FinOps in Action

FIA - Kevin Mueller
===

Kevin Mueller: [00:00:00] we're trying to figure out, You know, is there a way we can. Look at our system, it costs money inside the cloud.

It's variable. if I change things, am I getting better or worse? You know, we have performance and reliability and all those come first, but then it's cost

Intro: Welcome to FinOps in Action. I'm your host, Taylor Houck. Each week I'll sit down with FinOps experts to explore the toughest challenges between FinOps and engineering. This show is brought to you by 0.5, empowering teams to optimize cloud costs with deep detection and remediation tools that actually drive action.

Taylor Houck: Hello and welcome to another episode of FinOps in Action. I am really excited to be chatting with today's guest. He is an experienced engineering leader and FinOps practitioner.

With over 20 years of experience, he's led teams building incredibly high performance platforms, including an options trading system capable of processing over 36 [00:01:00] million transactions per second. He played a key role in developing the well architected SaaS lens at AWS, and is also an active participant in both the FinOps Foundation's unit economics and Kubernetes working groups Today.

He's a principal architect of Cloud FinOps at Presidio. Welcome to the show, Kevin Mueller.

Kevin Mueller: Hey, thank you very much for that introduction, Taylor. That was wonderful.

Taylor Houck: so excited to have you on the show, Kevin. And, uh, there's so many topics to get in with you. You've got a ton of experience. But before we do that, I would love if you could just tell the audience a bit about your, your background, your experience at FinOps, and how it led you to where you are today.

Kevin Mueller: Yeah. So, uh, You know, I was a, a software developer for years and then manager of software developers, and then I, I went to AWS, uh, around 2016. And I was working on a team called SaaS Factory. And so we were working with partners of AWS on how to build SaaS better on, uh, AWS. And one of the topics in that is, uh, was, uh, cost per [00:02:00] customer, You know, for a SaaS platform to maintain your margins, You know, unless you're scaling your infrastructure with your customer usage. Um, You know, it's hard to tell how much money you're making and then, You know, margins are so tight on, on SaaS. You have to understand what does that cost per customer. And so that really enthralled me and uh, really got me, uh, going deep on that. And so, uh, that's kind of how I got my start in FinOps And so. I was working with a partner of AWS and I talked to 'em about cost per customer, and they happened to be a cloud financial management platform, and they, they took the ideas and then they built a, a unit economics inside their platform. It was one of the first ones in the industry. And, uh, and that's kind of my background and how I got started.

Uh, And so since then I, I, I worked for a cloud financial management platform. I've been here with Presidio for two and a half years doing, uh, managed FinOps for our customers. Presidio is great. Uh, You know, they're a, a leading technology [00:03:00] service and solution provider. Uh, from found it to cutting edge ai. Um, Presidio is a trusted technology ally for our customers and, and they just love us. And the, the people here have been so wonderful to work with.

Taylor Houck: Yeah, it's, that's an amazing experience, Kevin, and I know that the work you guys do at Presidio is, is amazing, but I'm particularly interested in, in getting started talking about unit economics. You were mentioning that when you were back with AWS, you were talking about cost per customer, and I think that, You know.

Over the past few years, unit economics has come up so much in the FinOps industry. I'm curious, where do you think that the industry actually is with it? Are people doing unit economics correctly?

Kevin Mueller: You know, uh, You know, unit economics is a pet peeve of mine and it's. There's no middle ground. Uh, most people are not doing it at all, and then there's a few people that get it, and then they're doing it really, really well. Um, And so, uh, You know, it, it's amazing, [00:04:00] You know, unit economics is, You know, one of the capabilities inside the FinOps Foundation framework. Uh, but I find most people just aren't mature enough to get there and do it. Even though it's the best way to understand you are profitable.

Taylor Houck: Let's maybe start at the very basics, Kevin, because I think people throw this term around a lot. Unit economics. What does it mean to you?

Kevin Mueller: Well, it means to me is, You know, we're all in the business to make money, right? You know, the companies are there to support the stockholders and to make money. And You know, us as developers and those engineers, You know, we are given tasks to develop something. That provides value to the customers of that product. so to me, unit economics is saying, okay, how do I measure the efficiency, the output [00:05:00] in the terms of value so I can relate that to profit. So that's unit economics. Now there's unit metrics as well. and in that sense we're trying to figure out, You know, is there a way we can. Look at our system, it costs money inside the cloud.

