Ep#151 Accurate #FinOps Reporting Builds Trust within the Organization with Youssef Ibrahim

September 3, 2023

Episode Summary

#awscloud #cloudcost #costoptimization

Welcome to the Jon Myer Podcast with Jon Myer, and guest Youssef Ibrahim, a Cloud Economics Lead at Block, discussing various aspects of cloud economics, #FinOps, and financial management practices in the cloud space. Youssef, with over five years of experience in this field, highlights his journey and insights into managing cloud costs effectively.

Youssef emphasizes the importance of meeting teams at different stages of cloud maturity, tailoring strategies to their needs, and fostering a FinOps culture. He touches on both centralized and decentralized approaches to managing cloud costs, using a mix of rate optimization and usage optimization initiatives. He talks about implementing AI and automation tools to enhance efficiency, generate scripts, and facilitate optimization efforts.

The conversation delves into the significance of accurate reporting, highlighting how trustworthy and transparent reporting is the foundation of effective FinOps. Youssef shares his experiences with budgeting and forecasting cloud costs, acknowledging the challenges of estimating and tying growth metrics to infrastructure investments. He emphasizes the need to communicate assumptions and strategies in the forecasting process.

Overall, the discussion provides valuable insights into cloud economics practices, the challenges faced by FinOps teams, and the strategies employed to manage cloud costs effectively within an evolving landscape.

Are you looking to elevate your brand through compelling and highly professional video content? Myer Media has the solution for you, from start to finish we offer complete video creation services that include Customer Case Studies, Creative Content, Podcasting, and more.

To learn more CLICK HERE:

YoussefIbrahim_AshleyJenningsPhotography-4-e9e9a87991c5f27aed8b49d6bd38a8d1

About the Guest

Youssef Ibrahim

Youssef has been practicing FinOps for over 5 years at multiple companies, giving him a well-rounded perspective on Cloud Economics and Financial Management practices across various centralized/decentralized operating structures. Youssef has been standing up the Cloud Economics discipline at Block since 2022 and previously worked at DoorDash and Nike. He currently resides in Los Angeles where he alternates Cloud Economics work with soaking up the sun and getting outdoors.

#aws #awscloud #finops #cloudcomputing #costoptimization

Episode Show Notes & Transcript

Host: Jon

Hi everybody, and welcome to the Faces in FinOps podcast, powered by ProsperOps. I'm your host, Jon Myer. Faces In FinOps podcast is about highlighting thought leaders in the cloud, financial management space, and insights and how they're making an impact not only within their organization but within the broader FinOps community. Today's guest is Youssef Ibrahim, who's a cloud economist lead at Block. Youssef has been practicing FinOps for over five years at multiple companies, giving him a well-rounded perspective on cloud economics and financial management practices across various centralized and decentralized operating structures. Please join me in welcoming Youssef to the show. Ef, thanks for joining me.

Guest: Youssef

Yeah, thank you, Jon. It's a pleasure to be here. How are you?

Host: Jon

I'm doing great. So how's your day going?

Guest: Youssef

Oh, it's going well. Just excited to talk to you here. It's going to be the highlight of my day.

Host: Jon

I'm glad this is the highlight of your day. And actually to kick things off, how about you tell us a little bit about yourself and where you're from?

Guest: Youssef

Yeah, absolutely. So my name's Youssef. I've been doing this cloud economics FinOps thing for over five years, back when it didn't have a name. So I've been doing this for a little bit here now. I was born and raised in Chicago, lived internationally for a little bit, came back to the States for school, and now I live in Los Angeles after spending about seven years in Portland. So I've been doing cloud economics, like I mentioned for about five years. Previously was at Nike for a couple of years and did a pit stop at DoorDash doing some cloud economics work there. And now I'm leading the cloud economics discipline here at Block. So when I'm not thinking about cloud costs and pouring over the cost and usage report for AWS, I love watching sports, big Chicago sports guys, especially the Cubs play a lot of chess. And since I moved down to LA earlier this year, I've been trying to learn how to surf. I've been failing a little bit there, but it's just nice to be out in the water. And yeah, just keep myself active, playing a little golf, a little tennis, and generally trying to soak up the sun while I get it. So yeah, it's a little bit about me.

Host: Jon

Youssef, I love how you put a little thing in there I did a little pit stop at DoorDash. You did a little dash at DoorDash. Sorry, I had a little throw a little pun in there.

