Season 3 Finale

This is catalog and Cocktails world.

Alarm 31, welcome. It's time for catalog and Cocktails presented by datadot world that you're honest. No BS, non-sales. A conversation about Enterprise data management with tasty beverage is in hand. I needed some gas for a long time, Day dinner and Prada guy. I dated a Trojan and buy one cicada. Hey, Tim. I'm one of principal said something about world and we made it, we made it to the season finale season 3. It is Wednesday, middle of the week, end of the day and end of our season and it is time to celebrate the takeaways to take over to the takeaways, we got the bubbly. Yeah, celebrate not exactly a cocktail. That's okay, we're breaking. The rules cocktails. Believe you should switch it to cocktails and beer and yeah. Are you favorite beverage data and drinks?

Anyway, so what are we drinking? What are we toasting for? What? We're having some Bubbles and thank you to all of you. Cheers to you all for for watching, for tagging along, for jumping into the comments. I don't know if y'all know. But when you comment on YouTube and Facebook and Twitter and stuff like that, it shows up lies here, the next season, he keep on being a part of the conversation, keep on joining us cuz we love you that your what make the show possible going back? 95 episodes today, we started over two years ago, was it May 15th or something two years ago, did not think they would get here and Ashley, this is just the beginning. Like, we're going to take a break for for the summer because we are planning out what next season's going to look like. We have a lot of great guess coming up. What are the topics were going to be hitting you up the road so a lot more stuff to come. Let's go.

Here's to everyone, everybody's listen to us because it's all thanks to you.

Sound very excited today to go through a recap of what happened this season. What were the key big takeaways? You got? Six big themes. They want to make sure to cover with you all. Can't wait. So take it away. Let's Jump Right In.

Number what you think's, right? So first big topic, the modern-day to stack. It was a big theme throughout our season, sort of what is the modern State estate tax. How do we take advantage of it? What are the pieces of it? How does it bring value to our organizations into us as individuals? And I'll hit a few different things here. So first of all, Sarah Catanzaro amplify Partners talked about sort of this pendulum and I T that we've seen over history where there is things go more to the infrastructure side and then they go more to the analytics. I have an infrastructure analyst in the pendulum swings back and forth and in recent years you know go back 5 6 years there was a swing toward analytics and to be high. And so you saw like looker and heard of this next generation of the I too was going to come out in the pendulum really small in that direction and you'll see companies like thoughtspot back on the thing and that pendulum is swinging back in the other direction now.

And she talked about how very recently companies have been focusing a lot on things, like the data models, the documentation run to do the monitoring of the date of the start of the observability and things like that. And we're maybe previously there was a lot of focus on analytics, we realize that like, yeah, analysts are good but if you know, if you're building a house on a foundation of garbage that's going to be a problem, right? So you got to get your data in order and there's a strong realization of that in the market. We've lost sight of what we do with the data and we got to make sure we do that. Well, so that was really valuable there around thinking about why we been talkin about modern-day to stack. Why is it bigger than just like a fancy? Emily Hawkins data, engineering manager of data platform, over a drizly, I joined us and she talks about kind of what she thought were the core components of the modern-day to stack. She said, of course, DBT get that in there to manage your Transformations and that brings not just a code based approach but a version Don't approach to how you manage how your data goes from.

Is raw input into the stuff that's going to be used by those different analytics. Tools. Five, Tran write a modern integration to him snowflake a modern Warehouse Liquor, a modern business intelligence tool Daxter orchestration, right? And then all of that, which is interesting to something that could be implemented in two years and then in her experiences, right? So one of the benefits of monetary stack implementing it quickly. No BS, dude from materialize, co-founder and CEO. He talked about the, the MDS, the modernist at Cora's. What about how you need for the reporting? You want to be able to take action on decisions. He focused a lot around actions, the important and sort of the importance of sort of like data is coming into your organization in the modern day today.

Actually needs to orient itself as much around actually real time as possible. How do we take actions in real-time? How do we personalize and do other sorts of things in real time? That can be a big difference maker and that's where a lot of dust. Especially new parts of the modern-day stacked. Like these streaming data warehouses, like materialized are focused on you talked about how streaming can really be a couple of different things. It's not just sort of the incremental stuff that's coming in but it also can be the Deltas assertive. How things are changing and comparing changes over time. So it was good too. Kind of curious of a different perspective on what streaming can be. Any talked about how I should have just extracting load that's going to start to become more. More real-time like the Paradigm in the past has been much more around sort of like statues or fast. Just need a DDT is a little bit more of a bat faced approach. Usually, when you're applying it he's talking about how really we're going to get more and more into streaming architecture. That's going to be a big part. And one of the key use cases, he talked about

I was notification anytime you need to notify something alert. A customer needs to get a coupon in real-time. Anything that's notification oriented is going to be heavily built on this idea streaming.

Bob muglia came on and talk to us and some of what he talked about as well oriented around the modern-day to stack. He's the former CEO of snowflake entrepreneur investor and just a maven of the day to stay for a long time. And he said that the modern-day to staff needs to deliver analytics on SAS. It needs to leverage the public Cloud because that's what's going to give you scale and cost in the modern-day. The staff is going to give you the bass on sequel on us too. Well because that's what's going to be the sort of the lingua, Franca of the modeling industry the gist of semantics with in your stack, right? And and the. Essentially are the core tenets of modern-day to stack. Are those those three things? We talked a little bit about even though so I asked who I was for the core of the modern-day to stack that, that's not to say that SQL is perfect. In fact, it's far from it, SQL is not God's gift to language. He said and it's but it's not going to go away and

Maybe even you look at the English language. Write the English language is not a perfect language, but you know what, it's very popular. It's using a lot of ways and it's probably not going away anytime soon, right? So random random for their, what else? Modern-day to stack. Not only the CEO and co-founder of O Cara said that the modern-day to stack has actually an issue. So we talked about the different components of it. We talked about it for what it is and what it represents from from Bob Nan said that this issue round policy like policy is not unified today across the modern-day to stack and policy being like, who should access what, and you know, how do you manage the security around these different things policy is very cross-cutting, but what tool is responsible for policy? It's kind of unclear, right? And that's why over do care. As an example of the two Motrin. Trying to sell some things around policy.

