Tim Gasper [00:00:00] Hello, everyone, it's time once again for Catalog& Cocktails presented by Data. world. It's your honest, no- BS, non- salesy conversation about enterprise and data management with tasty beverages in hand. You can see I'm at this beautiful digital bar here, and we're always excited to chat with you all. I'm Tim Gasper, long- time data nerd, product guy, customer guy at Data. world joined by Juan Sequeda.
Juan Sequeda [00:00:22] Hey, Tim, I'm Juan Sequeda, principal scientist, and the head of our AI lab here at Data. world, and as always, it is a pleasure. Wednesday, middle of the week, end of the day, time to take that break, and let's go chat about data and have a drink. And today, super excited to have a guest that... I love it when we have a guest and they suggest a guest and we can reach out to them and they say yes, and we get there super quickly. It's Amy Raygada, who's a senior data solutions manager. Andrew Jones suggested Amy as a guest after our episode we did on data contracts and here we are. Amy, it's super awesome to have you. How are you doing?
Amy Raygada [00:00:56] I'm doing fine. Thanks a lot for having me today, and I wanted to share with you the drink that I made today, actually, it's a cosmodata. So, it's something invited by myself. It's a cosmopolitan but with a little bit of gin and some other stuff just to make it fun. So, if someone is interested, feel free to reach out and I can give you the recipe if you feel funny.
Juan Sequeda [00:01:17] I love how we just jumped into it, and I have to say that out of all the drinks we've had here on the podcast, and going to all data conferences and drinks people have, I cannot believe nobody's ever come up with a name called the cosmodata. This is awesome.
Amy Raygada [00:01:37] You see how I'm into the topic, right?
Juan Sequeda [00:01:39] I love it.
Tim Gasper [00:01:41] We should just talk about cocktails today. This could be a special episode.
Juan Sequeda [00:01:46] If we do talk about cocktails, I know we're somehow going to get into the data anyways, but... I know Amy is from Costa Rica and I know it's why you're there, so I was looking to my bar, what do I have from... And I don't have anything from Costa Rica, but what do I have from Central America? And I found Flor de Cana, which is a very nice decent rum from Nicaragua. Mix it with some bitters and a little bit of lime soda, nice refreshing solid drink ready to pick me up right here through my day. So, that's what I'm having today. Tim, what invisible drink you're having today?
Tim Gasper [00:02:21] Unfortunately, today I'm having an invisible Moscow mule with one of my fake copper mugs over here. I'm actually going out to dinner and drinks tonight, so I'm going to have something very delicious in commemoration for this show, but y'all enjoy. Take a sip for me.
Juan Sequeda [00:02:34] All right. Well, we'll cheers it. What do you want to toast for, Amy? What do you want to toast for?
Amy Raygada [00:02:39] For more data governance.
Juan Sequeda [00:02:42] All right, we'll get into that, cheers. So, we got our warm- up question inaudible. Today, we'll be talking about business goals and strategies. What's a personal goal you have?
Amy Raygada [00:02:53] Well, actually, I... I told you a little bit of my background. I was first into tech, data engineering, QA engineering, software engineering. I've been in this business for 15 years, and one of the reasons I moved into the product side, into the management side, is because I wanted to be this bridge and be able to translate things in between business and technical people because there's always a miscommunication. So, my personal goal now is I'm launching my own business where I'm going to provide this kind of consulting, training people, actually. My idea is also to train the people that you have in your company to be able to do this communication seamless, easily, and based on the experiences and pain points that I had in the past. So, now I'm in Topmate.io. I'm starting with some coaching, also career changes, which I get a lot because I do free mentorship also in the mentoring club. And my goal is, at some point, maybe go solo and keep hiring people to do a little bit more of changes around data in the world because it's a topic here to stay and I think we still have a long way to go.
Juan Sequeda [00:04:00] I love that there are so many people who are focusing on education, on helping, mentoring, writing books, so that's awesome. All the best of luck on this change, and it sounds really exciting.
Amy Raygada [00:04:10] Thank you.
Juan Sequeda [00:04:10] Tim, how about you? Any personal goals you're striving towards?
Tim Gasper [00:04:14] Yeah, I'll throw one out there, it's to write a book. I would love to write a book at some point, and I know Juan and I have had some ideas on some topics that we could write about. We hear a bunch of good ideas on this show, we've got some ideas of our own, so maybe an honest, no- BS data book is on the horizon.
Juan Sequeda [00:04:29] Oh, man, I was not thinking about that. There was another book I want to go write on knowledge engineering stuff, but that's the one that we have. We have probably 300 pages of notes of every single podcast episode because everybody knows when we do our takeaways, we write them, so there's a book or multiple books. inaudible like a cocktail book, a fun book, a serious book.
Tim Gasper [00:04:51] Yeah.
Juan Sequeda [00:04:52] Yeah, I'm with you on that, we'll do that. But anyways, let's dive in. Okay, honest, no- BS, how important is it to really have the notion of ownership in data?
Amy Raygada [00:05:03] Actually, that's the pillar. You want to have real good data and data governance, and also to be able to have data quality. So, where you don't have ownership, usually, everybody... It's like the meme of Spider- Man, everybody pointing to each other because everybody blames each other. So, the reasoning why ownership is so important is because you can really... This person should be accountable of this data source, of this dashboard, of this data set, of any asset that you create with data, and make sure this complies with the needs. Or if something is going to be changed, that needs to be communicated, because what happens usually is, okay, we're going to drop from the back- end system a column, but they don't notify anyone. And then all the dashboard gets broken, and there is days when you debug, and you don't know what the hell is going on. Because that has happened to me when I was data engineer and you don't know what's happening, people is blaming, all the business because, " Hey, my dashboard is broken, everything is on fire." And ownership provides a wall to avoid these kind of issues. And also, the fact that when you change things, for example data types or other kind of schema changes and the data quality gets compromised, you're able to spot them earlier in the game so that things in production doesn't get real impacted.
Juan Sequeda [00:06:22] One of the things to dive in here, and this is pedantic... I wonder how pedantic is this, is I think the term ownership is something people get... So, the word ownership is something people get too strung up on. And I agree, totally agree that we need to have the issue of accountability, and there's so many moving pieces. So, is it really the notion of ownership or is it the notion of... Who is the owner? Is it really just one person, one department, or like... Because I think there's a lot to unpack there, and I think that's what people get annoyed about. It's like, " Oh, what do you mean you own the data?" Maybe they feel like they're hoarding the data and stuff like that. What are the different types of ownerships and roles and responsibilities here? Because at the end of the day, I think we agree that we need accountability.