Guest: Youssef

Yeah, no, I did some dashes at DoorDash. I delivered some food out in Portland while I worked out at DoorDash. It was fun. Wasn't cloud costs.

Host: Jon

I think the value of FinOps and that people who are in it, don't consider themselves one of the original, but doing it five years is like a senior member in my mind.

Guest: Youssef

Thank you. I appreciate that compliment. I will say I've seen some stuff, so things surprise me every day still, and that's just the beauty of FinOps Something new is going to pop up a new billing construct, a new anomaly is going to pop up and you're like, hi, I haven't seen that before. So I'm always willing to learn because I don't know everything there is to know.

Host: Jon

I agree with you. There's always something to learn. Youssef, how about you? Tell us a little bit about your role at Block and what you're doing.

Guest: Youssef

Yeah, absolutely. So I lead the cloud economics function at Block. I joined about a year and a quarter ago, so late last spring in 2022. Similar to DoorDash, I was the first full-time cloud economics hire, which is a very interesting place to be in because you have a lot of folks that have been doing some sort of cloud financial management in addition to their other jobs, whether they were an engineering manager or a TPM or what have you. So standing up the discipline from, not quite zero, but standing up the very formal discipline has been interesting, and I've experienced that at both DoorDash and Block. So as of today, our team has grown from a team of one to a team of two. Shout out to Jonathan Guy on my team. He's been great so far. And what we do here is we are tpms cloud economics, TPMs, and what's been interesting is over the five years that I've done FinOps I've had different titles.

Guest: Youssef

I've been a cloud governance analyst, I've been a cloud cost analyst, I've been a cloud economics, and TPM, I've reported to engineering, and I've reported to the cloud business office. I've reported to a TPM org. And no matter what the title is, no matter what organizational structure I report to within engineering, the work is still the same. We are the cloud economics cloud cost experts, and we're the ones taking a look at the bill, trying to deduce variances, trying to deduce patterns and trends, and being the conduit between finance and engineering leadership and individual contributors to try to make our cloud environment more stable, more efficient, more modern, and all of the other fun things that this industry provides for us. So all in all, I do a little bit of everything. We do centralized rate optimization. We work with decentralized teams to look at usage optimizations, and really what we're going for is trying to foster a FinOps culture at companies at this company in particular that didn't have a strong cloud economics function or culture. Before.

Host: Jon

You said for our audience, what is the TPM?

Guest: Youssef

Oh, great question. TPM is a technical product manager, and that's just the job code that we're in. So what our role is we work with other PMs, whether they're project managers or program managers, and try to make sure that all of the cloud economics initiatives, all the optimization initiatives, tagging for example, that get surfaced, it gets prioritized on folks roadmaps, and we do a lot of partnering together and communication to make sure that folks are aware of some of the org-wide initiatives that we have and how that pertains to their actual engineering or product teams. So as I mentioned, right, the titles are always a little different. At DoorDash, I had a cloud cost analyst title. It was pretty much the same work, making sure that engineering teams, for example, knew about the tagging initiative that we were rolling out and what changes they had to make to their repository to start ingesting the tag values from our internal systems.

Guest: Youssef

So it's a grab bag title from the cloud economics perspective because we support the entire company, we support all of Block, which is made up of Square Cash Title Afterpay. So we have to be proactive communicators. We have to be able to get a lead and a bead on what work is going on in what organization at any given moment. And really, it's almost like cat herding. You're trying to get everyone together and working on the same initiatives because if you have, for example, say you have an e c two optimization opportunity and one team can save $2,000 a year, they may look at that and say, oh, that's a rounding error. I don't need to save $2,000 this year. But if you have 50 teams that can save $2,000 a year, if you get the entire organization to buy in, all of a sudden everyone's contributed to a hundred K per year savings. So really making sure that everyone is on the same page and is optimizing the same way and is focused on the same priorities, it's a great way to increase the magnitude of the savings and the impact that you're able to have. If we're just two people and we're managing a large cloud budget, we have to figure out ways to get more folks with skin in the game to act on these optimization initiatives.

Host: Jon

I like that. How would you rate your company's FinOps maturity from the crawl, walk, and run? Are you in various stages throughout it?

