 theCUBE's live coverage is made possible by funding from Dell Technologies. Creating technologies that drive human progress. Hello everybody, welcome back to day four of theCUBE's coverage of MWC23. We're here live at the theater in Barcelona, wall-to-wall coverage. John Furrier's in our Palo Alto studio banging out all the news. Really, all week we've been talking about the disaggregation of the Telco network, the new opportunities in Telco. We're really excited to have AT&T and Snowflake here. Dave Wittington is the AVP at the Chief Data Officer at AT&T. Roddy Trenum is the Assistant Vice President for Channel Performance Data and Tools at AT&T and Phil Kippen, the global head of industry telecom at Snowflake, Snowflake's new telecom business. Snowflake just announced earnings last night. Typical Scarpelli, they beat earnings, very conservative guidance. Stock's down today, but we like Snowflake long-term. They're on that path, the 10 billion. Guys, welcome to theCUBE. Thanks so much for coming on. Thanks, Dave. Dave, let's start with you. The data culture inside of Telco, we've had this, we've been talking all week about this monolithic system, super reliable. You guys did a great job during the pandemic. Everything's shifting to landlines. We didn't even notice. You guys didn't miss a beat, saved us. But the data culture is changing inside Telco. Explain that. Well, absolutely. So, first of all, IOT and edge processing is bringing forth new and exciting opportunities all the time. We're bridging the world between a lot of the OSS stuff that we can do with edge processing, but bringing that back, and now we're talking about working, and I would say traditionally, we talk data warehouse. Data warehouse and big data are now becoming a single mesh, all right? And to use cases in the way you can use those, especially, I'm taking that edge data, I'm bringing it back over. Now I'm running AI and ML models on it, and I'm pushing back to the edge. And I'm combining that with my relational data. So that mesh there is making all the difference. We're getting new use cases that we can do with that, and the volume of data is immense. Now I love ChatGPT, but I'm hoping your data models are more accurate than ChatGPT, you know? I never know. Sometimes it's really good, sometimes it's really bad, but enterprise, you got to be clean with your AI, don't you? Not only you have to be clean, you have to monitor it for bias and be ethical about it. We're really good about that. First of all, with AT&T, our brand is platinum. We take care of that. So we may not be as cutting edge risk takers as others, but when we go to market with an AI or an ML, or a product, it's solid. Well hey, as telcos go, you guys are leaning into the cloud, so I mean, that's a good starting point. Roddy, explain your role. You got an interesting title, channel performance data and tools. What's that all about? So literally anything with our consumer retail, consumers channels, all of our channels, from a data perspective and metrics perspective, what it takes to run reps, agents, all the way to leadership level, score cards, how you rank in the business, how you're driving the business from sales, service, customer experience, all that data infrastructure with our great partners on the CDO side, as well as Snowflake, that comes from my team. And that's traditionally been done in, I don't mean the pejorative, but you know we're talking about legacy, monolithic sort of data warehouse technologies. Absolutely. We have a love-hate relationship with them, it's what we had, it's what we used, right? That's, and now that's evolving, and you guys are leaning into the cloud. It's dramatic evolution, and what Snowflake's enabled for us is impeccable. We've talked about having, people have dreamed of one data warehouse for the longest time and everything in one system. Really, this is the only way that becomes a reality. The more you get in Snowflake, we can have golden source data, and instead of duplicating that 50 times across AT&T, it's in one place. We just share it, everybody leverages it, and now it's not duplicated, and the process efficiency is just incredible. But it really hinges on that separation of storage and compute. And we talk about the monolithic warehouse, and one of the nightmares I've lived with is having a monolithic warehouse, and let's just go with some of my primary traditional customers, sales, marketing, finance. They're leveraging BSS OSS data all the time. For me to coordinate a deployment, I have to make sure that each one of these units can take an outage if it's going to be a long deployment. With the separation of storage compute, they own their own compute cluster, so I can move faster for these people because if finance, I can implement his code without impacting finance, or marketing. This brings in CICD to more reality. It brings us faster to market with more features, so if he wants to implement a new comp plan for the field reps, or we're reacting to the marketplace where one of our competitors has done something, we can do that in days versus waiting weeks or months. And we've reported on this a lot. This is the brilliance of Snowflake Founders, that whole separation from compute and data. I like, Dave, that you're starting with sort of the business flexibility because there's a cost element of this too. You can dial down, you can turn off compute. And then of course the whole world said, hey, that's a good idea. And if VC started throwing money out at Amazon with Redshift, said, oh, we can do that too. Sort of, you know, we can't turn off the compute, but I want to ask you Phil. So it looks from my vantage point, like you're taking your data cloud message, which was originally separate compute from storage, simplification