EP. 33 How Etsy Prepares for the Future of Cloud with Dany Daya

Matt Pacheco
Hello everyone and welcome to Cloud Currents, the podcast that navigates the ever evolving landscape of cloud computing and its impact on businesses. I'm your host Matt Pacheco from TierPoint and I head the content marketing function. Today we're thrilled to have Danny Dyer, the Director of FinOps and Cloud Strategy at Etsy. Danny brings over two decades of experience bridging the gap between business and technology at Etsy. He's been instrumental in leading the company's cloud migration, optimizing costs and spearheading initiatives in sustainability and generative AI. In this episode, we'll explore Etsy's cloud journey, dive into challenges of implementing cutting edge cloud tech like AI and machine learning and all those great things, and discuss how to balance innovation with financial responsibility in today's competitive e commerce landscape. So, Danny, thank you for joining us on Cloud Currents today.

Dany Daya
Thank you very much for having me, Matt. Super excited to be here and let's dig into it.

 

01:17 - Etsy's Cloud Journey and Cloud Migration Challenges

Matt Pacheco
Yeah. Want to learn all about you and what you're doing over there. Can you walk us through your cloud journey from beginning to where you are now?

Dany Daya
Yeah. So Etsy made the strategic decision to migrate to the cloud, I believe in late 2017. We kicked off this migration in 2018. I think 2018 was a critical year for us and the first year for peak season, which is the holiday season for starting Thanksgiving for a lot of e commerce companies. Etsy is one of them. It was a pivotal moment because this is when were serving traffic from the cloud infrastructure and no longer from the data centers. The migration continued in 2019, I believe we concluded that in 2021, but the bulk of the migration was concluded in 2019. And since then we have been trying to leverage the latest and the greatest that the cloud can bring to help us achieve our goal in serving our community of buyers and sellers.

This is from leveraging more efficient architectures to try to do sub cost optimizations here and there. And now recently of course, with the whole Genai advancement in AI again, we're trying to be on top of all of that and the cloud has enabled us to do all of that.

Matt Pacheco
What originally interested you and or got you interested in cloud?

Dany Daya
Like personally, this is a very interesting question. So I come from engineering background. I started my career working as an electrical engineer and slowly I pivoted into the business side of engineering. So to me cloud was the ideal fit because it has tons of engineering technology requirement and this is what it's all about. And yet it's a huge importance, important piece of business. When you look at it from a financial perspective. So to me that was the great fit where I can exercise and use my engineering background, work with engineers, which I love and enjoy, but also with angle focused on strategy and business and financials. So this is how I ended in the cloud. I started with AWS on the flip side and then I shifted to the customer side with Etsy.

Matt Pacheco
Really interesting. Can you talk a little bit more? And you kind of talked about this at the beginning, but give us a little more about Etsy's cloud migration journey and then possibly some of the challenges you faced while going about that journey.

Dany Daya
Yeah, that again, that was a very interesting, long, challenging migration. I think to me one of the biggest challenges. Again, I'm not going to be touching too much on the engineering challenges are experts in engineering who can talk more about them. But like, from my perspective, if we kind of look at this intersection between technology and business, I think the biggest challenge, the first big challenge was changing the mindset and the culture. Historically, when you talk about data centers, you have great developers doing whatever they want to do, but they are limited with the capacity of the data center. When you go to the cloud, you have theoretically infinite amount of resources, which is great to unlock innovation, make you do great stuff. But also it could lead to major financial aspects.

So this aspect of teaching or providing education to developers about what's changing and when you are pushing code or you are spending resources, it has financial issues or could potentially lead to financial conversations. It could be positive. But this whole introducing this whole financial aspect to engineers was a major challenge. On the flip side also, for our partners on the finance and the finance organization, it's the same challenge. Historically they are used for data centers being capex, pretty expected, easy to forecast, they know what they are dealing with. All of a sudden the whole operation model is shifting into opex where things are not only more expensive as a line, as a dollar item, but also it's much more variable. So it's hard to predict. It's. You are giving power tons of engineers to do whatever they want to do.

And again from a financial hat, you need to forecast that you need to report on that, which is stuff that they needed to change their mindset about. There needs to be room for ambiguity and understanding that this is no longer the old days. So I think the cultural and educational shift to me was the most interesting challenge and the one that we are still working with till today. We fixed it a lot, but still it's a journey.

Matt Pacheco
So are you in multiple clouds Are you in the hybrid cloud? What's your, what's your infrastructure?