Amy Raygada [00:07:12] Yeah, I actually created a framework for one of the companies that I worked before where we decided to go in a smaller scale. Meaning, for example, having technical owners, which could be back- end engineers that owns an API. Of course, we put data contracts there that it's... That's why I'm connected with Andrew because we are also very fan of this topic. We have also business owners, and the technical owners could be a data engineer or a back- end developer, for example. And then the business owners usually are the BI developers or let's say data analyst, or it could be a businessperson who is really involved with the data. Because, for example, we have people from monetization teams that they really know how to do SQL or these kind of things and they know how the business works. And that's when they get accountable for this because they need to provide, for example, even in metrics, how metrics are calculated. What are the correct formula they need to use? Because me, as a data engineer, I can say, " Okay, maybe I'm going to calculate revenue this way or leads this way," but it's not what business is thinking about. We need input from them as well so when you show them that most of the issues are happening because they are not providing enough requirements or information, and you fix it and they see it how it's really changing the game for them, they are more willing to own. So, I think one of the keys to make them own the data without feeling, " Oh, it's an obligation. I don't really want to do that. It's another hat I need to wear and it's more work for me," is actually trying to find a problem or something that you can solve. And then from there on, you see if you own this, then you can get what you're looking for. And they are more prone to do it. It's what I told you, Juan, the other day, it's like preaching and going and telling, " Do you accept data governance and ownership of your savior?" And if they accept and they are into it, they will be the ones giving the pamphlets to the others eventually, and more people will come to you.
Tim Gasper [00:09:16] I love that. You're trying to create this network effect of the value of ownership when you do that. There are some people though who are very anti- ownership, and Juan, I think you were talking about the term, and some people get worked up around what's the right way to call about it. There's also some people who are just like, " No, ownership is bad because then people try to hog the data and they try to keep it to themselves, and it becomes siloed and things like that." What do you think about that? When people say that, are they thinking about it the wrong way in terms of ownership?
Amy Raygada [00:09:49] Totally because that's not the case. I think it's more about collaboration because there could be more than one owner for one dataset. It could be, for example, let's say a data analyst and a business analyst. And they own it together because they need to define goals, but usually, you don't define these goals alone. You go to the POs, you go to the business stakeholders, you have a meeting, and then you define how this data is going to be set up for the good usage of it or for having real insights of that data. So, I think to be siloed is not really. I think... Well, at least the way that I see it is that you need to structure this in a way that foster more collaboration. That ownership jealousy, my boyfriend dataset that I love, and I don't want to give to anyone, right? It's more about just collaborative stuff. So, I have done many templates. I will be actually selling those on Topmate soon. For example, we do data value proposition. We found at some point... I see Juan-
Juan Sequeda [00:10:54] Sorry, sorry, I have to pause. There's a T- shirt right there, my boyfriend dataset I don't want to give to anyone. Oh my God, inaudible.
Amy Raygada [00:11:02] Yeah, it's like this old meme of the crazy girlfriend, remember? With the eyes like, " Oh, I love this dataset." But yeah, the thing is I think it's super important that you first evaluate what data is important and not, because then you will have people having owning maybe 10 or 15 datasets. But what happens if you actually curate this and you really understand what value is bringing to the company? We had many dashboards with no owners that maybe was used once, and business keep asking and asking and asking, asking for things. So, I made them write a template. I mean I created the template and I'm like, " Why are you doing this? Who is impacting? Have you seen a dashboard or any other data product that has already this in the past? What is the business goal? What is the refresh rate?" And that's how we can start creating products that are meaningful and usable for many other people. And we can assign it to the right domain or to the right people as an owner so we deduplicate. We have less datasets and people is less prone to say, " I don't want to own this." And it's more collaborative because they need to speak between domains, between businesspeople. So, I think that's the key where the governance actually comes in place to be able to have a better ownership.
Juan Sequeda [00:12:26] Or because I'm taking notes here, your template form is why are we doing this, who is it impacting, have you seen something else that already does this? What is the business value? What is the refresh rate?
Amy Raygada [00:12:39] Uh- huh, who is using this?
Juan Sequeda [00:12:42] Yeah, this is-
Amy Raygada [00:12:43] Because we want to know which other teams could be... For example, let's say sales, let's say, I don't know, logistics. Let's say many teams are seeing this already. Or maybe logistics needs only one metric from that dashboard, but we can reutilize already that metric for a different dashboard for them, so you don't have to redo the whole thing. Because at some point, you see in some data warehouse they have multiple metrics with the same name. All of them they calculate it differently, and you don't know which one is the right one. So, we foster one single source of truth, let's say that way, so everybody needs to agree on how they are going to calculate this metric, all the teams involved. And after that, well, I have other templates like the data product value proposition, you know more also about-
Juan Sequeda [00:13:28] Keep going, keep going, inaudible. I love it, keep going.
Amy Raygada [00:13:31] Yeah, the data product value proposition is more about the design of the product on technical perspective. So, any developer, any data engineer, any BI developer come there and understand what exactly is going on there. And we display it in our data catalog. Also, in Notion, at the beginning when we starting, it was more in Notion, now it's more on the data catalog. And also, I have another one which is the metrics library. So, if you want to create a metric, we have some other fields. People think twice because they have to fill out so many things to always give whatever they need and not doing it twice. That's a good way to do it. It's like teaching kids.
Juan Sequeda [00:14:19] How do you avoid folks going rogue and doing things without following this process?
Amy Raygada [00:14:23] It needs to be done. I mean it's part of the process that we have, otherwise, the developers won't create a new data product. It's as easy as that. It sounds bad, but that's the only way sometimes to enforce it. It's the same with the data contracts. We have a policy. If you want access to this dataset, you need to like and subscribe because we have it in GitHub, so every time there is a change, you will get a pull request. You will need to speak with the data product owner to understand what is changing and how is this going to affect you. Otherwise, they won't care if you are changing something or not. You need to be notified, so you need to subscribe to the GitHub data contract to be able to access this dataset. So, you can explore it in the data catalog, but until you're sure you need it, you ask for this permission.
Tim Gasper [00:15:12] So, this is interesting. I was going to ask you a question about ownership, we'll come back to that in a second. What is this setup that you have around GitHub data contracts? How does that work?