now, data sharing, automated governance, security, ultimately the marketplace. Taking that same model, break down the silos into telecom, right? It's that same, sorry to use the term playbook. Frank Slutman tells me he doesn't use playbooks, but it's a pattern match, not a pattern match, but he's a situational CEO, he says. But the situation in telco calls for that type of strategy. So explain what you guys are doing in telco. I think there's, so what we're launching, we launched next week or last week, and it really was three components, right? So we had our platform, as you mentioned, and that platform is being utilized by a number of different companies today. We also were adding, for telecom very specifically, we're adding capabilities in marketplace so that service providers can not only use some of the data and apps that are in marketplace, but as well, service providers can go and sell applications or sell data that they had built. And then as well, we're adding our ecosystem. It's telecom specific, so we're bringing partners in, technology partners and consulting and services partners that are very much focused on telecoms and what they do internally, but also helping them monetize new services. Okay, so it's not just sort of generic snowflake into telco, you have specific value there. We're purposely, we're purposing the platform specifically for the telco market. Are you a telco guy? I am. You are, okay. So there you go, you see that snowflake is actually an interesting organizational structure because you're going after verticals, which is kind of rare for a company of your sort of immaturity, I'll say. I don't mean that as a negative. Dave, take us through the data journey at AT&T. It's a long history. You don't have to go back to the 1800s. So thank you for pointing it out. We're a 149 year old company, so Jesse James was one of the original customers and we no longer got his data. So I'll go back, I've been 17 years singular AT&T and I've watched it through the whole journey of where the monolithics were growing when the consolidation of small wireless carriers and we went through that boom and then we've gone through mergers and acquisitions. But you know, Hadoop came out and it was going to solve all the world hunger and we had all the, you know, all the aspects of we're going to monetize and do AI and ML. And some of the things we learned with Hadoop was, you know, we had this monolithic warehouse, we had this file-based structured Hadoop, but we really didn't know how to bring this all together and we were bringing items over to the relational and we were taking the relational and bringing it over to the warehouse and trying to, and it was a struggle, let's just go there. And I don't think we were the only company to struggle with that, but we learned a lot. And so now as tech has finally emerging with the cloud, companies like Snowflake and others that can handle that where we can create, you know, you were discussing earlier, but it becomes more of a conducive mesh that's interoperable. So now we're able to simplify that environment and the cloud is a big thing on it because you could not do this on-prem with on-prem technologies, it would be just too cost-prohibitive and too heavy of lifting going back and forth and managing the data. The simplicity the cloud brings with a smaller set of tools and I'll say in the data space specifically, really allows us, maybe not a single instance of data for all use cases, but a greatly reduced ecosystem. And when you simplify your ecosystem, you simplify speed to market and data management. So I'm going to ask you, I know it's kind of internal organizational plumbing, but it'll inform my next question. So Dave, you're with the chief data office and Roddy, you're kind of, I mean, you all serve in the business, but you're really serving, you're closer to those guys. They're banging on your adoraforce. Absolutely, I try to keep the 130,000 users who may or may not have issues sometimes with their data metrics away from Dave and he just gets a call from me. And he only calls when he has a problem. He's never wished me happy birthday. So the reason I ask that is because you described Dave sort of the Hadoop days and again, we love hate with that, but we had hyper-specialized roles we still do. You know, you've got data engineers, data scientists, data analysts and you've got this sort of, you know, this pipeline and it had to be the sequential pipeline. I know Snowflake and others have come to simplify that. My question to you is, how is that those roles, how are those roles changing? How is data getting closer to the business? Everybody talks about democratizing business. Are you doing that? What's a real use example? From our perspective, those roles, a lot of those roles on my team for years because we're all about efficiency. We cut across those areas and always have cut across those areas. So now we're into a space where things have been simplified, data processes and copying. We've gone from 40 data processes down to five steps now, we've gone from five steps to one step. We've gone from days now, take hours, hours to minutes, minutes to seconds. Literally we're seeing that time in and time out with Snowflake. So these resources that have spent all their time on data engineering and moving data around are now freed up more on what they have skills for and always have the data analytics area of the business and driving the business forward and new metrics and new analysis. That's some of the great operational value that we've seen here. As this simplification happens, it frees up brain power. So you're pumping data from the OSS, the BSS, OKRs, into Snowflake. Scheduling systems. You name it, if you can think of what drives our retail and centers and