Latest / Signal & Noise: Bringing Clarity to Media, Data, Tech and AI / Vibe Analytics: The End is Nigh for Analytics Tech? Brett and Rio chat with Adam Greco.
Transcript
- 0:03This is signal and noise, bringing clarity to data, tech
- 0:07and AI. Welcome, Adam Greco to Signal
- 0:15and Noise. It's great to have you on the
- 0:17podcast. Thanks for having me.
- 0:19It's great to be here. So a little, a little point of,
- 0:24of, of interest that I thought I'd mention is, is one our, our
- 0:28audience certainly wants to hear your background and your words,
- 0:31how you came to be one of the, the top data in digital
- 0:34analytics and product analytics experts in the industry.
- 0:38But a little point of, of similarity that we have is we
- 0:41both studied at the University of Sevilla.
- 0:44I was going to try to pronounce that in Spanish, but I know I'd
- 0:46be like the only but, you know, I'd totally kill it in the 90s.
- 0:50So you know something that that ages.
- 0:52Did you overlap or? Around the same time.
- 0:55I don't know. I was there in January to May of
- 0:5891. I was in 94 right for for a
- 1:02year. I, I was, I went through the UNC
- 1:04Chapel Hill program. I was at UMass Amherst, but I
- 1:08studied with their program and we were in at the University of
- 1:10Sevilla, that old building. Yeah, you probably studied in
- 1:13the tobacco factory. The tobacco factory.
- 1:15Yeah, I. Actually did not.
- 1:17I was crazy and didn't realize that studying abroad was
- 1:20supposed to be a party. And I actually took an intense
- 1:24advanced business class, business in Spanish, in Spanish.
- 1:28I learned a lot about the agrarian Spanish society and it
- 1:31was about a 45 minute further walk from the tobacco factory
- 1:35and I lived over by the in Los Romedios by the fairgrounds.
- 1:39So I had like an hour and 15 minute walk and then I had to do
- 1:412 times a day because we had to come home for the siesta.
- 1:44So yeah, yeah, yeah. I live in the.
- 1:46Natural Tobacco factory was or an old one to be converted.
- 1:50Yeah, it's an old one that was converted into the University of
- 1:52Seville. Yeah, and it's a spectacular
- 1:54building, a beautiful city. I was like 15 minutes walk from
- 1:58that. Yeah, so I go back every couple
- 2:01of years cuz Seville's one of my favorite places and I still have
- 2:04been pretty good about keeping up my Spanish.
- 2:07That's awesome. That is awesome.
- 2:08I once actually did a digital analytics presentation in Lima,
- 2:12Peru for 45 minutes straight in Spanish and I almost passed out
- 2:16afterwards but my head, my head hurts so much.
- 2:18Oh my God, you use a little bit more than the average 4% of your
- 2:21brain power, right? Like I found that I at the end
- 2:24of every day in Sevilla, and I took some classes in all of them
- 2:28are in Spanish, but some of them were in Spanish for second
- 2:30language speakers. And some of them were like
- 2:31native classes. And the native classes in it,
- 2:34you just, I always say, I always use this analogy.
- 2:36It's like if the elevator door opens and you're not paying
- 2:39close attention, you'll see that the elevator's already left and
- 2:42you're just looking at it. You're just completely lost,
- 2:44right? So it was the glutton of her
- 2:46punishment because after Seville I moved to Florence, Italy and
- 2:49learned Italian. That's remarkable.
- 2:52Well, there we go. So, yeah.
- 2:53So tell us a little bit about yourself and let us let our
- 2:56audience that doesn't know who you are, no where you where you
- 2:59coming from. Yeah, so I'm a Chicagoan native
- 3:03and I got into the digital analytics space in the early
- 3:072000s when I started ran running a website for the Chicago
- 3:11Mercantile Exchange. And in the early days of the
- 3:13Internet, no one knew why we had websites.
- 3:15And we're all told we had to build a website, but weren't
- 3:18told why. And so I just did a little bit
- 3:20of research and found this web analytics industry and took a
- 3:23chance on a little startup vendor out of Utah called
- 3:26Omniture and ended up being the 5th customer of Omniture and
- 3:30really liked that product and dove in with both feet.
- 3:33Started learning all the cool things you could do, and
- 3:36ultimately kind of knew more than the Omniture people.
- 3:39So they hired me and I became the 165th employee of Omniture.
- 3:43Was there in the Consulting Group for many years.
- 3:45Moved to Europe, started the Consulting Group there, started
- 3:48blogging about the product. Ultimately wrote the book on
- 3:52Adobe Analytics, The Literal. Book or the playbook, right?
- 3:57Yeah, yeah. The actually the main book back
- 3:59in 2012 on the product that's kind of going away right now.
- 4:03But the the main Site Catalyst product, after it was renamed
- 4:06Adobe Site Catalyst, went to work at Salesforce and became a
- 4:10customer again of Adobe Analytics and Digital Analytics.
- 4:13I was the global head of digital analytics for Salesforce
- 4:16Worldwide, then did consulting for 10 years and helped fix
- 4:20really bad Adobe Analytics implementations around the
- 4:23world. And then after 20 years in the
- 4:26Adobe Analytics world, wanted to learn a little more.
- 4:29And so I joined a startup called Amplitude and helped them kind
- 4:33of build out their products so that there could be a little
- 4:35more, more options out there other than just Adobe and Google
- 4:39Analytics. And that was really fun for four
- 4:41years and built some. And amplitudes known for product
- 4:43analytics, right? One of the years in product
- 4:45analytics. Exactly, Yeah.
- 4:47So Amplitude was in product analytics, Adobe was marketing
- 4:51analytics. And you know, at the time I had
- 4:54thought Adobe should get into product analytics.
- 4:56They didn't want to get into product analytics, but Amplitude
- 4:58was willing to get into marketing analytics.
- 5:01So I went there and then we threw in experience analytics
- 5:03with session replay and and it actually worked out really well
- 5:06because our Forester rating for Amplitude just came up recently
- 5:11and I think Amplitude was either ahead or tied with Adobe in the
- 5:15Forester wave, which is. Yeah, they were completely in
- 5:18the lead. Yeah.
- 5:20So it's pretty cool to see that. And of course, ironically, Adobe
- 5:24then got into Adobe Analytics in the end because of what we were
- 5:28doing in Amplitude. But yeah, so that was it.
- 5:30And then I was there for four years.
- 5:32And now I'm a product evangelist for High Touch, which is a
- 5:36composable CDP and AI decisioning platform.
- 5:39And so it's a big shift for me. So for the first time in 25
- 5:43years, I'm not in the digital analytic space.
- 5:45So I thought it'd be kind of fun to talk about it because I could
- 5:48speak much more freely. I don't have a a horse in the
- 5:50race, as they say. Yeah.
- 5:52And totally fair. And that's one of the reasons
- 5:55why we wanted to have you on that.
- 5:56And we, we had a conversation, I think it was a a couple months
- 5:59ago, right? You had sent me some articles
- 6:01you're writing on, you had mentioned that you can speak a
- 6:03little more openly freely about it now.
- 6:06And you know, we talked about some of the, some of the things
- 6:08that are happening to analytics because of what's going on with
- 6:12AII mean we're going to call this vibe analytics, a lot of
- 6:15what we talked about. But I mean, I, I think more
- 6:17generally just about what's happening analytics because of
- 6:20this. But before we jump in, Adam, I
- 6:23mean, I mean, when we ran into each other, you were at
- 6:25Amplitude as a product evangelist.
- 6:27You're now a product evangelist at high touch.
- 6:30It's, it's a newish title or let's say occupation.
- 6:36I mean, I, I, I think Brett and I both consider ourselves
- 6:38product evangelist to a certain extent, right?
- 6:40Love it. Maybe before we dive into vibe
- 6:42analytics love, maybe you could spend a minute and tell us like
- 6:45how you define a product evangelist.
- 6:46What does a product evangelist do?
- 6:48And like, like what? Like what did you, how much of
- 6:51this was you figuring out this role versus you finding out
- 6:54about the role? Love to hear a little bit about
- 6:56that. Yeah, it is a unique role.
- 6:58And this is probably my third time doing this role, maybe
- 7:03almost fourth because I worked for one little kind of startup.
- 7:06And I think the the idea of the role is interesting because I
- 7:10think of it as a marketing role. But I kind of fell into it when
- 7:15I was at Omniture because when I started blogging in the early
- 7:19days of blogs with Omniture, what I found is that if you
- 7:24educate the market, and most of my blog was educating, here's
- 7:28the cool things that you can do with Omniture Psych Catalyst.
