Transcript 0:00 Welcome everyone to a new episode of the FinOps Weekly Podcast, and today we are going to talk about one of the most difficult challenges that enterprises figure out today, and it's the hybrid cloud and the hybrid FinOps challenge. 0:12 So when, uh, the cloud expertise and the cloud FinOps is not enough. And for helping us today with this challenge, we have, uh, Jeremy Chaplin from Flexera. Jeremy, how are you? I'm very good, Vic-Victor. 0:25 Thank you ever so much, and, uh, excited to be back on FinOps Weekly Podcast. Um, so yeah, very much looking forward to our conversation this afternoon. Yeah. You are... 0:34 I think you are the first one that, uh, uh, repeats in our podcast, which is, which is amazing, and you have that- Very honored. [laughs] Yeah, yeah. It's always... 0:43 There is only one first one, so, uh, this is, this is great. 0:47 And we are going to talk about something that I think worries, uh, most of the companies, especially the big ones, uh, that have a hybrid cloud, so something part on-premise, something on private cloud and on public cloud. 0:59 So, you know, what do you think, um, to start, uh, with direct question is, what do you think are the main, uh, challenges that arise from having hybrid cloud FinOps instead of cloud-only FinOps, where you only have the CSPs that we all know, and now we have to interact between the on-prem or the private cloud and the public cloud for, for doing this optimization? 1:23 It, it's, it's a real challenge, isn't it? 1:25 And, you know, already I think when I, I, you know, started getting involved in FinOps and the focus was on that, you know, collaboration with a broad range of straight stakeholders, public cloud was touching everything, right? 1:39 All the way down to engineering. 1:41 Um, and I came from a kind of ITIL, ITAM background where we probably didn't, you know, end up speaking to engineers at the end of the day, and so, so that was a kind of paradigm shift a little bit in terms of the breadth there. 1:54 But now as we see, you know, the framework, including other scopes of, you know, private cloud, data center, uh, SaaS management a-as well, I think that we're, we're increasing even further the breadth of the engagement and collaboration that we're trying to achieve. 2:10 And so doing that across, you know, the inform, optimize, operate, you know, uh, life cycle ultimately becomes more challenging. And I'm sure we'll get into that. 2:20 But the, you know, the challenges in tooling and visibility and the different levers and things that ultimately we need to pull or at least bring together to have an aggregate understanding presents, you know, real challenges mo-mostly, I guess, in organizational collaboration, uh, as much as anything else. 2:37 Yeah, that's true. 2:38 I think, like, whenever you, you know, change the paradigm from, you know, from a cloud perspective to something that licensing, which is, uh, what we are going to talk about a bit, it's so different than the cloud and so different than the SaaS platform, which, uh, work in a totally different consumption way, um, that, you know, you need to, to put a different, [chuckles] a different perspective of the whole FinOps that, uh, is out of, of cloud, right? 3:04 Totally, yeah. And, you know, everybody kind of likes, [chuckles] likes to stay in their lane, don't they- Yeah... for some degree, right? 3:09 I mean, you know, we spoke earlier just before we started about people that have spent a lifetime doing license management, and they understand all the nuances and terms and conditions and the t-taxonomy and, you know, all of the three-letter acronyms in the Microsoft licensing space, for example, that, that, you know me as a FinOps practitioner and, and heavily focused o-on public cloud would never have an opportunity even to scratch the surface i-in terms of understanding. 3:36 And so it's how do we leverage, I guess, the expertise that those other folks have developed, um, you know, and be collaborative rather than trying to kind of reinvent the wheel or redevelop an understanding myself. 3:47 I, I would never be successful u-unless I brought in the right people, um, to work with to, to leverage their expertise in, in their domain as well. That's true. 3:56 I think maybe we are seeing like, uh, a more unspecialization of FinOps, uh, since we are expanding, uh, what the, what the practice is doing, like from cloud to all the different, uh, steps. 4:08 Now we are integrating like with other roles that before weren't mentioned in the context of FinOps or were finally mentioned, such as you mentioned, uh, the ITAM or the even the SaaS, like SaaS could be even different or the AI part, so you can do FinOps for AI very good, but not FinOps for cloud. 4:24 So let's just start with, with these domains, and let's go step-by-step so that people understand what are the difference. 4:29 So starting with the licensing, which is one of the topics that, uh, we have started, what do you think are the fundamental difference between cloud FinOps and, uh, ITAM or this, uh, license management for, for controlling the cost? 4:43 Yeah, no, great question. And it... You know, in the public cloud space, particularly as it's kind of, you know, been technically focused, um, we think about provisioning infrastructure. 4:54 We think of this kind of mentality with a pay-as-you-go mentality on, right? 4:58 I can put down a credit card, and I can access all of the things that I need, and I c- I think the, the cloud vendors love that, and of course, we love the flexibility and the agility that that affords. 