Guest: Youssef

Yeah, no, it's a great question. It's something I think about quite a lot. We're in various stages around our different maturity levels. We're currently in the process of modernizing our tech stack and moving the majority of our services to a W s. So we've had to do a fair bit of education and opportunity. There's a lot of opportunity around the migration to educate folks about how to use the cloud efficiently. So I would say in terms of our actual spending, we're probably at a run, we're spending quite a significant amount on a W Ss in terms of discipline over the last year and a half. We're probably a lot closer to walking with visibility and allocation. We still have some progress to make with tagging, but that's underway. We've established a holistic hierarchical tagging strategy that fits most of our business units, and in some areas, we're probably still at a crawl, I don't think there's any shame in that this is a marathon, not a sprint.

Guest: Youssef

So you're not going to get to crawl to walk to run at every phase of FinOps or cloud economics all at once. There will be some things that lag. And what really kind of affects that to me is I've seen centralized and decentralized organizational structures and each has its pros and cons, but when you're trying to support multiple business units, and each business unit might be at a different stage of their cloud maturity level, much less to say their cloud economics maturity level, you have to tailor your approach to make sure that folks are getting what they need and you're not trying to overload their plate with all of these opportunities or initiatives. I'm not going to ask someone to set up auto-scaling based on a super-specific metric. I might ask them if they have autoscaling set up initially or if their workload is good for autoscaling.

Guest: Youssef

And it's knowing the difference between crawl, walk, run and what each organization needs at each stage that allows us to be more effective because we're meeting folks where they're at right in the beginning when you're moving from the data center to the cloud, the cloud billing model is very strange, right? It's variable. It's OPEX-reliant. You're not thinking about fixed capacity in the data center to support Super Bowl Sunday. You have to think about how you autoscale because, on a random July Tuesday, your demand isn't going to be the same as a Black Friday for credit card transactions or the Super Bowl for credit card transactions. So bringing folks along for the ride is more how I think about how we get from crawl to walk to run, and knowing that it's a long process is what we're trying to prioritize and focus on right now.

Host: Jon

I like how you said you're bringing folks along for the ride, and you touched on three of the main keys for implementing FinOps between the centralized, decentralized, and hub, or a variation of both. Are you using both? I think you've mentioned it where you're using centralized and decentralized at the same time throughout the organization.

Guest: Youssef

Yeah, so Block I would say, is a pretty decentralized organization. The business units function relatively independently. So Cash and Afterpay and Square and foundational engineering all have slightly different environments. And so when I think about centralized versus decentralized operating structure, it's a lot different, I would say, than another company that may have one engineering organization and one repository and one way of deploying. We have a bunch of different ways to deploy and spin things up, we have a couple of different Kubernetes environments, for example. So when I talk about centralized versus decentralized in the context of block, what I'm talking about is what can the cloud economics team manage centrally and try to propagate centrally, and what do we need to partner with application teams on our engineering teams on in a decentralized format so they can do some of the optimizations? So it's a pretty standard FinOps definition of rate optimization versus usage optimization in the operating model that we have tried to implement here at Lock and are working on improving every day, cloud economics is responsible for central rate optimization such as RIS and savings plans, trying to get to the spot, trying to get to storage steering for three and helping folks implement that versus usage optimization where we may work with an application team that's deployed into Kubernetes to reduce their idle costs by reducing the delta between their requested Kubernetes capacity for CPU and memory and what they're utilizing themselves.

Guest: Youssef

So the difference to me is what can the cloud economics team or centralized FinOps team do and manage themselves versus what do we need the teamwork and the collaboration with the application and engineering teams to do? And then again, bringing them along for the ride. It really kind of sucks as an engineering team, when you hear the cloud economics team tell you, Hey, you're over-provisioned and you're wasting 80%, please fix that. And the application team, maybe they're new to the cloud, maybe they don't fully grasp what the cloud economics team is saying, and for good reason, because there wasn't a lot of detail, that request carries a lot more weight when we come to an application or an engineering team and say, Hey, we noticed you're a little bit over-provisioned on your E C two. We took a look at your hourly data and we see that it's pretty static.

Guest: Youssef

We took a look at your Datadog monitors and we've seen that you spike in your utilization at 2:00 AM every day. Is that a batch job? If so, here are a couple of ways you can optimize your batch job to be long-running rather than all at once. With this change, you might be able to move from an R6G eight XL to an R6G to XL. The engineering effort is probably going to take a week or two. These are the tests you can run. And oh, by the way, you can reach out to us during our office hours and you can reach out to the cloud infrastructure team here in their Slack channel. If you provide all of that detail upfront, that team is a lot more willing to enact that change rather than sitting there looking at the slack from the cloud econ team and saying, Meh, I have other priorities. I have other fish to fry. I'm going to ignore this. So really again, meeting them where they're at in a decentralized model and giving them as much information as you possibly can as a Cloud FinOps representative is the key to making it work in a decentralized model.