- 7:31And I wasn't a salesperson, I was just a consultant who worked
- 7:34with lots of companies. And I was literally just
- 7:36sharing, here's a feature of the product.
- 7:39And here's how I have used this feature.
- 7:41And after those blogs became so popular, then we would have like
- 7:46a conference every year. It used to be in Utah, then it
- 7:48moved to Vegas. And I would have a session where
- 7:51I would basically take everything I did in my blog post
- 7:53and put it into a PowerPoint and share all these tips and tricks.
- 7:57And my session would have like standing room only.
- 8:00And people would be like taking all these notes and they would
- 8:03always come up to me. They said your session was the
- 8:05most useful of anything. I've learned because I can go
- 8:08back and do stuff. And so in a weird way, I kind of
- 8:12fell into being a product evangelist.
- 8:14My nickname in the Omniture days was Omni Man and I was like on.
- 8:19Twitter Did you have a little outfit with the well?
- 8:21It's funny because now Omni Man's really big because my son
- 8:24has showed me the new Invincible series and the.
- 8:27Is there an Omni man in the new there?
- 8:28Is an Omni man who's the main character of this new very
- 8:32popular series on Netflix called Invincible and it's a whole
- 8:35comic book thing that my son has told me about so I was a little
- 8:39worried I'm actually glad now I didn't get like sued for
- 8:41trademark but but ultimately when I did that at Omniture,
- 8:45then when I worked with this little product called Apollo,
- 8:48which was this cool little thing I was trying to build the ease
- 8:51implementations then I went to amplitude and I like literally
- 8:54reached out to amplitude said you don't know who I am but I am
- 8:57going to be your next product evangelist and then I did the
- 9:00same thing for high touch. I reached out to them and said I
- 9:02really like what you're doing and I'd like to be your
- 9:03evangelist and so. Yeah, it probably had to define
- 9:07it for them I'm. Assuming yeah, there's.
- 9:08I don't think there's a job description that generally is
- 9:11out there, ironically. I'm starting to see some.
- 9:13I think that's. Changing, yeah.
- 9:14Now we're starting to. And later this year I'm
- 9:16actually, if all goes well, I'm going to be publishing a book on
- 9:19product evangelism. So I'm going to write a book on
- 9:21what it is. But just to answer your
- 9:23question, I think the main thing is, is it's a little bit of a
- 9:26cheerleader role for both the industry, the product and the
- 9:31company that you work for. It's a way that you, it's a role
- 9:35in which your job is not to sell a particular deal, but to really
- 9:40kind of like help shape the industry and help understand,
- 9:44help the company work for, understand what's happening in
- 9:46the industry. So it's very much like a liaison
- 9:50between the company and the industry.
- 9:52And so the role that I spent a lot of time doing is writing
- 9:55blog posts about what I hear from customers and then try to
- 9:58extrapolate that to what's happening in the industry.
- 10:01I speak at a lot of conference events just like I know, I know
- 10:05for sure Rijo, you do and basically try to like hear
- 10:08what's happening at conferences, read what's happening online and
- 10:11try to kind of just see where things are going and where we
- 10:14fit in. And so that's been a, it's a fun
- 10:16role. It's a kind of unique role.
- 10:18And I think you need to have a a lot of experience to have this
- 10:23role. It's not like a role you could
- 10:24do right out of college. I happen to have a very large
- 10:27network of people that I can reach out to, to ask questions.
- 10:30And IA lot of times will just have meetings with people just
- 10:32to hear what's going on. And when you hear enough of the
- 10:35same things, you start to identify some trends that are
- 10:38happening and then you can share those internally and externally.
- 10:42Cool. Well, I, I love it.
- 10:44I think you've done a great job helping define this role in the
- 10:47industry. And you know, as I mentioned, I
- 10:49think both Brett and I, we're not officially a product
- 10:51evangelist, but I think part of it we both do day-to-day is and
- 10:53as well as many other people. So you know, a great job.
- 10:57Kudos to you for helping be a trailblazer in this respect,
- 11:00Adam. And then I guess getting back to
- 11:01the main topic at hand, Vibe Analytics.
- 11:04And I mean, the reason why we wanted to have you on is
- 11:06obviously your your deep background in digital web and
- 11:08product analytics. And it's an interesting topic
- 11:12because I think that right now we're looking at, because of the
- 11:15way technology is going, like the way we we can look at
- 11:20analytics and use analytics really could change.
- 11:23And a lot of people are thinking it's going to change because of
- 11:26this trend called Vibe Analytics.
- 11:27Now, I know this is a new trend and you could probably define it
- 11:30different ways, but it really, in my opinion, is probably going
- 11:33to change how we use data and how we interact with analytics
- 11:35tools. And generally speaking, as we go
- 11:37to more agentic, agentic UIS, maybe the way we use SAS is
- 11:42going to evolve as well, analytics being a part of that.
- 11:45So I think with your background being an early pioneered
- 11:48Omniture dropping force and amplitude and someone has
- 11:52really, I think had his finger on the pulse of digital
- 11:55measurement like writ large over the last couple decades.
- 11:58I think this is going to be a lot of fun.
- 11:59I think it's going to be good discussion and to the listeners
- 12:02out there, and I think this is important because we really
- 12:05don't know where things are going.
- 12:07I think some of us are starting to have beliefs, but we're
- 12:09really not sure how will, will analytics be more interactive?
- 12:13Will it be more vibey? Will we talk to platforms and
- 12:17UIS and will they interact with us?
- 12:19Will the concept of the dashboard even be around in a
- 12:21few years? I think these are, it's a really
- 12:23good question and maybe we should start there.
- 12:26I mean like vibe analytics. I know maybe it's not your
- 12:28favorite term, but I'd love to hear like, how do you define
- 12:30that? And like, like have you, how
- 12:33long have you been hearing it? And like, what is your opinion
- 12:35on like the the state and evolution of analytics as a, you
- 12:38know, as a discipline as well as a technology?
- 12:41Yeah, I think we are at an inflection point.
- 12:43So I think it's a good time to talk about it and it's always
- 12:45fun to kind of pontificate of what could happen in the future
- 12:48and try to make some educated guesses.
- 12:51I think the way that I like to approach things is always to
- 12:53understand how we got to where we are today.
- 12:56So really quickly, your listeners probably know this
- 12:59because they're pretty advanced, but honestly, digital analytics
- 13:03really started as web analytics. And that really started because
- 13:07of advertising, because digital ads were out there and we were
- 13:10basically buying ads to and we needed some place to send them.
- 13:14You know, a lot of times they couldn't just go to like a
- 13:16Google search list and just buy something.
- 13:18So we had to send them to a website.
- 13:20We wanted to know how often that was working and which ads were
- 13:24working. And so the original early days
- 13:26web analytics was really a third party, a non ad agency referee
- 13:32version of what's really converting so that you didn't
- 13:34just trust your ad agency to say, hey, you gave us money and
- 13:38of course it's converting like so that was kind of how it
- 13:40started. But then once you had JavaScript
- 13:43tags on your website, you could track everything that people
- 13:46were doing. And so it kind of got into path
- 13:48reporting and looking at KP is and all this and so on.
- 13:52And I think digital analytics has stayed pretty similar for
- 13:56for a while. You mentioned there was this
- 13:58convergence that happened where product teams thought of it a
- 14:02little bit differently in terms of product analytics.
- 14:04And a lot of times they were working with mobile apps.
- 14:07And when the iPhone came out, a lot of people are too young to
- 14:10remember this, but we had websites and then the iPhone
- 14:13came out. And when the iPhone came out,
- 14:14there was a mad rush for everyone to get an app and
- 14:18product teams became the thing to make apps.
- 14:21And these product teams really were like, an app is different
- 14:24than a website. So why would we put a Google
- 14:27Analytics or an Adobe Analytics or a Web Trends on it?
- 14:30And then a whole new crop of vendors like Mixpanel and
- 14:33Amplitude and a bunch of them came up.
- 14:35And basically the world split between I have one analytics
- 14:39tool on my website, another on my mobile app.
- 14:42And even though the same customer eventually was using
- 14:45both of those platforms, which made no sense.
- 14:47And so I think we finally got to the point where all of the
- 14:49digital analytics vendors. Have they've converged again?
- 14:52They've unified, yeah. Yeah.
- 14:54So whether it's Adobe, Google's still working their way through
- 14:57it because they've got, they've got a different mobile app than
- 15:00they do with their GA 4. But it's kind of coming
- 15:02together. But I think we're getting to the
- 15:04point where where all that analytics has converged.