5:09 I don't n- need to separately go and buy a, a Windows server or a SQL license from Microsoft to get up and running i-if I don't need to. 5:17 The same for Oracle and the same for, you know, the plethora of things that are in marketplaces, uh, you know, across the cloud vendors. 5:25 Um, but the reality is, for most enterprises at least that have been in business for many years, they, they own those licenses, right? 5:32 They've had a whole procurement process and probably a team of people that are managing license compliant, you know, from an audit perspective, from a, you know, cost-saving perspective. Are we using these efficiently? 5:45 Have we got licenses that kinda overlap with each other? Have we bought things that, you know, we, we could reconcile, uh, you know, down to, to single licenses and so on? 5:54 So there, there's a lot of, um, different lenses that we can look at that problem through. But historically, at least my experience says those in the public cloud space typically haven't thought about it too much. 6:08 Um, but the reality then is, you know, that actually we did some research, um, you know, internally. 6:15 We saw of the s- cloud spend that Flexera had under management, over five percent of it was on two products, which was Microsoft Windows Server, Microsoft SQL Server. 6:25 So of, of all of the cloud spend, five, five percent pay-as-you-go o-on those, uh, operating systems and databases, and another twenty percent of our customers' spend is, is in marketplace.And so if we're not working with our software asset management counterparts, right, the people that do manage that procurement cycle, the people that do know the nuances of, you know, what they own and how I might use it in public cloud, th-there's basically a quarter of your cloud bill, you know, you're ruling out from an optimization perspective. 6:58 Um, so yeah, that, that's really kinda where I see the challenges are in terms of gaining that understanding, bringing those teams up together. 7:06 You know, I, as a public cloud practitioner, won't know the nuances of Windows licensing to know whether I can use it in public cloud or whether I can't, and how many CPUs I need to count for a SQL license and so on, right? 7:20 It's pretty complicated stuff, so. Yeah. Yeah, indeed. 7:22 And also, uh, something that, um, brought up in a, in a conversation that I had with, with another practitioner is that, um, talking with these, uh, people that have a different role may help you even improve as a FinOps in your cloud role. 7:37 And I can put down a specific example, which is, um, ne-negotiating, negotiating contracts, uh, with cloud providers because they are very used, like item people or licensing people, procurement people are way more used to li- uh, license management and license negotiation that when you are in this, uh, contract management, like the Microsoft program or the AWS program, the ECP, they will be so much... 8:00 Probably they will be much better than you the, the doing this negotiation that you can leverage them to, to lend you a hand on, on this, right? Absolutely. Yeah. 8:09 No, it's a great point, and that, that whole contract negotiation, particularly with, uh, we use the term mega vendor, right? Yeah. 8:15 But someone like Microsoft, where they're doing public cloud, they're doing SaaS, they're doing on-premise licensing. In order to have a successful negotiation, you need to know the holistic view, right? 8:27 What am I spending on all of those domains? Because ultimately, you know, it's the, the aggregate of that cost that is going to that vendor, and you should leverage that for negotiation. 8:37 Um, there was a really interesting example that came up recently, obviously, [chuckles] anonymity, uh, anonymity for the customer that, that this was about. 8:45 But they had negotiated an all-you-can-eat kinda license agreement for certain key products from, from Microsoft, and yet we, you know, we saw a significant amount of pay-as-you-go spend in the public cloud space. 8:59 So they've, you know, spent a lot of time [chuckles] in the procurement cycles and negotiating to get this all-we-can-eat licensing, but then they're not leveraging it, right? 9:06 They're paying again, uh, whi- which is not untypical for, for us to see. Yeah. Um, it, it's, you know, uh, still relatively early days, I guess, right? 9:15 So procurement getting involved in, in the kind of public cloud, uh, conversation, you know, EDP negotiation also- Yeah... 9:22 becoming something that we're seeing partners, you know, provide a role for their, their customers for as well, so. Yeah, definitely. 9:28 And I think it's, it's getting there, um, like the, the expansion I think for FinOps is very early. 9:34 So until, uh, as I know, like, till enterprises are able to integrate something like a practice or a technology like coupling with cloud, it takes time. 9:44 So it'll take like it's ten [chuckles] years or, or more to, to have everything like mature and figured out, and then from there we are starting the maturity level and then evolve. So moving on to like the, the... 9:58 a second cost driver that is very different from, from cloud, which is SaaS. You know, SaaS, it's probably one of the highest, uh, hot... 10:06 or the hottest topics from past year I think in FinOps, especially on the, on the FinOpsX, uh, Barcelona and, uh, on San Diego. 10:13 That, uh, you know, is driving a lot of costs, especially, um, certain, you know, observability tools, data analysis tools, et cetera, that ingest a lot of, uh, information, and they drive a lot of, uh, costs. 