jer Tippets (05:44.179)
Yep. Yep. Right.
Jer Tippets (05:52.677)
They can't do it. just imagine that behind the scenes is... When it matters going on behind the scenes, you're giving your perspective, you've been a consultant, you do business travel. If your wife's like mine, she does leisure travel. So when your personalization engine comes into play, they're trying to give you business and pleasure at the same time when it comes to this. It's just chaos. It's absolute chaos. But I'm going to find that person. I know people at Marriott and I'm going to find out for you. I'm going to fix it.
Rio (05:55.246)
Well someone's hoarding data. That's probably what's happening. Someone's hoarding data.
Brett House (05:57.564)
Ha ha ha.
Rio (06:17.741)
Yeah. Some data sharing, not data hoarding. So so so Jair, we're thrilled to have you on for a bunch of reasons here. I mean this is gonna be I we could tell when we first started talking before the we even hit record button is gonna be a fun, fun conversations living up to that so far. And you know, I think this is timely, as as Brett mentioned a minute ago. This is a this is a topic that I I think it's kind of a topic that people
Brett House (06:18.794)
Yeah, yeah, I need I need some data stitching.
Jer Tippets (06:25.135)
Thank you.
Jer Tippets (06:35.006)
Thank you.
Rio (06:44.469)
I don't appreciate as much as they should, right? They take these personalized experiences for granted. They take everyone talks about personalization, every everyone talks about dynamic creative, about all these other things about decisioning, right? But none of that works if you don't have the data right. Also I love the fact, by the way, that you came from the creative side and went and went over to tag. We've had so many great people on this pod who actually who started on the humanities or s or something or something, let's say not technical, then found themselves in a
successful long-term technical career. So great. A lot of our hosts are very and in fact both both Brett and I both have but yeah I was at Journo. We both have political we both have we yeah, we both have like, you know, definitely not technical backgrounds, although we've worked in tech for many years. So but when you but I but I guess f my perspective, what was really interesting about this is when you look at how these companies run, how marketing teams actually function, optimization decisions, personalization engines.
Brett House (07:23.424)
Yeah, writers, b yeah, the literature designers.
Jer Tippets (07:24.776)
love it.
Jer Tippets (07:40.169)
Mm-hmm. Yep.
Rio (07:43.67)
Even these AI models that everyone's trying to build and deploy right now, data is the foundation of that. And data is only as good as what you're collecting and how you're organizing it. So as everyone rushes to embrace AI, rushes to every, every, every organization I talk to talk to now is thinking about decisioning. That's the big thing now. As everyone tries to get this right, I think this is a conversation that's maybe in some way sh sh more important or at least more foundational. So welcome to the pot.
Jer Tippets (08:10.537)
Well, thank you so much. And I love the word that they use there, it's foundational because it is so important for people to talk about data early in the process. I hate it as a consultant, in the years past, we've done consulting where you're brought in at the end of the project and like, we build this great new app. Now just slap on analytics and give us the best. And you're like, well, what's your purpose? What are you trying to get out of your data? And I love what you talked about. There's people from lots of different paths and lots of different ideas and data is just something that exists.
in the world. don't understand that it needs to be created. It needs to be structured. It needs to be strategized on. And so many people think data exists in the world and they can just throw an AI agent on top of it and make sense of all this chaos. And the truth is that you need to have a fundamental strategy when you start your project. What am I going to do out of this? Right? Like I've been brought up so many projects where like we want to know
Brett House (08:57.334)
Yeah. Yeah.
Jer Tippets (09:02.723)
X, Y, and Z. I'm like, how is your analytics built? Are you product analytics? Are you marketing analytics? Are you customer analytics? What is your perspective on why you collected your data? And like, no, we just collect everything. And I'm like, well, then it's bullshit. It doesn't mean anything if you don't have any strategy behind it, right? And so.
Brett House (09:15.948)
Yeah. Yeah. Yeah, and and strategies gotta define, right? It's gonna define governance, it's gonna the rules of the road in terms of what we need, what we don't need, and then how that data's structured and and governed, right?
Jer Tippets (09:28.417)
100%. 100%. Yeah. It's so fundamental, it's usually thought about so late in the process that it becomes useless at that point. And no one knows what they're going to do with it. Because they're hoping some magic tableau dashboard is going to give them the secret answers of increasing ROI by 25%. But I'm like, there's no strategy, guys. You can't just throw garbage in a pot and hope it's going to turn into something magical. It needs to have strategy and governance behind it. Yeah, go ahead.
Brett House (09:49.356)
Yeah, have
Yeah, have you have you found in your career that that that you've seen sort of the rise of this as an important you know, back to that sort of kind of the theme and thesis of this episode is that that it's r data collection and normalization and everything that goes along with that is a much is a much more strategic function versus kind of an in in like how they used to think about it is like a back office implementation task, right? 'Cause there's all these downstream impacts across whether it's whether it's content production, advertising production, marketing, analytics, product products
Rio (10:09.676)
Agreed.
Jer Tippets (10:14.535)
Yeah. Yo, yeah.
Jer Tippets (10:20.275)
Yep. Yep.
Brett House (10:22.169)
stuff, right? I mean have you found that your role has become more central? I mean are you seeing an evolution in how early you're brought into the process to to affect positive change before things break or or if you get brought in too late?
Brett House (10:41.845)
We lost ya, we lost ya.
Rio (10:46.015)
Your audio is
Brett House (11:01.279)
No, your your speaker is off. Some somehow the audio is not picking up the mic.
Rio (11:05.633)
Yeah, your mic the mic cable might have got disconnected.
Rio (11:18.327)
Yes. It's a l it's it's not very loud, but I can hear you.
Brett House (11:19.059)
A little bit.
Brett House (11:29.739)
Yeah, either try to turn up or you could jump off and jump back in. oftentimes that does and just make sure that your audio settings are correct. And we can edit this part out.
Brett House (11:58.169)
Okay, we are yeah I would jump back out. Jump back in. Jar is frozen.
Rio (12:05.088)
No, he's totally frozen. I don't know if you kicked a cable or what happened, but
Brett House (12:09.324)
Well yeah, something happened and he he was moving around and then he and then the screen went blank.
Rio (12:14.57)
Yep. yeah, while we're so by the way, the guy from
Just heard back from that guy Patrick. He he can talk tomorrow Eas Eastern times, eleven to twelve thirty tomorrow.
Brett House (12:26.583)
Yeah.
Rio (12:36.95)
Which I can do.
Brett House (12:38.006)
Yep, I think that I think that works. Let me just 'cause I think that was right around the same time as our
Rio (12:44.68)
Actually actually you know what? Can you do one to three?
Brett House (12:48.961)
Yep.
Rio (12:50.836)
One to three eastern.
Brett House (12:51.924)
One to two. Can you hear us? Are you audio connected?
Jer Tippets (12:54.441)
I can hear you fine. Can you hear me fine?
Brett House (12:57.216)
Yeah. You just I think you gotta turn up it's it's not like I've got it I've got it turned up pretty loud. It's not picking up very loudly, so you either have to
Rio (12:58.188)
It's
Rio (13:04.434)
Yeah, either move the mic closer or maybe switch the mic.