Chad Sanderson. Join the show is the head of product data platform over Convoy and he talked about how the modern-day to stack is valuable, right in and he's implemented a sort of a version of the modernist act within his own company solving real problems. But he said that the way that a lot of people in the space in the database to talk about the modern-day to stack is a little bit disjointed, kind of doing things a little bit out of order. When we talked about the latest I can we try to implement it just if you follow us, if it was a marketing, a thing. And he said that right now, if you just kind of like getting all the data and sticking it in the lake and then just trying to use her to stay like Alice democratize that and let's just Transformer transform it. However, it is that you're building the modern-day to stack on a foundation that is like a swamp. It's like liquid in your end of things you do on top of it. They slowly sink into it, and they're lopsided, and things like that. If you're moving too quickly, the questions about like what you transform and he's going to transfer, go out, you know.

Do the transformation who owns the transformation. Those things don't get an answer properly and things get really messy. So you really do have to approach minority to stack with us. An idea of, how do we do a sound Foundation? How do we approach this in a Smart Way? Sarah krasnick, the founder of version of all that I owe talk to us about her own, modern-day to stack and she actually talked about for the Enterprise, modern-day to stack and Amsterdam honored. As her perspective actually was a little different than some of the other folks that we talked to. She talked about Ashley bed in Enterprise, modern-day to stack rather than taking the vendor or based approach to do, right, and snowflake by panel, that kind of stuff that actually open-source can make a lot of sense in the Enterprise landscape and that every new vendor, you take on that you bring on a going to be a lot of time and especially in terms of financial investment and that opens first actually can be a very viable solution. If you were larger Enterprises and you can throw the resources at it, that requires to sort of up to

Sadness rid of open open solution and that for startups, right now, if you want to move way quicker, open source isn't really as viable for you, but still something like, d d team. Actually, make a lot of sense for you. She did mention know that starting with d, b, TD may be hard because because you have to learn this framework, and it kind of adds complexity of tacos. That's too. Well maybe just starting with us and behind your, a transformation language. That is a SKU. Well could be a good start on your way to adopting BB&T. So she also talked about you know he should catalog your knowledge. Every bi dashboard starts with a question you don't always need to build things of scale. You can just start from start of data and and do what works right do it works. First before you have to scale it up if you scale a happy system. That's a problem though. So you do have to make sure that you think about resiliency that if you need to claim data bankruptcy that you do that, but

Ideally, you don't have to do that and you should ahead of time to think about what is databank ruptcy for you. So that way you can avoid it instead of having to go bankrupt. Actually think about what if your team right now to just cut in half, what's going to happen leadership? You think about that? Now think about that chaos Monkey kind of approach. Like what happens if a chaos Sonic in? What if your hat hacker team went away? What if your database went out of the right thing about those types of things? And she talked about things like the metrics player as well but we'll talk about that a little bit more. You not coming section Luke, slott Wednesday, the VP of data analytics at Elijah's said that an important aspect of the Monday to stack. That's a little under talked about is the catalog right? Where are you keeping track of the discovery and the governance around sort of your your stack? You know, we've got catalog up here in the oven in Vale podcast here so we can catalyze the prequel and as part of that, you want to have a canonical model both of the semantics as well.

The date of the physical data itself to help you really drive is from the slicing and dicing that different analyst need to do for various business questions. Then he plugged the data, virtualization is also another thing that isn't often talked about in the context of the modern-day to stack, but actually can be a really effective way to get access to lots of different day. Bring it together to solve distance questions without having to do. Lots and lots and lots of more complicated.

Enter round out our conversation around monterde to stack. We had Sunday of Mohan who's the owner and principal of Saint James Mo and salt Tang, talk with us. Sunday was great. It's always a pleasure chatting with sanjiv and he really talked about the fact that in the modern-day to stack and in the store data space, in general, he's got this micro-segmentation going on or like, every layer of the staff is getting sliced in half and a quarter is a Nathan. It's like we're, we're not observability where, you know, alerting for data. And we're just waiting for data from your Warehouse in my so things are getting really sliced up and that's making it confusing for the market with an N. Even with it, sliced up there are still like 20 vendors and every box claiming that there should have been that box so that I can be challenging to navigate right at work. I G where it's just like with, like all the logos and he said that ultimately customers just care about future-proofing, right?

Not just following where the through the technology winds are blowing and they care about the business problems. The tranks all. So it's important that we don't forget about that. Like we get excited about the technology and it's exciting. There's exciting stuff happening in the space, But ultimately we're trying to solve people's problems and help them future-proof. He said that the core modern-day to stack for him was the building blocks of Integrations, like five, Tran transformation mytvt, and the warehouse like Snowflake. And he talked about two approaches that you can go towards making that modern, do you stack best-of-breed and then integrated together or an integrated solution where maybe they already have all those different pieces together that might be more expensive but then you don't have to worry about actually the integration of that and then finally he said that the metadata is the glue. He said that matter data should really be the leader that drives our security or privacy your access your transformation. Meditative should be the driver that actually helps for things like moving to the cloud and doing cost reduction.