Amy Raygada [00:15:22] Yeah, I mean depending on the type of domain if you have a stream align or consumer align, it works like this. So, for example, if there is a back- end engineer, there is a contract between the API, and we use Pub/ Sub, for example. So, for Pub/ Sub to ingest the data, there are different mechanisms. We check the schema, we do a lot of data quality, we check all the infrastructure, and everything that is working. So, there are some tests that need to pass in order to do the pull request. Then the data engineer and the data product manager will receive a notification, " Okay, there are some changes happening here," and this guy, the guy from the back end won't be able to push to production or anywhere. It's just a pull request that they need to approve in order to move forward in the chain. But there is a constant communication because the back- end engineers understands that this data is being consumed by the data teams. And then, for example, if you're a consumer of the curated dataset, once all this data is ingested and cleaned and ready to be used, there is another data contract there in between the data producer and the data consumer. And then if there is some changes created by the data engineer, it's the same thing. There is a pull request, and all the people who are subscribed will be, " Okay, there is a change that is going to happen. Please, approve if this is good for you or not." We are trying to work now in something a little bit less technical, in the sense, for some other people that might not be that technical to understand the data contract, so they can read it more easily because in GitHub it's not for everyone. So, the data catalog that we use there, putting some things, but also, we are thinking in some other software from Agile Map. I don't know if you're familiar with them as well. But we are seeing how we can make it more accessible for non- technical people, so everybody is more accountable as well.
Tim Gasper [00:17:25] That's interesting. So, you're basically creating these data products. You're representing them in GitHub and using GitHub's pull request and change approach to manage notifications, do some automated checks as well upfront, and then get approvals before changes to data products happen.
Amy Raygada [00:17:46] Exactly. Yeah, exactly. I mean I remember when we started with all these topics of data management, governance, and all that, and I was reading more and more about data contracts. So, I spoke with our data architect who was not really familiar with the topic at the moment. But then he investigated, he was we needed in our life. So, he went through and he implemented it in this way, which I think it was brilliant because it's totally automated. We now also check, for example, the usage of a dataset, automatic crawling, and if it's not being used, then we drop it because we are also showing to the C- level that we are able to save also money. So, it's connected to a lot of automation in the data platform to make sure we are optimizing, and we get more support from them. If you go to the right people like the CFO, then you get someone who is vouching for you who's going to vouch for the other projects.
Juan Sequeda [00:18:42] Let's dive into this issue about... Not, the issue, this topic about the usage of data. Really, like you were saying, you keep track of the usage. So, yeah, what does that mean, and if I zoom out, I think I love what you're saying that you're connecting the usage, dropping things if it's not being used. Then you're saving money and then you're showing at the end finances kind of things. You're like, " Let's unpack this much more," but how you're keeping track of usage, what does usage mean, and how are you connecting that to the actual business value, and you can provide evidence that, " Yeah, keep investing in us."
Amy Raygada [00:19:17] Yeah, I mean we had a project last year where we were able to save around 200, 000 Swiss francs on our Google Cloud based on doing project cleaning. And this was a lot of projects that were created by teams because everybody used to have access. So, access control was something very important that we had to start putting in place. And not only that, we also had a lot of datasets that were maybe duplicated, that were only testing ones that just keep going there. We, of course, have deletion policy for certain datasets depending on GDPR, and these kind of things in Europe, and the Swiss law. But in general, what we tried to see is how often is this dataset being used, and if it's connected to a metric that is really being used by someone. And this comes because of the reasoning that I told you before. We have several dashboards, some of them never been used, or being used just one, and that's why we have these mechanisms of avoiding waste. Because the ESG also things are coming into place now and are really important. We wanted to really save also and be more sustainable with data, so this is a way to make sure that we are only checking which... I mean if it's not being used in one year then we go and we ask, " Okay, what's going on? Do you really need this or not? Can we drop it?" And that's how it happened the first cleanup, right? But then the consequent ones is going to be... I think we have it now every quarter that it's been scanning or something like that, just to see how they are working. And depending on the teams, they will say, " Okay, we need it or not," because we have different business units, and they are free to work in certain way. But we are putting in top the cloud governance, which is also quite important to make sure we are complying with all these cloud savings, efficiency, and data products mindset.
Juan Sequeda [00:21:16] I'm curious, do you ask people if they're using it, or do you just pull the plug and see if they complain? Because I've heard both.
Amy Raygada [00:21:25] We have done both because sometimes they're, " Yes, yes, we need it, we need it," but they never open it. And then you're like, " Really? Let's see what happens, and if they don't complain, we know it wasn't really needed. Mostly in the old platform before we migrated to the cloud, but yeah.
Juan Sequeda [00:21:45] All right, so for folks who are listening right now, what would you suggest? Should they go ask or should they just pull the plug?
Amy Raygada [00:21:54] I told you something about the Game of Thrones and the politics, and it's quite... Before you get the hate from everyone, my advice is go ask and be nice. " Hey, how are you?" If you want something from them and you want to gain them and be part of your cult or data cult, go and knock the door, ask them, and explain the reasoning behind. If you go with a small, maybe with a cocktail, with a cosmodata, yeah. And then you will see if you need to earn the trust of these people, you go and ask. If you know that this is already a bridge that was burnt, don't waste much time. But be careful always with the Game of Thrones thing because it's a reality, it's happening, and you need to be closer of your enemies as they say. Keep your friends close and your enemies closer, right?
Tim Gasper [00:22:49] Yes, the art of data war.
Amy Raygada [00:22:52] Exactly.
Tim Gasper [00:22:55] No, this is fascinating. I'm still digesting all of this, and Juan, what are you thinking?
Juan Sequeda [00:23:01] So, I wanted to dive into the politics part because you said something really important. Something I've learned in a somewhat sometimes painful way is politics. And I think one of the things a mentor of mine told me, we equate politics as something negative. And politics is not... There is negative bad politics, but politics at the end of the day is how a group of people get together to be able to move towards something that we agree on and make that change. So, I think it's really important. My initial mindset is saying, " You know what? Let's just pull the plug and see people complain." But then you're going to make enemies, especially if you're coming in, you don't want to start that way. So, you want to, " Hey, let me see what you're doing and blah, blah, blah." I think that's a very important point you made. Now, what other things... Let's talk about politics because I think there's this whole issue about... I mean I call it the politics of data, but it's also the politics of knowledge about, " Oh, this is what... What is a customer? Well, there's only definition of customer. My customer is definition, your customer... My metric definition is better than yours." It's like how do we manage all that agreement? How do get to some consensus, and what happens if we don't? So, there's all this politics of data and knowledge. I'm curious if we can unpack that a little bit more because I know that's something that we were talking about before.