Rio (13:15.084)
So while doing that. So how about like l one to two thirty tomorrow?
Brett House (13:19.552)
Yeah. That's fine.
Rio (13:24.076)
Okay, cool. I'm gonna C C U.
Brett House (13:28.042)
Heads up, recording will continue in a new track.
Brett House (13:38.74)
New inputs, however, the d the transition there is the chance. Yeah, so we're now we're not hearing you at all.
Yeah, it's it's super it's just super d do you have airpods in?
Jer Tippets (13:48.694)
I'm putting AirPod in. I'm gonna s switch over for my other one.
Brett House (13:50.792)
Okay. So yeah, you might want to jump out and then just say, I am using headphones and make sure it picks your AirPods and then
Rio (13:57.259)
Yeah, just just log out, log back in. That's the best bet here, Chair.
Yeah, lo log out, log back in. That just and just say 'cause it probably doesn't is not picking up your headphones and you just have to reinitialize it.
Brett House (14:17.889)
Well now now your internet's Yeah, you're we're getting a s a super lag in internet. Actual recording is higher quality, it does say that, so
Rio (14:18.312)
Really ch
Brett House (14:30.505)
Yeah. This sounds better. Here tri just tri Yeah, that's fine.
Jer Tippets (14:38.468)
I'm sorry guys. Alright. My apologies.
Brett House (14:39.135)
Alright. it's alright. Yeah, that's good. You're you're yeah, you're a little you must be on just Wi Fi not hardwired. That's what I'm or are you hardwired?
Okay. Yeah, no, it sounds better now. Let's yeah, so so yeah, so basically the last question, I think I got the full question out about data collection is a strategic function, not a back office implementation task. and and how critical it is to the downstream of different use cases, you know, analytics use case or marketing, you know, personalization content, dot dot dot. So yeah, maybe maybe you start there and just you know, like we can just cut it cut everything in between.
Jer Tippets (15:16.202)
Yep, yep. Yeah, absolutely.
Rio (15:19.98)
Yeah, we'll let it all that spot out. Don't worry about
Jer Tippets (15:21.932)
Well data's so interesting now these days, right? Because it's used for so many different things. What's interesting, what we used to be able to do is collect data for data analytics. It went to Google Analytics or Adobe Analytics or what was you know, one of the platforms, but now we're being pulled into so many different streams at the same time. It's going into our personalizations engines, it's going into our marketing attribution platforms, it's going to all of our marketing channels, it's going into our back end data systems, it's going to AI. The downstream effect of so much data that we have is becoming so complicated.
Rio (16:31.574)
So JR, I bet you're called into a lot of discussions or projects where someone's building AI model or trying to build a decisioning engine or implement personalization and then they that either the tagging was wrong or the structuring of the data that was tagged and collected was wrong. is that something you see pretty frequently and like what do you like what do you do about it?
Jer Tippets (16:50.09)
Well, especially at Hyatt, right? Hyatt has a very long and storied history with with with data. They're a very forward thinking, technology focused company, and their data collection is decades old. the some of the code that I'm looking at is over fifteen years old, which is really, really interesting because most implementations that I work on is, hey, we implemented this with a guy, you know, two or three years ago and things changed. We have fifteen years of history that we're trying to build on and trying to keep that data, you know, accurate and and and and flow and make sure we don't have any changes in the
Historical data is so important. So when someone comes in and says, Hey, we need to change how we're focusing this on AI to understand are they using AI in search or how do we use our search AI feet features better, or how do we create more personalized experiences there, they just want to bolt on another aspect to our data collection strategy, and that's so difficult because going back to it, and we'll go back to a hundred times. Identifiers is the hardest thing trying to stitch a user, the correct user, across multiple experiences, both app and web and in property to be able to.
Brett House (19:01.557)
Yeah, we should dig into Rio, we should dig into the the one built on fifteen year old code and and what Hyatt's doing about that because I've talked to a lot of engineers and product leaders that are like, just d don't keep bolting on to old code.
rip the code out and replace and rewrite it almost immediately with with the with the help of AI. And I think so that's one question is how do you kind of manage that? And then I think the other is there's a big first party data question here, which people have you know put up on a pedestal, but it's it's it's not the same as permissioned, accurate or even useful data. you know, to the point of of, yeah, Marriott's got first party data on me, and my wife, but but they're not using it in a way that's
Rio (19:16.042)
Rewrite it.
Brett House (19:43.837)
improving the customer experience. In fact it's sort of a it's a broken app website login experience because of their inability to stitch these things together. So so how how
Rio (19:54.433)
And Gier, we did not we did not plan to like be be attacking your competitors on this. Like this this was just is just house.
Brett House (19:57.922)
Ha ha ha that that's just Yeah. That's just me riffing it's just a personal experience of of like I realized there was an identity resolution problem at the core of this and it and it continues to follow me around. So so yeah, so so how do you deal with that? 'Cause you guys amass a huge amount across all of your locations globally of of data. There's different levels of permission, right? There's different levels of quality, and then the fact that you've got this fifteen year old code in some places, how do you kind of manage all that?
Jer Tippets (19:58.902)
I'm not attacking them. I'm not attacking them.
Yeah. A hundred percent. Yeah.
Jer Tippets (20:25.75)
Well and and it's when you say you have fifteen year old code, like the developer inside of me says, What the hell? Fix that. Rip that out right now and fix that. But the reason we keep it around is 'cause it works, right? It works. It works well for our customers. And that's the reason why we still have some of this like we go in there every single sprint and we say, What can we get rid of this time? What w what can we refactor? What can we modernize? And there's things that can modernize, there's things that are improving over the platforms, but at the end of the day you're like, But I don't want to touch it because it works. And it's what we've reported on forever and that's important to what we need to do.
Brett House (20:35.775)
Okay. Yeah.
Jer Tippets (20:55.744)
And I think that's important thing as well to say, do you get rid of it just because it's old? Or do you get do you keep it because it's useful? And that goes back to the story before of w how are we using the data and what's useful and what's not. And so all those things when you come into into play, like I I my kids don't give what I do at all, right? They have no understanding of what I do. They don't even kind of care. I try, their eyes they're the they
Brett House (21:00.448)
Yeah.
Rio (21:17.14)
Not just your kids, brother. It's like my yeah, I think my daughter's no idea. She once she said, I asked her what do want to do when you grow up? She said, wanna do your job. I asked her why. She said, Because I can board board the airplane early
Brett House (21:29.087)
Yeah, I I can get l tons of Marriott points and tons of United
Jer Tippets (21:31.468)
Exactly. No, I love it. I absolutely love it. Anyway, my kids don't know what I do, so I try to dumb it down for them. And say basically what I am is I am basically a tradesman. I am a plumber. My job is to take data from A and take it to B and make it useful, right? My kids don't know how our city plumbing system works. All they know is when they turn on the water, water's supposed to come out. And that's the same thing when it comes to data as well. People don't really care that much about how the sausage is made. As long as it's efficient, as long as it's privacy compliant, as long as you're not gonna get sued for
Rio (21:31.828)
Yeah, I I get f I I get treated well in the hotels.
Brett House (22:17.025)
Yeah.