Dany Daya
Kind of like we are again nowadays. Like it's becoming really interesting when we talk about the multi cloud. So we are hosted primarily one cloud infrastructure, but we do use a lot of vendors, a lot of SaaS providers. So in a way we do operate in a multi cloud topology. But if we want to do use the legacy term, yes, we are one single cloud.

Matt Pacheco
It's interesting how you may think you're single cloud but you're actually multi cloud. It's, it's, yeah, it's really interesting how that works.

Dany Daya
It's oversimplification because when you are traffic from two separate clouds, it's a completely different topology. We are not there. But if you think about it like in bits and pieces of the infrastructure, we are leveraging this topology but at a substantially smaller scale. And for us it's really about like what makes sense. It's like when I think about what makes sense and this is the term that people at Etsy love or hate me for saying it day in, day out. What is the roi? What was the roi? Let's talk roi. It's not about the investment alone, it's not about the cost alone, it's not about the return alone. Let us put those two together and then we can make educated decisions about what works, what makes sense, what doesn't.

Matt Pacheco
Great answer. And as I'm jumping into questions about finops and cost optimization, I'll get to strategies and all that. But I'm curious, how has Your role in FinOps evolved since joining Etsy? What has changed in that world?

Dany Daya
Oh boy. So interesting. I joined as a TPM supporting cloud migration with a known focus on cost in the future. But I think what changed to me, the biggest thing is that this is what I believe was the good decision. We tailored FinOps practices within Etsy to be very closely aligned with the engineering culture. At Etsy we have a very strong engineering culture. It's phenomenal, it's amazing. So we started working as engineers and for example, the simplest example I give is like rather than me going to an engineering manager or a developer saying that hey, you have a cost spike.

I would, me and my team members would look at the cost spike, would do some digging into what happened and rather than going to the engineering manager with hey, there's a cost to spike, we aim to go to them that hey, there is this piece of infrastructure, for example, cluster of virtual machines, their utilization dropped and their count increased. So this is what's leading the cost spike. Can you help us understand that? So to me it shifted from a typical TPM role. Then we kind of tailored or took finops to the industrial or the industry definition of finops and tailored it to what worked for engineering culture. And the way I think about it is that we operate in this space between technical program management, the intersection between technical program management, vendor management as well as engineering.

Dany Daya
So we have to have weird all of these hats all the time. So to me we started with one and now we operate with a blend of three. So this is the biggest change.

Matt Pacheco
Very interesting, thank you for sharing that. So we recently did a report with 400 IT decision makers in the mid market. So various companies and respondents often list cloud budget, cloud optimization as some of their biggest challenges when modernizing IT infrastructure. And a lot of our listeners are very interested in the kind of what strategies can I use to approach these things. So Kate, would you be able to share some of the strategies that you found most effective in optimizing costs at Etsy?

Dany Daya
Yep, absolutely. And I want to. We touched on this in details at the beginning, but like probably the first fundamental one is setting realistic expectations and education. There is going to be a cultural change, a flip in the mindset of how we are approaching this topic. This is something that both sides of the equation, finance as well as engineering needs to be comfortable with. It's a journey, it takes time, it takes education, it takes a lot of meetings. So this is number one. Then you go into do you have visibility? Do you have visibility into what's going on within your infrastructure?

And this is where it could be a blend of finops practitioner working with engineers to kind of build this visibility and then do depending on your organization, showback or chargeback in a way that enables your end users or basically your engineering organization or developers to understand where the cost is coming from. Once we like the moment you expose this data to engineers, you start seeing with a bit of education you start seeing a lot of aha moment like oh, I didn't know that this will cost me that much. I didn't know that this will cost me that much. Oh, I did this. I didn't autoscale this cluster or I didn't auto snooze this machine. So there are these type of fundamental out of the box optimization that you can definitely take advantage of and they are pretty quick, easy wins.

Then you kind of start going into a bit more challenging or more advanced topics around. You were saying like modernizing your infrastructure, you moved from, let's say from the data centers to the cloud. Most probably you did lift and shift. Is your new topology or architecture cloud friendly? Is it? Are there stuff in the way the different components of the stack are interacting with each other that can be optimized to become more cloud friendly? And when I'm talking about optimized, I'm not talking only about cost. I'm talking about the blend of cost, performance, latency. A lot of things could go there. And this is one area where it could potentially unlock the doors to kind of ask the question, hey, what's out there with the cloud? Can we.