Like, one of the things that he notices that as people are moving to the cloud, some companies are realizing that the I thought they were going to save money by moving to the cloud. But actually they're spending more money by moving to the cloud and then they were on their way. So that's something we're metadata can actually help you saw that. And then finally, he said that using a graph for your metadata can be a very effective way to analyze the relationship centrality. Those different aspects around your metadata and that in general having graph-based metadata. As part of your staff is very important as part of the body, do you stack? So that's a little bit, text heavy, right? But I think it's interesting to hear about the technology side of the equation. That was a long one. I think that was long on purpose because it was the most popular one that we had. I think the second popular topic there is about semantics, layers metric Slayers knowledge layers. So it's going to happen to this.

Black. That should be a thing because this is uses metric system in different places. And those men in the pros at those metrics are defined and so many different places. That's what we need to go to keep track about trying to keep track of all these things. We ask the questions that are, that are small deviation. So it's like a, if you ask, what is a user go, ask user to different people and figure out how their deviated, what is revenue on go, ask how they deviated from it, right? If ended, what is the most important thing for your business? For example, is Airbnb for them. It's booking. So, how do they Define bookies? If it's over? It's right. How do you find these things? So I think that was an important aspect to consider their Lopez. She's at, she has issues of principle. I think it impolite visors. Data chick on Twitter. @M start. We talked a lot about data modeling. One of these things was simple to go model. It's really simple but really probably just half of the stuff is really simple. We really need to be able to understand a trade also. What? We're going to go model and Justin exception, it's not really an exceptional.

Right? Because you got a bunch of bunch of these different except it's around there. One man's look up table is another man's universe or thinking about these things, right? We always talk about, hey, do you have one of those values me and so forth, right? That was something I was a big, big, big top of it. Comes up. What are your best practices? Consider your trade-offs carefully. That means that we really need to start documenting what these trade-offs are and what are the decisions were the the semantics, the meanings behind that. That's why I can't afford to keep track of all these things. Let's check the shape of the data, right? So, if if, if we look at the ages of people, we have a bunch of people who are 200 years old. Well, that's a problem right there. Let's get the business folks and the Sons of Anarchy involved in the beginning to understand what those trade-offs are. And I think also we need to start thinking about modeling with with resilience in mine and how that's going to happen. Every good design decision looks at clock

Benefit and risk during 4 take away, Bob. He's bringing up against semantic players, and now, it's so Samantha. Clear for him is how you model the business. And I think my favorite destinations crisp and clear and concise concise a metric, is a function, applied over the relationships between business entities. The metric is a function over a semantic layer. That semantic layer is a relationship between business entities and making these business models executable. That's what you're able to go to. In an hour with the knowledge graph. By the way I'm Bob, I love his ability to simplify these Concepts and make it so clear and I love that Bob coming from this deeply, relational Warehouse, or my background is saying that the future is knowledge Grass at Symantec layers. That's a lot of validation, that this is an important thing that we need to look at Chad Sanderson.

Topics that we had that brilliant episode. Where does this logic? And he says, there needs to be a knowledge layers semantic layer that needs to be a very important thing is the pass, right? It's Enterprise ERD. Diagram seems we have we're doing this in the past but those usually were pretty pictures or PDF files or whatever. Like the next day when you go to today is to make that more excusable, going back to Bob's point is that these business models are going to beat make executable. And then the issue today is that these ERD diagrams Wright, architect this up front. This is an art that has been lost and why is it lost? Because the start of model is that we don't have time to argue that the data, right? We need to move very, very fashionable, build Services. When you make money fast and that should be your focus.

But if you look at the like the band companies, right? They started that way right at Builder apps and they made a bunch of money that leaked. And they said, we're going to go clean this up later. So you have to reverse-engineer the business Concepts that were happening with all this it with all the work that was happening. But guess what? Most companies, the orange tech companies like the fangs. So you should not be following the product, the prophecy that they did. You should be able to go spend time and go do that with that transformation, modeling up front. So why did it work before me? We've been doing all these semantic Marley's before and it was speeding and slow. Like we need to figure out a ways, how to do this fast and interrater, we need to be agile about it. We can do Version Control, we need to use cloud, we need to make sure we're not boiling. The ocean thing is that people take the path of least resistance, so we really need to make this easy. So it's still something that I think there's an open field here to be done. I think I'm really looking for what's coming up next

Another awesome conversation with Nick handle, who's the CEO of transform? His definition of a semantic layer, very specific, your mapping tables to the classes. The main kaun by quarters and products, every single row of that table or individual instances of that class and all those columns are the attribute. Basically, you have those boots wide tables that table need to concentrate. Their now, what happens today is that the abstraction, is that such a lower-level? So we don't really understand this. So the semantic layers and abstraction. So it's at the, at the layer of what the end-user no metrics today, how they're being Define are just a bunch of SQL queries, that are just super complicated over that lower-level abstraction, but with the semantic layer, right? We're now having those metrics those calculations over at a higher level of structure. I always say, computer science. At the end of the day, is this understanding, what level of abstraction do you want to go work on? And we need to have that I've dated two and and why

Why are why are we looking at metrics right now? Is because they, we now have a bunch of compute, like we can go do this and have easy ways of scheduling jobs. So again I'm going back to Serious point where we're going to this pendulum. I think we've gone so much, we've got to the pendulum of focus and infrastructure and now we're going back and focusing on the Samantha from the metrics there. So I think if we're going to be more data-driven, we need to be able to provide this data in that high-level abstraction. Today will be celtel. Tgtgtgtg going on, right? We agreed that his self service without a semantic layer, is truly not possible. So users need to be able to go interact with the semantic layer. My personal dream is like, any, if you have a semantic layer, your metric screws, should Lily be really small, Snippets of queries, SQL, queries, or whatever. That's what's going to be a success right there. And hey,

I need to get people in the room to be able to understand what the stuff being. So it is that people process. And why would I would have talked about the knowledge first approached and guess what catalogs? Yes, they need to be able to manage and count all those metrics and live at the knowledge graph conference with him. He is a share of the knowledge graph conference as we talked about, obviously, don't have a problem, cannot be solved easily with sequels that you need to have a much more expressive language of its price of their mommy's thing. That's were Knowledge, Graph come in. And ultimately, the goal is to automate decision-making and the one of the best ways to go start with, this is look at metadata management. I think this is something that we go say that. That's why Bob was saying your first kind of application on an autograph. Should be of governance of metadata management because the modeling around that is kind of a little bit more simpler than just the entire world over there. Another aspect could be also MDM which probably she was calling.