Host: Jon

Are you using some type of AI or automation tooling to help you achieve some of the results that you're suggesting to these different business teams or companies?

Guest: Youssef

Yeah, it's a great question. Right now, we don't have anything in production currently deployed. However, my team and I, we've thought about what we could do here. There's a bunch of different opportunities you could take. One thing we're trying to do is we have a large environment. We have over 5,000 AWS accounts, and over a thousand different individual-defined applications. We want to reconfigure an easy button. That easy button if you click it for your application, takes you to a report in the cloud optimization platform and cloud visibility platform that we utilize, and it shows you your report. It might show your optimization opportunities, it might pull in information from our observability and monitoring tool that shows your utilization, and we want to automate that so folks have their dashboard. Again, going back to crawl, walk, run. That's way more of a run deployment or a run project, but we're figuring out what the crawl and walk stages look like in terms of manual reporting, manual observability metrics, and manual things before we start automating that.

Guest: Youssef

And then with AI, I have a couple of different ideas that I'd like to see us start to explore and start to implement in the next coming months and years. The first one is for the folks that are maybe less technically minded, such as myself, I don't have a formal engineering background using AI to help generate and write simple scripts for automation, and I'll use EBS. There are a couple examples with EBS such as, can you get AI to write you a script that upgrades from GP3 to G3, or can you get open AI to write a script that terminates, sorry, excuse me, it's snapshots your unattached EBS volumes first and then terminates the snapshots. Getting AI to write those scripts, at least at a rudimentary level before you have to account for all of the other nuances of your production environment in a company or an enterprise is a great way to reduce a lot of engineering churn and overhead without sacrificing a lot of cycles. So that's an area where that could help me in a skill that I'm not as proficient at. Kind of bridge the gap in talking with engineers or folks,

Host: Jon

The difference between a mature and immature FinOps team. Let's talk about some of the biggest mistakes you might see from an immature, and then are there any mistakes even mature teams make within FinOps?

Guest: Youssef

Yeah, I've made a lot of mistakes, right? Five years. It's a lot of time to make some mistakes and I've made some mistakes. A couple right off the bat are buying risks that are less flexible for a team without proper vetting. And that team went and changed the underlying instance type four months in, and I was having a Deadweight RI for eight months. I have done that. I've owned up to it. It happens and you learn from it. Another mistake that I've seen immature FinOps teams make myself included is again, putting the cart before the horse, and I'll keep going back to crawl, walk, run, generally pretty tough to do run-style initiatives before you've done the crawl and the walk. Case in point, we have a large Dynamo DB environment and we were working with AWS to try to find five or six different optimization tactics, removing global tables as the default and moving to regional tables, looking at infrequent access settings.

Guest: Youssef

As I was presenting these optimization opportunities and recommendations, I found that folks didn't have visibility into what these meant for Dynamo db, even though they'd been running Dynamo for a long time. That was a case of myself putting the cart before the horse and saying, Hey, you have these six optimization opportunities, you could save 30, 40, 50%. And folks do not know that their DynamoDB was over-provisioned and underoptimized at first. So it's, you have to understand if you're working with folks that are new to FinOps and cloud economics, you have to bring them along for the ride. AWS is billing as great as it is, can be pretty arcane, right? If you're looking at regional data transfer, you have to know that data transfer dash regional dash bytes is for us East one while all of the other regions have the regional prefix in front.

Guest: Youssef

So you have to do a fair bit of explaining and getting folks on board so you can get to some of the more complex initiatives. So what I would do if I were the first FinOps hire within my org, which I have been now twice, some of the first things that I would do is really focus on visibility and allocation, figure out what opportunities there are for centralized rate optimization that don't require a lot of partnership with a lot of different engineering teams, and then figure out what the long-term cost allocation strategy is. So for visibility and allocation, one of the things that I did was set up a monthly spend review and at first it was pulled from Cost Explorer and hacked together with SQL queries on Athena where the CUR was living. And then it kind of builds out to what does the audience want to see, what does leadership want to see, what do the engineering managers want to see?