It's variable. if I change things, am I getting better or worse? You know, we have performance and reliability and all those come first, but then it's cost right there. And so a grand scheme, that's what unit economics means to me. And unit metrics means to me.

Taylor Houck: It seems so easy in principle, but it's obviously not so easy because as you mentioned earlier, there aren't so many companies who are doing it properly. Where do you think implementing unit economics goes Wrong.

Kevin Mueller: Well, a lot of people try to boil the ocean right away, right. And You know. You have to get your allocation [00:06:00] right and you have to understand what your shared services are and what non-shared services are. And so they go down into the low level details. And what I like to do is kind of start high and work down versus starting low and working up, You know, take, uh, whatever your product is, You know, have a product. should be doing one major thing. And then that product has probably various features that do various things, and then there's shared services that kinda support that. So what I like to do is just start at the beginning. You know, pick a metric that represents, You know, how you generate profit or value, and then start measuring that.

And you might have to get something from logs or get something from finance, but you have to collect this metric. You have to be able to put it someplace. You have to be able to inject it into a system that is normalizing all your spend and you can do allocation with. [00:07:00] then you have to build a system to kind of break apart that and do the division. You know, so you wanna say cost per customer is, A lot of people say, well, You know, I spend a million dollars and I have, You know, 10 customers, You know, so. Uh, my cost per customer is a hundred thousand dollars. Well, well, that's just peanut butter spread. You can't do anything with that. You know, you have to be able to match some type of usage to that value so that one customer you can see is driving more spend versus another customer. And now we can make decisions. So the hardest part is just starting. start high. Get your pipeline working as far as getting the delayed data, collecting it, and doing your division. you've got that, You know you have success, right? And of course it's not gonna be perfect and it's not great, but we've started. And now you can go from there and say, okay, [00:08:00] of my product, I have these five different features. Uh, maybe there's some add-ons. Now we start breaking it down and, and pick one of those and say, okay, how can I break that apart and then break down the next one and then the next one. now you're really getting to that point of saying, okay, look. Now I have unit metrics as well as unit economics to figure out what the cost of everything is. And now engineers can say, okay, what customers are driving costing me more than the others? And now I can ask why. Right? And maybe what features is costing more than another one? And I can ask why. You know, team A built something that's really efficient and Team B maybe didn't and now they can work together and figure out why.

And so. When we talk workload optimizations and, and saving money, You know, you got table stakes things and you have, You know, like 0.5, there are deep waste detection. But now we can get to engineering led optimizations. [00:09:00] That's where you can save money and make decisions. And starts with getting your allocation right, You know, picking, uh, your products and the metrics and starting and just starting.

Taylor Houck: It just gives you a measure of productivity of your cloud spend, right? And it's showing you the effectiveness of your investment into technology and into the cloud. But it's not just one metric, right? Kevin? I'm sure you've seen across different customers and, and even within one customer, maybe even within one application, there are multiple metrics that are important to measure spend against.

I am curious to hear from you and your experience. What are the metrics that you've seen work best, and if there's any examples of unit economics working well, I, I'd love to hear about it.

Kevin Mueller: Yeah. So, uh, we, Presidio has a customer called, uh, um, Q2 and they have, um. Uh, a, a large Kubernetes workload and they do something special with, [00:10:00] uh, SQL databases. You know, they wanted, uh, their customers are called a financial institution. There's a lot, a lot, a large bank. Their B2B, their customers are financial institutions, so they want to be able to get the cost per, uh, financial institution.

And so. Uh, part of what they're breaking down are these, uh, SQL statements. And so, uh, they can go into logs and do the times of queries, right? And they can calculate, what queries, You know, the same queries being run against various different customers and they can break it apart about query length by customer or financial institution and then really see they can change SQL statements. To make it more efficient and less expensive. Right. And then they're doing the same thing where they're using, uh, like mm. Pod start times and end time, You know, processing information. So that's a metric to break apart this [00:11:00] big container organization system they have and break it down into little parts and then they can combine these two metrics to kind of calculate, You know, a cost per financial institution. So we have compute, we have database, and then they can do the same thing with storage. You know, the size of objects, the number of objects. Uh, that's kind of one large example that we have going on, You know, right now. Now, um, things I've seen successful with are, uh, companies that are using Snowflake.