City management. We always talk about the word, ontology, and we've got it for so long, but afraid of the word, ontology. But hate, let's not be afraid of that right now because that's how we did find the semantics of meaning and honestly, a Knowledge Graph without ontology is completely naked and actually the knowledge graph conference. The keynote was Bob moogly, and I think if anybody is kind of disgusting about knowledge graph, just go send them are podcast episode. The keynote address of the key know that they have their is the CEO of the former CEO of snowflake. A relational data where data warehouse company is telling the world that the next wave is knowledge and Knowledge Graph. Like this is a big, big thing. You should be paying attention to hear is that, hey, the future. The future should not be about asking an agent of the soft rated for information. It should be about empowering the agent to make decisions for you. So that was an interesting part of the conversation.

It should be something where it does it for you. And it's like, that's because you need to understand what the objective. What is the objective function of your company? You want to go maximize that. So this is really thinking, kind of Exile. I really want this conversation and then our last conversation about cut it on knowledge and semantics and modeling was the one that we just had dinner last week with our guests, should to block these people have changed their opinions about data models. Right? I think with the modern-day to stack today and would like to send ocean of 1/2 speed Feast, let's go to fast. Things we go do all this ad. Hoc modeling all these, we do this just for this particular use case and that's a problem because we start having this e l t t t t t is Nick was saying all the time. So we really need to start thinking about is understanding, what does use cases are and again understanding what the main concepts are. So how do we get started?

Fergus has had a great approach the same. You know what that at her query that is being used for the most popular dashboard and try to understand it reverse-engineer it, I guess what again something that's coming up. It's like, we need to have the right people in the room. We need to have the right business, stakeholders and let's go talk about what the customer thing is. And if you tie them in the room, and let's go to figure out what this business work for the business process. Go to draw the bubbles in lines on the Whiteboard, and see what people think about these stuff. And I think this is another topic that has come up. Recently is business literacy, I think it's a, what's just once you get the crew would be the date in January and elections. You near like they're talking to those and users. And those though, the subject, matter experts, there actually understanding more the business. So we need more of this business letter business data literacy. It's exactly. And

Something I will never forget is when we are talking about hate. What are the skills that we need? Not talking about? I was expecting the what are the technical skills and it end up being and if I don't want to, I'm going to stop myself saying soft skills because I don't, I don't like that where this is like this is like the foundation of the skills that we need to have in this industry. We need to be empathetic way to be curious. I mean, those are the next level is going to be focusing on to take Noah to the next level. Do you need to be able to get so excited to get a kick out of how I help the business actually solved that problem, which I know how we're making money. You're saving money. That I was empathetic to be able to go understand it understanding mean, semantics. So, yeah, I like how he is a discovery motion to really understand what's going on and at the end, what are the problems that we talked about? And I need to get the t-shirts

Ocean don't bore the ocean. We need to start with the use cases. We say this over and over again, but we don't do this. Yes, Bottoms Up approach to understand. What is what are those are the ad-hoc? We're going to try to take to reverse-engineer this and we was just off work. It's fine. It's not going to be perfect. When we go in her eighth and follow that dry principal don't repeat yourself. All right. Number three leadership. Damn, that's pretty hot and are moving out of the next leadership. So the third big scene that came up here was around leadership in data and and more broadly. And going back to the store, a cut and sorrow. She said that it is not enough to just put data into the boardroom, right? You need to put data people in the boardroom. You need to put the data analyst at

Who built the spreadsheet and built the slides in the room in the boardroom so that when people are like, oh, that's a good question. I wonder what you mean by that, all the slide is there. They're going to be able to weigh in and help you, all right? And maybe even answer the question and maybe they understand a little better than the people in your ship. Is, how do we bring data? Not just be more data-driven and will be more did a little people write. So really important, hire more sea level data leader. She kind of said that like, it doesn't have to stop at the chief data officer. Let's Emily Hawkins going back to her. She said, you know, she was really thinking about like a lot. How, how do I delete a ship has changed and the importance of data leadership, being able to trust the data team and building that trust. And that trust actually is a really cute factor in this room.

The success of data, in an organization, how do you Foster trust? How do you make sure that data leaders, create a culture of trust in the data? And actually, their VP went to CEO either? Will she kind of talked about the story about how, like, why like, why are we moving this new paradigm of data? And it's still, these leaders are getting elevated up. In the organization who want to use these new date approaches, you want to use these new monitors. Cools and so that sort of the new generation of leadership. The next people moving to, this leadership positions are changing the way they were thinking. And in her case, there was this story, she gave about this VP moving to the CEO position and making the changes in that way Dora, the senior director of data strategy in architecture or red Striker talked about the importance of leaders in data, setting the foundation. She said a my building a ranch or a three-story home, right? So thinking about like depending on

What it is that you're trying to do here with your data with your organization around it? Are you going to take a different approach and you really want to think about, like, like what am I doing? What am I trying to do here? What is the foundation of building? And what is the story that you're trying to tell her rounded it up?