- 15:07But I think at the end of the day, digital analytics has
- 15:11always been about going into an interface so that you create a
- 15:15bunch of precanned reports or maybe you have the ability to do
- 15:19some kind of almost like pivot, pivot table reports and really
- 15:24reporting on what happened yesterday, what happened last
- 15:27week, what happened last month. It's been a very reactive thing
- 15:30with the goal of identifying insights that you can then go
- 15:33back to the website or mobile app team and say, hey, I think
- 15:36if we change this or change that, maybe things will get
- 15:40better for our users, will sell more stuff, and then we can go
- 15:43back into analytics and see if our theory was right.
- 15:46So it's been a very reporting focus to very reactive now that
- 15:51we're entering into the world of AI, as we've all heard, like
- 15:54anything that is very kind of routine, those are the things
- 15:59that AI is starting to replace. So I'll pause there just to make
- 16:02sure we're like we're all on the same page of kind of where we
- 16:04came from and where we've gotten to now.
- 16:07And then we can dive into what I think are going to be some of
- 16:09the changes because of AI. Yeah.
- 16:11It's interesting you mentioned some of that stuff and like
- 16:13about I think 10 years ago no one would have considered a
- 16:16website a product, right. But I think that's most people
- 16:20probably would now and then looking at how are people
- 16:23interacting with products, whether they're mobile apps or
- 16:25websites or what have you. I think understanding that and
- 16:28using some of the same whether segmentation and personas and an
- 16:32analysis I think is very important.
- 16:34So I I have seen that convergence over the past couple
- 16:37years. I know, Brett, you had a couple
- 16:38questions about analytics creature.
- 16:39Do you want to ask right? Yeah, yeah.
- 16:41And I was, I was sort of thinking it's, it's a, you could
- 16:43put a frame around. I think you're starting to hint
- 16:45at this, Adam, like the criticality of data analytics
- 16:47for organizations, whether you're a brand or their agency
- 16:51or you're a publisher, right. It's obviously critical to
- 16:54survival, to differentiation, whether you're differentiating
- 16:57content or you're standing out amongst the noise of advertising
- 17:00out there. Do you tend to think of things
- 17:03through kind of an analytics maturity curve type of
- 17:07framework? And I've seen these frameworks a
- 17:09few times in my career where you think what you described as like
- 17:13your basic reporting to me is like descriptive analytics,
- 17:16right? And then you move to more
- 17:17diagnostic and then predictive and then prescriptive.
- 17:21And then AI starts to get into the self learning.
- 17:24Sort of categories. And it's just like this
- 17:26evolution of sophistication. I mean, how do you coach
- 17:29organizations to or traditionally coached
- 17:31organizations to think about that?
- 17:33Yeah. And I mean, when I was a
- 17:34consultant for many years, I was kind of disappointed with where
- 17:38people were on that maturity curve because I can't tell you
- 17:41how many organizations, even large organizations, are basing
- 17:45decisions based on vanity metrics, which is just like,
- 17:48awful. They're a big problem media, I
- 17:51mean, for. Sure.
- 17:52Yeah. I mean, I think, you know, I've
- 17:54tried to work hard to work with companies to understand, like,
- 17:57what were the key questions that they wanted to be able to
- 17:59answer? And the way I would pitch it to
- 18:02them is, I would say your business is now moving online.
- 18:05And one of the one of the things that sucks about a business
- 18:08being online is you lose this kind of human touch.
- 18:12You can't read people's faces when you're talking like you're
- 18:14in a store. A person who's in a store can
- 18:16basically say you hold up a sweater, they could tell
- 18:19immediately if the customer's like, no, I don't want that.
- 18:21They can kind of do that, but it's hard to do online.
- 18:24And I think digital analytics is a great way to basically listen
- 18:29to your customers, but you're listening implicitly using data.
- 18:33But it's infinitely harder than just reading someone's face and
- 18:37get and having a conversation with them.
- 18:39And most companies are just not really good at that.
- 18:41But you have to be able to figure out what are the
- 18:42questions you want to answer. And a lot of it is kind of
- 18:46making changes, but I was shocked at how most companies
- 18:50were doing just the real basics with digital analytics.
- 18:53And I, it was great for me as a consultant because I would come
- 18:56in and I'd say on a scale of 1 to 10, like you're like a three
- 18:59and I can get you to a 7-8 or a nine.
- 19:01And I would get paid big bucks to do that.
- 19:03So it was great for me, but it was bad for the industry.
- 19:06But even the best digital analytics program or
- 19:09implementation is still pretty reactive and we didn't have AI
- 19:14back then. So getting back to Rios point
- 19:16with like Vibe Analytics, where I think we're going to start to
- 19:20see is that AI is going to be able to analyze massive amounts
- 19:26of data and actually proactively help you make changes to
- 19:30websites and apps faster than a human could.
- 19:34And some of those will be good, some of those will be bad.
- 19:36I think a lot of the humans are going to be needed to kind of
- 19:39like review what the AI is about to do and make sure that it's
- 19:44not going crazy. But I think that ultimately the
- 19:48Nirvana for digital analytics is to absorb all the data as
- 19:52quickly as possible and impact a change because of the data and
- 19:56then measure that change proactively.
- 19:59Instead of having a team run a dashboard, have a meeting, talk
- 20:05about it, and then wait four weeks for a developer to make a
- 20:08new change to a page, then wait another four weeks to see if we
- 20:12have enough data to determine if that was a result.
- 20:15And then, I mean, all of that I think is going to be compressed
- 20:19down to like minutes. And I think that's the ultimate
- 20:22goal. And I think the current digital
- 20:25analytics tools that are out there are not.
- 20:28Ready for that. And I think there's going to
- 20:29probably be a whole new generation of of analytics tools
- 20:33that are going to come out that are going to kind of leapfrog
- 20:35that. Yeah, it's like skipping the
- 20:37first couple of steps in that in that sort of maturity curve,
- 20:40right. It's the, it's the, it's the the
- 20:42capturing of data, the diagnosis of data getting quickly to the
- 20:46ability to predict what's next or to prescribe a recommendation
- 20:50or suggestion. But you're saying the tools
- 20:52today aren't quite ready for that reality?
- 20:55Yeah. And I think the sad part, the
- 20:56thing that always frustrated me about digital analytics, and I
- 20:59love the industry and I don't want to crap on it, but the
- 21:01thing that really bugged me is that a lot of times if you think
- 21:06about what you have to be to be successful with digital
- 21:08analytics, you've got to do an implementation.
- 21:11You've got to get people who are good at analysis.
- 21:14You've got to build reports, dashboards, make sure you have
- 21:16the right questions, come up with the right hypothesis of
- 21:20what you should change, make a change, and then have that
- 21:23change be positive. And at that point, that's, I
- 21:26just rattled off probably about 8 or 9 steps at that point is
- 21:30when you can start to think about getting ROI.
- 21:33And that's why I think it's been challenging.
- 21:35It's like everyone needs digital analytics and wants it.
- 21:38But when I really talk to companies and find out, are you
- 21:41getting true ROI from it, It's hard to get that ROI because
- 21:45it's hard to be repeatable with the data and continually coming
- 21:48up with new ways that you're, you're improving the bottom line
- 21:51or saving the company money. And I think that this new, I
- 21:55think AI is going to shrink that time, time frame so that you can
- 22:00get to ROI faster. Like for example, I was talking
- 22:02to a friend of mine on the other day where I was saying, imagine
- 22:05a world where you literally have an AI and you point it at a
- 22:09website URL and in 5 minutes it tells you, here's the exact
- 22:13analytics implementation that you need, here's everything you
- 22:16need to track. We've already put the tags on
- 22:18the site. We're already starting to
- 22:19collect the data. You just have to push go and
- 22:22then the minute you push go, an hour later it says, here's the
- 22:26things we're seeing that you need to change on your website
- 22:28and it's proactively telling you all the things you should change
- 22:30in the website, in the mobile app.
- 22:32And I think we're getting there. I think companies like Amplitude
- 22:34are really pushing and trying to get there, but I think it's
- 22:37going to be a couple years before, you know, it's truly
- 22:41automated and and going like that.
- 22:43It's interesting what you said about the dashboard.
- 22:45I think it's always been kind of the the challenge with analytics
- 22:48is ultimately in order to consume analytics, you need
- 22:51dashboards. But then people get lost in
- 22:53these dashboards, right? You create all these different
- 22:55dashboards, all these different pages on them, and they show
- 22:58different things, like understanding what it means can
- 23:00be challenging. Yeah.
- 23:02Totally so when I was. It's a lot of data, not a lot of
- 23:05insight or easily accessible insight.
- 23:08Yeah, I'll give you an example. When I headed up analytics at
- 23:10Salesforce, I took the job and literally they told me
- 23:14afterwards that my job was to keep these 4 dashboards updated.