Um. Yes, you can get to cost per data warehouse, but about cost per query, per data warehouse? then there's a way to fine tune queries so that I can say, okay, this query is now running it against customer one versus customer two. And so now I can say, okay, what are my most expensive queries? Let's go optimize there. now, why is this customer [00:12:00] charged, You know, making this query so expensive, but all these other customers are not. What was it about that release we did last week that made our cost per query go up or go down? Right? These are the questions you engineers can start asking. And You know, engineers, they really don't wanna see that they got their performance, they got reliability and, and everything else to do, but. If you wanna save money, you gotta take that extra step,

Taylor Houck: This right here. This is the important part, right? This is where you flip from visibility into action. It's so easy, I think, in FinOps to measure all of this stuff, put it in a dashboard and. And look at it and feel happy that it's there, but you need to drive towards action. And that's what you're getting at, right?

That's when you're looking at, okay, if I have all these unit metrics, now I can identify the queries that are more expensive relative to all the others, and that's where I can focus my time and attention. Same thing with specific customers.

Kevin Mueller: Right. You know, and these queries are something, [00:13:00] You know, that you're doing over and over and over again. So you, you can save 5% here, but You know, that might drive 20% margin on your customers. And then, You know, it's not just about engineering now it's about product. People, product people, they gotta come up with the products, they gotta come up with the future and they gotta come up with the pricing. Well, how much should I charge for this? If, if I don't know how expensive it is? I don't know how to figure out how much to charge, and I don't know how to protect my margin, right? Imagine salespeople, they can go in and, You know, uh, a customer comes in and does a proof of value and within 24 hours. You know how much it's gonna cost you to serve this customer for the next year.

Okay? Um, You know, this is the contract value I want to get to. I'm willing to go down to here, but I can't go any lower because it ruins my margin That's gonna ruin, uh, my SaaS metrics and it's gonna ruin me from going from a series A to B2C [00:14:00] to D funding, right? Because you have unit economics. not just engineering, it's product now it's finance, it's sales as well.

Taylor Houck: This actually brings up a really good point that I wanted to get to, and it's thinking about the scoping of your cost allocation and of your unit economics. Right? Because a lot of people that I talk to when they are thinking about FinOps and and cost allocation and chargebacks and show backs, they go to allocating cost by team or by business unit or by owner.

But what you're talking about, it sounds like you're more framing it around products and features and applications. How do you think about this distinction?

Kevin Mueller: Uh, You know, you nailed it right on top of the head. You know, everybody. Wants to do allocation ownership. You know, I have an application. This application is owned by this team, and it goes up to a business unit and, and you need that, right? But that usually does not correlate to how do we make revenue or how do we provide [00:15:00] value, right? And so I think about, You know, products and features and it's a little bit loosey goosey and it's kind of harder. to associate those things. But you have to try to do that because now your product, people can talk about products and features and finance and say, okay, we make, we sell this product and that product, and we make this revenue and that revenue.

And now I can say, what is the cost of that? And so think you have to do both allocations. Yes. You need ownership for a lot of things. once you get to unit economics, you need to think beyond that. And to me that's products and features, services, microservices, allocation, You know, You know, application that feed into that feature, that feed into that product.

And you need to do both allocations.

Taylor Houck: Yeah, so what you're saying is it's, it's not necessarily one or the other.

Kevin Mueller: Now what happens though is, is engineers usually push back and it's like, Hey, I build this [00:16:00] thing over here and it's a shared service and it really doesn't support a product. And I, I like to challenge that and say. You know, about this company is driving a profit. I mean, that's how you stay.

Uh, the company stays valuable. That's how you, um, they can pay you as an employee and that's how they give the, the bonus to you, right? And so it all feeds into profit. And I don't care if you have this shared service, it is supporting a product somehow. And you have to figure out, okay, how do I. know, show back this shared thing, back to a feature or something like that. And so, You know, you have to do a hundred percent allocation. Uh, but some things are gonna be features, but then some are shared services, and that's just a, it's a just a feature to me, right? But those shared services are usually supplementing these features somehow, and you have to figure out whatever that unit [00:17:00] metric is that feeds back into that. And so, You know, a lot of times it's like query time, it's message count, uh, those type of things. Um, You know. Um, new bank is a big customer out there. They've talked about how they have, You know, cost per 1 million credit card transactions, You know, and there's probably various features that feed into that.

And so even though your shared services, you're, you're supporting a feature that's supporting a product.

Taylor Houck: I'm curious to get your take and this, it can't help to come to my mind because, You know, when I started in my FinOps journey, I was working for a software company, a SaaS company, right? So development and cost of goods sold was inherently software and it was easy to allocate cost back to customers, things like that.