Jer Tippets (23:01.264)
later and then you're like, well are what are you gonna do with them later? Is it AI? Is it products? Is it marketing? How are you going to leverage this data? And that's where I think the at the opportunities really come into play to say how are we going to use AI moving forward? One of the things that I'm nervous about is auto tagging because so many platforms are coming out and saying we're just gonna like Shopify. My my wife has a little flower shop here in town and we use Shopify. Like we're gonna auto tag everything. And I'm like that's amazing. Everything's gonna be done with AI. It's gonna be perfect. And it's all garbage. And I'm like
Rio (23:25.354)
You can rip through everything with AI and all and auto tag it. Yeah. I ev everyone's saying they do that now, yeah.
Jer Tippets (23:31.178)
It none of this works. It it's not customized in the way that it that works in a way that I know how to act on this data and make it useful. It's kind of cool, makes really pretty charts. My wife loves the really really pretty charts when things go up, but it really at the end of the day the customer needs to be able to say, this is actionable and this is useful and this is valuable to me. And that is that's always the trickiest part.
Rio (23:49.793)
So, JR, like you you brought up a couple of things I think are really important there. Number one was like the mentality of let's collect it all, right? So I think organizations have really like for years, since for as long as long almost as long as I've been doing this, right? But that's been one of the monsters. Let's collect everything, collect everything. Whether they do anything with it's another story. But that and but then I think the counterforce to that is privacy law. And like what should you be collecting, what can you be collecting? And if you collect it, you do need to ha
There there's responsible ways of using it and you do have to inform people how they're using or at least be able to. The people want to be i it create a D SAR or ratio with D SR, you need to be able to delete them. This is a place by the way I've seen Telium has been very strong in terms of collecting permissionable data that then becomes actionable. Love your thoughts on that about like how that's impacting your role.
Jer Tippets (24:36.396)
Privacy is amazing and I love privacy. only because I've seen so many developers who collect everything to an obscene amount. Like I I I've seen I'm one of those guys that keeps the console open in my browser when I'm on the web 'cause I love to see what other people are doing. I love seeing what they're tracking or what their marketing tack tech stack looks like. I love seeing what other companies are doing. And so many problems out there. If there's a privacy out there out lawyer out there that wants to attack a couple of companies, I got a I got a hit list that you can hit right now.
there's so many companies that don't know what they're doing with it be because it's so hard, right? It is true. It it it is true. And I'll be honest, there's been projects in the past where I go back and think about it like crap, I fired that pixel using custom JavaScript and I didn't follow the right load rules to trigger off privacy compliance to make sure there was happening. Even though on the front end on the on the homepage it is, but there's that one page, that one experience where it's not gonna happen, that's a problem. And so to your point, T Lium is like really there's a lot of vulnerabilities in there, right? And
Rio (25:09.984)
Funny. It's true.
Brett House (25:32.267)
Yeah. And it's a vulnerability. Yep.
Jer Tippets (25:36.226)
And and it's not a privacy is on or privacy off. It's like every event, every step of the way is very, very tricky to do. but I've seen developers who just collect everything. Like there was one the other day, I had a button click, they were tracking the button click, and there were over seven hundred data points they collected on that button click. Every single cookie, every single local storage, every single DOM attribute, every single thing about my history was pulled up on a button click, which was so excessive on the amount of data that they were collecting on that. And the funny thing was I already opted out of data collection across board.
Brett House (26:08.511)
Yeah, and d does it actually op you out of those seven hundred and there's actually some so a lot of academic we did some did some research and and there was a a academic r study at the end of twenty twenty three, twenty twenty four, a little bit outdated on Google Tag Manager, that found GTM present on about forty two percent of the top one one million websites in the US. I think it was actually global, it was US plus Europe. and and out of those forty two percent,
like half of them had hidden data leaks, incomplete disclosures, third party you know, cookie issues, right? There were some EU data protection issues that weren't being addressed. I mean to your point about a button click collecting f seven hundred data points, how do you actually build a system that manages that appropriately and and governs that that complexity correctly?
Jer Tippets (26:55.936)
Well it's it's so hard because even then, right, you have your data strategy, but you need to go layer higher and start with your legal strategy of what are we going to allow and how much risk do we want to take. I was I was sitting with some executives a couple jobs back, and they said, you know what, privacy compliance is the cost of doing business. If we if we collect it and someone sues us, that's the cost of doing business. My har
Rio (27:16.95)
Can be an expensive cost.
Jer Tippets (27:18.412)
They were a multi million dollar international company. That was it was in their it was in their risk you know line items in the budget. It was it they were totally fine. But it just made my heart sink because I thought, damn, that's the price that they're putting on the data is they're willing to take a half a million dollar hit because the data they're collecting is worth more than privacy compliance. And then how do you leverage that out with a with legal to say, here's our level of acceptance, and then pass that down to the developers and hope the developers know what the hell they're talking about.
Brett House (27:18.744)
Yeah. That's for sure. Especially in in Europe.
Rio (27:59.768)
Well well well chair, I think that's I think that's an important point you just made about like attorneys don't know, right? And then the thing is the people who are being asked to do these assessments. So I remember like there was a one company I was asked to do an ad tech assessment and I told them at the end of it, I don't think you're gonna be compliant with C C P A. I don't I'm not an attorney. I didn't write the law. I've read it, I think I understand what it's trying to do, and I think that you're gonna be in trouble, but you should hire a professional
who's actually do this, but they couldn't really find attorneys who had the specialization. So they ended up saying, you know, forget it. And then they ended up getting getting fined, you know, a lot of money actually. It was very expensive. And then and then but I but I do think and just to make a f a little a a joke about this too, the company that we should investigate is next door. I actually don't think it's possible to unsubscribe from nextdoor dot com. Just throwing it out there. I've tried unsuccessfully dozens of times.
Jer Tippets (28:46.486)
You know what? I so many I emailed.
Brett House (28:47.112)
Interesting.
Brett House (28:51.061)
Yeah. I'm I'm not sure it's it's possible to unsubscribe from the text that I keep getting from the Democratic Party. th they have got they have got my mobile ID and I mean I have unsubscribed and reported spam. I mean probably twenty five or thirty times, maybe forty times, and they just keep coming at me from f you know, any sort of political affiliation and you're getting you're they're f they're
Jer Tippets (28:56.556)
Rio (29:07.213)
Yeah, they're th they're spammers. It's crazy.
Jer Tippets (29:09.772)
Yeah.
Well, and it makes sense, right? Like you have one candidate who gets a list from the party and they copy and paste in Excel and they email it to their friend and then they copy and paste and email it to their friend. So d there's no way to do any kind of compliance when it comes to that, because when you opt out, you opt out from person A, but not person B, C or D that they emailed it to. There's no way for to make this work, and it's not just the Democratic Party. I've seen that across the board. Like how many times has a company implemented a new email platform and they just imported everybody? And whether or not they they hit that checkbox to import, you know, did they offer
Brett House (29:34.197)
Yeah.
Brett House (29:38.902)
Yeah.