- 23:18And I was like, that's not what I want to do here, but they were
- 23:21spending 90% of their time on that.
- 23:23So I automated all of those dashboards and use this cool
- 23:27integration with Adobe analysts that use Microsoft Excel.
- 23:30And basically everyone got an e-mail every Monday.
- 23:32Here's the dashboard. And one person would spend an
- 23:35hour and say, here's a couple things that we think you should
- 23:37know. And then the rest of our time
- 23:40was all spent trying to figure out what things we could change
- 23:43on the website to improve our lead generation and so on.
- 23:46And so we moved from I, I, I have this idea of a thermostat
- 23:50versus a thermometer. We were like a thermometer
- 23:54organization that was just like reading the temperature, but the
- 23:58the difference between a thermometer and a thermostat is
- 24:00that you could change the temperature with the thermostat.
- 24:02And so I wanted us to spend 90% of our time on thermometer on
- 24:06thermostat projects when we were spending 90% of our time on
- 24:10thermometer and that mindset. Really sounds familiar.
- 24:14Because that's that's when you start to get more value.
- 24:16Yeah, it's interesting. I mean looking, I guess
- 24:18returning to Vibe Analytics, I mean, I guess the it's really
- 24:22based on the hypothesis that at some point very soon, but as we
- 24:27go to a more agentic LED or Gentic first web, right that SAS
- 24:31is going to change instead of knobs and controls and panels
- 24:34and dashboards. And you know what I mean?
- 24:37That's a lot of what as digital marketers we spend our time
- 24:40doing is going into these platforms and changing things,
- 24:43right? The you know, the to your point
- 24:46before, if you really look at vibe analytics, it's based on
- 24:49the hypothesis and that's going to happen.
- 24:51And then vibe analytics is going to be more like interactive.
- 24:54You're you're interacting with these SAS platformers more the
- 24:56way you're talking to people, whether it's text or speech.
- 24:58And if you really take that a step further, does this
- 25:02ultimately result in almost a de specialization of analytics,
- 25:05right? Like Will Will, are we going to
- 25:06end up in a place for literally not non analysts, non-technical
- 25:09people? You could argue this would be
- 25:12good, right? It's more democratized.
- 25:14We'd actually be be able to use this without deep training.
- 25:16What does that do to the profession?
- 25:18Yeah, it's a great point. And that's what I think is going
- 25:20to be a monumental shift there. So there's a couple different
- 25:24ways. There's a whole cottage industry
- 25:26right now of employees and consulting firms and agencies
- 25:31whose sole job is to tag, implement analytics and then do
- 25:37dashboards and tell you what's happening with those dashboards.
- 25:40All of those jobs are gone. They are all going to go away in
- 25:43the next year or two. You are going to go in and just
- 25:47literally ask a question through a chat bot and it's going to go
- 25:51in and just tell you the answer. Just like today, the idea of
- 25:55reading through 2 pages of Google results seems really like
- 25:59antiquated and you just assume you're going to just get the
- 26:01answer from AII think the same thing is going to happen.
- 26:05And I think the profound impact of that, I think the one who's
- 26:09the people are going to be hurt the most are people whose whole
- 26:12career is tied towards these mundane like building
- 26:17dashboards, building spreadsheets, or domain
- 26:20knowledge of a tool. Like I keep telling people if
- 26:23you're an Amplitude expert, a Google Analytics expert or an
- 26:26Adobe Analytics expert, like you're that's not going to be
- 26:30around in five years. Like the implementations can be
- 26:32automated, the data collections to be automated and the
- 26:35analysis, a lot of that's going to be automated.
- 26:36So you need to find another job and I do think it's going to
- 26:40have a profound impact on the digital analytics vendors.
- 26:44And this isn't the reason I left the digital analytics industry
- 26:48completely. But there is a part of it is I
- 26:51wonder what does the future look like for the Google's Adobe's
- 26:56amplitudes in the future because of AI.
- 27:00And here's how I'll explain it. If you use an off the shelf
- 27:04digital analytics product, any of them, Mixpanel, Amplitude,
- 27:07any of them, you're basically collecting data, sending it into
- 27:11that particular tool. They use a proprietary database
- 27:14to make sure that the reports come back really snappy and then
- 27:17you can do your analysis. Well, most companies are backing
- 27:22up that digital analytics data into a cloud warehouse.
- 27:26And so you're going to have the digital analytics data mixed
- 27:29with all of your other customer data.
- 27:31And so the theory that I have is even though these vendors are
- 27:35all doing the right thing and putting AI in their tool, if you
- 27:39are going to do a chat bot AI to say an Amplitude or an Adobe
- 27:42Analytics, like you'll get an answer.
- 27:45But you could probably do the same thing direct to the
- 27:48warehouse and get a similar answer.
- 27:51And if you don't, if the UI becomes an AI chat bot, then why
- 27:56do you need the pre canned reports that come out-of-the-box
- 27:59in all these digital analytics tools?
- 28:01And if you don't need the pre canned automatic report, then
- 28:05why do you need to be sending them any data at all?
- 28:07Why do you need to pay them for data collection?
- 28:10Why don't you just send all of your event data right to the
- 28:13warehouse? And someone is going to build a
- 28:16really cool AI on Snowflake and Databricks and AWS and Google.
- 28:21And why don't you just ask your question there?
- 28:23And so I wonder if the digital analytics vendors get squeezed
- 28:27out in the next couple years and if Snowflake Data breaks, if
- 28:32those become the digital analytics vendors of the future
- 28:36because they have all of the customer data and then someone
- 28:39just has to build a. Yeah, Yeah.
- 28:41They're like the Super highways of of sort of data management.
- 28:44We noticed that when I was at New Start, which became we, we
- 28:47were acquired by TransUnion is the biggest problem that we had
- 28:50with our big brand clients was data fragmentation.
- 28:53To your point, they had data in all of these places which are
- 28:56either at the analytics platform level, at the warehouse level
- 28:59like GM as an example, I had 60 different data warehouses.
- 29:03So half of our job from a data analytics perspective, which led
- 29:06to things like MTA and MMM outputs was really data
- 29:11unification, data normalization, sort of re platforming, which
- 29:15inevitably went to the cloud, it went to the snowflakes, it went
- 29:18to the data bricks. And guess what?
- 29:20We were building identity resolution and data in all of
- 29:24that. Stuff data cleansing.
- 29:25All within native build applications within snowflakes.
- 29:29But what you just said is, is, is funny.
- 29:31It, it, it's kind of why I moved from the digital analytics world
- 29:34to the composable CDP world, because I see that digital
- 29:39analytics is, is really a subset of customer data.
- 29:43And you probably want to analyze all of the customer data.
- 29:46And while some companies will try to shove all of their
- 29:49customer data into like an Amplitude or an Adobe Analytics,
- 29:53like, I think my gut and what I my bet is, is that these cloud
- 29:58warehouses are going to be the logical place for that.
- 30:01And so if I want to analyze all of the customer data, I can't do
- 30:05that with a digital analytics tool, but I could do it with a
- 30:07CDP and a CDP. It's been kind of an interesting
- 30:11like evolution. It's interesting you say that.
- 30:14I mean, because we had a guest on from a VC firm a few weeks
- 30:18ago. He talked about how he felt SAS
- 30:19was generally safe with AI disruption because you still
- 30:22need databases, right? You're still going to need
- 30:24certain functions that SAS tools have even if the UI changes.
- 30:27But I guess your point is maybe changing my perspective on this
- 30:30because if you don't need data collection anymore, if that's
- 30:32all automated with AI, right? If the the database architecture
- 30:36to store it because you're in the cloud data warehouse instead
- 30:38of let's say some intermediary analytics tool with the
- 30:42Hatterson's database, database's data structure.
- 30:45If the and even the dashboards themselves can just be done
- 30:49using that's using text and prompts on the fly instead of
- 30:52precanned dashboards. Like you've basically taken away
- 30:55every, like you've taken away everything that an analytics
- 30:58platform offers, right? And it's just you just it's
- 31:00either done by AI or it's stored, stored in the cloud data
- 31:03or else it's really interesting. But I'll give you another
- 31:05example. I'm sure you all have used Power
- 31:07BI, Tableau, and Looker. We're using Power BI, we're
- 31:10using all three of those, so. BI tools.
- 31:13Have you ever sent data to a BI tool?
- 31:18You don't. You send data to a warehouse and
- 31:21then a BI tool sits on top of a warehouse.