When I moved on, I, I went to NBCUniversal and a lot of our cloud and technology spend was supporting internal use cases and applications, right? Tech was no longer a revenue driver, it was a cost center. How do you think about unit economics [00:18:00] outside of traditional SaaS companies?

Kevin Mueller: Yeah, so I want to frame this to be more of, we're generally talking about a cloud where there's variable

Taylor Houck: Right?

Kevin Mueller: Right? And so every day have to make trade-offs between, You know, speed and performance and uh, security. And then I think cost is one of those things as well, right? So though. You might in, in your example where you're just providing services and stuff like that, you can use unit metrics to say, well, how much does it cost me to do some widget? You know, at that time you're making architectural decisions and that's gonna go up or down. And so imagine a graph, You know. As, and this doesn't matter if it's cost per customer or unit economics, You know, your cloud spend is rising and everybody's like saying, it, it's too expensive.

Spend less. But [00:19:00] if you can show that your unit economics is going down while your cost is going up, it's like, come on now. I'm, I'm so efficient. Bring on more customers because. Because I'm going down, I'm making more profit, or I'm, You know, providing some return and investment to the customers and it's going down, on the fuel.

But if you make an architectural decision and that reverses your cost is going up and your unit economics is now going up, you, you might not have the best decisions. Now there might be a compliance issue, there might be a performance issue. That takes precedence, but at least now, You know, and you can make those trade offs.

Taylor Houck: Yeah, and it really gets into the point that cloud spend is not inherently bad, right? Because you're spending a lot of money in a given month or a given year. That's not a bad thing. This is a productive asset, right? But it's our job to ensure that it is [00:20:00] as productive as it can be.

Kevin Mueller: Right. You know, and You know, and that's one way software engineers should, can get promoted if they. Embrace unit economics, and then they can show they're efficient and they're getting better over time. You know, kudos to that person, promote him to a manager so he can do it across more things, right? So engineers, I think, should embrace this and do this because they can show how successful they are.

Taylor Houck: Now I, I want to segue into a topic that is, is not always the most glamorous. It's been coming up forever and it is important, especially when it gets into actually tracking and measuring your unit economics, and that's tagging. Okay. How have you seen tagging be put into place, and what are the challenges with tagging?

Also, you talked about allocating to teams, the business units also to products and features. How many tags do you need to make this stuff work?

Kevin Mueller: So, You know, I've, I've had a shift of [00:21:00] thought, uh, lately. You know, I used to think, okay, maybe you need five or six tags out there. Um, but I've changed my tune lately. I, I worked with one customer and, and they had one tag. Um, and their allocation was kind of in ServiceNow or backstage or some other type of CMDB. And so if, if you have one tag and then you have some source of truth that can map that to your ownership or to your products and features, then you just need that one tag. then you have to build a system that can pull that source of truth out and then feed in whatever your You know, cloud financial management tool is, uh, maybe you create cost categories.

If you're using AWS only, uh, maybe you use something else like, uh, the teams feature in points five, which allows you to group spend together and you can feed that in. So now you only need one tag, and if you change ownership or you change mappings, don't need to retag. [00:22:00] so I think that helps. Location because you don't have to retag and there's such a fear of retagging. Um, I don't understand why people fail at tagging. You know, everybody should be doing, You know, uh, infrastructure as code and it's just set at once and then you can kind of forget it. I was working with one customer and for a long time they couldn't tag but. A new CTO came in and he said, we're stopping everything and doing this, and within three weeks they got, went from 20% tagging to 80% tagging. So it can be done. I just don't know why there's a mental barrier there. I think it's just education and, You know, whatever. It just, it strikes me crazy, but you can go multiple tags or you can go a single tag.

You just need a source of truth of where that allocation is.

Taylor Houck: Yeah. I think that a lot of the, the challenge, it also comes down to this concept of virtual tagging, which I think is, is probably gonna go away if I'm being honest, because there's there, there's a lot of thoughts that I have on this [00:23:00] in general. But before we dive too deep. Into that. I, I wanna talk about Kubernetes.

You know, I know that you're a, a big expert in Kubernetes. Been working on it for a long time. You mentioned in the intro you're part of the Kubernetes working group at the FinOps Foundation. It, it, it's not always so, so easy when it comes to Kubernetes. There's a lot of, You know, shared spend and there's a lot of resources that could be attributed in, in multiple ways.