Jer Tippets (30:11.148)
interacting with you because I'll be honest like there's companies that I'm nervous to walk in their stores because I know the cameras are there, I know they have facial recognition, I know they know exactly what I they're building a real-time profile on me in a physical world. I've seen those pitches, I've seen those demos, those things happen every single day, and I don't have a way to opt out of that. Same thing online. If I don't trust you to use my data properly, why would I give you my money? Why would I give you any of those things? And I get it. I am the 0.05% of people who understand this process
Brett House (30:29.845)
Yeah.
Brett House (30:38.454)
Yeah.
Jer Tippets (30:41.038)
in detail. My wife gives her email literally to everybody. She does not care. If you give her 10% off, you have her email address. And her inbox has like 56,000 unread emails and she doesn't look at email anymore because she's like it's just full of garbage. And I'm like, well you because you gave your email address to everybody and you refuse to opt out. But it's the same thing. And you can't unsubscribe, right? They're ask everyone's asking for money for for because it's it's the November's coming up. So I totally get it. And that's where the complexity of what we're trying to do and that's why everything that I
Rio (30:59.243)
And you can unsubscribe.
Jer Tippets (31:11.711)
There's an arc, right, in every developer's in every developer's lifespan where you start to learn things and you're like, I can do everything. I can build the coolest process. Like the first time I learned how to do an API, I'm like, I'm gonna take over the world, right? Because I can do all these amazing things. And I I see the same thing with AI now. It's like everyone's like, I can build my own my own app. I can take over the world. I'm like, you guys are just on the developer curve. You're gonna crash and burn in about two years, like everybody else does. But the important thing is everyone goes on this curve where you start to learn and the opportunities.
Brett House (31:24.895)
Ha ha ha.
Rio (32:30.893)
Yeah.
Brett House (32:30.942)
Yeah, and and there's a cost there's a cost to doing that. To your point, it might be a line item on a budget on a budget sheet, a PNL as a you know, but but I think Gartner I read a recent report from Gartner that said that
It cost the the average organization about thirteen million dollars a year for just to bad for bad data quality. Meaning if you're tagging everything and you're auto tagging and you're collecting seven hundred data points per button click, you're going to amass a bunch of poor quality data. There might be a large percentage of that that's really good quality. but point being, how do you how do you how do you think how does Hyatt think about that in terms of data monitoring during your collection process, data governance, and then unification across the
fifteen hundred, what is it? F fourteen hundred and fifty hotels, right? I'm assuming there's a centralized data warehouse, an ERP system, everything that feeds into one central
Jer Tippets (33:21.044)
Mm-hmm. yeah. Yeah. Yeah.
We we definitely have a lot of talks about that. And that's one of the tricks that we run into is that high's a growing brand. We acquire new brands all the time. And trying to bolt on existing systems into our enterprise system is very, very challenging. And so not only do we have our historical fifteen, twenty-year-old systems that have set up a certain way, we have to bring in these new systems as well and make them work with what we currently have. And so there's a lot of conversations all the time about how do we how do we make this work, right? And we have the maximum.
Jer Tippets (35:23.566)
It we can add new things if it's interesting and see if it works and if it doesn't then it's gone. But let's keep our consistency in play here. I think it's where a lot of companies l lose out on that is they don't have the discipline to focus on consistency for long term plays. It's like health. You can do any ninety, you know, ninety day tick tock health challenge that you want. It's not gonna really change your life, right? It health is something that you have to do over and over again for forty or fifty years to really see where the benefit comes into play. Data's kind of the same way. You have
Brett House (35:45.099)
Yeah.
Rio (36:13.185)
So Chair, we ran into each other a couple of years ago. I remember this is at a Telium Digital Velocity Summit in San Diego. I know you were speaking and I I I had a session too on that it was Snowflake, it was a good one. And I remember the topic you talked about with the work you're doing at Hyatt with Telium. I it w it was it was it was a good talk. It was about, I mean, the theme was treating customers like humans. And I think that works well with what you were just talking about, about giving people those delightful experiences.
Taking into consideration their preferences for privacy and what's what needs to be shared. I mean, obviously it seems it seems kind of obvious on one hand, but building high quality digital experiences can also seem kind of transactional. So wondering like, what is the philosophy at Hyatt and like how does that parlay into your data collection, data structuring, tagging, and all the things we're talking about? How does it all work together?
Jer Tippets (37:05.45)
It's complicated. It's messy. I have job security for a while, which I really, really like. but yeah, there there's
Brett House (37:07.718)
Yeah.
Jer Tippets (37:14.664)
Every good enterprise that has the focus on the customer. And hospitality is so interesting because you're if I'm if I'm online selling a pair of shoes, my experience with you is maybe 15 minutes, right? You you search on Google, you find my website, you like my price, you buy your shoes, you're gone. I might be able to build a long-term relationship with you, but you're not gonna stick around that long. Like I can't remember the last time I I don't know if my shoes came from Sierra online or I mean from Sierra stores or from Dixon Sporting because I can't really remember. They're just they're snikes. I don't really care, right? But a hospitality brand is different because
Brett House (37:53.205)
Yeah, it's the li it's the lifetime of like guest interactions, right? Which is everything from search to bookings to loyalty, all that stuff, yeah.
Jer Tippets (37:56.236)
Yep.
Brett House (39:17.364)
Yeah.
Jer Tippets (39:18.418)
all sudden their KPI strategy is completely different for no reason whatsoever. Consistency and accuracy is the most important thing here. And I think that's that's it's such a boring answer. Like that is like not the answer anyone wants to hear. Is solid consistency.
Brett House (39:30.965)
Well, and does it have an impact does it have an impact on down funnel metrics? Like does it affect customer lifetime value, long term value, due to the fact that you're you're more effectively communicating with with folks that are return customers or existing customers. Now, so that's your that's your sort of let's say your first party, you know, existing customer base, that data data store. How do you guys think about
net new customer acquisition, right? And leveraging d you know, incremental data sources like third party data sources or even leveraging your first party data to do a lot of what the the I I've been hearing a lot of hype about you know digital twins and synthetic audiences and modeling to to come up with some predictive capacity to say, hey if we target people of this type according to these types of profiles that we've modeled on like a good set of seed data, we're more likely
Jer Tippets (40:12.032)
Yeah.
Brett House (40:24.596)
to hit the right people that are likely to become customers and so it drives down our customer you know, our CAC, right? It potentially increases long term value of the new customers acquired. How are you guys thinking about that whole acquisition path in data workstream?
Jer Tippets (41:09.274)
just have the best time of your life. That's that's the experience. It's the best customer experience you possibly can when you're there. And so how do you find those people? That is a my my oldest son, he's 24 years old. He is not buying an all-inclusive, you know, trip to Cancun because he can't afford that. That's just that's just not what it is, right? But then how do we get not yet, he will, but how does he get to that point, right? He starts out at the lower tier so when he needs something he's at our entry level properties and then he works up and then we give opportunities to for to for him to be able to have a better relationship with the brand. So he trusts how
Rio (41:25.773)
Yeah, not yet. Yep.
Rio (42:24.664)
Well what do you what do you think about the future of C C D Ps? I'm curio I know it's a sidebar, but I'm curious to get your take on that 'cause we we've had a bunch of guests from C D Ps on here. We've written about it quite extensively. Love your thought on Adam. You can keep going with the with what you're talking but I'd love I'd love maybe as a get your commentary on that's interesting.