- 31:24Yep. Why is digital analytics any
- 31:26different? Why because of digital
- 31:28analytics? Are you sending data to a
- 31:30special third party SAS vendor? Whether that's mixed panel,
- 31:33Adobe Amplitude, you name it. And it's mainly because there's
- 31:37certain reports that are really difficult for BI Tools to do
- 31:40that are time series like path reports and, and persistence.
- 31:44But I just imagine in a couple years AI is going to be able to
- 31:48solve all that. And so I'd say 9 out of 10
- 31:51companies I talked to are, are backing up all of their digital
- 31:54analytics data in the cloud warehouse anyway.
- 31:56And they're paying twice. They're paying for all this
- 31:59event data one in like an Adobe. And then they're duplicating,
- 32:03they're duplicating it. And if you were going to sick AI
- 32:06at something, would you rather sick it on something that has a
- 32:10portion or all of the data? Now, initially it'll be bumpy
- 32:15because if you ask a digital analytics type question through
- 32:18AI to a cloud warehouse, it may not understand that this is like
- 32:22a digital analytics question. But how long is it going to take
- 32:24for AI to figure that out and be like, hey, this smells like a
- 32:27path report that they're looking for and it'll just come back
- 32:30with the right visualization. And I'm not the.
- 32:32Organization the taxonomy of the data which is going to be but,
- 32:35but even then, I mean AI probably starts to handle that
- 32:37better overtime too. We had a question a couple weeks
- 32:40ago about digital asset management.
- 32:41I mean, do you, is it more or less important with AI because
- 32:45you know, if you're using AI to classify meta, you know, to,
- 32:48let's say, identify and write metadata for things, do you
- 32:50actually need it stored in the dam?
- 32:52Because one of the big things the DAM does is offers the, the
- 32:54taxonomy in the storage, right. For actually grabbing things, do
- 32:57you need AI? So I guess that's interesting
- 32:59too. I mean, like if everything's
- 33:01going to cloud data warehouse and using AI to, to grab it,
- 33:05stuff it in there and organize it.
- 33:07I mean, I think that's very interesting.
- 33:09Yeah. And I'm not saying this happened
- 33:10tomorrow. It's just I'm looking at how
- 33:12fast AI is accelerating. And I think that there's a
- 33:17chance. And I think that the people who
- 33:19I know who are in the industry need to start thinking about
- 33:22like, maybe I'm wrong, but if I'm right, what does that mean
- 33:25for their job? I do think that it'll end up
- 33:28saving a lot of companies money because I think they're going to
- 33:31say I don't have to then pay $2,000,000 a year to Adobe
- 33:35Analytics. I'm just going to take all my
- 33:37data, skip Adobe Analytics, send it right to the warehouse, and
- 33:40basically be able to answer 80 to 90% of the questions that I
- 33:44answer today, but just do it right off the warehouse.
- 33:47And but they'll still need some application that actually do
- 33:50that search for them, right? Even if it is AI that's AI
- 33:54powered. And then actually give them the
- 33:55answers or some kind of dashboard.
- 33:57They may not dashboard, but some kind of analysis, right?
- 33:59But they'll still need something to do that.
- 34:01But and the the advantage of the of the data warehouse is whether
- 34:04it's Azure or AWS or data bricks or Snowflake is they're starting
- 34:07to sort of eat into large portions of the brand and
- 34:10publisher ecosystem. So the majority of both sides of
- 34:13the buy and sell of the of the industry for advertising
- 34:16specific are in these platforms. So when it starts to lead to
- 34:20data collaboration use cases beyond first party data, right,
- 34:24you need to have an easy way to do that.
- 34:26And your point solution platform that's specifically built for,
- 34:30let's say web analytics isn't gonna have the same capabilities
- 34:34of ingesting and data sharing and data collaboration, which is
- 34:38what the Snowflakes, it's like we used to do a ton of
- 34:41partnerships where a brand and a publisher partner would be
- 34:44sharing data in a privacy sort of compliant clean room type of
- 34:48environment right within a native building Snowflake and
- 34:51that can't happen. That couldn't happen in our
- 34:53separate individual data warehouse.
- 34:56It had to happen. Let's take it even further.
- 34:58Let's say that you're a Company A and you want to partner with
- 35:01Company B through a data clean room.
- 35:03But Company A is using Amplitude couple B Company B is using
- 35:07Adobe Analytics. How are you sharing data?
- 35:09But if all of them end up in these cloud warehouses, then you
- 35:13could have some data clean rooms where you can, you know,
- 35:16securely share and partner with people.
- 35:18So I think it's going to be interesting.
- 35:20And I think if I were the digital analytics vendors, I
- 35:24would be thinking how do I make my product work, but make it
- 35:27work in a composable way off of the warehouse.
- 35:31That to me I think is the future.
- 35:33There's a couple of vendors who've started doing that, but
- 35:35until the big vendors do it, it's not going to be a trend.
- 35:38But I think you shouldn't in the future, a company should not
- 35:42have to send a separate version of data to a third party SAS
- 35:47tool and also have it in the cloud warehouse.
- 35:49They should be able to just do it composably off of the
- 35:53warehouse. And that's happening in the CDP
- 35:55space very rapidly. And I think it'll eventually
- 35:57work its way to many other SAS products.
- 35:59And since we're talking about digital analytics, I think it
- 36:02might work its way to digital analytics in the next couple
- 36:04years. Yeah, it's interesting kind of
- 36:05the zero copy future composable feature.
- 36:08And I think you're with clean rooms.
- 36:09I mean like I mean like snowflakes already doing this
- 36:11already. Clean rooms probably not going
- 36:12to be another destination, right.
- 36:13It's just going to be part of your of your cloud data
- 36:16warehouse environment that has certain privacy rules applied to
- 36:19it where you can collaborate data and do your control and
- 36:22conjoined analysis, analysis and other things in there.
- 36:25Right. And both parties can can protect
- 36:27their data too and know that their data is not being usurped
- 36:29or used for other purposes, all that sort of stuff.
- 36:31Right. Yeah.
- 36:33So realistically, so it sounds like we're talking, we're
- 36:36hinting at Rio and I had a little conversation before we
- 36:38talked about this and we kind of geeked out on this.
- 36:40But we're it looks like we're coming up to like another
- 36:42innovators dilemma type of situation in the industry where
- 36:46there's going to be a bunch of players whether.
- 36:48It's for the analyst companies. Definitely, yeah, that that are
- 36:51that are going to be listening to their customers.
- 36:53We talked about the you know, if you haven't read the Clayton
- 36:54Christian, sure, sure. You have the Clayton Christian
- 36:56suit book, right? You listen to your customers,
- 36:58they're not going to tell you where the puck is going
- 37:00necessarily. They're going to say I, I, I'm,
- 37:02they, they're looking at what's right in front of them today,
- 37:05next quarter, the quarter after. That yeah, yeah.
- 37:07No one knows what it's going to look like, right?
- 37:10So, so do we have a like if there's a lot of successful
- 37:12CDP's out there, there's a lot of successful composable
- 37:15software companies out there, do we see a future where they're
- 37:18going to be the an innovators dilemma problem where they're
- 37:21just not going to change fast enough and they're going to be
- 37:23replaced by a whole new. Yeah, I actually wrote, that was
- 37:26the title of a blog post I just put out about a week or two ago.
- 37:29And I do think that that's, you know, we always go through these
- 37:33cycles. And if we start with digital
- 37:36analytics, I think that there's hundreds and thousands of
- 37:39companies that are using a current analytics tool.
- 37:43And I think those vendors are going to come to a pivotal point
- 37:46and decide are we going to just kind of like put ourselves out
- 37:49of business and come up with a new AI approach.
- 37:53And I think what I'm seeing so far right now is more that
- 37:56they're putting AI as a way to like be like a new interface to
- 38:03speed up doing the things that you do manually work is kind of
- 38:08really starting from scratch with their with the AI first.
- 38:12So I think that's definitely happening in the analytic space.
- 38:14And I think it's going to be interesting.
- 38:16I've already seen a couple new startups that don't have the
- 38:18baggage of a lot of old stuff. I mean like if you look at the.
- 38:22Baggage or the tech debt? Right, yeah, Adobe already has
- 38:24two analytics products. They've got the Adobe analytics
- 38:27one I wrote the book on. They've got a new one, customer
- 38:29journey Analytics CJA right. So they've got CJ.
- 38:32So they've got 2 analytics products.
- 38:33So now what do they do? You know, because they're trying
- 38:36to get everyone to CJA. But is, is like, should they be
- 38:39getting people over to an AI version of, of analytics and
- 38:42said, or is AI going to just be a, a quick little, you know,
- 38:45front end to what they already have?