What are the challenges that you see in, in Kubernetes and how have you seen it best, uh, resolved?

Kevin Mueller: Yeah. Well thanks for saying I'm an expert, but uh, um, I know a lot of things, but there's people that know more than me. know a little bit more than somebody else, so as well, but.

Taylor Houck: It's always the case.

Kevin Mueller: so I was working with one company and I, uh, You know, they broke apart Kubernetes spend and they used metrics to do that, and I maintained the agent that, uh, was pulling the metrics and sending it in, so I got to get my hands dirty in there and stuff like that. Um, You know, Kubernetes is a, a wonderful platform and there's, [00:24:00] there's a problem is there's several different ways you can allocate spend, You know, with labels, by workloads, by name spaces. Uh, just pick your, whatever that you want to do. But I find interesting in the FinOps is everybody like, Hey, I have this shared source.

How do I break apart? Give me visibility. Okay, here it is. Well, now I have visibility. How do I optimize it? And I'm like, oh, wait a second. You know, to me there's visibility and then there's optimization. I've yet to find a tool that can do both. Uh, pick a tool. There's a lot of tools that are great at visibility. There's a lot of tools out there that can help you bend, pack your, uh, requests and your limits and, and, and look at those performance metrics. Uh, get an optimization tool. Uh, don't try to confuse the two.

Taylor Houck: So. When it comes to optimization in Kubernetes? 'cause I think this is a topic that just keeps coming up. What, what's the starting point in your [00:25:00] opinion?

Kevin Mueller: Well, You know, at the end of the day, You know, Kubernetes manages pods and pods can communicate back to Kubernetes of, of, of their needs through this thing, concept called requests. You know, so. Reserve this much memory for me, reserve this much CPU for me. And You know, engineers, You know, they're running great code.

They're putting it out there, but You know, they're always, they want to be safe. And so they generally set the request high. And so now Kubernetes has to say, okay, I need this much memory, this many CPUs, let me spin up some more nodes. Well, the only way to save money in Kubernetes is to reduce the nodes and reduce your network traffic. don't know of any other way. It's that simple. So what can I do to reduce my nodes? And that is manage your request. And then there's a whole topic about should you set limits, should you not set limits? And You know, there's a lot of big [00:26:00] debates about that one. But start with your request and visibility tools should be able to say. Look, here is my, I'm requesting two CPUs and I only use a quarter of A-A-C-P-U over the last 90 days. probably Overprovision and we should probably do something about that one.

Taylor Houck: Backing up just a little bit from here because I, I know that you also have a lot of experience with event driven architectures. How do you make the architectural decision in the beginning of the day for whether you should even go the Kubernetes route?

Kevin Mueller: Yeah, I mean, that's a tough one. You know, when I was AWS, You know, I, I learned about Kubernetes, I learned about, You know, Lambdas and, and You know, and I was doing event driven architecture prior, and, And so I spent a lot more times in the Lambda area and, um, You know, trying to skip Kubernetes but keep getting brought back into Kubernetes, um, You know, you really need to spend some time. [00:27:00] Thinking about your architecture and the response time you need, then make a judgment call from that. Um, technologies are needed and one is not preferable over the other. But, You know, unit economics is a nice way to find out, measure which one's more effective for you. Uh, You know, a, a lambda architecture with, uh, You know, some event bridge and everything else like that.

Or, You know, do you need continuous on all the time? You know, API gateways, You know, with a container. Uh, so I don't, I don't know enough to kind of say. One is better than the other. They're both valuable. Use unit economics to figure out which one costs you the best.

Taylor Houck: I mean, that's it right there. There, there isn't one that's that's better than the other. And that's how this kind of loops back to this entire discussion, because unit economics, it's not just about visibility, right? It should be a decision making [00:28:00] framework that helps you make the optimal decision for your cloud infrastructure.

Now, Kevin, we've been talking about Kubernetes as a shared resource, right? But I think there's an even bigger shared resource coming at us, like a wave right now. Uh, let's talk about ai. All right? You told me during our prep call that you think of AI as more of a shared resource than even Kubernetes.

What, what did you mean by that?