Jer Tippets (42:39.092)
Absolutely, yeah. Let me talk about that just here at the end here, because I it ties really, really closely with my view of what a C D P is. C D Ps are amazing, but they are so complicated because everyone is different. And a C D P gives a promise that we're gonna profile everyone and put them in these nice little buckets and so you can know who to target and how to market to them to get them to the next step of your journey. No one really fits in those buckets is the problem, right? Everyone has a different experience. And then so e even a Hyatt, when you're looking at a customer, we're like, Well, are you doing for business or pleasure? Are you are you traveling with fans?
Family to go to Disneyland or are you strictly for business and going to Cincinnati? Well, on Monday you might be going to Cincinnati, but Friday you might be going to Disneyland. And so how do we create that the right messaging to get you to get to the right hotel at the right time? It's very complicated. And it doesn't matter what your business brand is, everything is complicated in that space. And so CDPs are really, really, really interesting and are really really useful in certain aspects of it. I've seen the CDP promise far too many times where they're We're gonna profile.
Everybody and we're gonna bring in third party data and augment up your your customer profile and you'll know exactly who everybody is. And I'm like, that's great. I have millions and millions of customers. I don't have enough buckets to do that. So I love C VPs for what they can do. They can get some really interesting analysis, some really interesting behavioral things, but I don't know how valuable they are outside of fifteen or twenty use cases. You can't do personalization.
Brett House (43:41.643)
Yeah.
Brett House (44:07.456)
Yeah, and are those are those use cases mostly c existing customer use cases versus net new acquisition use cases?
Rio (44:37.101)
Wait, wait, so this is for web personalization? I'm gonna they wanted real time personalization?
Jer Tippets (44:39.028)
It was pro web personalization. They they wanted to know what the propensity of the customer was to buy before they even landed on the site. And I get the value of it. I'm like, that would be so amazing to have. And it's like, so what we're gonna do is we're gonna we're gonna bring in this data from a third party data source and we're gonna create these profiles before they even come to our site, and then we're gonna be able to stitch it in real time to get it done.
Brett House (44:46.987)
Hm.
Rio (44:57.749)
We're gonna try to match them somehow when they arrive with you know, I wonder what the match rate was even, right? That's
Jer Tippets (45:01.328)
Zero. It was zero. We had terrible success with it, right? It sounded so cool and when he would pitch it to the executive board, it sounded so cool. But the real timeness of it all, those identifiers we're talking about earlier, can't couldn't be stitched in real time. It couldn't be happened fast enough to know who the customer was before they even acted interacted with the site. And it
Rio (45:03.885)
Yeah.
Brett House (45:03.914)
Yeah.
Mm.
Rio (45:27.309)
Yeah, I I don't know how that would work. That that is that I mean it sounds cool, it's like a minority report, but like yeah, that's totally impractical, yeah.
Jer Tippets (45:30.622)
It sounds cool. And millions of dollars were invested in that. And there were so many pitches to the executive to the executive board about how this is going to change the company. And it drove no value whatsoever. Because it didn't matter to the customer about what the personalized experience was. They came to the site because they wanted the product and they were gonna do their own research. And even no matter what these third party sources said, the first party data was the most valuable, 'cause it actually showed their intent. You can have, you know, four hundred data points about a customer from a third-party data source, and if they come to the site and
Brett House (46:03.475)
Yeah. Well no, but if but if you've got a C D P so if you're collecting all this data and you know and you can through predictive analytics, machine learning, you can predict that
You know, these there are thousands of combinations that can be used to to to predict whether someone's going to be a future h you know hi hi based on your i this massive sh amount of data that you have on your existing customers. That that go that yeah, that goes into a that can go into a model for targeting purposes so that you're that you're trying to hit people that that that hit a certain threshold of like, you know, you're checking the boxes of these fifteen, these twenty, these hundred and fifty attributes
Rio (46:25.133)
Yeah, once you know who they are. Yeah.
Brett House (46:41.079)
or what's predictive of person entering in at the the lower tongue lower rung level or or at the higher end property level. Right? I'm assuming th you're you guys are kinda saying that's a pipe dream, but isn't that what audience targeting is really about?
Rio (46:53.109)
R right, but Brett but but Brett, you think about it, like but Brett, you think about it. Someone has arrives on a website, you're gonna have maybe y you have some cookie information, you probably have an IP address, and maybe you have a temperat browser on it. You can you can have some information, but the ability to match that with any accuracy to a profile which which then you would use to personalize the experience, which is probably gonna be wrong. I mean you're like zero percent again, I'm not surprised. I mean I I don't think that would be more than five, ten percent best case scenario. If you have
Brett House (47:04.895)
Yeah.
Rio (47:20.651)
Great third party data. There's not that much great third party data out there. Like the ability to match that with any accuracy and then just say we're gonna then personalize. What if you get the gender wrong? What if you get the state wrong? What if the offer like you're you d y it's a much better risk you're gonna screw it up and have a terrible experience this person they'll bounce, to your point, G.R. So I don't know. So I just think
Brett House (47:36.315)
Yeah. And do and doing that all in real time at the arrival of a site, there's a promise there, but but very i and do you think that applies to to other
Rio (47:44.013)
Well that sounds like this what a SaaS salesman promised more than anything else, right?
Jer Tippets (47:47.2)
yeah, it was definitely a SAS salesman. And he made a great commission because they paid a lot of money for those tools. But but to your point, Brett, I mean, how hard is it to get a third party anonymous identifier stitched when when Marriette can't even get your email address figured out when you log in? Figure out who you actually are, right? And and the truth is, when you have one customer, yeah, that makes tons of sense. It's super easy. But when you have millions of customers going in real time and you only have so much budget for this kind of project, it's so hard. It's so complicated. And the problem
Brett House (47:47.305)
Ha ha ha.
Brett House (48:02.3)
Yeah. Yeah.
Brett House (48:11.818)
Yeah.
Jer Tippets (48:16.974)
Promise is our imagination is so big that yes, you can do all these amazing things and AI is gonna make it so much better. But I'll I'll tell you right now, the campaigns that I've seen that are more effective is when you just know what kind of ad they clicked on when they came to the site. If you're clicking on an ad that is the most simple thing. If your ad has a red pair of shoes, show a red pair of shoes and that's gonna be a lot more effective. It's it's it's a simple stuff, but the
Brett House (48:31.828)
Yeah. It's i the most simple of of data points, right?
Rio (48:40.238)
Yeah, you have the UTM parameter, you which ad they served up. I mean that's it's pretty simple, right? Yeah. That's
Brett House (48:44.488)
Yeah, so so so would you say back to that point about synthetic audiences, would you say because there are a lot of AI jockeys out there that say, What do I need third party data? We yes, we know third party data is dirty, noisy. like why why do we need that when we can come up
with kind of excellent predictive capacity of of you know, who's going to respond to X, Y, or Z, right? do we think synthetic audiences are useful for net new customer acquisition, or do we think it's more of a an academic exercise of of predicting who's likely to buy?
Jer Tippets (49:16.8)
I think it's really, really interesting and I think it's definitely worth the test. I personally have not seen a lot of success from synthetic audiences. I haven't seen it actually I've seen some really cool reports and some really cool studies on it, but I haven't seen a lot of dollar science. And that's the interesting thing with is it's it's very academic. It's
Brett House (49:24.565)
Yeah.