Amy Raygada [00:24:28] Yeah, I mean I have a master degree in digital transformation, and one of my teachers once said, which is true, and I love the analogy. It's like a chess game. So, you just move the pieces, sometimes you need to remove some for you to move forward, but then you keep moving towards the main goal. And for me, one of the things is to make people feel comfortable and be friendly and show actually that you really care, and you are listening. It's not about only, " Hi, I'm from the data platform team, this is the new process you need to follow." Because they will immediately say, " Hey, hey, hey, no." I actually gave a talk last November, I was getting some questioning at some point at the beginning like, " Why do we need data mesh? Why are we doing this? Because we have multiple data platforms, why this data platform is better? I don't want to use yours." This, and this, and that, and this negative or backfire that you get. But when you go there and you listen what kind of problems they have and you understand how you can contribute to them a little bit or tell them, " Oh, look, I'm interested in what you're doing. I'm doing it this way. Can I see how you do it because maybe your way is better, and we can introduce it in our process?" And, " Oh, yeah, sure." And they come like this, " Oh, I didn't know you did it this way. Maybe I should try as well." And these kind of dialogues are the ones that bring this politics, this knowledge sharing together. Or for example, I used to organize this data conference at my company where we have all the data teams. We were around 70 people altogether because it's the only way they will share, and they will say what they were doing in their own business unit, how they were doing it. And fostering this communication in between everyone, people is happy. It's like data gathering and all that, so people really feels like the data platform team cares about everybody being integrated and being part of it instead of the data platform is here to ruin everything we were doing separately in our teams, and they want to be the boss, right? So, I think it's more about teaming up with people instead of just putting them a mandate on what they need to do. Otherwise, you lose them, and reclaim that trust is quite hard because the team will get bad reputation, and you don't want to do that.
Tim Gasper [00:26:59] Yeah, these political dynamics are interesting. Do you see sometimes a situation where you have a really... They could be frustrated or just excited, but the passion is high, coming from a stakeholder around the data. There are some cases where, actually, the owner should be that person because they're the one who's so passionate about it. Actually, instead of spending so much time negotiating with them, maybe they should be the owner. Does that situation happen?
Amy Raygada [00:27:28] Yeah, yeah, I actually have a quite good example, I told it to Juan the other day. We have this team from monetization, and they were the biggest complainer ever because it was like, "Oh, I need the LIPS metric, and we have five different ones, and this is not really working. And we are giving bad data to our customers, and bad data to finance, and bad data to us, and I don't know what's going on here." And data team is who to blame as usual. And at some point, I also work as a data quality engineer for quite some time, so I'm a huge advocate to it, and I was like, " Okay, let's..." We were using dbt, so I'm like, " Okay, let's Great Expectations and some other things to be able to do better testing. But from you guys, I need you to help me with best cases when we do the metrics definition." Or these meetings that I told you, we find the formula and how you want the metric to be calculated. I also need tests to see what is the deviation. What is also the intervals that are proper or not, or how we can support you to have the numbers the most accurately possible? And we were migrating at that time from Microsoft SQL Server to Google Cloud. We didn't want to do lift and shift, but because of the time and we were hiring some vendors to help us out, we decided to go lift and shift. We noticed there were a lot of issues there with their calculations and the way the data models were created, and we needed help from them. From the businesspeople to give actually real quality. So, they were, " No, why do we need to do this? I don't want it. I think this is a waste of time, blah, blah, blah." At that time, I spoke with my director. He's really supportive, so he escalated with the director of the monetization team, and they had to help. They were very like, " No, I don't want." They were in the meetings, long faces and all that. And then basically, at some point, when they were escalated, they had to work with us. And when they got the first results of the metrics with the quality they wanted, with the test cases, with the alerts, when everything went wrong, they were so, so happy, so satisfied that they were the ones coming for the next set of metrics being migrated. " Ah, I want this test case. This is not working right now. I provide you here with Excel and the calculations. Please test with this, please do that, please this." And they started to own the metrics. And now they are the owners. It's the business who owns this. It's not even a data analyst anymore. I mean they built it for them, but they own the entire data, and if there is something wrong, they need to provide the context. Why is this not working? Because otherwise, how a data analyst, a data engineer, a BI developer will know that their logic is wrong if they know business but not that deep as them? And it was a long process. As I always say, digital transformation is something that happens, or data governance transformation, it won't happen in six months or one year. It take two, three, four years depending on the organization up to five years. And only for doing these metrics and coming to this state, it was probably around six to eight months for a few set of metrics. And then you had to convince other people other product managers that were brought into data, and they didn't want to as well. But it's not about that they need to know SQL or anything, but at least to understand you understand the business, you need a dashboard. That's what you need to do, for example, right? Just provide what context do you need so we can do it better for you. I mean sometimes you have to go to the hard way, but once they understand... I told Juan, they are now the biggest fans of this. They work with other monetization teams from the other business units talking about it, and one of the guys came to me and told me, " Hey, what you did for them, I actually want for us as well. How did you do it? How can I get this?" And I'm like, " Oh, no, we did this and this and that." And he was, can you speak to my data team? Because I want the BI manager and everyone to implement the same because we want the same amount of quality. We want the same harmonization of metrics. We want everything." So, you see they start advocating for you, but the hardest part is to come and get the first door open.
Juan Sequeda [00:31:53] Okay, this is a fantastic story, and actually, the way I think about this is you are selling yourself. You're an entrepreneur. Here's this new thing you want to go do or you want to change people are doing things, and you're trying to get your first customer. And you want your customer to advocate for you, and that's the story that I'm hearing right now, and then that's the sign of success. And part of that is all the politics. I mean this politics is just like sales and marketing and, " Do it this way, it's better, but let me understand you." You have to be very empathetic. Understand what you're trying to go do. So, I really love this story about this. What I do have to ask is we always love to go tell great cool positive stories. What's not working? What has not worked? What are the failures that you've seen that we need to learn from?
Amy Raygada [00:32:52] From this specific case, for example-
Juan Sequeda [00:32:54] No, just in general.
Amy Raygada [00:32:56] In general?
Juan Sequeda [00:32:56] Just in general, the stuff that you're like, " Don't do this because I have done this in the past and I've always failed with that." I'm curious to dive into that side.