Kevin Mueller: I, I, I do, You know, you have all these shared models out there. You have shared licensing, You know, um, you're building GPU farms, You know, to, You know, process multiple teams and everything else like that. So. I mean, Kubernetes is a huge shared resource, and I think AI is e even worse. And, and, and there's a lot more moving pieces to AI than there is in Kubernetes, You know, from. Just the training, the models. Is it SaaS? Is it pass? You know, [00:29:00] am I what licensing agreement I have? And trying to pull all that together in a single pane of glass is, is really hard. Um, And so. And, and then you start talking about products and features. It, it gets even harder to, to do than, uh, Kubernetes and some of these other shared resources. Um, And so, You know, AI from a Enos person, You know, I'm really fascinated with it. I'm really starting to dive deep on that one. I'm still just scratching the surface, but, um. Yeah. So, and, and with the unit economics, the, the most people, when they do to do ai, they're trying to solve a problem. Like, I want to get more productive, I want to increase conversion rates. Um, these are things that are adding value to the company, and you have to tie your AI usage to that value. And if you can't, you're just experimenting at that [00:30:00] point.

Taylor Houck: I mean, this is exactly it. And by the way, you were saying a. Earlier that it's like you're, you're new to it. Well, the thing is, everyone is, is new to this. There's no one with 10 years experience managing the cost of AI workloads at enterprise scale. Um, but it's funny because in the conversations that, that I've been having with, You know, big AI spenders right now, this is the main question is what is the value I'm getting out of this investment in ai?

How do I measure. The productivity gains because at this point, it, it, it, it's kind of intuitive that it is making you more productive, it's making the team more productive. But how do you actually measure that and how do you make sure that your investments in AI are going to bear fruit at the end of the day, especially now that, You know, for the past couple years it's been easy to talk about AI but then kind of toss it to the side as not a big deal.

'cause it's still less than 5%, less than 2%, maybe even less than 1% of your overall cloud spend. But the [00:31:00] growth that we're seeing right now, man, it's, it's exponential.

Kevin Mueller: It, it is, uh, huge and just gonna go higher. You know, but You know, when I'm advising people on FinOps for ai, it's, you have to understand why you're doing that, You know, and it is that you're trying to increase customer engagement by 40%. Okay, well, how do you measure that today and how are you gonna measure that tomorrow? I mean, the fact that you want to increase customer engagement, um, somebody in the finance or product knows how to measure that. And so now, okay, I keep measuring that same way, and I introduce this ai. Now I can. Track the cost. And because I'm allocating to these products and features of customer engagement, I can say, well, this is what that cost is. And then however they're measur measuring customer engagement, is it going up or down? And if it's not moving the [00:32:00] needle. You've just wasted a bunch of money, right? Uh, and if it is moving the needle, okay, great. What is that worth to you to increase customer engagement by 40%. You know, is that a dollar or is it $10?

Okay. Am I using $25 to save that? $10? Okay. you get there, now You know what to fund and what to stop. And again, that goes into the unit cost, and it's even more important with ai.

Taylor Houck: Now let's talk about the other side, right? You're talking about measuring the productivity that AI can have. Well, let's talk about harnessing that productivity, right? I'm talking about AI for FinOps. How are you seeing. AI changing the, the work of a practitioner. And how would you, how do you expect that AI will be used within, uh, modern FinOps teams in 2026 and beyond?

Kevin Mueller: You know, think just talking about AWS, You know, the Kerr, You know, these billions of rows of [00:33:00] information with all these columns. You know, only a few people really know the insides and outs of that. And as a FinOps person, I'm pretty good at that. I'm not the best, but I'm pretty good at that.

But imagine an engineer and a finance person. You know, that's a foreign language. Uh, AI is gonna democratize that and make it easier for people to ask questions about the data. You know, everybody's putting that into their platform, and that's really good. now we have this concept to MPC that, You know, uh, the neutralize the normalized data brought in from all these places.

You know, you have to normalize ai. SaaS vendors and You know, you get this PPDF, but You know, it's hard to scan that. But now you can put it all in there and ask AI questions and organize it for you as a FinOps practitioner, now I get to spend less time ask, answering, You know, questions that, uh, most people are making that are basic.

'cause [00:34:00] now finance people, engineers can ask those questions. Now I can get onto bigger and better things like unit economics.

Taylor Houck: It, it moves you from the, the analyst number crunching question, answer role, into a more strategic function that's driving, You know, the, the investment into the cloud.

Kevin Mueller: Yep. And then, You know, uh, You know, a lot of times I'm building dashboards. Now my speed at that is a lot faster because I, I know the data a little bit. I kind of know where I want to go. I can interview somebody, figure out what they want. And I could build that faster than trying to be a, a bi expert in QuickSight or BI or whatever else is out there.