10:27 So for SaaS tools, what are your, your perspective, which are like, uh, what are the challenges, uh, to analyze and to do FinOps for, for SaaS tools? Yeah. Again, it's an interesting domain, isn't it? 10:40 And, um, there's a lot of confusion I think often when we talk about SaaS, particularly in the FinOps space. So the number of people that I've spoken to and said, you know, "How do you manage SaaS?" 10:52 And they immediately think about marketplace, right? That would be the initial k- you know, that's the SaaS [chuckles] side of things for a, for a public cloud, you know, FinOps practitioner. 11:02 And then, you know, historically here at Flexera, well, what was SaaS? Well, it was your Office 365s, your Salesforce, your ServiceNow, right? All of those typical SaaS platforms. But, you know, I think, uh, the... 11:15 Well, certainly, obviously the, uh, you know, the AI aspect has blurred the lines between what is SaaS and what is, is public cloud. 11:25 I'm quite interested actually to see the FinOps Foundation come up with data clouds as, as a domain or a taxonomy, 'cause I would've called those, you know, Snowflake and Databricks actually SaaS applications, and I, I think that they do themselves. 11:39 But there becomes then a difference, right? 11:41 So, you know, what was historically, you know, in the typical business SaaS domain was possibly, you know, per unit pricing on a per user basis, consumption-based pricing possibly, or, you know, outcome-based pricing as well. 11:58 Um, but that's shifted, right? The public cloud was always a price times quantity thing, and so I think we're now seeing the data clouds, if you treat them as SaaS, ultimately then becoming completely different, right? 12:10 How much data am I indexing? How many API calls am I making? How many queries are you running against these services? 12:17 And then, you know, not just the data clouds, but then the, the SaaS-based AI services, Meta and others. 12:24 Well, how do you start to measure and understand, you know, both the price and the quantity of those kind of things given all of the different dimensions that are, that are being introduced? What, what about you, Victor? 12:35 I'm interested in your view on, uh, the whole data cloud piece as well. I didn't see that one coming. [chuckles] Yeah. So the data cloud, yeah, it's, it's a different thing. I think it's super important. 12:46 There are, like, companies built upon data, um, especially now that, you know, everyone... 12:51 I think now everyone can get, like, a lot of data, um, from, from the internet with all the AI scraping, all the, the information. So there is a lot of information, especially with IoT devices, all these things. 13:04 And the, the thing, I think there is a-Paradox here is that there are data tools 13:13 that work like SaaS, um, that are embedded within the, eh, cloud providers, which is like BigQueries, um, eh, QuickSight, if you can do it that way, and on Microsoft, I don't remember what it is, but I think there is one and I don't remember one. 13:33 Um, but there are... All the providers have these tool because they use it. 13:38 [chuckles] Uh, because BigQuery is probably the backbone of, eh, most of what Google does, and, you know, eh, AWS, of course, with, with QuickSight does, does a lot of things. 13:47 So they are probably BigQuery especially is the biggest, if not, I think one of the biggest cost drivers for these type of data clouds. I think it's probably the biggest one. 13:59 So you have the interaction between SaaS and cloud here, which is a very different and difficult thing to manage because you can have like, uh, the SaaS tools working on like on-prem or similar things or custom, let's say, deployments, and then you have it on cloud, which is Databricks on each of the cloud providers. 14:19 So that's like quite a challenge. I, I don't know how, how you are, you know, thinking about it, but for me, it's like quite a challenge to figure out. It... Yeah. 14:28 So a, a lot of those vendors are making their services available via the public clouds as well, and, and, uh, you know, in some degrees, it makes it easier. 14:37 So if I've already got tooling that gives me visibility of all the cloud services, then I can start to see some of those, you know, AI-based or s- what we would call SaaS-based or data cloud-based services in my own cloud bill. 14:49 Um, but often that's not the case, and it's then, you know, it's a whole new bill to ingest. 14:54 Uh, it's a whole new, you know, um, uh, a whole new domain to understand how is it billed, how is it priced, what are the levers I can pull. Can I normalize it into a common taxonomy? 15:05 Is it, you know, focus compliant and all of th-that good stuff? 15:08 And so those are the challenges that we're, we're trying to solve here ultimately is, uh, it almost res-- It, it almost looks like the, um, process we went through with Kubernetes, right? So it was- Mm-hmm.... 15:21 just another abstraction of spend that we already had, and I needed to get another layer of granularity kinda to the, uh, namespace or even to the container level to start to understand and break down that bill, and we're just repeating the process here, but with a whole load of different kinda metrics and measures, um, you know, be-behind the data consumption. 15:40 So, uh, it is certainly interesting times as we look to solve that problem. Yeah, definitely. 