Are the cloud providers providing new technologies that are easy to use that will help us do our job much more efficient? And the big example that comes to my mind always is like virtual machines, cloud providers, almost all of them, every year they are releasing new type of virtual machines that are optimized for specific type of workload. And this is the beauty with the cloud. Again, if you're starting to have like a cloud architecture, you can sandbox these, you can test these on a small scale, and if that works, you can go to the full stack and this is where you can start seeing this concept of performance cost benefits. Again, this is one example.

And then obviously there areas that will take years to be fixed, especially when you're talking about big architectural changes in the way your stack integrate with each other, in the way your data is configured, your data lake, like these things will take years to be able to do an appropriate modernization. But it's like the way I would advise everyone to do it is like just to start with the basics and grow with the organization. Like, as the organization grow and matures in understanding these kind of evolve to leverage the latest and the greatest in the cloud.

 

14:57 - Balancing Innovation with Cost Management

Matt Pacheco
Like your company, many other companies often have new features and new products that they're giving. For instance, at Etsy, you're delivering new features for sellers, for buyers, and you're doing it on a global scale. So you have to constantly innovate. How do you balance the need for that innovation, those new features, with all of the things you're doing in the cloud as it relates to costs?

Dany Daya
This is, this is a fantastic question. I love it. So it's, to me and to Etsy, it's all about roi. To be able to have an ROI in place, you need to have an understanding of the cost of this new feature that you are developing as well as the potential benefit of this feature. And both of these are very complicated tasks. But once you build the plumbing once for those, you can easily leverage it again and again for new launches and new products. So for us, for example, when we started cost attribution or assigning costs, were primarily focusing on basic tooling like labels. Now we find ourselves that we are able to pinpoint every single launch, how much it's touching each of our component of our infrastructure and what is the cost at each level.

So that we are able to say that this feature is costing us X per year with some assumptions around normalization about traffic, etc. But we are having this information readily available for all of our launches. And by the way, this took us five years to be able, five or six years to be able to get to that level with high precision. So it's not something that is available out of the box. When you are talking about complex engineering architecture. Similarly speaking, when you are measuring the return, like different companies have different ways of measuring the returns. For Etsy we do a lot of AB testing and so at the stage of A B testing, while we have the infrastructure to capture the return, we are also able to tell the cost. So we are able to have this conversation about what is the roi.

And this is where the conversation becomes interesting. There are obviously situations where the ROI is extremely high, so no brainer, let's launch. But also there are some edge cases where, or kind of like in the gray zone where it's like it's not a clear cut win. However, this is viewed as an investment that will enable us to do more in the future. Or we know that the first level of launching is not going to be as optimized as possible. We know because the priority to launch, to measure the return and then you focus on optimization. So this is where it becomes important to not only look at the numbers or what is the roi, but to take it a level or two deeper to kind of understand what is the impact and what is the narrative behind this data point.

And we find ourselves launching stuff that initially might not look ROI positive but within quick iterations we will make it huge ROI positive. So it's really about a conversation behind this data point. So this is how we do it. And again, it's not always straightforward conversations, it's not rosy all the time. This conversation could get tense because at the end of the day you are trying to balance a very sensitive and complex topics.

 

18:48 - Approaching AI and Machine Learning

Matt Pacheco
I agree, but that's good guidance. I mean that's a good strategy. Follow up question because I'm going to get into the AI section because with Innovation. When we talk innovation, we're talking about new features like using AI. And before I ask you about, I'm going to ask you about AI and finops. So first I want to ask how Etsy is approaching the implementation of generative AI within the platform.

Dany Daya
At Etsy, we have been leaning heavily. In fact, going back to the initial question, we didn't talk about it, but one of the main reason we took the decision to migrate to the cloud in 2017 is AI and ML capabilities. Because it was clear to us that AI and ML are the future and to be able to leverage them at scale, you will need the resources. And these resources become extremely challenging with data centers. So one of the reasons was to go there. So it's something in our mind like for the last seven years and we have building on it, we have built our ML platforms, our AI on top of it, the application layer, all the support system, enablement system, the data pipelines that's required. So all of these were built slowly and surely over time.

And we have had major wins with some of these launches over the years. And again in, I would say like two years ago, the whole buzz around the Genai became much bigger. And again we kind of tried to leverage genai as much as possible, whether in the marketplace or in the internally for admin related tasks. And what I love about the initial framework we talked about ROI is that the same concept applies for gen AI. Like what is the ROI we're talking about, what is the benefit? Of course here we are also touching more on legal regulations because this is again we don't want to expose ourselves, but we need to be very candid about all of these topics. So again, like we started the AI journey pretty much in 2017 and we have been like leveraging the latest and the greatest.