Taylor Houck: I am so excited for what's to come as it pertains to dashboarding and creating, You know, visualizations on top of data, because I've already seen what Claude can do in terms. It can literally just generate for you HTML and CSS, and you can host it on a la like behind Lambda and it'll cost you nothing.

And then people can [00:35:00] just have access to these, these custom visualizations that then, hey, you wanna look at this in a slightly different way. Just query it in natural language and it will adopt to you.

Kevin Mueller: Mm-hmm. No, it's gonna be, uh, fantastic and open up, uh, so much more. Um, You know, unfortunately. Uh, as a FinOps team, maybe my FinOps team's a little smaller because we can do more and, and that's, You know, but that's okay. But allows me to get away from the mundane things and do more of the exciting things, is what I'm excited about.

Taylor Houck: Yeah. Again, it just makes us all much more strategic and there's gonna be so much more cloud spend in this AI era. There's gonna be no shortage of. Of things to do and places to add value as, as an individual. But, uh, on that note, You know, Kevin, there's, there's still a lot in this life that, that AI and, and GPUs can't do.

Right. Uh, taking care of a, a farm is one of that. Maybe you wanna tell us a little bit about where you're sitting right now and what's going on all around.

Kevin Mueller: Yeah, so, uh, I, uh, I [00:36:00] moved to Northern Virginia and I'm sitting on about almost eight acres of land. And, uh, my wife wanted chickens, so now I have 27 chickens. I got four peacocks and five goats. Um, uh, I'm sitting in my, uh, what I call my office. She. So I built a 12 by 16, shed about 150 feet away from the house, so I still commute out of the house every day.

I get into my office shed and I got starlink and I can. and talk to everybody, and I can look outside at some nice trees. And then the chickens, they come up and they want to be fed. And, uh, AI can't do that. Uh, I would like a robot to do the lawnmower. That would be really nice. And, and my wife wants a robot to carry the 50 pound bags of feed out. Maybe AI can help me a little bit.

Taylor Houck: That's so amazing. Did you, did you always want to do that? Have you, have you had chickens and, and farm animals your whole life, or is it a recent endeavor?

Kevin Mueller: No, I just, uh, we just, uh, You know, we had to move. I [00:37:00] switched jobs and, uh, You know, we had to move and, You know, we just said, Hey, we're just kind of tired of this, uh, suburban life. uh, my wife's sister-in-law, uh, had a little bit of land and stuff like that, so we're like, um, I, I circled it back and I said. They gave me a choice of five cities to go to, know, New York, San Francisco, and I'm like, if I move to any of those, I'm gonna have like an hour, hour, 20 minute commute. And so we chose Virginia and I said, this is where I gotta work. Here's an hour commute. You know, find something. And we found a, a little slice of heaven, uh, that's really nice.

Surrounded by trees. I, I can't see any of my neighbors or anything else like that. And, uh, we just have a lot of fun.

Taylor Houck: That's amazing. I think that's something so many of us, uh, strive for and, and what a good life and AI is not taking that, uh, not taking that from you. So that, that, that's amazing, Kevin.

Kevin Mueller: It's wonderful.

Taylor Houck: Awesome.

Well, this has been, uh, such an amazing episode. Uh, Kevin, before we wrap things up, where, where can people find [00:38:00] you and, and reach out if we're interested in learning more?

Kevin Mueller: Yeah, You know, I'm on LinkedIn, so I think I'm just LinkedIn slash You know, Kevin Mueller, uh, one word. Uh, you can find me there. Um, You know, uh, I think that's the best place to, to get ahold of me. There is, uh, and then, You know, uh, add a note if you do a connection or follow, uh, I would love to reach out and talk FinOps, top unit economics. Uh, just reach out to me.

Taylor Houck: Awesome. Kevin, this has been incredible. Thank you so much for coming on the show.

Kevin Mueller: Hey, thank you for the invite. I really enjoyed it. Thank you.

Taylor Houck: And thank you to our audience. If you've got something out of today's episode, please share it with someone that needs to hear it. This has been another amazing episode of FinOps in Action, and we'll see you next time.

Outro: That wraps up another episode of Fit Ops in Action. Thank you for joining. For show notes and more, please visit fit ops in action.com. This show is brought to you by 0.5, empowering teams to optimize cloud costs with deep detection [00:39:00] remediation tools that actually drive action.