- 38:47And I think underneath the tech debt of a lot of these legacy
- 38:50vendors, whether it's Adobe, sales force, you name it, I
- 38:53think there's just a lot of legacy technology and you can
- 38:56put a cool little AI veneer on it.
- 38:58I wrote a blog book. Staple it on.
- 39:00Top you're like a lipstick on a pig and something yeah, yeah,
- 39:03yeah. But like I think that's not
- 39:05going to cut it in today's world.
- 39:07Like you have to come up with a whole new way of thinking about
- 39:09it. And and I think the CDP space is
- 39:11further ahead than the digital analytics.
- 39:14So we could use that as an example.
- 39:15Like you had you've a bunch of vendors, Adobe, Salesforce,
- 39:19Helium, the Action IQ, many of them are being acquired because
- 39:22like their technology is kind of doing it the old fashioned way,
- 39:25the non composable, the non warehouse way.
- 39:28And they probably had to decide, do we want to start over or do
- 39:32we just cash out? And I think a lot of them are
- 39:33cashing out. And I think the big legacy
- 39:36vendors that we run into the most Salesforce than Adobe, like
- 39:38they've really great products, but they're just heavy.
- 39:42They take a long time to implement.
- 39:44They are quasi warehouse native composable, but not truly built
- 39:49from day one. And they're carrying over a lot
- 39:52of this kind of tech debt. And we're just seeing more and
- 39:55more companies that right now in an unstable economy have to like
- 39:58save a little bit of money. They they want to invest in
- 40:01their warehouse. They don't want to have multiple
- 40:03data copies of customer data for both legal and privacy reasons.
- 40:08And so all the winds are kind of blowing towards.
- 40:11Composable and warehouse 1st and at first when when high touch
- 40:15first came up with this concept like they were laughed at and
- 40:18the the analysts didn't cover it.
- 40:21A lot of the legacy vendors said it was dumb and now all of them
- 40:24are falling over themselves to position themselves as
- 40:27composable and as. Yeah, And can we and can I think
- 40:31it might be helpful for the audience.
- 40:33You know, my brother talks a lot about this.
- 40:34He's the president of Elastic Path, which is a composable
- 40:37commerce company just to define kind of the the imaginations of
- 40:41how that works. Like when you have ACBP and it's
- 40:44composable, it's and it's it directly integrates to a native
- 40:48warehouse environment like a. Snowfall, yeah, The idea in
- 40:50general of composable or warehouse first is that you are
- 40:54not storing data in an extra place.
- 40:57You are sitting on top of an existing data store, usually a
- 41:00cloud warehouse, but you're offering the same functionality.
- 41:03Now in ACBP, the key functionality is identity
- 41:06resolution, the ability to build audiences, the ability to do
- 41:10analytics on your audiences and then to activate audiences to
- 41:14either add networks or e-mail or maybe even just update customer
- 41:18records in Salesforce or other SAS tools through reverse ETL.
- 41:22And so those, that's the the feature set.
- 41:25But how you do it is very different.
- 41:27And if you, the easy answer is if you ever have to store any
- 41:32data in a vendor, then they're not composable.
- 41:35That's just, that's just, you know, an easy litmus test.
- 41:37But also you should. The other part of composability
- 41:40is being able to purchase different features with.
- 41:42That I was going to ask. The whole suite.
- 41:45So like for example, it's very difficult to buy.
- 41:48You can't really buy Adobe real time CDP without buying AEP.
- 41:52Now who knows, maybe Adobe will disagree with me on that, but
- 41:55that's not what I've heard from a number of companies.
- 41:58And so you have to do your profiling and your identity
- 42:00resolution. That's all in AEP.
- 42:02And then real time CDB sits on top of AEP.
- 42:05But like we just believe in our core is that the warehouse, the
- 42:11should be the cloud warehouses should be the center of truth,
- 42:14should be where all of your customer is.
- 42:16It's the most efficient, most cost efficient.
- 42:18And they are adding features to snowflake and data bricks at an
- 42:22infinitely greater pace than helium.
- 42:25Adobe, Salesforce are in their CDP.
- 42:28Like I don't think I would want to compete with data bricks,
- 42:30snowflake, big quid in that world.
- 42:33Yeah, data warehousing world. They're just it's just the
- 42:35economies of scale. Yeah.
- 42:37That you can. It's like if you think of what's
- 42:39in China, they have what's the cars they have that are
- 42:41competing with like Tesla, I forgot the name of it, the the
- 42:44electric cars in China. BDO yeah, BDO like they just
- 42:47have that down. I kind of think of as like
- 42:50snowflake and data bricks are like the BDO like they're just
- 42:52they've got it down. It's cost effective.
- 42:54And then you've got Tesla, which is a great I have a Tesla.
- 42:57It's a great car. But like it's more expensive and
- 43:00it's a little more proprietary and so on.
- 43:02So I think that's kind of the analogy.
- 43:03And I think the world is moving more towards the BD.
- 43:06Yeah. So, so it's like if I have an
- 43:07identity resolution use case today, I can I can sign on with
- 43:12vendor A that's got a composable capability in that area.
- 43:15And then just, you know, add on just like you do in a classic
- 43:18SAS world where you're like, hey, this features grayed out,
- 43:20this features grayed out, add on, add on additional features
- 43:23and capabilities over time. And I'm not paying for the
- 43:25suite, right. And getting a kind of a a
- 43:27volume. Based discount because yeah,
- 43:29you're not, you're not using and we have a lot of companies at
- 43:32high touch that have they figured out their own identity
- 43:34resolution. So they just send us the
- 43:36resolved identities and they don't have to pay for that part
- 43:39of the product. But most packaged legacy vendors
- 43:42like you pay for everything whether you use it or not.
- 43:46Oh, it's BYD by the way. Yeah, BYD, yeah, but.
- 43:49BYD yeah, I think. The main thing is people want
- 43:51flexibility right now. And I think especially in a
- 43:53real, in a weird economy, people aren't looking to drop $10
- 43:57million a year. Like they're like, hey, I want
- 43:59to just have the flexibility. So if I don't like something, I
- 44:02can change it. And that's the other benefit of
- 44:04composability, and this applies to CDP and analytics is if you
- 44:07don't like, like let's say you're using Cubit, which it
- 44:10happens to be a warehouse native analytics tool.
- 44:13If you don't like it after a couple of months, you could
- 44:15switch. If there was another warehouse
- 44:17native one that you like, you could switch very easily because
- 44:20you're not storing any data in it.
- 44:22You're just mapping it to the schema you already have in your
- 44:24warehouse. So it means that you can swap
- 44:27vendors in and out very, very quickly and very easily.
- 44:30And so you're not making these big commitments to one big
- 44:33monolithic. Well, well, yeah.
- 44:34But then the vendors are stuck in the situation, you know,
- 44:36coming from the SAS world, stuck in the situation where it's all
- 44:39one time revenue recognition versus, you know, multi year
- 44:42deal recurring revenue, which is a death now.
- 44:45So how do you prevent these companies?
- 44:48Yeah, how do they? Do going out of business because
- 44:49they don't have any ARR, everything's one time revenue
- 44:52because people. They're going to get crushed,
- 44:54right? Short term, they get crushed
- 44:55even if like, like let's say they have their iPhone moment.
- 44:57In fact, I'd want like be curious, like when do you think
- 44:59that'll happen? Like let's say like when, like
- 45:01when we get to a point where it's either an incumbent or I
- 45:03think you're right, Adam is probably going to be a new like
- 45:06a startup that just says, you know, we don't have the tech
- 45:07debt, we don't have the baggage, we don't have the the trained
- 45:10customers on our old platform. Let's just try something new.
- 45:12And everyone's like, wow, this is we never knew we wanted a
- 45:16touch screen, Brett, to your point before.
- 45:18But like we we, we wanted one, right?
- 45:19And everyone what? How far off you think that might
- 45:21be? I'm we're seeing it now.
- 45:23I mean, high Touch is growing at 140% right now.
- 45:27And most of the kind of package legacy CDP vendors are either
- 45:31growing at 2 or 3% or there's like segment just came out of
- 45:35their port where they're actually -5%.
- 45:37And I think it's just that model.
- 45:39It just is it had its time and I think it's time for it's the
- 45:44world to show shifting. And I think it's the non the non
- 45:47composable. I think it's both non composable
- 45:50is going away. And I think AI and if you put
- 45:54those two things together, it's like a double whammy on a lot of
- 45:58these legacy vendors. And I think like Salesforce is,
- 46:01I love Salesforce used to work there and they're doing their
- 46:03agent force stuff. But I think most people I talk
- 46:05to are like agent force is is just kind of a veneer on top of
- 46:09the same products they've had for 10 years.