15:45 And I think maybe in these type of tools, um, probably FinOps like analyst or whatever you, you want to call that role, have a advantage because they are used to deal with similar, um, things for analyzing the billing data. 16:01 Uh, well, BigQuery was basically, um, I'm not sure if it is still, but for GCP, BigQuery was the, like, the export. You need to do it to do it on focus. So you need to deal with these type of people. 16:14 I, I recall, like, Anderson Oliveira's, eh, work on, on AWS for, uh, for Azure with Focus. So everyone in FinOps is kind of related to these tools. 16:26 So there is probably, like, an easier development here from, from FinOps, and they could probably handle better than, for example, compute optimization or, as you said, Kubernetes. 16:36 I think Kubernetes, you need to have domain expertise because it's very difficult, um, to do the optimizations, to do the cost allocation, for example. If you have, like, shared workloads in Kubernetes, 16:48 you need to know what you're doing. Um, so I think maybe here, uh, FinOps, like analyst, could have an advantage. What do you think? Yeah, definitely. 16:58 Um, I was watching actually, um, a, a recording of one of the sessions I think that had been done on, on AI, which is, of course, one of the areas that really spans all of these domains, um, from, from Mike Fuller and others at the foundation. 17:13 Uh, and they started out saying, "Look, you probably know more than you think you do," right? 17:17 [chuckles] So, you know, that, that's the starting position here is having done this, you know, for public cloud, you're in a really good space, right? 17:24 You understand the complexity, you understand the, the nuances, and you understand the terminology. 17:30 Um, I think it, it, it just, you know, for me, just means that you have to be open to a new wave of learning almost, right? And just to kind of again understanding the levers of, you know, what, what is... 17:44 How is it licensed, right? How is... Is it true SaaS? Is it per seat? Is it per user? Is it per query? Are we going into AI and it's about data and model optimization? 17:53 And so it's just a willingness to jump in again and kinda learn all of these new termin-terminology. 17:59 But actually the, the framework itself, for good reason, you know, is well positioned to help you through that process, right? 18:06 So it still becomes an inform, optimize, operate conversation, and the, the domains and capabilities are no, no different really. 18:14 It's just a, a, a different abstraction or a different domain or, you know, the, the KPIs and costs of, uh, of things that we need to start to learn become, become different. 18:23 So it's really, uh, embracing change [chuckles] as always. Yeah. 18:27 I, I think it's going to be like, especially with AI, you need to embrace change because, uh, and you need to be, be prepared for because otherwise you are going to be disrupted every, every few months. 18:37 So, um, moving on to, to AI, but, but before on that, I want to remind, uh, the people that are listening that we have, uh, a webinar on the November 13th, uh, six PM, um, Spain time or five PM UK time about that we are going to dive deep into all these, uh, cost challenges for hybrid FinOps, that you are going to, to be here, and we are going to have Gerhard, which is an expert in one of the topics that we have discussed, which is licensing and all this, uh, hybrid cloud, uh, thing. 19:09 So I will leave the, the link in the description below, so you can join, and we'll do a deep dive on SaaS, on licensing, and of course, what we are going to talk about, which is AI.Um, so everyone, uh, can join in the, in the link below. 19:23 And, you know, um, moving on to the hot topic, uh, everyone talki- talks about now and which is AI. 19:32 Um, so in your perspective, you know, it's been a very, uh, talked about topic, but in your perspective, like AI, uh, costs, how, how they are differing from traditional cloud costs that you have seen, and what are the, the complexities you have seen either from customers or from your personal experience, uh, on this, on this topic? 19:54 Yeah. It's, it's an interesting one, and, you know, I think this has been the real catalyst for the convergence of all of these domains, right? Mm-hmm. 20:02 So, um, just looking at the numbers from the, the FinOps Foundation, uh, i-in that, um, webinar, as I say, that recording that I watched, right? 20:10 So Gartner projecting gen AI spend to get to sixty-four four, uh, six hundred and forty-four billion this year, right? 20:18 It's [chuckles] massive growth here and ultimately, you know, the, the audience, the practitioners in the foundation have also articulated that it's not just a public cloud problem. 20:29 Um, so they are using SaaS, they are using data center, they are using p-private cloud, they are using the data clouds to do this. 20:36 And so the first one is just recognition that it is a driver behind how do we solve this hybrid cloud problem, right? 20:43 And it, and it may be the thing that forces organizations to say, "Yes, I need to bring a SaaS persona in. 20:49 Yes, I need, you know, my SAM or ITAM person in the room, because the only way we're gonna solve this AI problem, which we're being, you know, challenged to do ultimately from, from realizing value, is if we, if, if we get together." 21:03 And whilst we've tried to kinda do that for ITAM and SaaS separately, the momentum maybe hasn't been there for all organizations. 21:11 So I, I think it's a real catalyst, um, you know, to change and to, to better bring these things together. Um, but it also then becomes an, an interesting problem that starts... 