Amy Raygada [00:33:05] Of course, and it's something that I like always to share because I've been in many conference and talks, and they usually say rainbows and cats and unicorns and rainbow bright or something like that, and it's not true. It's quite hard, it's demanding, it's tiring. You feel frustrated most of the time. When you see something good that happens, you smile and you feel like you made a lot of progress, which is true, but then the next day is again the same. So, for example, we started with the data mesh thing. One of the difficult parts I will say we did it wrong. I mean the thing is... Well, it takes a lot to get the right use case. And then you need to hire people and do a lot of things, but we said, okay, we have an architect, and we build up this super cool shiny data platform. Everybody's going to come to us and say, " I want to work on your data platform." Wrong. No, they don't. They are like, " I don't need it, I don't want it. Why you want to do this? This is not my thing." I want to continue on Snowflake, I want to continue in AWS, I want to do this and that. We don't care, we don't want to collaborate." And this is something that I learned, and I will do different wherever I go, is to start actually with the mentality change from the stakeholders, from the business perspective, from the data team first. Selling that part at the beginning and take also them in consideration for the new platform for the things that are going to be set up for them. It will be much easier than actually put some people to use whatever you build up. Even though it could be the coolest, but if they don't want, they don't want. They want to feel included in the process. You cannot let them outside of the process because it's a big, big, huge mistake, and that's something I learned out of it. Also, I learned that if you as the data platform team and people is already resistant because of these issues, the best way to do it is listening as I told you. Because if you come, " Hey, look, this is the new process. We're going to put everything on this data catalog." They will be, " I don't want to use that." I have a case in one of the places I worked before where we were doing a POC of a data catalog, and we have several options. And we were doing POC at the end with two of them and three of the teams we decided to go for one. And first, we said, " Okay, everybody agrees on this one, because the other team was saying, " Yeah, we like it." And when we were about to sign the contract, they were, " Actually, we went for the other one. We didn't like that. We already signed the contract two weeks ago." Things like that because everybody has separate budgets and everything, so it's interesting the way politics works because when they feel pushed to something, they will do the opposite. It's like having teenagers. It's like a teenager, so you cannot do that. You have to talk to them very direct, very polite, very nice. Listen to them first, tell them, " Oh, yeah, you're right, cool, cool, cool." Give your perspective, and then little by little. It's super difficult. I feel like the mom always. My teammates told me the mom of the team or this and that because I'm always doing these kind of things, so yeah.
Tim Gasper [00:36:31] Yeah, yeah. You can't just be the cool parent. You also have to be the respected parent, right?
Amy Raygada [00:36:36] Exactly. Exactly.
Juan Sequeda [00:36:40] There's another T- shirt right there, it's like dealing with teenagers.
Tim Gasper [00:36:44] Yep. So-
Juan Sequeda [00:36:46] The politics and families and parents, there's so much in between here inaudible-
Amy Raygada [00:36:51] Exactly.
Tim Gasper [00:36:51] Yeah, it's a really good analogy. Maybe to continue that analogy a little bit here, sometimes you've got a lot of teenagers. You've got a lot of people that you got to interact with, and they're combating with each other and things like that. And so, to turn that into a data situation, a lot of times you've got a lot of different stakeholders, different use cases. Amy, as you think about your templates and your experiences, how do you choose the right use cases to focus on as you're addressing data across the organization?
Amy Raygada [00:37:24] Yeah, for example, I want to share one more mistake to come to this one actually is when we were doing the use case references for data mesh, for example, it took me around maybe four to five months to find the right use case. Because we started bottom- up instead of top- down because we wanted to show something valuable to the business, to the C- levels to get the sponsor because data mesh includes hiring more people as well, investing in infra, and a lot of things. So, I found things with finance, found things with HR, found things with some other logistic teams, found things with customer service, some with one of the business units. Because we have central services, and the business units focus on different marketplaces. And we did it in one of the marketplaces, but the thing is the people move so fast. I was not an advocate of this, for example, because I wanted to go more and have workshops, brainstorm sessions with the teams, introduce them what is data mesh, how it works, blah, blah, blah. But the management was rushing in the sense of I want the data mesh now, I want the first data point released so we can show it, and it's going to be inaudible completely quickly and seamless. It wasn't the case because you really need to start the data mesh or any data governance initiative by, again, as I said, going to the people first and explaining. And then I changed the rhetoric, and we don't speak anymore about data mesh, but we speak about only data governance or data privacy or cloud governance. And we divided into packages instead of just a whole buzzword because everybody's, " Okay, data mesh is a buzzword that is on LinkedIn, that is on Medium, blah, blah, blah." And everybody, as I said, cats, rainbows, happiness, everybody's happy with the domains. It's not true, it's difficult, and a lot of people is resistant out of it, so it's better to start little by little in the sense of dividing these packages for the use cases. So, one of the things that I have learned in my entire career is that you will have to identify how you can bring impact monetary to be able to get the buy- in top- down. So, what we did next in our team was create data apps for revenue and reducing costs. So, this project that I told you of cleaning up the projects, saving some money out of it. You have to build relationships with the right C- levels. Let's say CFO because he likes money. If you save money, he's going to advocate for you. A CPO or a COO, for example, for HR we automated some of the dashboards for them where they used to have an intern that manually will go to Excel and transcribe numbers from one system to Excel and then create some things. Very prone to errors, and now they get everything BI EPi and using Looker, for example. And they were, " Oh, wow, I love it. This is so easy that we have this intern or these people that were doing this manually to do something else, and we can have a nice dashboard for the meetings or for the presentations that we have." So, now the COO also is advocate of it, or the head of HR, for example, in this case. Because we need to understand the value of process as well. So, one of the two keys to get this top- down buy- in is get this revenue part, communicating with the CFO, or people who is more into finance, and then also understand the value of these processes and efficiency where they can free up some hours or time or people to do something else where they can produce more and doing as much automation as we can. So, if you find these allies to bring these things to the board, it's much easier for you to be selling this and to get this executive buy- in or this support from them. So, now, for example, we got it after much time working and trying to see how it works, and they are much happier. But things that you should do before going into this is understand where are we. Where are you in the organization of terms of resources? If we need to hire more or not, if we can optimize this, and if we are in track to deliver whatever we can with data because data teams are not only delivering governance and these kind of topics, they are delivering products for the business itself. So, yeah, that's something that I wanted to share as well.