And now we are also experimenting with Gen AI. We've had few successful launches that leverage Gen AI in our marketplace. We have internal use cases leveraging Gen AI. And again for us the goal is not to use AI or Gen for the sake of using them. It's using them to achieve our goals or expediting our goals in serving our.

Matt Pacheco
Community of buyers and sellers really good. So you talked about the ROI of AI. How do you evaluate the ROI of AI, especially considering sometimes it can be pretty expensive. How do you go about that?

Dany Daya
The fundamentals are the same. What's the return? What is the cost? And then let us have a conversation like these are the fundamentals. To me, yes, the investment piece could be very expensive and a lot of time, their return might not be there. But this is the beauty of the gen space is that all the major LLM providers, they are continuously launching new models. So for a lot of the quote unquote, like legacy or basic use cases, you don't necessarily need to use the latest and the greatest in the LLM space you can pivot into basic models or the basic models that are offered, which are usually substantially cheaper and they are much more reliable and available.

So again, this is where it's the beauty of having the engineering muscle of trying quickly, trying, failing quickly and reiterating because eventually yes, the cost might be at the beginning, might be very hard to observe, but the reality is with time you can leverage other technologies or sometimes older ones that can do the job for the same price or less, most likely less. So this is how we are approaching it. And yes there are situations where it's very expensive and it doesn't make sense, but also there are a lot of incidents where we found it to be like a huge ROI positive investment.

 

23:17 - Using AI in FinOps

Matt Pacheco
That's good to hear. So we talked about the potential uses for Etsy's new features with AI. Let's talk a little bit about if there's an opportunity within FinOps to use AI and where are there some benefits of doing so.

Dany Daya
What I'm getting to, at least for a company like Etsy, like again as I mentioned, one of the beauty yet challenges of my role and my team's role at Etsy is that we need to not only look at financials but take it a layer or two deeper to do some research, some investigation about what could have happened. And with where Geni is today, we have tons of bots or agents that could be configured to kind of hey, you saw a spike in cost, you detected a spike based on specific parameter. Take this information and now go to our knowledge base, to our GitHub, to our Slack, to our drives and search for terms or for conversations or documents that happened around the same time of the spike minus few days and see if you can come with a result.

And my goal of that is like rather than us as practitioners doing manual job to do this, all of research and finding the information is leveraging AI and gen AI to kind of do this research for us in the back end and provide us with its understanding of what could likely be the cause. So rather than me spending three hours trying to debug what's going on, I might have the answer in few minutes. So this is one way that will enable us to scale and capture way More than the other kind of like.

And this is where it becomes my vision where we are very like I would view this to be like 12 plus months out is like can all of this process be packaged in a way that any engineer can interact with on an ongoing basis or any depending on their seniority or the area that they focus on that bot or that agent is trained on what is likely to be the case. So this is how I am thinking about it.

I am much more optimistic with the first one because at least at Etsy we have all of this already available like part of our leveraging and we touch base on it like leveraging gen AI for in Etsy we use it to boost productivity of administration and one of the areas is having robust search engine internally that can touch base on all the areas where our knowledge could be living. So we already have that, we use it heavily. So can we just build an agent and tell it take information from here which is like a spike or whatever it is and run it in our database. So I'm quite comfortable with saying that this one is something we can ship with the appropriate engineering resources over the next six to 12 months.

Making it more of a self service function is a bit in the future. But again with the pace at which this technology is evolving, frankly speaking I think it's doable.

Matt Pacheco
It sounds very exciting and I hope that can be useful for you guys in the next 12 months.

Dany Daya
I'll keep you posted out.

Matt Pacheco
Awesome.

Dany Daya
We're just, we just scoping the project and getting the resources so we'll see where we get there.

 

27:13 - Navigating Regulatory Complexity with AI and Data

Matt Pacheco
Sounds great. So with all this innovation and AI it creates, it often creates a lot of data and with that data creates come some complexity and understanding how Etsy is in an interesting. It's. It's a global solution. People can access it from many countries around the world. How do you navigate some of that complexity when it comes to the regulatory landscape, especially AI and data privacy?

Dany Daya
That is an amazing question and it's one of the biggest ongoing challenges we are going through and just to kind of also paint a realistic picture for those who are not necessarily familiar with this topic is that the regulations are catching up. So whatever regulations we have today might to change in the future or be more strict in the future. So this is one thing we keep in our mind is that how are we balancing or how are we shipping code or products in a way that will enable us to pivot back in case regulation changes. Like sometimes you find stuff in Europe are ahead from the US step or two. So it's like, how can you have this mindset of.