You want to understand the, the what the why? And the how and what's in it for you? Around your data and leader is player really key role and asking those questions, pretty ultra-rare. Those questions are going to be asked and overall communication is key. Really pressed on the importance of communication between leaders from L to everyone in the organization & Beyond going back to Luke. Linsky, he talked about how it's so important for data leaders to take a Socratic approach, you need to challenge those in your organization to use data rights that person who's been in the organization for 20 years and and still doesn't know how to use Excel. You need to you need to get them going like hey man, have you someday they'll learn today though? You got to do it and instead an expectation within the organization that folks should get their hands dirty with data and you as the leader should get Hands-On, did he talked about how he himself?

His hands dirty with the bi tools, and with the data itself, he's in there he's involved. And that's the only way you're going to demonstrate that you're an expert and that you're setting the the tone you're craving that culture. And so that you teach people to fish and people aren't going to learn how to fish at that expectation. Isn't there? She talked about how it exacts need to dig deeper into the Y themselves before going to the right. So rather than just immediately saying like hey whatever, south of mind, my questions, like they did it in the day to see him. Of course is like

Work on that, right? Like think critically about those questions. Go a couple klicks before you actually go to that data team because you'll learn a lot yourself and you will actually bring better questions to the data team. There going to be more efficient there. Going to be more effective. You talked about how the Chief Financial Officer is is is usually very data-driven. They need to be very data-driven, but it shouldn't just stop with them. It should be the rest of the leadership as well. It's getting involved in that data and that the CFO and all the leaders in the organization should get training, they should get training like analyst, do and learn how to use these tools and learn how to ask these questions, learn about things like bias, learn about things like statistics, right? So, and he asked, and he said, you need to ask. Why, why why? Why? So, three wise, five wise? How many wise asked me, why? Right? So I think that's a really important takeaway from Luke.

May I do have one more thing is Luke, Luke has a fantastic leader who in and at the snowflake Summit couple weeks ago. You was talking about how from the leadership, the CEO. The equality is a priority and they tied the equality to like 25% of the bonuses of everybody and they were making, they were showing all these arguments that if I have the right quality is what how much when I'm going to make it has back playing which one you can lose like that, is tremendous leadership. I really really applied the folks, Luke and the folks at 4. Wat, is there an example of what is daily, meditator and governance to business results? Literally, what makes money would save money? That's an example, Steve Perry, director of data, and analytics at genius, Sports joined us and we talked about Pastor syndrome and how you know, the data Community is so fast changing and in some cases you know people can move up into leadership positions very quickly.

it's easy to wonder, like

people are looking to me for the answers, but the answers are not always clear. And how do you handle that? And how do you be smart about and how do you always be learning right in the importance of leaders to both always be learning and be recognizing of that and also embraced imposter syndrome. Like it's okay to not know something, it's okay to admit it. And it's okay to not bully ocean, right? Because sometimes we can get a little over and vicious. Our eyes can be bigger than our stomach unless we have a little bit of humility, he mentioned the importance of leaders to pick their battles. Is it worth buying on that Hill? Can you get the internal support? Can you get these opinions from others in the organization, who may think a little bit differently, right? And even if you're wrong, you know, if you are honest and you're open people want to help you write. They empathize with people who are trying and are trying to solve the problem. You talked about have come

Stations with smaller communities. Find those smaller communities within your company, find them in the broader Community, right? Whether it's in Social, whether it's meet up, some things like that, talked about the importance of working finding Piers, how much value and information you can get from your peers and the value of one-on-ones. I know that's something that that one, and I talked about a lot as well, as the value of one-on-ones. Having one-on-ones with not just people within your company but across the industry, smart people getting together and talk and write any talked about the importance of company culture. If you are not part of a small community, find that Community allow people to be open, create that culture of openness and do things like crate guilds in your organization and he said the data expert is like a life expert, so they don't exist.

That's another good quote right there. Nobody's nobody's truly quite expert but we try CEO and co-founder of world joined us because we were supposed to join us, right? So and it finally did from David. World and he talked about Sir this challenging economic environment and how data can be really, really important to get yourself out of that and the way that you're going to be effective when things get challenging by leaning. Into your data, is really understanding your customers really understanding the users, knowing their pain. Don't just get stuck in your bubble or the jargon, right? And he gave a really great and it go to really good story around, Sam Walton.

You are really really focus on knowing the business knowing the metrics, knowing your customers and avoiding the use of bulshit metrics and information and how that can be a big difference maker when you're building your company and especially challenge navigating these challenging times companies that focus on being data-driven of the ones that can weather pain can create Focus, right? And so, if you mean into the challenge, a lot of times that will create a both the answers, as well as the cultural resiliency, you will come out. The other side stronger when everything is going great. Sometimes things can get kind of sloppy, right? And so, even when things are going great, sometimes it's good. We talked earlier about. So the chaos monkey or something like that, you got to really be thinking about that, right. Do you truly know your facts, right? Do you really know how many customers you have? Do you really know the pain that you're end users are having maybe you should go hang out at the actual retail location.

I just as Leaders, but also just estate professionals, right? We can learn from our stakeholders and then he also talked about public benefit Corporation and which stated our world happens to be one in the importance of having a company that has a protected Mission. And for those of you that don't know much about public benefit corpse or about the corpse definitely, how about a Google search for that. It's super interesting. Just as oriented around profit as a C Corp, but really focus on that benefit Mission which we think is a really important idea and a way to empower better leadership. And then finally here and I'm from matillion talked about information by us. What are the things that we see in the data that, you know, is potentially crap, right? Sometimes people people on Minecraft and it's real. And sometimes people are claiming crap, and it's not real, and they're just, you know, and they're and they're just causing a fuss about nothing, right? How do you, how do you know, like, how do you know what's real and what's not? How do you surf?