online, all that data, scheduling system, chat data, call center data, call detail data, all of that enters into this common infrastructure to manage the business on a day in and day out basis. How are the roles and the skill sets changing? Because you're doing a lot less ETL, you're doing a lot less moving of data around. There were guys that were probably really good at that. I used to joke when I was in the storage world, like if your job is managing lawns, you need to look for a new job, right? So, and they did and people move on. So are you able to sort of redeploy those assets and those humans? These folks are highly skilled and we were talking about earlier. SQL hasn't gone away, relational databases are not going away and that's one thing that's made this migration excellent. They're just transitioning their skills. Experts in legacy systems are now rapidly becoming experts on the Snowflake side and it has not been that hard a transition. There are certainly nuances, things that don't operate as well in the cloud environment that we have to learn and optimize, but we're making that transition. So just within the Chief Data Office, we have a couple of missions and Roddy is a great partner in an example of how it works. We try to bring the data for democratization so that we have one interface. Now hopefully no, we just have a logical connection back to these Snowflake instances that we connect but we're providing that governance and cleansing and if there's a business rule that the enterprise level we provide it. But the goal at CDO is to make sure that business unions like Roddy or marketing or finance, that they can come to a platform that's reliable, robust and self-service. I don't want to be in his way. So I feel like I'm providing a sub-level of platform that he can come to and anybody can come to and utilize that they're not having to go back and undo what's in sales for their service now or in our billers. So I'm sort of that layer and then making sure that that ecosystem is robust enough for him to use. And that self-service infrastructure is predominantly through the Azure cloud, correct? Absolutely. And you work on other clouds but it's predominantly through Azure. It's predominantly in Azure shop. That's the first party citizen. Okay, I like to think in terms sometimes of data products. I know you mentioned up front your gold standard or platinum standard you're very careful about personal information. So you're not trying to sell, I'm an AT&T customer, you're not trying to sell my data and make money off of my data. So the value prop in the business case for Snowflake is it's simpler, you can do things faster, you're in the cloud, lower cost, et cetera. But I presume you're also in the business, AT&T of making offers and creating packages for customers. I look at those as data products because it's, you know, it's not a, I mean, yeah, there's a physical phone but there's data products behind it. It ultimately is but not everybody always sees it that way. Data reporting often can be an afterthought and we're making it more on the forefront now. Yeah, so I like to think in terms of data products. I mean, even if the financial services business, it's a data business. So if we can think about that sort of metaphor, do you see yourselves as data product builders? Do you have that? Do you think about, you know, building products in that regard? Within the Chief Data Office, we have a data product team. And by the way, I would be disingenuous if I said, oh, we're very mature in this, but no, it's where we're going. And it's somewhat of a journey, but I've got a peer and their whole job is to go from, especially as we migrate from cloud, if Roddy or some other group was using tables three, four and five and join in together, it's like, well, look, this is an offer data product. So let's combine these and put it up in the cloud and here's the offer data set product or here's the opportunity data product. And it's a journey, we're on the way but we have dedicated staff and time to do this. I think one of the hardest parts about that is the organizational aspects of it, like who owns the data now, right? It used to be owned by the techies and increasingly the business lines want to have access that you're providing self-service. So there's a discussion about, okay, what is a data product? Who's responsible for that data product? Is it in my P&L or your P&L? Somebody's going to sign up for that number. So it sounds like those discussions are taking place. They are, and we feel like we're more of the, and CDO at least, we feel more, we're like the guardians and the shepherds, but not the owners. I mean, we have a role in it all, but he owns his metrics. And even from a higher perspective, we see ourselves as an enabler of making whatever AT&T wants to make happen in terms of the key products and offers trade in programs. All that requires this data infrastructure and managing reps and agents and what they do from a channel performance perspective. We still see ourselves as key enablers of that. And we've got to be flexible and respond quickly to the business. I always had empathy for the data engineer. And he or she had to serve us all these different lines of business with no business context. The business knows good data from bad data and they just pound that poor individual and they're like, okay, I'm doing my best. It's just ones and zeros to me. So it sounds like that's, you're on that path. Absolutely, and I think we do have refined, getting more and more refined owners of, since Snowflake enables these golden source data, everybody sees me and my organization, channel performance data, go to Roddy's team. We have a great team. And we go to Dave in terms of making it all happen from a data infrastructure