Tim Gasper [00:42:21] I think that's a really important takeaway here because I think that it is easy to assume that maybe a different approach, or that a better approach is, " Hey, let's go talk to people at the ground floor, do more of a bottoms- up approach," and that certainly is one strategy. But another strategy is to really interface with the leaders of the company, understand the strategy of where things are going, and really align to making more money and saving money. And between those two things, you have to find a way to both have an impact and also tie to the key metrics of the organization if you want to build momentum.
Amy Raygada [00:43:04] Yeah, that's basically the low- hanging fruit. If you find the right project and you say, " Okay, I'm going to focus on this because it's going to bring savings, and this one is going to bring process- oriented stuff and efficiency," then the next project, you will have support. It will come here so you do one, two projects that bring value, then another two that might be improving things for you and your teams, and then another one that brings value, kind of iterative. If you iterate in this, you're always in this momentum as you said. They will be, " Okay, okay, okay, okay."
Juan Sequeda [00:43:46] I'm curious, I'm bringing up this comment here from our LinkedIn user who's actually our loyal listener Mark Kitson. How do you connect the OKRs, the business goals to all the data work that's happening in your experience? This is something that I see all the time is starting bottoms- up, the issue why that usually is not succeeding is because, from the bottom, they're not really connected to what are the business goals, the objectives of the organization. That's why you want to go start from the top because they get that inaudible down. In your experience, how are you managing connecting the very specific business goals, OKRs, and stuff, to the data strategy?
Amy Raygada [00:44:37] That's a really complicated answer, actually, because for example, in my organization, we don't have OKRs into the main organization, but more into the business units. And each business unit needs to work on their goals plus working on the data. And that's where I told you the governance comes into a second place in a way because they need to deliver whatever the business is asking for, right? And that's something that is managed by each data team. But in other organizations that I have worked on, usually, for example, if one of your OKRs is optimizing resources, or as I said, saving money, you can still attach to it and show some value and find the right use cases for doing this. Or for example, if it's more sales, then you can create a data product that will show you which funnels were more successful, for example, and how this can improve the sales in X amount of percentage by the marketing team knowing where to invest more. So, it's more about understanding the needs and see the opportunity. What can you do with this data that will bring value to them, and hopefully to... There are some of them that it's going to take time to see actually value. The data governance initiatives, you're not going to see a value added until years. But in the meantime, you need to do something that shows something. And basically, if you can go through these through sales, to cost efficiency, to processes, and just deliver little by little that is attached to the OKRs in general in companies, I think that's the best way where you can show some value while doing what you really want to do to improve the processes for the data teams. Because it's again, it's a Game of Thrones. It's basically doing things on the side like Cersei and move.
Juan Sequeda [00:46:39] That's the theme of this podcast episode, it's a Game of Thrones. Look, there's so much stuff I want to go through more and we want to hit our lightning round, but there's one topic I want to touch before we go to our lightning round, and this is the honest, no- BS rant here, and I want you to go inaudible. You've already touched about this. We all love to talk about the beautiful things like rainbows and all the unicorns and all that stuff, and we go to these conferences and we hear about this stuff all the time. But there's just so much blah, blah, blah, blah, blah at conferences. What is your call, your rant and stuff for people presenting at conferences, people at conferences listening to that stuff? How do we deal with all the blah, blah, blah bullshitty stuff that we're hearing? How are we all more honest, no- BS about everything here?
Amy Raygada [00:47:33] Gosh, I hate it. I was last year, actually last spring, in a Google event about data mesh, and there was these people from one big supermarket in Germany talking about data mesh and this and this and that. And the lady was talking about, " Oh, we have these domains and we have back- end engineers and this, and everybody's happy. And we have the catalog and this, and discovery, and metadata, and blah, blah, blah. And now it's so seamless and blah." But she's just talking about the final product, if you could say that. And then I was with one guy from another company next to me, and we looked to each other and I'm like, " I don't think that could be that easy because she was talking from a period of six months or something. I'm like what the hell did she do? I need to know. I need it in my life, how much is it? I buy it. And then the guy and I, we finished the thing, and we start to ask questions, and other people as well because they were also quite intrigued and how they made it happen that fast. And she was, " Oh, no, no, actually, this was already there for some years. I just arrived to this company a year ago." I was like, " Ah, okay." I was, " What were your pain points? How difficult was it for you and this and that?" And she started to get all these questions and she got super uncomfortable about it. And I think that's so wrong in so many levels because when I give speeches in different conferences, I speak about how hard it is to really get buy- ins, to convince people, to go and preach to get the doors closed in front of your face, and you have to go and knock maybe 10 to get one open. I always talk about those things and what worked and what doesn't work because I don't want people to go through the same things I did, and I want to hear from people the same so I can fix also my own mistakes and try something different and better. And you hear either sales pitch, either my company is the best in the world, we are doing the best data forever. That's why sometimes I get annoyed on conferences because our very little non- technical talks because technical usually tends to be better if you have hands- on and stuff like that because it's just a lot of BS everywhere. And at the end, you ended up wasting 45 minutes of your time or maybe an hour listening to blah, blah, blah, yadda, yadda, yadda, and nothing really that adds value. And when I speak, I want to add value to the people and to learn something, and move to the next level.
Juan Sequeda [00:50:09] So, what I'm taking away from this, and this is my interpretation. One, if you hear all the rainbows and unicorn stories, be freaking skeptical and go talk and figure out what the heck are they doing?
Amy Raygada [00:50:24] Yeah, I'm like, "Bye, I'm out."
Juan Sequeda [00:50:24] And then, actually, I think the call here is for people giving talks is like, actually, don't only talk about the positive things. Talk about the... Call it the negative things because that's how you know how things are real and that's when you'll connect with everybody.
Tim Gasper [00:50:37] Yeah, the hard stuff. What sucked, right?
Amy Raygada [00:50:40] Exactly. Yeah, that's why I like to mentee into my presentations, and I ask, " Who knows what is data mesh?" Or, " Who is confused about the topic?" For example, and then I get... I'm like, " Me too, so no worries, I was the same." Or, for example, " How long do you think digital transformation takes? Six months to one year, one year to two to three?" There is a lot of people who thinks it's maybe six months, some people between three to five, and it's crazy. And then you're like, " No, actually, it's going to take a lot, so don't get discouraged if in six months you don't get any results because it might not happen. It might take you years until you see something tangible.