- 46:11And I think, I think. They aren't exactly easy to use.
- 46:14Yeah, I think we're entering the golden age of where new vendors
- 46:17going to come up left and right because AI makes it so easy to
- 46:20start a new company, start a new product, and you can get things
- 46:23going. So you know, Scott Brinker's
- 46:25chart, I wrote a blog post about what I call micro composability,
- 46:29where I actually think we're going to end up with instead of
- 46:3115,000 vendors, we may end up at 45,000 vendors because it's
- 46:34going to be like each part. Like you could take identity
- 46:38resolution and there may be a vendor that says I'm going to do
- 46:41a really good job of probabilistic identity
- 46:43resolution and you could plug that into your warehouse.
- 46:46Instead of having to buy a whole identity resolution package,
- 46:49you're just buying a part of identity resolution.
- 46:52So I. Think.
- 46:54I mean, how much users is it to start a company now, right,
- 46:57instead of, you know, millions of dollars in teams of
- 46:59engineers. Maybe you only need like I was
- 47:01talking to like one designer. I'm good friends with that.
- 47:03You know, he he's got one engineer part time.
- 47:07They're able to vibe code really incredible applications.
- 47:10I mean, you know, obviously in order to in order to actually
- 47:12get things production ready, I mean it probably takes some
- 47:14heavy lifting, but he's able to get move the boat, move it to
- 47:16like 5 yard line, right? You can get to that product
- 47:20market fit, you know, you can find out really quickly instead
- 47:23of having to take a year to find out if you've product market
- 47:25fit, you could find out in a couple weeks there.
- 47:28Right away, right? It almost seems like a an app
- 47:30based SAS model where it's like you're buying individual
- 47:34applications. My question with somebody like a
- 47:35like a high touch and composable is, is, is how do you increase
- 47:39switching costs, right? So brands, if they're only using
- 47:43one feature, how do you get them to lock into long term
- 47:46relationships when they say? Well, that's the best part about
- 47:50it, about composability. You can't.
- 47:53And it sucks as a vendor. But here's the, here's the good
- 47:55news. If your product continues to
- 47:58deliver value and if you actually iterate, and I think
- 48:02that's what we're finding is we have like 95% retention rate.
- 48:07And basically because we just keep cranking through new stuff.
- 48:10Like we just launched a bunch of AI products.
- 48:12And I think that you have to be able to always add new value.
- 48:16And not to badmouth other vendors, but there's a number of
- 48:18vendors out there that they get a customer, They've had the
- 48:21customer for 10 years. If you talk to the company, how
- 48:25much value do you get from this monolithic vendor?
- 48:27They'll be like, like, we don't get a lot of value, but like the
- 48:30idea of getting rid of it is really like not pleasant.
- 48:34And I think there's a lot of companies right now that for the
- 48:37last 10 years have been in this, like, if I could get rid of
- 48:40this, I would get rid of it. But it's a pain to remove.
- 48:44But once you move to the warehouse is the center and you
- 48:47move to a composable mindset, you don't get trapped or locked
- 48:50into any monolithic vendor. You could still use a couple of
- 48:53those vendors products, but you don't have to and you could swap
- 48:57them out anytime because most of your logic and your data is
- 49:00stored in the warehouse that you own.
- 49:02So I think it's going to create much better SAS products.
- 49:06I think customers are going to be happier and they're not going
- 49:09to feel trapped. And for us, it's at high touch.
- 49:12We get really scared how easy it is to switch from one composable
- 49:15CVP to another composable CV, which keeps us on our toes to
- 49:18make sure we're always delivering really good products
- 49:21and services. Yeah, it is back to the
- 49:23innovators to live a topic, right?
- 49:24But that's really, this is going to facilitate force faster
- 49:28innovation, less of that sort of sitting in your laurels to your
- 49:31point, right. It's like it's like mini
- 49:33computers or discs but but amplified in a sense because
- 49:37because the switching costs are lesser.
- 49:40Argument, I know we've really diverted.
- 49:42We've diverted from our whole digital analytics conversation,
- 49:45so happy to go back if you want to, but this has been fun to
- 49:47just kind of riff for a little bit.
- 49:49Yeah, well, I guess I mean, I love, I love the local
- 49:51composability topics. So like, no, no, no issues with
- 49:54me. I mean, but I guess, you know,
- 49:55going back to vibe analytics, analytics a little bit.
- 49:57I mean, like, as someone who's been in the trenches for, for,
- 50:01you know, for a long time, like what would be your advice to,
- 50:04let's say, anyone who's in analytics or wants to get into
- 50:07analytics? I mean, like with all of these
- 50:09big seismic changes coming with, you know, the potential
- 50:12potentially like extinction level changes for some of the
- 50:15legacy vendors and a lot of really cool startups, maybe
- 50:18many, many more startups coming up with that really are changing
- 50:21the game use, you know, whether it's 5 analytics or just using
- 50:24AI to change how people interact with, with dashboards in
- 50:28general. Like what would be your advice
- 50:29to anyone in the industry? Yeah, I mean.
- 50:32It's a little. Scary if I were like, I have a
- 50:34couple kids who just graduated college and I don't know that I
- 50:37would tell them to go into, you know, digital analytics right
- 50:40now, just knowing what I know. But I do think there are still a
- 50:44lot of opportunities. One, there's going to be this
- 50:47conductor role where you are kind of the conductor to AI and
- 50:51helping it figure out what should it be, you know,
- 50:53verifying that it's tracking the right things, that the data
- 50:56quality is good. So there's that role.
- 50:58But I also think that a lot of people don't understand how
- 51:01important from storytelling and, and turning insights into
- 51:06something at a company that can influence change is.
- 51:10And so like I give presentations for a living and I attribute
- 51:14that to a lot of my success when I was in the digital analytics
- 51:17world as a practitioner is being able to take data and wrap it in
- 51:23a story and say, This is why? Because really, if you think
- 51:26about it, you're going to someone at your company and
- 51:29saying, I'm asking you to change the homepage, which is a very,
- 51:32very heavily trafficked page for a very big business based on
- 51:35data. And you've always got an
- 51:36executive who's going to be like, no, I, my gut instinct
- 51:39says we shouldn't do this. And so you've got to be able to
- 51:42kind of do the politics. You've got to be able to tell a
- 51:45narrative story that says, let's picture a world where this is on
- 51:48the home page. And once they do this, they do
- 51:51this and I can show you because we did an experiment that we
- 51:55think we're right because we, we tried with 10% of the population
- 51:58and then we did see this result. But it would be a big change for
- 52:01us. And we think this is worth the
- 52:03risk. And if it doesn't work, this is
- 52:05our contingency. And being able to tell that
- 52:08story, like AI is not going to be able to tell the story like
- 52:12that in the near future. Now, if we get to a world where
- 52:16everything is completely automated, then I don't know
- 52:18what that means for data storytelling, but I think that
- 52:20there's going to be a need always to, you know, interpret
- 52:23data and insights and communicate that within the
- 52:26organization. But I think a lot of the other
- 52:28grunt work is going to really go away in a couple years.
- 52:31Yeah, it's like elevation. It's like an elevation of data
- 52:34science and data analytics, right?
- 52:36Cuz it, cuz it's funny, We, you know, we talked about this in a
- 52:38couple of episodes. Rio, How how you know, you know,
- 52:4110 years ago to become a computer engineer was a great
- 52:44career path, right? You, you'd be building stuff.
- 52:47It was a huge, it was a huge need with the, with the.
- 52:50There's always need for more. Engineers and now AI has
- 52:52replaced that. That's the most poorly recruited
- 52:57role out of College of, of any kind of role in the, in the, in
- 53:01the, in the business world right now.
- 53:03Now we're saying data analytics and portions of data analytics,
- 53:06especially the grunt work are going to be disintermediated.
- 53:09But there's this role for the orchestrator, the conductor, the
- 53:12strategist to come up with the plan and deliver the insights to
- 53:17humans. Yeah, right.
- 53:18Yeah. So yeah.
- 53:19How? Many of the analysis.
- 53:21I don't know, but we'll yeah, fewer the.
- 53:23Analysis the thinking and then the the presenting, right, the
- 53:26storytelling is, I mean, I think it just, it's not just in
- 53:30analytics, but I think a lot of occupations, a lot of
- 53:33professions, right? Like I, I mean I've talked to a
- 53:35lot of people at consulting, right?
- 53:36I think that if AI is going to take notes for you and do a lot
- 53:40of the grunt work, right, that that means that, okay, you're
- 53:44not going to build for that stuff anymore, but that means
- 53:45you can have more time to think about things.