perspective. So we do have a lot more refined. This is where you go for the golden source. This is where it is. This is who owns it. If you want to launch this product in services and you want to manage reps with it, that's the place you go to. So the Chief Data Office doesn't own the data per se, but it's your responsibility to provide the self-service infrastructure and make sure it's governed properly and it is automated way as possible. Well, yeah, absolutely. And let me tell you what, everybody talks about single version of the truth, one instance of the data, but there's context to that, you know, that we, and we are trying to take advantage of that as we do data products is what's the use case here? So we may have an entity of Roddy as a prospective customer and we may have an entity of Roddy as a high value customer over here, which may have a different set of mix of data at all. But as a data product, we can create those for those specific use cases, still point to the same data, but build it in different constructs, one for marketing, one for sales, one for finance. That's, by the way, that's where your data engineers are struggling. Yeah, yeah, of course. So how do I serve all these, you know, folks and really have the context? Common story in telco? Absolutely. Or are these guys ahead of the curve a little bit or where would you put them? I think they're definitely moving a lot faster than the industry is generally. I think the enabling technologies, like for instance, having that single copy of data that everybody sees a single pane of glass, right? That is, that's definitely something that everybody wants to get to. Not many people are there. I think, you know, what AT&T's doing is most definitely a little bit further ahead than the industry generally. And I think the successes that are coming out of that and the learning experience is starting to generate momentum within AT&T. So I think, you know, it's not just about the product and, you know, having a product now that gives you a single copy of data. It's about the experiences, right? And now how the teams are getting trained. Domains like network engineering, for instance, they typically haven't been a part of data discussions because they've got a lot of data, but they're focused on the infrastructure. So by going ahead and deploying this platform for platform's purpose, right? And the business value, that's one thing. But also to start getting that experience and bringing new experience in to help other groups that traditionally hadn't been data centric. That's also a huge step ahead, right? So you need to enable those groups. A big complaint, of course, we hear at MWC from carriers is the over the top guys are killing us. They're riding on our networks, et cetera, et cetera. They have all the data. They have all the client relationships. Do you see your client relationships changing as a result of sort of your data culture evolving? Yes, but I'm not sure I can... It's a loaded question, I know. Yeah, and so we want to start embedding as much into our network on the proprietary value that we have so we can start getting into that OTT plate. As a, us as any other carrier, we have distinct advantages of what we can do at the edge and we just need to start exploiting those. But, you know, because whether it's location or whatnot, so we got to eat into that. You know, historically, the network is where we, we make our money and we stack the services on top of it. It used to be star 69. Yeah. If anybody remembers that. Of course. But, you know, it was stacked on top of our network. And then we stack another product on top of it. It'll be in the edge where we start providing distinct values to other partners. I mean, it's kind of a great business that you're in. I mean, it's a really good connectivity, you know? And so it sounds like it's still to be determined where you can go with this. You have to be super careful with private and for personal information. Yeah, but the opportunities are enormous. There's a lot. Particularly at the edge, looking at private, you know, networks or, you know, just an amazing opportunity, factories and, you know, name it, hospital, remote hospitals, remote locations. I mean, connect the cars. Connect the cars are really interesting, right? I mean, if you start communicating car to car and actually drive that, I mean, that's, now we're getting to Byzantine fault tolerance people. This is like really. That's not the place. That's a whole nother conversation. You scare me as much as we actually are. Yeah. So how's the show been for you guys? What a, what a big take away from this. It's a tremendous experience. I mean, someone who doesn't go outside the United States much, I'm a homebody, the whole experience, the whole trip, city, mobile world, Congress, the technologies that are out here, it's been a blast. Any things, any, you know, top two things you learned, advice you'd give to others, you know, your colleagues. Out of, out of, in general, you know, we talked a lot about technologies today and we talked a lot about data. But I'm going to tell you what, the accelerator that you cannot change is the relationship that we have. So when the tech and the business can work together toward a common goal and it's a partnership, you get things done. So, you know, I don't know how many CDOs or CIOs or CEOs are out there, but this connection is what accelerates, it makes it work. And that is our audience, Dave. I mean, it's all about that alignment. So guys, I really appreciate you coming in and sharing your story on theCUBE. Great stuff. Thank you. That's a lot. All right, thanks everybody. Thank you for watching. I'll be right back with Dave Nicholson, day four, Silicon Angles coverage of MWC23. You're watching theCUBE.