Guest: Youssef

And then looking at reserved instances and savings plans where at one company I was able to reduce the underutilization of one RI by changing another convertible RI to a different instance type so the unused RI could pick up the slack there, things like that didn't need to make any infrastructure changes. It's just taking a look at the billing constructs and seeing what fits your environment. And then reviewing your cost allocation strategy. It's pretty much an audit of your cloud environment. Are you using single-tenant accounts multi-tenant accounts or mono accounts? I know some people refer to them as mono accounts. Are you utilizing tagging?

Guest: Youssef

Is your account naming strategy work as a pseudo-tagging strategy? What can you get to with your account names? What can you get to with tagging? How do you look at cost? Is it all in one payer account? Do you have different payer accounts? Things of that nature. And once you've done that audit and you've started to set up a monthly spend review, you're unlocking the visibility for folks and folks can start to see month-over-month trends, they can start to see the effect that cloud economics may have. They can start to envision new opportunities as well. There's a lot that you can do in the first couple of months to kind of get to an improved state. Again, knowing that this is a marathon, not a sprint, there's a long way to go.

Host: Jon

What were you doing before joining Block and how did you get into FinOps?

Guest: Youssef

So I had been working in infrastructure strategy for about a year, two years at the first company. I joined out of college and I enjoyed it but didn't know what I was going through. I was figuring out what infrastructure engineering looked like. I was doing QBR, I was doing data center reconciliation, and I had a one-on-one chat with my manager and my manager mentioned at the time that she needed someone to take a look at our cloud costs and start looking at this thing called cloud economics. And we didn't even know what the term was at the time. And I told her point blank, I said, Hey, if you don't mind me learning on the job because I'm starting from zero here, I would love to explore cloud costs. I knew that we had migrated to the cloud. I knew that folks were starting to use it, but I didn't know what it was at all.

Guest: Youssef

I kind of knew what EC2 was, but it was still a very foreign concept to me. So lo and behold, my manager considered that. She had a chat with some folks in our department and she came back to me and said, Hey, you start on Monday, welcome to the team, come up with the title. And that's how I became a cloud governance analyst focusing on AWS costs. And that was about five years ago. So it was pretty fortuitous. I saw an opportunity and I took it, I'm grateful to my manager at the time for believing in me and giving me a chance to do something I knew nothing about. And we're here five years later.

Host: Jon

Five years later, and you're a cloud economics lead at Block. What do some of your typical weekdays or months look like from a routine for FinOps?

Guest: Youssef

Yeah, it's a great question, and I'm sure other folks will resonate with this. There's not a typical week or month or day. So I can talk about some of the things that I do regularly, but waking up every morning, I don't know what's going on besides what I see on my calendar. So every month, we take a look at our coverage for reserved instances and savings plans. We try to maintain industry standard coverage of about 80% plus on our savings plans. And for ris, the less flexible they are, the more judicious we are when purchasing these ris. So we do that every month. Another big thing that we do every month is our cloud spend review, where we take a look at not just our public cloud providers, but our major SaaS providers, which I'll not name here.

Guest: Youssef

And we review the cloud costs there. We look at the highest variances to spend. We take a look at the largest initiatives that are currently in flight, and what we try to do is increase our visibility into what folks are doing in the cloud. I'm sure those of you with very large cloud bills may know that sometimes it's a mystery as to what this random six-figure charge may be or things of that nature. So our cloud spend review seeks to highlight those largest variances and changes to our cloud spend. Currently, we're in the middle of our annual planning process. So a lot of my days currently are filled with budgeting and forecasting conversations, making sure that we understand what net new incremental initiatives are coming down the pipeline and what our organic growth rate targets should be. And as of recently, I have shifted more into a management-type role.

Guest: Youssef

So a lot of my days are currently spent with meetings, and I would love to get more time on the keyboard and do some deep diving into our cloud costs. But I'm a pretty stellar team that can do that for me, and we trust each other to cover each other's backs in terms of what we do. So we're a two-person team and we cover a lot of ground in terms of meeting with teams, meeting with finance, meeting with procurement, figuring out the different personas and what they need, and trying to get ahead and be more proactive instead of reactive as it comes to managing our cloud infrastructure.

Host: Jon

How critical is it to have an accurate reporting?