And this is where it helps us that we are a global company because we have already presence in Europe and we know what's going on there. So even if stuff in the US are not there yet, we are adhering to what we think is going to be the more stringent rules in the future. The other challenge here is that a lot of this is pretty new. Like a lot of time we are meeting on an ongoing basis with our attorneys to kind of explain to them what actually is happening with this technology. What actually data. It's touching because sometimes, like you don't want to go too extreme in applying guardrails when the use case does not require that. So there is a lot of back and forth and education.

Like now it's coming to my mind, like the conversations about bridging gaps that we used to have with finance six years ago, now we are having it was legal to kind of make sure we find the sweet spot that everyone is comfortable with. So to me, it's an ongoing discussion. And again, the other thing is like, there is also ethical practices that we adhere to. Like the regulations might have like some gray areas, but at the end of the day, we also focus on ethics, on sustainability, on being fair and equitable. So all of these aspects also weigh into how we are thinking about governing the use of AI. The flip side to that, which is again just to be transparent, is that these guardrails could be viewed as limiting and slowing down innovation.

So how can you come up with a version of ROI that works in this legal technology intersection? It has been an amazing evolving journey and we're enjoying the ride.

Matt Pacheco
Yeah, I was just gonna ask you about the intersection between compliance and innovation. Cause we talked earlier about how innovation and cost management kind of cross. I was curious about compliance and innovation, but I seems like you answered that.

Dany Daya
It'S really a sweet spot that you need to try to remain within. Like you don't want to skew your practices in one direction or the other. So it's really juggling back and forth to find what is the sweet spot. And also how can we scale? Because it's one thing if you are having one or two use cases, and it's a substantially different thing when you are having like hundreds of use cases. So it's like how to balance all of these and at what point you will introduce friction to validate that everything is where it's supposed to be. Because if you introduced it very early on in the process. Like if you think about it, people will start with ideas. These ideas could lead nowhere. These ideas are using dummy data. So the exposure with this type of use cases is pretty minimal.

However, this could evolve very quickly in being a production ready or externally facing product. So where do you draw the line in the sense of okay, this is okay to do this within these guidelines, but then you will have to do more due diligence when you reach to the stage where you are confident that this is going to make it to production. So it's really all gray area that we are trying to navigate.

Matt Pacheco
Great. It sounds like sustainability is top of mind. Can you tell us about Etsy's CloudJewels initiative for measuring energy impact?

Dany Daya
Yes. So again, this goes back to 2017, 2018 when we started doing the migration to the cloud. Even before migrating to the cloud, Etsy was did have environmental goals. And when I talk about environmental goals, it was literally touching everything from the infrastructure to the actual offices, to even people commuting in and out of the office to shipping. Like we look at it from an ecosystem perspective. If we zoom in on the infrastructure piece of the stack. Like if we look at before and after, we used to be data centers and then in the cloud data centers, it was pretty simple because you receive an electricity bill that tells you exactly how much energy you are consuming. And were able to also divide it by clusters or racks. So it was pretty straightforward to say that this workload is using this much of energy.

When we move to the cloud, none of that is existing and we're like, we're not giving up on our goals. Let us figure it out. Of course we tried working with cloud partners, but for good reasons, they will not give this data because this is secret sauce. So went into saying, okay, let's try to analyze this. When we think about energy, it's really amount of usage multiplied by the unit or the energy per unit of usage. And the usage itself, like let's talk about virtual machines, is pretty much documented and measured by the cloud provider. Like you know how much resources you are using. So the question becomes, how can we come up with how much energy this specific resource is using?

And this is where one of my colleagues, Emily Summer, championed this work, did tons of research and went through tons of data sources and resources and consultants to kind of come up with a methodology to kind of say, how are we estimating the per unit usage of resource? This is how cloudjul was born. Because were able to find this data, were able to get it certified, so we started using it, but it was all like manual effort. And we are extremely happy to see today that a lot of cloud providers are providing you a version of this data readily for you to use. So for us it was like a period of 2, 3 years where Cloudjules did amazing job for us because there was no other data source to rely on.

And now the cloud providers are getting more closer to providing more and more of this data and they are also being on top of it because the industry is shifting from energy to carbon, which is all great. So that was an interesting journey that we also took as part of the migration.

Matt Pacheco
Do you think that will improve what you get from the cloud providers over the next few years? Do you see them putting the effort into that?