21:22 And, and you made this point really at the start here about, you know, how do we go and buy this stuff, right? How do we negotiate with a vendor? 21:31 And i-in many cases that nobody knows how to even estimate what they're consuming, right? You go and ask [chuckles] you know, "How many queries are we gonna make for Copilot or, you know, ChatGPT or whatever?" 21:43 No-nobody knows how to start to even, you know, make a guess at these. 21:47 And so we're often reliant on the vendor to give us some sort of indication, "Oh, you need this tier of, you know, consumption or queries," or, or whatever it might be. So the problem actually starts in buying it. 21:59 How, how do we even go around procuring, licensing ourselves when all of these different metrics become part of the conversation? 22:06 It's no m- no longer I need twenty licenses because I've got twenty people that are gonna use it. 22:12 It could be actually that you've got AI that is using its own license [chuckles] for, for these things when you're, when you're starting to automate stuff. 22:19 So a ton of, ton of complexity that we're seeing there as we start to think about the breadth. But a-a-again, I would encourage people to think about that from a SaaS management perspective, right? 22:31 If you're not managing the SaaS, uh, environment that you have today for business SaaS, when we start thinking about the AI element of all of those platforms you're already using and data clouds and other things, the complexity of what we now would, would typically say SaaS management is, I think is gonna balloon ultimately. 22:49 Definitely. And I think that, you know, I think AI changed the paradigm especially for SaaS consumption in like from a per, as you mentioned, per use, uh, per user, uh, perspective to a consumption-based. 23:02 Like we are now moving on consumption base, which I think it's what happened to procurement when, you know, moving to on-prem to the cloud. 23:09 That is something that is way more difficult to estimate, especially for the future. 23:13 Like if you tell, I don't know, uh, one hundred, uh, thousand company, uh, employees, uh, tell me like estimate how much AI, uh, will be spent in next year, uh, which is probably what they are going to be asked about for procurement, they will know. 23:30 Like, how are, how are you going to, to know about that if the models are changing, um, the pricing of the comp... the models are changing every time the consumption is changing, so the context is longer. 23:42 Um, so all these cost drivers are going to change, right? And [chuckles] it's, it's super different and super difficult to, to know what's going to happen for companies that, uh, they need to know this. 23:55 Like they need to do their estimates. They, they, they, you know, we are, uh, you know, in, uh, we are in the middle of Q4, so that's going to happen, uh, right? 24:04 What, what, you know, what you can say to, [chuckles] to these people? The... I mean, let's face it, right? Budget and forecast in public cloud has been hard enough, I think, right? Yes. 24:14 If I see, you know, most of my customers and their level of maturity there, it's been really difficult to get, you know, the organization to get their head around P times Q. [chuckles] Yeah. 24:25 And that's when the price and the quantity are relatively, you know, relatively well understood, uh, you know, uh, uh, items themselves. 24:32 So yeah, I think we-we're gonna see, uh, again, organizations going back to the coral walk run, right? 24:39 And so the, the initial stages is just defining, well, what, what makes up the price and what makes up the quantity for the, you know, the services that we're consuming there. 24:48 Um, and a real kinda back to, back to basics start, uh, as they kind of iterate over that. It, it... 24:54 I think ultimately it's part of the problem in terms of the whole value proposition of AI is, well, how do we define that value, but also how do we measure all of these different things that go to make up, you know, ultimately the cost element, as well as the complexity in measuring the business outcomes because they're, they're [chuckles] often very non-tangible things, you know, that, that you're looking to measure. 25:15 Yeah, that's true. And I think you, you brought a, a very interesting topic, which is like the complexity. 25:20 You know, the complexity that we are getting now, especially on, on the cost level for, for things like how do you control this and-And, you know, we, we can talk about the, the paradox that it brings, which is like now you have to control more things, so you may need more tools, uh, for the different providers, especially, you know, big companies with multi-cloud cost, with hybrid cloud cost, like on-prem c- uh, cost, the SaaS cost. 25:44 So what... 25:46 Can you explain us like why having all these tools, providing you visibility and trying to, you know, figure out things, it could probably lead to le- le- uh, less clarity to, um, to a FinOps to be able to see, okay, this is my picture. 26:02 So why do you think it's, it's going on? Yeah. Again, it's really... 26:06 it's been an interesting space, and I've spent a lot of time, you know, i-in the domain and obviously seeing native tools from the cloud vendors grow up and become, you know, really quite capable tools. 