Guest: Youssef

Well, it's P-Zero. If you don't have accurate reporting, no one's going to trust you. And if no one trusts you, no one's going to believe in you. And if no one believes in you, it's a lot harder to get them to act on cloud economics initiatives or complex multi-quarter projects. So the main thing that we look at when we do reporting is, I know there's a couple of different schools of thought we show folks their total utilization cost in a net amortized fashion so they can see the discounts and they can see the effect of the centralized rate optimization. I know that there have been some schools of thought out there that say, Hey, we show folks undiscounted on amortized costs, so we show them their true consumption without any of the other effects, but without diving too deep into discounts and all of that, certain products have deeper discounts than others.

Guest: Youssef

We want to ensure that folks understand they're utilizing the right products and services at any given moment to maximize both their performance and their price to get to a measure of efficiency. So when it comes to accurate reporting, we make sure that our finance team is aligned with what we're showing out in terms of our total utilization cost. We try to differentiate between what the business ledger may book for any month's cloud costs versus what an actual team might see for their total usage costs for any given month. But if we see an inconsistency or an irregularity in our cost reporting, or if we see a report showing $10,000 one day and $9,000 the other day, our hackles are raised, we're in problem-solving mode, we're in troubleshooting mode, we're trying to figure out what this delta is because if we have inaccurate reporting or incomplete totals, that is our source of legitimacy.

Guest: Youssef

And if we can't show legitimate totals, if we can't show legitimate variances, folks won't trust our information. And that erodes the core principle, which is that we are being trusted to show back word salad if we don't have legitimate totals if we don't have legitimate variances, if there are errors in our data, people don't trust us to accurately bubble up cloud costs. They don't trust our visibility and then they don't trust our allocation and then they don't trust our optimization, benchmarking, and performance. So at its core, having accurate reporting is crucial to everything else that we do in cloud economics and FinOps.

Host: Jon

You're working on budgeting and forecasting right now. Do you find it difficult to accurately budget and forecast cloud costs?

Guest: Youssef

Yes, absolutely. As we are in the middle of this data center to cloud migration, we have tried a bunch of different methodologies to try to estimate data center cloud costs or data center costs into cloud costs. How do we translate compute capacity in the data center to EC2? How do we estimate data transfer if we have a database and a set of tables in the data center, what does that look like in our cloud strategy? If we're deploying to multi-region or if we're doing multi-AZ single region? Those types of cost estimates are relatively hard, and I won't lie, sometimes we get it wrong. I feel fortunate that our team members and one of our teammates have developed a pretty comprehensive model in trying to forecast what our compute and database capacity looks like in the cloud for our migrations. But other than that, when you're trying to forecast preexisting cloud spend, there are so many other factors that you have to consider.

Guest: Youssef

So really what we try to do is we try to caveat a lot of our growth with our assumptions. We have to be crystal clear on our assumptions, so if an assumption changes, we can make the updates in our cloud model. And we are trying to figure out how do we tie our budget. How do we tie our growth not just to our infrastructure cost and our infrastructure trends, but how do we tie that to our top-line, top-level business metrics? How do we tie it to our business's growth so we can show that an investment of $1 in infrastructure results in X amount of increase in whatever top-level metric we have, whether it's revenue or profit or what have you? So yeah, it's very difficult. I like to think it's a lot more art than science in terms of budgeting and forecasting cloud costs.

Guest: Youssef

And really what we try to do is bucket and aggregate our costs into some higher level of order buckets such as business unit or cost center, and it depends on what your organization looks at and looks forward to. So for us, what we're trying to do with the cost center is bucket all of these applications into a single cost center. So if one application grows, it's not that application's out of budget or over budget, that cost center owner can sit there and try to balance it out with maybe another application that has to slow their growth or they can deprecate another aspect. So it's really all about opportunity cost, and that's the kind of visibility and kind of strategy we're trying to bring to our budgeting and forecasting

Host: Jon

As a cloud economics lead for Block. What are some of, or what's the biggest challenge you're facing right now?

Guest: Youssef

It's the culture shift. 18 months ago, folks weren't asked about their cloud spend. There were a lot of applications still in the data center. Folks were not responsible for the costs for their application because it was a fixed cost, it was depreciable, and that was not a question that was kind of bandied about moving to the cloud has a whole host of advantages compared to the data center. We don't need to go over all of them ad nauseam here, but one of the other factors in moving to the cloud is folks have a lot more control over how much they're spending and costing the company. So really the biggest challenge we're facing, and this is the first and foremost thing on our minds, is how do we get folks who historically have never cared about costs, to care about their costs, we have to bring them along for the ride.