Brett House (49:32.052)
Yeah, it d it does seem academic, yeah. 'Cause and I think it to your point, Rio, it breaks down once you have to start matching these potential audiences to actual real identifiers that represent somebody's, you know, either a hashed email or or a correct household address or an IP address. That's where things start start to break down, right? Those identifiers that actually are persistent and linked to people.
Those things are often ina inaccurate or unavailable. Is that is that your point, Rio, you think?
Rio (50:02.945)
Well, yeah, I think that I I think real time personalization real time web personalization being its own kind of weird use case that everyone was trying to sell for a bunch of years. I mean, you know, ad Adobe being the main culprit here, right? And then but had theoretical value, but doing it was so difficult, it was such a heavy lift, the value you could extract out of it was so minimal.
Brett House (50:15.508)
Yeah.
Rio (50:26.455)
that it wasn't worth it and a lot of people bought things thinking they could do this and then tried it and it didn't work and then actually never ended up doing it. So it's like okay, do you r is it something you really need? In theory, if the data were in good shape, if the person if the matching were somewhat accurate, it could be a game changer. But just the ability to do it was so hard. I think people stumbled with yet with the execution. But looking but if you can match people to profiles, especially of first party data where you actually know people, incredibly valuable. But now look at the synthetic data, I mean I GR I think I
more agree with you. I I've seen some interesting things like synthetic focus groups. I think are actually super cool. Now, I I would I would love to see some studies comparing those to actual normal focus groups, which I I've never been a fan of either, by the way. So I yeah.
Brett House (50:59.572)
Yeah.
Brett House (51:05.001)
Yeah. It's sort it's sort of responsiveness and engagement metrics based on a whole bunch of inputs, but you know, and you don't have to actually go and spend the money and the time.
to bring actual human beings together and you can you can get an idea of what this campaign creative, et cetera, et cetera, is gonna look like and and deliver. Right. Brea, I was I was thinking is that I had this weird analogy in my head. What do they say that all that all thing that there's like a what is the de evolution cycle that everything i inevitably ends up as a crab? Right? Is that entropy, right? Where the the
Rio (51:31.926)
Entropy?
Also everything breaks down, right? Like it and the universe, yeah.
Brett House (51:35.934)
Everything everything breaks down to like a crab like creature, crab like mollusk. I'm sort of thinking everything in in media and marketing kind of breaks down to like maybe it's all about just direct mail. Maybe we're all because direct mail is is i i you know, it's a weird you I think you s yeah. Yeah, yeah.
Rio (51:48.065)
Ha ha ha.
It's making a comeback. The that and smoking they're making a comeback, Brett.
Brett House (51:55.616)
'Cause I mean direct mail is one of those things where i it's like all of this promise of profiling to the h to the nth degree, at the end of the day when you have a household address and you've got an individual that you know lives in that household. Well, of course, you know, based on postal records, some of that does get outdated. There's a great degree of likelihood, despite the cost, that you're gonna reach that person in that household with a with a direct mailpiece.
Jer Tippets (52:18.048)
I just bought a roof. for the ro the roofers sent us a mailer in the in the in the mailbox and that's how I found my roofer. I s went online, I researched thirteen, maybe fourteen different roofers, I looked at reviews, I did their you know, looked at their Instagram, all those all those things. And ultimately I bought I bought from the guy that sent me a thing in in the in the mail with a twenty percent off coupon. That that's that was it. Yeah.
Brett House (52:39.433)
Yeah, it's
Rio (52:39.693)
Well, okay. Well there you go. I mean you heard it here single noise. Direct mail is still still li alive and kicking. Slowly dying but still alive. but but looking at AI and how it's impacting like your profession and the way people are operating, I love your comments before, Jared, about okay, people think okay, we're gonna use this AI to rip through and collect ta collect data from tags and normalize it and we're gonna even use it for you know, to just tag things at a scale humans could never do, right? Or like, you know, like couldn't practically do because you wouldn't want to hire so many people or like
So like how how is AI impacting and is it has it been positive? Has it been negative? It sounds like there's some there's some definitely some downsides to it. Love to see like what like what are you seeing today?
Jer Tippets (53:19.574)
So I
I haven't seen a lot of positive yet, and I'll tell you why. And not that I'm against the industry, because I I I'm very pro AI. I love I I love the I the the the promise of what AI is. What I don't like is people who don't understand what AI is thinking that they can just fix the problem with AI. I have had a lot of coworkers, a lot of people in my industry lose their jobs because of the promise of AI. a lot of people think, I'm just gonna plug this into AI and get a report and it's gonna be just as good as as an analyst. Analysts have contacts.
Rio (55:02.079)
Really solid like digital team there by the way. Like some really smart people work there. Yeah.
Jer Tippets (55:04.082)
Amazing digital team. they've taken something so boring as looking at historical records and turned it into a billion dollar company. Like it is just phenomenal what they've been able to do. My favorite story there is that an executive asked three different teams to come to the table with with a report. I i I c what the report is doesn't r it doesn't really matter. The important thing is that each team came with a different view of the data 'cause they interpreted to what their team does. So they all came to the table, they didn't have the same definition of what success was, and then he said, Okay, let's talk about this and
it was chaos because no one was talking the same number and he lost his mind. And and the story is that chairs were thrown 'cause he was so mad. He's like, Why is this so flippin' hard to get the right number? Like this is not like complicated stuff. That's a human aspect of it. Exactly.
Rio (55:42.071)
No kidding.
Brett House (55:45.981)
Yeah. That that's the human aspect, right? It's the human bias, it's human incentives that are driving an output that's you know, even if they're not aware of of of the bias, right?
Jer Tippets (55:56.138)
100%. Yeah. Yeah. Absolutely. And so going back to what we talked about earlier, your data has to be accurate and actionable. And just because you can put a something into AI and get data out, doesn't mean it's accurate or actionable. It's gonna pretty. I a hundred percent believe that, but I don't know if it's right.
Rio (56:12.269)
Well, you you you're gonna get a probabilistic response that's may be close to the truth, right?
Jer Tippets (56:16.936)
And maybe close enough, but every single time enough to replace your entire team, I don't know. And that's where the expertise comes into play.
Brett House (56:22.664)
Well, but yeah, yeah, yeah, the expertise, but can can you build some of that expertise into a common knowledge graph that's that's enabled and leveraged by the entire organization, right? Like to me that's the the future of the AI first organization, there's a lot of talk about that, right? Is the idea of how do you centralize this knowledge graph? It's almost like an identity spine in in the marketing world, right? and and where where people are are going to the same source of truth
Right? And that source of truth could be multidimensional. in order to generate queries and responses and actionable insights and that sort of thing. I mean, are you guys pushing towards that direction? Is that something that that Hyatt's doing to cut to Yeah.
Jer Tippets (57:02.002)
Every organization is every organization is going for that at varying levels of of success. And it goes back to what we're talking about before is you still have have governance, standards, definitions, of what those things mean. And if you have a developer who's just been tasked with this ticket of creating of importing their entire data warehouse into their locally hosted LLM, is he gonna have the right definition of what a session is? Is he gonna have the right definition of what a visit is? Is he gonna have the right definition of of how to cross stage multiple identifiers? Maybe.