26:17 But we always kind of defend from a Flexera perspective against that because typically they don't do multi-cloud very well and, you know, they don't in-interface with other business systems and they're not necessarily, you know, as configurable as third-party tools might be. 26:32 But when you then start to think, well, all of those other domains, right? What, what is it that you use for visibility into your data center or private cloud, right? How do you know what the costs are? 26:44 Where does the usage data come from? How is it, it exposed, right? This is [chuckles] definitely an area where Excel typically is the tool of choice, right? The most popular FinOps tool on the planet. 26:56 And then we get the SaaS element, okay, and there's gonna be another set of tools, right? 27:00 There's, of course, market leaders in, in the space of doing SaaS management as well because they have their own sets of different criteria, right? 27:08 We're looking at licensing, we're looking at compliance, we're looking at, you know, whether there are unauthorized users where we've-- they've left the organization, and they're still using my Office license or, or whatever it might be. 27:18 And we're looking at it from a period-- a perspective of, in a year's time, I need to renegotiate and understand what my true consumption is and if I'm using the right type of licenses. 27:28 Whereas what we're trying to do now is bring together that kind of typically license-based discipline and then the, you know, the P times Q, the public cloud discipline and pl-- and provide visibility of that. 27:40 And, uh, it's a massive tooling challenge. 27:43 We're fortunate enough to have a consistent data layer across the tools that we have in all of those spaces, but it's still difficult, of course, to start to normalize these things and report on, on them in a, in a meaningful way. 27:55 Um, you know, I have discussions with our product managers from SaaS Manager, and I'm saying, "Well, you know, my FinOps tool is probably the right place to put data cloud data like Snowflake and Databricks because it's a, it's a P times Q thing, and that's not gonna make a lot of sense when you compare it against, you know, Office licenses and Salesforce," right? 28:14 [chuckles] So, um, it, it's really difficult. How do we bring this together? Which ultimately we need to when we start to say, "Well, what's the unit cost?" 28:22 Or, "How do I budget and forecast based on a particular service that I'm delivering when its costs are spread across all of those different hosting domains?" 28:30 Um, so exciting problem to solve, but I think it, it's definitely a challenge, you know, to, to bring those tools together. Yeah, that's, that's tool-- That's cool. 28:39 And, and I think that, uh, you know, what you mentioned about, uh, the visibility and the different, uh, you know, the different context, I think what, what each information should be put, I think makes, uh, a lot of, um, a difference between having, uh, you know, something that is just pure data and something that is, you know, knowledge and, and insight. 28:59 So I'm going to, to dive into this because it's, it's a topic that, uh, you know, I've been seeing a lot lately. It's like, you know, what do you think 29:09 makes the difference between having visibility on information and having like an actual insight or an actual, you know, action item that you can do? What do you think makes that difference for... 29:23 especially in, in FinOps for, for hybrid? Yeah, that's a great question. So the visibility piece is, you know, arguably, well, it's hard [chuckles] for a start. 29:33 But, you know, visibility for me is not just ingestion of the cloud bill, but it's actually contextualization of the bill or SaaS or data center or data cloud, whatever it might be. 29:45 So how do I get to a showback perspective, right? That, that would be visibility solved for me. And, and of course, the all of the metadata challenges and everything else be-become a problem there. 29:57 But until you can get to that point, just ingesting it, whilst you might say, "Okay, well, the bill is visible," do you have any insight, right? 30:04 The insight comes from the business context that we try and lay over the top of that. Um, and you know, which, which I understand of do- of course, of doing that in public cloud. 30:14 Tagging is probably the hardest part of the, of the job for, for a lot of people to get right. 30:20 Um, but if we're not looking in at, at those costs in the context of my business and what I'm using those resources to deliver as business outcomes, the, the insight is purely maybe useful from a procurement perspective, i.e., this is the, the aggregate of my spend. 30:36 But actually to take any meaningful decisions and, and, you know, lead that into business drivers and other things is, is very difficult. 30:43 And then, of course, the problem then becomes compounded because metadata, business context, units of consumption, they all start to differ as we then go into SaaS and data center and everything else, right? 30:55 So n-now we've got four sources of metadata that we need to normalize and, um, I'm hopeful that things like the Focus standard are really gonna help us deal with that, right? 31:05 If we can get that common taxonomy across sources of data, that's at least gonna get, get you started because normalization is again a, a really big problem there alongside the metadata. 31:17 We, we do have some customers that are, are using Focus for their on-premises infrastructure, for example, which I think is fantastic, right? That's- Yeah... exactly the way to go. So- Yeah, I thinkYeah, yeah, yeah. 31:29 I'm, um, I think that that was the goal of, of Focus, right? Or, or that's when, when I understood, uh, you know, the Focus, I... when I, when I studied it, um, that was the idea. And I saw it like-- 31:42 I first saw it like, you know, OpenAPI for big cloud billing data- Mm-hmm. -which is like I, I, I come from a developer perspective, so I was easily for me to, okay, eh, why m-m-- does, eh, why didn't this appear before? 