Guest: Youssef

We have to implement a crawl, walk, run model. We have to let them know that, hey, if you're over budget the first time, that's fine. We'll work with you, we'll get some education, we'll get some training. Cloud economics will also seek to understand why you went over budget in case it was a matter of us under budgeting or something like that. And just shifting the mindset to consider fixed costs versus variable costs and all that comes with it. Folks that might spin up a box to test and capacity on, may not know to spin it down because they didn't have to In the data center. That's a behavior that we are seeking to shift and change. And so in all the data center to cloud culture shift is our biggest challenge right now. We have some great enthusiastic folks who are diving headfirst into AWS training and trying to get their certifications and understanding what the differences are between deploying to Dynamo DB provision versus Dynamo on Demand versus RDS Aurora, or even thinking about Aurora Serverless. So folks are curious, and I genuinely believe they want to do the right thing. We need to provide them with the right resources, and the right strategies so that they can fully embrace the cloud variable cost culture and move away from the data center fixed cost culture.

Host: Jon

I think the culture shift is key because you indicated in a data center that cost has already been allocated. I can run those servers as long as I like because they're already powered on. I'm already using that virtualized infrastructure where within the cloud it's that I'm just paying for what I use, and it's a mindset that I have to turn something off when I didn't have to turn it off. Youssef, I'm going to wrap things up with a couple more questions. One of those is, do you have any pieces of advice that you'd like to share with those who are trying to get into the FinOps world, who are in the FinOps world or just the broader FinOps community?

Guest: Youssef

Yeah, absolutely. If you're curious about FinOps, if you're coming at it from an engineering perspective or like me, more of a finance, math, or economics perspective, there's a lot of great literature out there that you can just start reading. AWS's documentation is very thorough, and you'll have to do some digging to find what you need there. But there are also some great blogs, some great foundations, and some great organizations out there that have published a lot of FinOps material that I would highly recommend you start reading and reviewing. There are meetups and there are groups that you can join as well to start learning about FinOps, and it's a pretty active community on LinkedIn as well. So if you're looking to break into the FinOps org, try to start reading, and formulate some questions. If you're working and if you're fortunate to have access to AWS or GCP or Azure environment, a cloud environment, find out who's responsible for the cloud costs at your organization and go talk to them.

Guest: Youssef

Go figure out what their biggest problems are right now. It's an area where I felt very fortunate in that when I broke in five years ago, I was able to ask a lot of questions of folks in my organization of folks that worked for vendors that had a cloud economics or FinOps footprint. I ask a lot, a lot of questions, and I would highly recommend that anyone looking to join the FinOps community start asking a lot of questions as well. It's the best way to learn, just start asking questions and then start learning by doing.

Host: Jon

That's some great pieces of advice, including LinkedIn. Well, who are some of the most influential people or practitioners in FinOps that you should follow?

Guest: Youssef

Well, I always get a kick out of Corey Quinn and Mike Julian out of the DuckBill group. Their content is wonderful. It's written in an easily digestible format. If you're familiar with Corey Quinn, he's got a unique voice as well when it comes to being out there and talking about Cloud FinOps as well. So I do like the Duck Bill group as an organization that talks about cloud cost in a really interesting fashion.

Host: Jon

All right, Youssef, my last question for you. This one's going to be tough, so hang tight as I answer. You might have to think about it.

Guest: Youssef

Oh no.

Host: Jon

Yeah, I'm going to put you on the spot. Where would you be right now if you didn't need to work?

Guest: Youssef

If I didn't need to work right now, I would not be in front of this computer, I'll tell you that much. No offense. No, I would be somewhere outside close to water, be it a river a lake, or an ocean. I would have a book paperback because I'll probably be trying to disconnect and some sort of snack like chips and a nice cold drink. And if I'm next to the water, maybe on a surfboard, I'll keep failing how to surf out there.

Host: Jon

I don't think you've thought about this one down to the chips now. You're making me hungry a little bit.

Guest: Youssef

It's almost lunchtime.

Host: Jon

Well, that's awesome. Youssef, thank you so much for joining us for this awesome episode for Faces and FinOps.

Guest: Youssef

Thank you, Jon. It was a pleasure.

Host: Jon

All right, everybody. Today's guest was Youssef Ibrahim, who's a cloud economics lead at Block. This has been another awesome episode and discussion around faces and FinOps, powered by our good friends at ProsperOps. Be sure to hit that like subscribe and notify because guess what? The latest episodes will be on our new YouTube channel and a blog post on Until next time.