Juan Sequeda [00:51:19] All right, this is a great call. I think it's for people who are sitting at talks and things. Be obviously very skeptical and go talk to people and just... I mean, honestly, we got to call people's BS on things. And yeah, if they feel comfortable, at the end of the day, that's how we all were, and hopefully, the next time somebody's like, " Well, I'm not going to do that talk again because I don't want to be called out on that stuff." And then also, for people who are giving talks, let's be real. Let's talk about what's working and what's not working, and let's empathize with each other. Wow, so many notes and stuff, and we got a bunch of great T- shirts we want to go build now.
Amy Raygada [00:51:58] You have to send me one at least.
Juan Sequeda [00:51:59] Going back to our original inaudible that should be one of our personal goals. Start our T- shirt store.
Amy Raygada [00:52:08] And maybe, yeah, the geek cocktail brand.
Juan Sequeda [00:52:12] Oh, yeah, yeah.
Tim Gasper [00:52:13] Yeah, the geek cocktails, and I want-
Juan Sequeda [00:52:15] I've been secretly adding more water to mine, so it's-
Tim Gasper [00:52:17] Oh, there you go, balance it out. I really would like to have a Game of Thrones T- shirt where it's got maybe the different kingdoms and all that stuff. We can work on something.
Amy Raygada [00:52:27] You can put the cloud and maybe some stakeholders here, and maybe data governance, and yeah.
Tim Gasper [00:52:30] Yeah, then there's the governance kingdom and they're scary. No.
Amy Raygada [00:52:36] The C- levels.
Tim Gasper [00:52:38] They're like the Iron Kingdom.
Juan Sequeda [00:52:39] Oh, please, oh, oh. I truly ask, I just think I'm going to call him out, one of my buddies we worked together, Stuart Carver. He's great friends and he's super funny. I tell him constantly he would be awesome having a stand- up comedy stuff like this. We're laughing here, I bet more people would laugh at this stuff, have a stand- up bit about Game of Thrones about data and stuff. Anyways, I digress. Let's go honest, no- BS lightning round. I'll kick it off first. If you are a bad data owner, should there be negative consequences? Should be reprimanded, fired?
Amy Raygada [00:53:19] Yeah, totally. Maybe not fired, but start with the process like, " Hey, dude, either you fix this or there will be consequences." Like I said, sometimes you need to escalate and that's how they understand, unfortunately.
Juan Sequeda [00:53:38] Oh, yeah.
Tim Gasper [00:53:39] Yeah. inaudible-
Juan Sequeda [00:53:39] inaudible lightning round, love it. That's it, perfect.
Tim Gasper [00:53:42] Yeah, and sometimes you need to escalate, that's a great... Nobody wants their boss to yell at them, right?
Juan Sequeda [00:53:49] inaudible.
Tim Gasper [00:53:49] All right, second lightning round question. Should all data have owners or just the most important data products?
Amy Raygada [00:53:58] I will say everybody. Why? Because it's not only about curated dataset that affects our data, it's back- end data. And maybe you don't want to have one person, but you should have a team at least, and this team needs to get certifications. Because people comes and goes, but if there is a service account with an email, you know this team in big companies, for example, you know that this team owns these datasets, and you can ask, " Okay, who's owning this API?" And then you can get someone because, otherwise, between a huge company how would you know who owns that? And it's longer to get this debugged and fixed. So, it's also about efficiency again.
Tim Gasper [00:54:42] Yeah, that's actually a good comment you made. What team gets the notifications? That's an interesting question to ask. If the tree falls in the forest, somebody should know. Who's supposed to know?
Amy Raygada [00:54:55] Exactly. Yeah, because at some point when I was doing my framework, my boss wanted to have one owner per dataset or per project. But we don't have right now connected in the system from HR with all the people, and we cannot connect when someone leaves or not. And what happen if Juan leaves the company, and then I'm like, "Oh, I'm going to look for Juan," and he's inactive for you don't know how many months. How would you know who to go? So, it's better to have it to a team level, and that's what I decided at the end. I told him sorry, but I think it's better on a team level because you know who to contact, and someone will know at that point, " Oh, yeah, no it's Tim the one doing this."
Juan Sequeda [00:55:35] Yeah, this is a good call right now. You can't govern the cake data product unless someone's also governing the flour, the eggs, and the butter.
Amy Raygada [00:55:45] Exactly. That's a good T- shirt as well.
Juan Sequeda [00:55:48] inaudible. All right, next question. This is an interesting one. So, do we need data product managers, or can we do the data product management and governance happen without them?
Amy Raygada [00:56:00] No, you need a data product manager, and actually someone who is really willing to learn. I can give you a little anecdote here about the data product managers from back end that were involved as domain owners and how they were reluctant to do this because they thought they needed to learn Python or SQL or Tableau or something like that. And it's not about that, but they bring actually the order into the chaos in the sense of let's organize the backlog, let's see what needs to be done next, what are the requirements? And they are the ones taking these templates that I told you to the businesspeople to fill that out together, to put this order here. They are the one that owns this domain and these datasets as well, someway somehow, so people can consume. And yes, you can have a data engineer owner, but tell me, if you are a businessperson, you go to speak to a data engineer, we go to the same thing. It's like speaking Chinese and Taiwanese, I don't know. Similar, but they just don't get each other, right?
Tim Gasper [00:57:04] Right.
Amy Raygada [00:57:04] So, you need the person in the middle. That's why I told you I moved to this because I wanted to be this translator because I understand very good the technical side and also the business. As a product manager, you don't need to be super technical, but at least if you... I'm actually offering as well some trainings for data product managers that are not into data, so they understand at least the basics. They don't need to be developers, but at least to understand what is a data engineer because I got these questions, " What is a data engineer? What is a data analyst? Is it the same?" And I'm like, "No, google it." No, I'm joking. I was like, " No, okay, let's..." And I explained it of course. But you need to have this patience.
Tim Gasper [00:57:40] You need to use translators.
Amy Raygada [00:57:42] Exactly, and you need to be patient and teach them, and then they will see it. Okay, it's as easy as being a product owner in a regular software setup. You just need to understand why do you have these people and what they can do for you. But you need them to put this order.
Tim Gasper [00:58:01] Well, stated. All right, final lightning round question. You talked about that very interesting approach around using GitHub for data product, data contracts, and the notifications. That's fascinating to me. Do you think that's something that all companies could implement?
Amy Raygada [00:58:20] I mean they can. Actually, with Andrew, we had this discussion, and with JPG also from the data contracts and standards, and they like this approach as well because it's quite simple. Companies cannot afford expensive software or stuff like that to have data contracts implemented, or maybe they don't have the technical capabilities to do it. This kind of implementation is easy and seamless, and it's something that worked because one of the things most important for me in a data contract, you can have it in Excel, I have seen that. For me, that's a horror story. I have seen that in Notion as well, I feel like I'm dying as well. But you need something to enforce it, and what is the simple way to enforce it? A pull request.