Brett House (57:08.029)
Yeah.
Brett House (57:11.975)
Yeah, yeah.
Brett House (57:27.336)
Yeah.
Brett House (57:31.036)
This is a multi functional team dynamic here that that has to be in play, right? Yeah.
Jer Tippets (57:33.94)
It's a big project to do that, right? And that and and hiring a team to say, Hey, go ahead and put all the code in there so you're obsolete in six months is a risky proposition because I don't know who you're gonna hire to b put themselves out of a job in six months. But theoretically, yes, all that is true.
Rio (57:51.105)
payment off they'll do Meta's found out you can do it, but you know, y it's debatable if that's been a smart strategy or not.
Jer Tippets (57:53.388)
Right, there's a whole lot of that. and you know, and and that's that's just the nature of the business. And and AI will mature and AI will get better and it'll have more data to draw from and it'll have better expertise in you know, in in in in multiple generations. But it makes me nervous. It makes me nervous that we're going to use AI to replace so much expertise and so much knowledge. Like in your meeting notes you talked a lot about what would I tell a future analyst or someone who wants to get into business analytics? What would
be your advice. And AI is the giant question mark to make. I don't know if they're gonna have an industry in ten years. I don't tell my kids in analytics, I don't really know if this industry is gonna exist in ten years in the same way it has in the for the last twenty. And it's it's a weird thing to say because
Rio (58:36.759)
Yeah. It's funny, yeah, like we we actually I know, sorry to interrupt, but I remember like that's interesting you said that and like I remember wanted to throw this out 'cause I know that you were you were not mature, and I'm imagining you ran into Adam Greco when he was there, right? part of the Yeah. Yeah, yeah, yeah. So he he's he he's OG analytics. Yeah, he's a friend of the pod. I I've known Adam forever, right? G really great person, like super smart. Kind of was like the Adobe Analytics, like the guru, right? Like w he actually liter literally wrote a book on it. So
Jer Tippets (58:48.951)
I know Adam Bird so well. Yeah, I Adam I'll linger.
Brett House (58:53.147)
He's been on the pod. Yep.
Jer Tippets (58:57.228)
Yeah, that's amazing.
Brett House (59:03.378)
Yeah, he wrote a book on the topic.
Rio (59:06.167)
But so we ha we had him on the pod to talk about the death of the death of analytics due to the rise of, you know, what we were calling vibe analytics back then. But I think the the conclusion being analytics is becoming like dashboards are slowly or maybe rapidly dying and it's more okay, like we I I want this, I'll ask AI, it'll create a report for me and I'll use it, that'll be done. I think that things are definitely moving that direction. I love your thoughts on that, based on what you're just saying, and floor is yours.
Jer Tippets (59:36.3)
My favorite quote a developer once gave me is that a a good developer, the difference between a good developer and a bad developer is which code to copy from Stack Overflow. I loved that quote back in the day because every developer and every analyst, it applies here as well, has access to so much information out there right now. And the difference between a good one and a bad one is knowing which data to pull and which data to act on. And that's the difference. And AI changes that perspective a little bit, but AI is also
Brett House (01:00:10.419)
Yeah.
Brett House (01:00:41.832)
Yeah.
Brett House (01:00:50.642)
Yeah, yeah. No, it yeah, and you have that experience, yeah. And it's like yeah, you're you're you're gonna like you said, there's a life cycle to this and that you're gonna see a lot of these small companies that have vibe coded and then and then sort of battle tested their their pro their prototype or their product. You're gonna see those products crash and burn.
'Cause they don't they 'cause they aren't built on the sound principles that sort of people that have the experience and the strategic acumen and the governance knowledge can build into these these products, right? That's that's and that's I think Rio what we've talked about a lot is that most people are saying it's the sort of data jockeying and the in the the mundane repetitive work that you can automate, the tasks, right? But it's the higher value work that people have to really lean into because that's where you're gonna have to add that's where you're gonna be able to add
More value to the process as the strategic value, the governance value, all that sort of stuff. and that takes experience. It doesn't, you know, so it so it really does challenge
the the earlier in their career folks that are coming in that don't have that experience, don't have that knowledge, those frameworks and strategies to be able to kind of manage this. And and I d I do think there's a risk with the democratization where everybody thinks they can they're a coder now. Every th everybody thinks they're a developer now because they can vibe code with Claude Code or otherwise. Right? That that's where things can get very messy without some top down sort of almost regulation within companies on how you manage this process.
process, right? Back to the the notion of like a centralized knowledge center that's governed and and can be implemented the correct way.
Rio (01:02:38.743)
Yep. And they will.
Brett House (01:03:04.274)
Yeah. And and context is key to this. Context maybe context is king, right? That's the one thing both technically and theoretically that seems to be the common underpinning, right? That you're you're creating you've gotta create context to the data that's being s that's being served into these these agents. Otherwise, they're gonna misfire. or they're gonna hallucinate.
Rio (01:03:26.069)
Yeah, Brett Brett Brett, I like that a lot, right? That's a really I think that's a very important point. You think about it. If dashboards and UI start to disappear, and I think they're going to destroy I'm saying that we won't have them, but I think people's time spent in them will be reduced because agents will be doing a lot of these things. Agents really won't care. If everything goes headless, agents really won't care like what systems they're pulling data from. I mean, and what what is really SaaS? It's a bunch of databases, it's a bunch of workflows, it's UI, right? So like
Brett House (01:03:52.788)
Yeah.
Rio (01:03:53.484)
As the UI disappears, everything's headless in these agents, people's time in platforms will be reduced. But the knowledge of what these databases and these workflows do and what's important is gonna probably become more important. So I actually think the people with the context, with the experience, someone needs to navigate this whole stuff. You can't just have a bunch of people who don't understand where everything came from and are just dealing with agents, they might get stuff back and it it won't be correct to your point, Jared. Like they're not gonna know how to even pressure test that.
Brett House (01:04:19.687)
Yeah. And that context needs to be programmed in from the data collection step all the way through the outputs, right? That that's where, you know, so it's so there's a technical component, I think, to context, right? which is governance and other things related to that. And then there's the the context that individuals have when when leveraging one of these tools, right? 'Cause at the end of the day it's a tool, right? You know
Jer Tippets (01:04:41.0)
It yeah. It's a tool. It's too and it's it's not that vibe vi vi coded apps and and programs can't adhere to those rules. They just have to be taught in advance that that's what's going on. And so that has to be part of the fun it's the same thing we we said since the beginning. It's about consistency, it's about standards, it's about rules, it's about understanding what the program really is, and that's what's gonna make anyone successful in this. AI is just another tool. It's not really revolutionizing anything, it's machine learning, it's the same thing we've had forever, it's just a fancy new name with it, they're really cool stock.
price attached to it, right? But it's not really that new. There's some really cool implications of it. There's really cool you know applications of it, but at the end of the day it's not really that new. And the the same button rules apply for 100 beers are still there. Yeah. Absolutely.