31:56 It's like it's something that is normalized and you know how to do APIs, which is probably one of the most widely done things in development, and you have a standard on how to do them. 32:07 Eh, why don't you have it for, for billing, right? And what is for me is, like, it really is a challenge. 32:12 It's like you have so many now out of the cloud, you have different cost drivers as we've discussing today that are super different in the way that they are counted. 32:23 The only thing that they have in common is that they, uh, you know, they apply to your budget. 32:27 So at the end of the month, they will probably, eh, remove some, some money from, from your budget apart from the value that they generate, right? 32:34 But, you know, that's having common like, but the way that they generate this, uh, you know, this quantity is super different from one to the other. 32:41 So if Focus is doing that like for SaaS, for on-prem, the moment that it can be done easily that way, 32:51 uh, we have a very powerful tool for, for FinOps in hybrid cloud, which is like probably, uh, hybrid cloud including like, uh, we discussed AI and SaaS is the definitive until the moment, eh, situation, right? 33:06 So it's the final, let's say, situation with you cannot get probably more, eh, you know, more complex than that right now. 33:13 Maybe in the future we'll have like another cost driver that will appear, and we need to count on like, I don't know, image generation or video generation or whatever that is different and goes to a different way. 33:22 So I think that's, that's, that would be the key, right? Like we have an easy way, which is difficult, like simplify this process is super difficult when everyone can like, "Okay, let's deploy this. Let's have this. 33:35 Let's count everything. Let's have a, a s- a situation." And then, okay, from this data, tell me what an insight is, how I can do this, how I can, you know, uh, analyze that data. 33:47 And I think what AI helps on there, I, I, I have to say that I use it for a lot of data analysis and, you know, having the insight is like, "Okay, thank you. This is what I w- I was asking for." Like, yeah. Yeah. 34:00 I hope that you are accurate, but I think that this is gonna be ideal if you are, you know, you can give me the number, right? Yeah. No, it's an interesting challenge. Uh, and yeah. 34:10 Well, AI, as you say, may be part of the problem. Um, we h-had a conversation with some of our product teams recently just about taxonomy for AI services. 34:21 And, and they surprised me because we're actually using AI [chuckles] to classify AI, right? [laughs] Yeah. We, we've got a model that looks at it and says that's an AI service or whatever, right? So [chuckles] 34:32 bit of a circular reference there, but, uh, i-interesting space. Um, I, I'm excited. 34:37 I think, you know, the go forward really, you know, now we've got the Focus standard, now we've got the framework that is applicable to all of the scopes. These are really good building blocks to build on top of. 34:48 Um, you know, continue to remember the principles of, of, of collaborating, of course, with other members i-is really the kind of key to solving these problems, so. That's true. 34:58 And, and you know, uh, you know, we've been talking about a lot about, you know, AI, all different tools or more technical things, but one of the things that I think it's one of the core, eh, drivers and the core drivers of success for, for a FinOps practice is the people, right? 35:13 So after all, people are the ones that are, you know, doing this implementation, developing all this stuff, uh, you know, learning the new things, adapting to the new model. 35:21 So, you know, for companies, eh, what is your, you know, your perspective on how they should, you know, structure themselves for, eh, solving the hybrid, eh, FinOps challenge, eh, on your perspective? Yeah. 35:37 No, it's great. A great question. And a-again, I think the collaboration is the key, but, you know, how do you, um, how you achieve that in an organization can be very different, right? 35:47 So small organizations, agile, nimble, whatever you want to call it, you know, small relatively number o-of people, then it's easy to kind of break down barriers and, you know, collaborate and sit around a chair or a room, you know, to have these kind of conversations. 36:03 But recognizing, of course, as you scale that out and you get to, you know, large multinational or government organizations, then, then often there isn't the process or the ability necessary to break down those barriers. 36:15 It becomes a lot more challenging. 36:16 And so I think we have to recognize as part of that kind of framework some of the complexities of, you know, the political landscape ultimately that you might be operating in, uh, a-and, and how you break down those silos, right? 