I love that. It was like a great but you know, it shouldn't just be about data literacy. It should also be about business literacy, talked about that, and he talked about a great example of red BMW where they had to really prove with the data that something was going to be effective before they would get the budget for it, but at the same time, they had a hackathon culture. So leader is really embracing must be data-driven, but combine that with agility. So that way it's not analysis paralysis, its analysis to go faster. So

Siri is a little erected there before we move back to commercial. So this episode is brought to you by, did it at World, the Enterprise data catalog for the modern-day to stack data. World makes data Discovery, governance and Analysis easy turning data workers into knowledge superheroes. So it's to learn more. Please visit the website at data. World. So over to you want on the next topic, I do want to pull up. One more thing that Karen said it was a really like is diversity, diversity in culture in a company. So important. Because that helps us to understand about the different types of vices that we could have asked. That's an important aspect about our fourth one. We talked to go live at, or from the WM and like, why do we continue to cycle? Where is a business value is up? And he's like, you was very specific. It's because we still live in this Technical Center world. And if we want to go control that, I mean,

We need to go have the shift from a technical to the social type of glass, but understand what that means and how it is actually contribute. We really didn't listen enough people and if you don't listen it's like you're like the IRS you come in to collect people's taxes up. Don't be the IRS. You need to be an enabler for this, right? I mean technology is always a number like talking about the the value of catalogs, like both. If you you need to be at, you need to have a way to and they will be able to go find discover things that they need. So they can be really, really efficient. So that's what the value thinking about these things we talked about there's all these industry best practices. But you know what are they really best practices? Really try to understand what makes sense for your organization. What is a problem that you're solving and what we really need to go do is that when we have the data team of the pissing talk to each other, we really need to be honest and no BS and say what's in it for me?

This is people saying, why should I go spend time with it? What's in it for me? And if they do team is not able to go answer that, guess what? They are. Not able to articulate the value. Articulate, the pair trying to go solve, they need to go do a better job around that stuff. That's what I need to. Go to bed yourself, inside the business and really focus and understand those uses use cases. This is why business literacy is so crucial right now and it's okay, if we're feeling, let's go where understand where were failing to go fix that stuff, right. I think I'm really under. Another important thing is, like, you need to be invited to the table. If you are being told, you have to be to the table there, like a sack compliance, got such persons to be there. Do you want to be invited? Because that means that people want you because you will have your bringing answers to their problems and hey, it's fine if you don't sell the problem, perfectly actually start small iterating, you'll get better and better. And

We'll talk more about privacy in a bit, but we also had 42. And who is the CEO of private Ai? And she brought up something really interesting when it comes to manage. Look at companies, like duck duck to go look at companies. I mean, like apple glass, privacy for them. Like that is a core part of their business, right? So you can, it's really opportunity to show how you're doing things with us. We can look at that, doesn't have to be a liability. I can actually be something that makes you stand out. And Chris at this is exactly because she had this great post about the value of the modern-day as giving and I literally just go to the basics is like, what is the value? You need to know how much time you're saving. How much money you're making? Like, we will all the day of work. I'm going to go do, we should be able to go to tie it to money that's being made. My it's being save time. I mean, I'll be more efficient. We need to be able to go show the

What are the requirements are on this day? Let's me to get good requirements, right? We need to be able to talk to more people, right? That's what we need to go pair the date of teens with the domains, we can understand. What that business is understand that context, what is important for your business? Every business has they have okay, or is he have their North started? The agreed-upon what is either Revenue number of users conversion rate, understand what that is and make sure that that's also your focus for for the date of work, you do it. So that was business value.

Tons of great take away, is there on business value of this item here. So second-to-last is governance, right? Governance, I think definitely, we all know that governance is required right in this came up and a lot of cases across a lot of our different guess. So Dora, Buddhist for example, talked about how we need to make sure that is part of governance. We are engaging with the right stakeholders who are the right people to bring to the table to talk about and engage with and ensure that we have the government's right to think a lot about the who when it comes to garment, see you're bringing the right people together. Maybe people who are already very involved in government in the organization. Although, maybe they don't quite know it, right? Who are those experts? Who are those two ribs when things go wrong that you can talk to you and get him? Alright. Governance is there to facilitate I think it was a huge name across the border on governance. There's a there's a huge shift in thinking about governesses of policing function to go.

Has an able meant function, and that's been huge because now it brings governance away from sort of security and I T in this world that were like, it certainly has one foot in that world and actually brings it into the analysis world, right? And brings governance closer to the date of teams that are actually working with the deed itself, right? Laura Madsen, who authored disrupting, get a data governance, must read book. You got to check it out. She has these great bold ideas and Anne and Concepts around how to be effective. While government will also be thinking governance and she talked about in the past as you heard me mention earlier, it said it was about command-and-control, it was more of a policing functions for the pie arrest function, right? That these are the exact definitions and things like that but that is so short term focused that is so slap the wrist focused, right? And ultimately would happen.