Juan Sequeda [00:59:03] I like that. I like the pull request. All right, so much stuff. Tim, kick us off. Take us away with takeaways, Tim.
Tim Gasper [00:59:12] Takeaways. Well, we started off with honest, no- BS. How important is the notion of ownership in data? And you said ownership is the pillar, it's the foundation. If you want to do data governance and data quality, you need to have ownership. And you said think about that Spider- Man meme where all the Spider- Mans are pointing to each other. That's what happens if you don't have clarity around ownership. And you mentioned about some of the templates and the frameworks that you have. You said there is technical ownership, and there's business ownership. And a lot of times you need both of those in order to have the right folks that really you know who's going to get the notifications, which I think was a good litmus test I think that came up a little later in our conversation today. And you had a really great quote, which you said it's not about silos, ownership's not a bad thing. It's not like your boyfriend dataset that you don't want to give to anyone, it's not a jealousy thing. This is really about collaboration, and ownership creates better collaboration, not worse collaboration. We talked a little bit about contracts, and the importance of data contracts in data governance as well as in ownership, and even using things like GitHub as a way to enforce that. Obviously, there's probably a lot of different ways you can enforce data contracts. It's a topic that's come up a lot on our show, but that can be really, really important to create dependability around data products. And we talked about usage of data. It's important to keep track of the usage and understand how data's being used by different stakeholders across the organization. And it's going to be something that allows you to really save money and minimize duplicate data and have better life cycle around data. And as a queue up to the next topic that I think Juan's going to talk about, you talked about the importance of understanding the usage of data, and understanding how data products are being leveraged allows you to really create connections across the organization. And puts you in a position to build trust. You really need to build that trust before you can go in and start doing your Game of Thrones games and the overall politics. So, Juan, over to you, what are your takeaways?
Juan Sequeda [01:01:10] So, I am super happy we talked about this whole issue, the politics of data and knowledge. And I like how you said, it's like chess. You need to remove pieces so you can move forward. You need to show you care and that you're listening, so empathy is really part of the whole politics. Don't be the, this is your new process and you need to follow because that's not going to work because people are not going to go follow, they don't like that. Tell people, I want to understand so I can contribute and help, and ask how are you doing that because maybe that's something I can do better too, right? So, it's all about gaining trust around this and also know that the biggest complainer about the data, maybe that's an opportunity because they should take that ownership, and that's politics right there. If it's someone on the business side, often they can also explain better why things aren't working and what's wrong, and they know the business and what needs to actually happen. And then they could actually start advocating for you, so it's that first big win that you want. We should also be talking about the failure stories and things like, " Hey, I'm going to build that super shiny data architecture and they will come." Wrong, that really doesn't happen. There's a mentality change that needs to occur with the stakeholders and the business stakeholders. Don't just focus on the technology and infrastructure. Start with the people and the process and build that inclusion and that process. And another thing that doesn't work that well is just bottoms- up to get to the right use case. That's something that we talked about in the failure stories. So, how do we get to that right use cases? Man, you really need to make those right relationship with the C- level. Talk to the CFO, understand how do you connect this directly to money to create apps for revenue, for reducing cost, and where are you in the org with respect to all these resources so you know if you're on track to deliver what you're promising? At the end of the day, the being thing here is think about this is a Game of Thrones. And then just the final comment here about conferences that we talked about. Be careful with all the BS in conferences. You hear those stories, all those rainbows and unicorn stories, that may be all BS. People who are giving those talks, don't waste people's time. Be real about it, and how do you know you're being real? Because you can talk about the hard stuff. About what worked, what didn't work, and why it didn't work, and how can you get that work? You really want to empathize with the audience there. Amy, how did we do? Anything we missed?
Amy Raygada [01:03:24] No, everything. Thank you so much for having me.
Juan Sequeda [01:03:27] inaudible. To wrap us up quickly here, what's your advice, who should we invite next, and what resources do you follow?
Amy Raygada [01:03:34] Yeah, well, actually, my advice here is to be patient and to listen. Basically, it's parenting as I said before. It's basically just listen and try to understand what the people needs, and then solve the issues that you see as a hanging fruit. Maybe for your next episode, you can invite... Well, you know the hype of AI, and everybody needs AI today and ML. I've been on that boat as well, but actually, you can maybe invite Ignas Mociunas. He's also a data product manager. He used to work with me at some point, and he also is into this AI topic, so it might be an interesting discussion if you're into this topic. And resources that I follow, well, I follow a lot of influencers on LinkedIn, but I had found already... Medium I don't do this anymore because that's also a lot of unicorns. But if you're interested, for example, in data mesh, I can recommend you I'm part of this community, Data Mesh Learning. Actually, they are quite good. I'm part of that group as well. We speak with no filters. We try to be real with all the topics. I have a monthly round table, next week is the round table about data contracts, actually, so it's free for everyone. We talk really open about all the struggles and stuff every month with different topics. And also, if you follow JPG for data contracts and Andrew Jones, that's a great resource as well. And Dr. Alexander Borek as well, he has also good inaudible for data governance, it's quite good as well. I recommend those because they also speak from all the sides of good, the bad, and the ugly.
Juan Sequeda [01:05:30] I love how you really like keeping it real, keeping it honest, no BS, so thank you for that. Amy, this was a fantastic conversation, thank you so much. Just quick, next week Tim and I are going to be live from Gartner Data& Analytics in Orlando. We're actually going to do things a little bit earlier like at 10:00 AM on Wednesday. 10:00 AM Eastern Time I think we're going to go do this or 11: 00. We'll announce it. We're actually going to have Cris Hadjez, who's the senior director of data governance at Norwegian Cruise Line Holdings, so that'll be an interesting conversation to know how data governance and all these cruise... How they manage the data on cruises, on ships, and bring that back to land. That'll be next. If you're at Gartner, Tim and I will be there, so please reach out to us.
Tim Gasper [01:06:10] Come find us.
Juan Sequeda [01:06:11] We always love talking to folks who listen to us. And with that, Amy, thank you so much. Thank you, thank you, thank you. As always, Data. world, thank you for letting us do this every Wednesday, and with that, cheers.
Amy Raygada [01:06:22] Thank you, guys, and cheers. My pleasure.
Special guests