Brett House (01:05:21.915)
Yeah, I mean the fundamental machine learning principles behind it. But but the compute power and the compute capacity and the speed
is arguably revolutionary, right? in terms of of right, yeah, the amount of of like data processing and decisioning that like the average person now has access to. I mean think about when we grew up, I mean this was the the mobile phone, your iPhone had as much power and compute as the as the original space shuttle. Right? Now AI, the average person's got access to LLMs, right? I mean the amount the amount of compute capacity that like per GDP that human
beings have now compared to 10 years ago, 20 years ago, 30 years ago, is exponential. To me, it's almost evolutionary. So I'm gonna I'm gonna disagree with you. I'm gonna say it is absolutely it's evolutionary and it's revolutionary. But I agree at the at the simple premise that that there is a machine learning underpinning that's been there for for 50, 60 years. I mean data science is behind all of this. So that's not radically changed. It's almost like the compute power behind it has an
Enabled these things to be scaled. Right?
Brett House (01:06:47.345)
Yeah.
Brett House (01:06:56.371)
Yeah, yeah. Streaming TikTok and
Rio (01:06:57.207)
Yeah.
Yeah. But but it's just thinking of younger people though. So so J like it's I mean my one of the w like a hypothesis Brett and I were kicking around as we were thinking about this episode was like I actually think and we both agree that because data quality, data governance, data structure, data collection are are more important now than they ever have been, right? I think a lot more important. I actually think there's kind of a new sexiness or importance to to a profession that's you're
It's been kind of overlooked. It's been thought as a back office thing, all these, you know, but I think it's more important now and it's getting more visibility within the enterprises, which is a good thing. So what would be your advice to to younger people who are interested in this, who think, this could be I could have a good career if I because I think they could, right? I and I think this is going to become more important. Yeah, some things can be automated with AI. We can collect and consume and tag things quicker, but w but the expertise is going to be needed by people. So what's your thought on like people who are interested in this or just starting out?
Brett House (01:08:52.701)
Yep.
Jer Tippets (01:09:25.862)
give you a career for twenty years anymore. Things are constantly changing all the time. And that's the only advice I give to anyone getting into the space is understand that it will change and you need to change with it or you're gonna be left behind.
Brett House (01:09:36.593)
Yeah. Yeah, and I th I think there's a there's a topic a great book called Range and I think that's one thing that I give my my teenage boys you know a lot of input on. Is like it it's better to be an inch deep, mile wide
Rio (01:09:37.902)
Good advice.
Brett House (01:09:48.891)
Right? And then go deep when you need to, but you you understand holistically a lot of different things, right? Because we've talked about in the pod the depth of specialization, the hyphenated job, the growth of hyphenated jobs, meaning you've got you're not specialized and narrow in your focus. You've got range and perspective. That's going to be critical because you can go deep when you need to go deep, but you've got to be able to bubble out and see things holistically. And that takes experience, it takes reps, right? People people are
more comfortable with specialization when they start out in their careers. Yeah.
Jer Tippets (01:10:20.98)
I don't know if you see behind me. I actually have a a a zoom back in the from way back in the day back there. And I I I keep it there on my little shelf behind my desk because it's a reminder that things always die. Doesn't matter how cool I thought it was fifteen years ago, things always die and things always change. And I I I I keep that zoom there for that reason to say what you're doing today will not matter tomorrow. And you have to adapt, you have to grow, or you have to give up. Those your choices. So yeah. I love it. I love it. I I think I think for life's perspective.
Rio (01:10:36.061)
Okay, I like that.
Brett House (01:10:44.325)
Yeah.
Brett House (01:10:48.038)
Yep. So should we s yeah, it's de definitely good advice. Should we go to to quick hits? Alright. So the fur the future of first cause you deal with a lot of first party data at Hyatt, the future of first party data and customer identity.
Rio (01:10:53.836)
Do it.
Jer Tippets (01:11:04.624)
so important. first party data when it comes to customer identity is the most important thing. We are constantly creating new identifiers for our users, not because we're trying to make things more complicated, but because we have new systems, we need to be able to have them in. And the way we're changing the way that we use identifiers is by creating more master IDs that can be backstitched over time so we can stitch across users across multiple systems and be able to have some sort of trace of who the user is across all the systems. We've tried forcing everything into one platform, that didn't work. We tried having a million different
Brett House (01:11:24.839)
Yeah.
Rio (01:11:53.195)
What's one piece of customer data every brand should prioritize collecting?
Rio (01:12:01.518)
What's what what's one piece of customer data or one signal that every brand should prioritize collecting?
Rio (01:12:34.05)
So a persistent ID.
Brett House (01:12:36.624)
Yep. And it sounds like it seems like email is the most persistent of IDs, right? You know, from a
Brett House (01:12:48.42)
Yeah. I mean people change email addresses and they have multiple email addresses which you can resolve, but other things like household, IP, cookie, mobile IDs, all those things are are
d definitely not persistent in any way. so customer customer so w so just thinking about sort of the modern Martech stack you've been involved leveraging and using and having your kind of feet in the f fingers in the soil of a lot of these platforms, customer data platforms, what what do you think that that future for from a modern Martech for consumer brands like Hyatt looks like?
Jer Tippets (01:13:20.958)
Mm-hmm. What does our Mortex stack look like moving forward? Is that kinda kinda is that is is is that your question?
Brett House (01:13:25.562)
Yeah. Yeah.
Rio (01:14:13.602)
What's one privacy myth marketers still believe?
Jer Tippets (01:14:17.928)
One privacy myth. that no one sees the banner. I love the myth that no one interacts with the banner, that people could go to websites and just dismiss it as quickly as possible.
Brett House (01:14:30.096)
Yeah, be you see the click through rates, they're so abysmally low. yeah, that you th you think is this really effective media? but
Brett House (01:15:23.078)
Yeah, so two myths. One, privacy's a waste of time and and that the display banner is is underrated, I think is what you were suggesting earlier in that in that point. Yeah. No, that the display the display banner, the display advertisements underrated, right? And that and that privacy pe you know, people shouldn't take that for granted because there's a larger segment of the population that is opting out or blocking cookies as they as they traverse the digital ecosystem.
Yeah.
Brett House (01:16:09.327)
Yeah. So data readiness, what what is every organiz what should every organization do before investing in AI? That's been a big topic of of this conversation.
Rio (01:17:05.068)
Will dashboards exist in five years? And if you want to go even more rad in the limb, will SAS UIs exist in five years?
Brett House (01:18:00.739)
All right. Well that was that was great. Hey yeah, that was fun. thanks Jared and and thanks everybody for joining us. and we you can reach us at www.signal and noise.ai, subscribe to our newsletter, you can also follow us on YouTube, Spotify, and Apple Podcasts, and we will see you next time.
Rio (01:18:02.99)
That's fine.
Rio (01:18:20.46)
And and and and before we go, big shout out to our I guess your our joint partner at Dynastilium. They've been a great, you know, great I know you you're a long term customer user of the of their tools and they are I as am I. I've worked with them many companies and they are a sponsor of the pod that we we love working with. They're just great people, great tech and big shout out to them for pulling us all together.
Jer Tippets (01:18:44.34)
Giant shot helium. Love it. Thank guys. This has been so much fun. Thank you so much.
Brett House (01:18:44.837)
Thanks everybody.