People go around you. They they they find other ways to get done with her going to do and governments fail, right? And she really talked about how data governance is a long-term activity. You have to be thinking about what works in a long-term building. Those relationships and simplifying. She pressured a lot on. You got to think about how you can simplify girl don't don't blow the ocean. Don't make it. So complicated, she talked about sort of four key aspects to talk about usage, right? What are people using lineage? How are these things connected together? How was the drive, right? Text him, how do you keep your data safe? And then she talked about quality, right? How are how how can you make sure that that this is the right data and that it's available on time at cetera, right? You just waste connection quality and then she talked about sort of component which is around the metrics. Are you really should be measuring all of this being very dated driven about it. And she said that if you want to start with a to cover,

First of all, if it isn't broken, don't fix it. Don't feel like you got to reinvent everything just because you read something and you like I should do that, right? If something is working, don't, don't don't change it. You might make it worse. That wouldn't be good, right? So I find the command and control issues, write, and figure out a way to democratize it. Make it more inclusive. Put people first be clear about what you need and articulate clearly what you need. So Clarity and communication very important. And then finally agility, right? You need to be able to focus on delivering the work. Iterate. Implement governance, one use case of time, measure success with governance, in days and weeks, not in months or years,

Bob muglia also talked about governance when he was talking with us. Governance he said is the bridge between the modern-day to stack and the future of Knowledge Graph. He's actually saying that governance is what's going to take us from modern-day to stack to this knowledge graph that we want to put it the center of sort of intelligence in our companies, as well as the broader world. And governance is where ultimately people are really struggling a lot today. It inherits a lot of the problems that the Upstream applications has and has to Wrangle with all of that. Try to make it, they try to understand it. And that's really, really challenged. You said that the first knowledge grass that you as a company, will Implement is up your meditator

Your first Knowledge Graph is your metadata and governance is your first data app on that. So that's really important. And that's one of the reasons why metadata management governance catalogues are really important part to wrapping around and try to make your data more effective and then he better. He said that you start with a catalog and you build up your data model and governance is ultimately what evolves into? And it's a really important data application that everyone is going to implement knowingly. Over to Kara said that policy is also really important part of governance getting access to data more quickly and compliance. Was that whole, you got to be safe, but you got to be fast, right seat? Belts are for driving faster, and installed analogy of you driving the car seat belts so that we can go slower know, the seat belts are so we can drive fast athlete. Breaks will also breaks breaks too, he talked about seat belts,

What, what are the rules of the road for a ride? Why do we have to have highways in green light and red light, right? So that way we can drive faster and that's important in our own organization. This goes back to her governesses police ain't right. If you take a policing approach then that's going to often results in slowing down. If you think in terms of safe enablement, right? How can we go? Faster safely. And that's how you get, two more of these rules of the road. He talked about the right policy being a business decision which is important. Cuz a lot of people often think about policy is being a via technical decision or something that the date of decision. It's a business decision, some of its legal and compliance with some of, it's just being a good citizen, a talk about Apple and how they really need privacy and policy security part of the brand. Whether you think it's working already, think it's reality, it's effective. It really helping your consumers understand. And it connects to what Patricia same said about leveraging, privacy and policy.

Security has a different Advantage. You said, there's a lot of different types of coffee to their simple policy just like marking data, that it's public, right? And intermediate policy, like, anonymizing date and more advanced things like differential privacy, where maybe an item by itself is not a problem, but that item in combination with an address in combination, with the first name in combination, with knowing the birthday of all the sudden. Now, you can identify versus decentralisation and how if you can do things as code around policy, then you'll be able to actually do a better job of centralized in your policy. Definition be able to have versions of your policy and being able to actually be declarative about your policy, which is not something that usually do a policy can be quite fuzzy can be black but it can be very gray and he, he really advocated know you should really try for one black and white no more.

In a PDF document, we need to make that executable right knowledge. That's why is your policy executable Patricia saying, talk to a bunch of about privacy. Definitely check out her episode of you Chris and learn more about data privacy and the nuances of that space. And she said that privacy is about giving control to the user's security is about keeping things safe. And that without security, you can't have privacy because if it is possible the day that you get in the wrong hands, that's not going to allow you to give that control and that, and that safety to the, to the user's, right? So those two things are intertwined, an important way. And there's a misconception about privacy. People do care about privacy, she advocated. And if you give them the option and you make it easy for them, they will choose privacy and they will choose to take control that, but you got to make it easy. You got to make it accessible right? And

She said that if you want to, what is should I think we asked her? Like what do you do for fun? And she was like, oh I like to read privacy policy is just something. Brought up an interesting conversation is like do you read the part privacy policy? Like maybe we should actually pay attention more to privacy policies and there's an opportunity for us as businesses to actually make our privacy policies, more engaging more interesting. Like we can do a better job of making privacy communicative and then, finally, Shane Gibson her friend over in New Zealand. Amazing, amazing conversation with James and he said that Al Jalil means so many things. It is ultimately, though a mindset, right? It is identifying the patterns that work with you, work for you and experimenting with them. That's following a script is not the end. All the all the documentation is important and that understanding the value of the end you

Sears in a customer's is so important and that if they change their mind, if you change your mind that you can be dynamic. And that's okay. But you acknowledge the consequences that we talked about the nuances of being agile, doing it in a smart way. He talked about how you want to have sort of like these stages like a factory and how you want to be able to sort of past the work to different stations and thinking about the cycle times of these different things. So it was very interesting to hear from him sort of how a chiller and governance can actually be something that is streamlined but also something that you enter a tan and involved in this Dynamic over time. He has been one of our first listeners. He would we were when we were doing live Zooms for the podcast, you would always join your having a, we were having our cocktail. He was having his his

Coffee in the morning in the future. And he also if you want to learn more later, he has an awesome podcast. Thank you for being one of our best and most passionate Watchers really appreciate it. All right. So the five head it is that it vinyl items or one more. Every freaking episode. I don't think it's all right.