Dany Daya
I think so. I think with the amount of pressure, lobbying and how important this topic is becoming, I think they will make this data readily available for folks to use.

Matt Pacheco
It's excellent to hear. I'm a huge fan of that. So let's get into the trends part of the episode where I ask you a few questions about things you're looking forward to over the next few years and maybe some advice from you. What emerging cloud tech are you most excited about? Personally, generic AI.

Dany Daya
Yeah, this technology is unbelievable. Like yesterday, OpenAI announced the launch of what they are calling it Operator. Etsy is one of the launch partners. But basically the essence of it is that you chat and say, hey, go find me a gift for my wife for Valentine's Day on Etsy. She likes a jewelry and this is my budget. And what it does, it just pops up a new window or tab and goes to Etsy.com applied the filters that needs to be applied, but in the search terms, clicks all by itself and you are just continuing doing whatever you want to do. And then it will show you like the recommendation of what you could buy for your wife. It's phenomenal to me. And I can see this also moving in more directions, like drafting emails, setting reminders. Like it's kind of like a personal assistant.

So my hope is that this will make us more productive and enable us to do more things with the limited amount of time that we have. So it's a jittery. My wife hates me about ChatGPT these days because she's like anything you do, you go to ChatGPT, she can't stand it.

Matt Pacheco
I mean the answers are pretty good and it's going to be really interesting when that's integrated with Etsy.

Dany Daya
This is the thing like it's this is one of the things that's always on top of, in the back of my head is like the answers in general are accurate, but you can never take them for granted.

Matt Pacheco
Yeah, that's true.

Dany Daya
So with the fact that answers are being very accurate, we also don't want to lose our human brains that will make sure that we are applying our judgment before using whatever outcome is coming from these technologies. So to me, it's an interesting space that I'm very excited about.

Matt Pacheco
Yeah, yeah, I agree. The, the verifying the information is always a big thing, especially when you try to do a recipe and it's telling, hey, use three gallons of milk to make these cookies. And it's like, oh, wait, hold on a sec. That doesn't seem right. No, that's cool. And, and feature wise, I'm sure a lot of people find it very useful to ask an assistant to help them find a gift for their wife for Valentine's Day or anniversaries. That's a, that's a great feature.

Grocery shopping, finding a ticket for an event. It's amazing. To me, it's amazing.

 

38:44 - Evolution of Cloud Strategy

Matt Pacheco
I love that. How do you see the role of cloud strategy evolving over the next five to 10 years?

Dany Daya
I think it's going to be a blend of reactive or proactive to the changes that are happening in the market. Like for example, were just talking about gen AI. How would this change the user behavior in 5 to 10 years? And from there I would work backward to say, okay, well if the user behavior is changing, then how is the shopping experience changing? So again, you move from trend to product and then it becomes like, okay, from a cloud perspective, what should we be ready for? So to me it's really trying to see when I think about strategy in this kind of relatively five to 10 years in the tech space, this is considered long term is like putting as much energy as possible into predicting what the change is going to be and working backward.

So, and to be prepared for all of that, being agile, moving fast, happy to make quick mistakes, learn from them, iterate. So to me it's like just be on top of what's going on and try to think five or so years ahead and work backward from there.

Matt Pacheco
That sounded like the next question I have for you about advice, because you've been giving a lot of good advice throughout this whole episode. If you had to give one piece of advice to someone listening now about any of the topics we talked about today, what would it be?

Dany Daya
Don't be overwhelmed with the amount of information out there. Anyone can be very easily overwhelmed. Maybe like give up. It's not the case. Everything takes time, especially in this space. It takes time, takes education. So give its own pace and work at your own pace. Just take actions. Just start somewhere. No one is going to start this journey in the ideal place. Wherever you are, you can take small actions to help you advance your journey. And keep in mind it's really what is the business looking for. So sometimes in some industries you can be more strict. In other industries you don't need to be strict. Other industries will be against being strict or putting too much guardrails. So keep a pulse about like the end goal is to help organization achieve their business goals. And this is where finops sit.

So just understand those and then tailored the practices around those.

Matt Pacheco
Thank you for that. And I want to thank you for being on the episode today. I think we had a great conversation. I appreciate you taking the time well.

Dany Daya
That was an amazing conversation. Thank you very much for having me.

Matt Pacheco
Yeah, we'd love to have you back. Thank you for being on and thank you for our listeners for listening to this episode of Cloud Currents. You can find us anywhere you get your podcasts or find us on YouTube. Stay tuned for the next episode and we appreciate you listening.