- 53:47But that means I used to always say this to consultants on
- 53:49projects. Like what does it mean?
- 53:51They're not paying you to like take notes and regurgitate it
- 53:53and put it build nice slides, right?
- 53:54I mean, and I think that's less important now than it was,
- 53:57right? Like, what does it mean?
- 53:58Figure it out. What is the impact?
- 54:01Yeah, I'm sure you can use AI to help you think through these
- 54:03things, but I think that's going to be out of my group.
- 54:04This can be more important, whether it's an analytics or
- 54:07applying analytics to let's say to business recommendations or
- 54:10whether whether it's do we need a new website, do we need a
- 54:12website at all? I think all of those things, the
- 54:15ability to think through and understand what what these
- 54:17signals mean is going to be more important.
- 54:19Yeah. And I think that AI is probably
- 54:21going to get to the point where it's going to recommend like,
- 54:23here's 5 different things that I think you should do to improve
- 54:26your conversion rate. And then you're going to have to
- 54:29kind of take those five things, think about it, work with your
- 54:32teams, work with your executives, and say we're going
- 54:34to stack rank these, we're going to try this one first, and if it
- 54:37doesn't work, go to this or maybe try these two at the same
- 54:39time. So yeah, it's going to be a lot
- 54:41of, I think it's like instead of being a player, a lot of people
- 54:44are going to start acting like coaches.
- 54:47Yeah. And, and, and it's just really
- 54:49AI is just accelerating the move through the maturity curve to
- 54:51bring that theme back right back to the, the prescriptive piece,
- 54:55because at the end of the day, it's like, what should I do
- 54:56next? Speed to insight, right.
- 54:58And if you got a lot of that decisioning and, and sort of,
- 55:02you know, data capture and data analysis happening
- 55:05automatically, then it's, it's just quicker for, for companies
- 55:09to be able to evolve and adapt to whether it's you're selling a
- 55:12product or you're selling content, right?
- 55:16Lu Pascala said this in our last interview, Rio, he said you're
- 55:19engineering to relevance. Well, what the hell does that
- 55:21mean? Engineering to relevance?
- 55:23That's really what it's all about, delivering relevant
- 55:26experiences, whether it's on the content side, the website side,
- 55:29the mobile app side, a product advertisement.
- 55:32And the only way you can do that is with data.
- 55:34And we've been promised this for like 20 years, being in the.
- 55:36I'll throw another crazy one at you.
- 55:39I could see a world in which humans aren't even needed to
- 55:43view to view the websites or mobile app because I'm starting
- 55:47to see people talk about building synthetic personas,
- 55:50which are AI personas that are trained on the current
- 55:54population. So imagine you have a new
- 55:57section of a website or a new app and before you launch to the
- 56:00public, you basically feed it a bunch of fake humans, a bunch of
- 56:06AI humans to use it the way a human would.
- 56:09And you might be able to use digital analytics and AI and
- 56:12digital analytics to say, hey, this page isn't working.
- 56:15This page is working. And basically almost use them as
- 56:17like a focus group or a beta test.
- 56:20And it doesn't cost you anything.
- 56:22You don't have to waste your advertising dollars.
- 56:24You basically. And you can do the same thing
- 56:25for ads. You might be able to like test 5
- 56:28different ads with an AI user group, see which one performs
- 56:32best, and then keep validating that with humans and eventually
- 56:35if it's always you know the same as the humans, then like you can
- 56:38just know exactly. Well, and after a while, do you
- 56:40even eat humans anymore for for like if you're just having
- 56:43Asians do things for you, you're.
- 56:44Gonna be playing a lot of golf. Right.
- 56:46Well. Well, think about it like if,
- 56:48if, if agents are going to go buy a plane ticket for you, book
- 56:51a hotel for you, You don't need to go to like the you know, I
- 56:54don't need to open up United app anymore to book, right?
- 56:56Asia does it for me. So like eventually, you know,
- 56:59not you're optimizing web experience.
- 57:03I mean, that's happening already.
- 57:04Like they're being optimized for AI and you can push, you can
- 57:07push content updates directly at API with like perplexity for
- 57:09examples. But I think at some point, maybe
- 57:13you have two webs, one for humans, one for AI.
- 57:15And at some point, as human traffic starts to dwindle, you
- 57:19know, to your point, maybe not only you're not only testing
- 57:20with synthetic humans, maybe you're just testing for AIS.
- 57:25Yeah, but someone has to buy the products and services
- 57:27eventually. Yeah.
- 57:28And Adam, I think you made a good point.
- 57:29I think we could, we're coming to a close here, but I think you
- 57:32made a really good point about basically synthetic data
- 57:35training sets, right? Because you have to have a truth
- 57:38set to train models on to say, this is exactly what this is
- 57:42exactly reflective of a human being's behavior on a website,
- 57:45or this is exactly reflective of, you know, in our business,
- 57:49we look at like advertising attribution.
- 57:52So we attribute ads. You have to look at an ad and
- 57:55say, OK, who's the brand? What's the product?
- 57:57What's the product category, what's the messaging, the copy,
- 58:01the visuals, everything. And you know, you have to train
- 58:05the AI on how to recognize those things so that it's repeatable
- 58:09and you need a truth set. And so synthetic data sets are
- 58:12oftentimes the best way to scale like a learning set that doesn't
- 58:17require, like you said, focus groups or sort of manual
- 58:21mechanisms to create a bunch of data to train on, because if you
- 58:24train on the wrong data, it's not going to do.
- 58:26Anything. I hope I never have to do
- 58:28another focus group. Well, this.
- 58:32Is great, I think. Is there anything else from a
- 58:35recommendation perspective in terms of the future of data
- 58:38analytics, Adam, that you think you'd like to tell our audience?
- 58:41Yeah, what we miss. Yeah, No, I think, I mean just
- 58:44to kind of summarize because I know we are a little bit all
- 58:46over the place. I think the way I'm picturing
- 58:49this is that the interface to do digital analytics is going to
- 58:54change whether that becomes a vibe analytics or AI chatbot
- 58:58analytics. I think that that interface is
- 59:01going to take over a lot of the grunt work of building reports
- 59:04dashboards. But then I think once we get to
- 59:07that world, I think that's going to force a different change to
- 59:09say, well, where are we pointing this AI at?
- 59:12Are we pointing it at a digital analytics data set only?
- 59:16Or are we going to point it at all of our customer data and
- 59:18have it figure out the kinds of questions that a digital analyst
- 59:21would answer keeping in mind the whole customer data set?
- 59:25And if that happens, I think we're going to see a major shift
- 59:28in the vendor landscape where the cloud data warehouses become
- 59:32the back end. And then the, the current
- 59:34digital analytics vendors will have to decide, do they build an
- 59:37interface that they think is worth enough that someone would
- 59:40pay for on, on top of the cloud warehouses?
- 59:43Do the cloud warehouses build an AI interface onto their own, you
- 59:49know, interface and they become the new digital analytics
- 59:51vendors? Or do a bunch of new startups
- 59:54build AI that answers the digital analyst questions
- 59:57directly on the warehouse? And I think that's going to be
- 1:00:00probably one of the biggest changes.
- 1:00:02And then from a personnel standpoint, I think the role of
- 1:00:05a digital analytics. Professional both implementation
- 1:00:08and analysis are going to really be shifting from doing the grunt
- 1:00:12work to doing much more of the coaching and conducting of an
- 1:00:17AI, doing a lot of the implementation, the coding, the
- 1:00:20analysis and the dashboarding so that you're really kind of
- 1:00:24making the bigger decisions. So I think the next two to three
- 1:00:27years are going to be really impactful for the digital
- 1:00:30analytics space. And for me, I kind of like being
- 1:00:34on the sidelines to be able to watch all this and not have like
- 1:00:37a heart attack every month. If like my job is going to go
- 1:00:40away because AI is, is removing the need for digital analytics
- 1:00:44products. But I could be totally wrong and
- 1:00:47people could be using Google Analytics, Adobe Analytics,
- 1:00:50Amplitude, you know, 10 years from now.
- 1:00:52So time will tell. But you know, you just, all you
- 1:00:55can do is look at the trends that are happening right now and
- 1:00:57kind of extrapolate that into the future and make an educated
- 1:01:01guess. And so we'll see who wins out in
- 1:01:04the innovator's dilemma in the digital analytics space.
- 1:01:09Is there a Gauss? Thank you.
- 1:01:10You couldn't have civilized it better right than that.
- 1:01:13Thanks, Adam. It was great.
- 1:01:14Great having you on your time signal and noise.
- 1:01:16A. 100%. Yeah, thanks so much for having
- 1:01:18me. And yeah, it's been great.