Datamash. Finally, The Best For Last comes up and she says it really you like she went off to the four things that the central light like ownership? I mean in different places Services infrastructure support form. Yes. We need a governance and full of course because I did a governance is like a very bad thing. Who would want to go centralized everything? Mother stuff is new. But what is Mew is putting these four things together making it a sting other. We all agree that datamash is putting these four things making that kind of making up conversation the big conversation. We had, I think, one of my, one of my favorite episodes on datamash is with Omer khawaja from Roche. Also Rock

Star at the snowflakes. Homemade, I was saying that there has to be a Omar fan club now, because everybody at the definition of a paradigm shift, a social technical paradigm shift, he is very explicit about these four pillars, and he like, he actually avoids, the word architecture, because when you say architecture, people jump in the technology and you really condemn people. And I'm so, do I and we say this all the time, if anybody is saying that they're going to sell your datamosh, please run as way as fast as you can from that. Vendor, please do that. Don't forget, run away. So, talk about domain, so if you come to the soccer World, this is hitting you for you, right? You know, about these domains and but sometimes, people can, I go to the other side of the spectrum to get all this? Talk about the domain and like, no, don't don't. Just go talk to people, right, by the way, it's not necessarily the same thing that happens in your org chart. It could it be if you need to go have a place to go start?

Go find the right? I p l go find those. He called it, the crazy Bunch or what are friends to worker. Says, like those crazy people wanted to be astronauts for your, for your astronauts. Find those right? Go mobilize them. All right, data is a product. You need those people who will say, I want to take ownership responsibility of the data for my domain. Who are those people I can you want to go find those folks want to take that workout ability and you want to treat that data's a child. You want to nurture it, people are going to be using it. They they're going to let me you don't want to have this just protected because he will protect all the time and let nobody go use it then you're protecting all the data. So that that's a really important thing about three days of product. S the Federated computational cover really three big words altogether. I love, is definition of it. Is go nuts.

In an organized manner, what is Federated competition of governance? Mean go nuts. In an organized manner. That is one of the best honest, no BS. Definitions out there and agree that we minimal things for data product, that needs to be kind of in a centralized and go put a write these things, right? A lot of the life science and the semantic Community talk about Fair, final accessible and drop a bowl and reusable data. We've been talking about that kind of ability boundaries expectations Downstream Casino with an explicit knowledge. And if you want systems able to go enforce it, no one will you you need to have a catalog to be able to go man's the government, run the stuff, but no one's going to use it if it's all of Emmanuel. So he needs us to the technology. Yes, we need technology comes in because we need to go to automate things. You will need it, but it will not be the first thing. It is an enabler. Actually, if you want to go get on this journey of the date of mash.

Start with all the technology, this whole episode that we had with Omar. We talk technology at the end for last couple of minutes because we said, hey, we didn't talk about technology. Let's go talk to let it be a little bit on a manual and then you figure out what he's going automated, right? When it comes to culture, it needs you need to have a decentralised coaster. If you were very Central, High School shirt, guess what am I going to work for you? So don't bother. And he went through this whole idea that whole section about how they had what he called an acceleration Workshop, you bring the stakeholders together, figure out who were the people on the same page and that domain, understand what they're, okay, ours are right. And those thick books will have the idea of what it what needs to change, right? They're going to start figuring that out there going to start discussing the green to figure out what the, what keeps people up at night. I literally a specific

What keeps you up at night? Go start after the school catalog. Those things literally you should in your dear catalog. He's not a resource called business questions. At this is a business questions. I know who asked them and what are the metrics that they're expecting about how they're Associated to the business question cuz that's what's going to drive? What date of price you want to go generate and how you going to make you. So your data and we are having to tell her, there's no need to go to the details of where this data is in what table in columns, I mean, yeah, identify what you need, but you don't need to go to those it to this level of detail and I really try to understand the whole slice of from the data to the business value all together. Or the people put, your privacy should be done from the beginning because we need to encrypt rest and in transit to lot of the risk happened in that process and we start talking about the sensitive data. We need to be able to go track these things. And then

Something that you want to date a product to say, hey, what are my contract? What am I supposed to take these were sensitive data? How can this be using so forth? Shane Gibson also talked about the supply chain of the data. We really need a document, how this stuff works today? What are the current tasks from the nose of the graph and how they're connected the edges in the grass? Start with that and focus on notes on the tasks that are not working? And that's funny. What you can figure out how to go start your whole process of beta match. Finally, Chad Sanderson, we're talking about data contracts. We really need to have a contract is an agreement between the engineer's, write the date of producers, and the consumers, the same, is this what you want? Is, what is what I'm going to give you agree on this contract. What are the escalade's? What are the establos? All right, you got it. And this is why when we've defined her ABCs of the products. Contract is one of those keeping a middle C.

Important.

We came to the end.

That was at 66 Barrack, takeaways stack, semantics metrics, knowledge leadership, business value, governance datamosh those are the themes of spring 2020 in the data world and an amazing quotes. Great takeaways I couldn't couldn't have thought of a better way to learn who wins the quotes of season 3 the winner of the best quote. So I think that's probably going to go to Sanjeev because he said, I never met a data. I didn't like

I did let the best one I have to say. All right.

Next Episode, it will be in August and August 24th. We were going to be live at the Gartner conference Orlando Baxley. So stay tuned. We have a lot of stuff. We also have a lot of t-shirts. Now that we've ordered, if you like her t-shirts, please reach out to us. We're going to start sending them out to the folks, will reach out to us to give it back. So let us know what it. What do you love about season 3? What would you like to see us do in season four and what's keep this conversation going? And enjoy your summer. You're going to see a couple of little we're going to have that whole thing's drop over to have some surprises. We're going to have more contact coming out. I do not have a special shout-out to Carly who is our producer. She is here every single week and making sure that everything goes smoothly. So, thank you, Carly Carly. And it's always thinks you let us do this. Every Wednesday at 4 p.m.

Check out and about world Enterprise data catalog for the modern-day this back. And with that Tim has a special thanks to the world for supporting Michelle, Harley berghoff for producing John Williams and Brian Jacobson show music and thank you for the entire catalog in cocktails and face.