36:30 It, it's a big part of, you know, I guess building a new FinOps practice. It's the stakeholder management and mapping and, a-and other things, and it becomes ultimately more, more important here. 36:40 So it, it's about change at the end of the day, and I think it's really looking at your organization and saying, "Well, how do we deliver change, and how do we manage change across what may be a, you know, a, a very large organization with multiple stakeholders and [chuckles] lots of politics and other things in play?" 36:58 Um, so seek, seek the collaboration but recognize, you know, the, the challenges that your specific organization, um, may hel- have in, in helping you achieve that. 37:08 We see, you know, organizations where SaaS or ITAM and FinOps all report into the same leader and, and they're very, very successful, but we see organizations where that's true and, and they're still struggling, [chuckles] right? 37:22 Yeah. 37:22 And we see th-that conversely, organizations where they've got separate ITAM and, and FinOps, uh, you know, teams reporting into different parts of the business, but they've been very, very, you know, uh, very able to, to kinda collaborate and work together. 37:36 Um, so e-executive sponsorship always gonna be the key here. 37:41 Um, you know, understand the mandate and maybe AI and, and the value of AI is one of those drivers to helping, you know, your exec help get alignment across the organization and, and pull people into a room, um, so that they can ultimately work, work on the problem together.Yeah, that's, that's true. 37:57 And, and I think you... what you mentioned is the, is very specific that, you know, each organization need to have, you know, the, the system and the context and the structure that, eh, works for them. 38:10 And it doesn't have to be, you know, a generic thing. You have guidance, uh, that you can, uh, you know, take a look and take, especially if you're doing the new practice. 38:18 But of course, you need to adapt to your own infrastructure or your own structure the way that you think so that, eh, you can develop the best practice for, for your company, right? 38:27 And, you know, talking more of a, in a practitioner level is that, you know, for people that are, you, maybe a company of them, their, their company is expanding to, to on-prem and to hybrid cloud. 38:40 Or maybe, um, there are new SaaS, uh, because of, like, an expansion of an MA-- an MIA incorporation, which is something that normally drives a problem to integrate, uh, different providers. 38:52 What is your advice, um, if you have one, for these people that may now encounter a t- a transition between cloud to hybrid cloud? Yeah. I mean, again, it's kind of share and communicate what you have, right? 39:07 So there'll be silos within that organization that don't know anything about the public cloud and the licenses that you're running in public cloud and the marketplace costs and AI costs that you have running in cloud, right? 39:18 So, you know, they're as kinda blind or blinkered as you are, right, in terms of visibility of the overall picture. And they're gonna be just as pleased to see you pop up and say, "Hey, [chuckles] you know, I can help. 39:29 I've got some visibility of the data clouds that we're consuming, right? Uh, you know, does that sit within SaaS or FinOps? 39:35 Well, maybe it doesn't really matter right now, but, you know, we'll get together and solve the problem." 39:39 So recognize that there's a lot of value there and find the, you know, the, the team members in other areas that are trying to solve that same problem, right? 39:48 You can do that collaboration, um, you know, at, at the bottom levels as much as you can have it, uh, you know, as a top-down mandate from the exec. 39:57 Um, but it's again about bringing the value that you have and sharing the data. 40:02 [bell dinging] I also think as a FinOps practitioner that you're arguably better placed than anybody else understanding the co- you know, the concepts of cloud and the complexities of cloud, but understanding the framework, right? 40:15 Because we're recognizing that that is actually a pretty good, you know, playbook for us to work through these things. 40:21 Uh, and so sharing your, your experiences there with colleagues, you know, in other areas of the business is, is super valuable too. Yeah, that's true. 40:29 I think the, the collaboration and the communication and being able to present the information properly, I think it's probably one of the soft skills that you value the most in FinOps because you are going to be talking about and, you know, sharing with a lot of different personas and being able to adapt, uh, the context and the pitch and the information that you share to the specific audience that you have. 40:51 Talk their language, which is one of the most, you know, shared in FinOps. I think it's, it's crucial and, and I think that's, that's super valuable advice. Um, so yeah, Jeremy, it's been a pleasure. 41:02 I think it's going to be... you're going to share a lot of value with, with us on the, on the thirteenth of November, right? Yeah. 41:09 Very much looking forward to that and appreciate the opportunity to come on for the second time to-today, Victor. A fabulous conversation as always. 41:17 So I'm very excited to continue that on the thirteenth, um, you know, with you and Gerhard as well. Yeah. 41:23 It's going to be a wonderful conversation, so if you don't want to miss it, please, uh, check the link, um, below so you can register. 41:30 Uh, Jeremy, it's a pleasure and an honor that you are the, the first repeater of, of FinOps Weekly podcast. 41:36 It's been a wonderful conversation, and I hope we have helped a lot of people with, with your insights on how to do this challenge, and more information on that will be the thirteenth of, of November. 41:48 So thanks a lot for coming, and see you in the next episode. Thanks, Victor. Pleasure. Take care. Pleasure. Bye-bye. [upbeat music]