Data products in a knowledge-first world

Loris Marini - Podcast Host Discovering Data

The average CDO has 2 years to turn data into business value. Today we talk about data-as-a-product, implicit and explicit knowledge, and the cultural revolution we need to turn data into knowledge at scale.

Also on :
Stitcher LogoPodcast Addict LogoTune In Logo

Join the list

Join hundreds of practitioners and leaders like you with episode insights straight in your inbox.

You're in! Look out for the next episode in your inbox.
Oops! Something went wrong while submitting the form.

Want to tell your data story?

Loris Marini Headshot

Checkout our brands or sponsors page to see if you are a match. We publish conversations with industry leaders to help data practitioners maximise the impact of their work.

Why this episode

The average CDO has 2 years to turn data into business value. The trouble is that most of the time data is fragmented and poorly managed. Can treating data as a product help to change this?Today I learn from two amazing data leaders Juan Sequeda and Tim Gasper from data.world and co-hosts of the Catalog & Cocktails podcast 🎉🎉We talk about data products and data-as-a-product, implicit and explicit knowledge, and the cultural revolution that's needed to turn data into knowledge at scale.


Data Management Marathon 5.0

When? October 12-13,2022

The Data Management Marathon 5.0 is a one-of-a-kind virtual event for every data lover, born out of a love for data management, knowledge sharing and storytelling. Organized by Thinklinkers in collaboration with the one and only Scott Taylor, this event features high-level speakers, influencers and an enthusiastic community of data professionals.Check out the event agenda and use our special promo code DISCOVER20 to get a 20% off the pro pass:


What I learned

Share on :

Want to see the show grow?

Your ideas help us create useful and relevant content. Send a private message or rate the show on Apple Podcast or Spotify!

Episode transcripts

**Loris Marini:** [00:00:00] So we'll pretend that I framed already the conversation and people know what we're talking about. So I'm here with one secu and Tim Gasper. One is a principle data scientist at data world is an expert on knowledge graphs, data integration, data catalogs, and Tim is a VP of product data.world. Together they host the catalog and cocktails podcast. The honest, non BS known salesy. Data podcast, which I absolutely love. And today we're gonna talk about a number of things. You'll hear words like semantic data, me, governance, data catalogs, knowledge graphs. So let's get into it. Juan and Tim, welcome to the podcast.

Thanks for taking the time.

**Juan and Tim:** Thank you for having us. Yeah, we're so excited to be here. We love the discovering data podcast. And we couldn't be more excited to join the conversation.

**Loris Marini:** Yay. Okay, so let's start framing the question as usual with the beginning of first five, 10 minutes is all about business and what the business is trying to to achieve. So the first question one is for you, I know that you just came back from the CDOIQ, the MIT conference and you wrote a really insightful post on LinkedIn.

You also published a special episode on the dinner catalog and cocktails podcast about that. What was the temperature in the room and what did you what did you learn.

**Juan and Tim:** So the main takeaway was everybody's asking about value. We have been investing so much on data science and people are like, how are we knowing that we are providing value? So that was the hallway conversations everywhere. When data mesh was obviously a hot topic, but data mesh was a hot topic in the light of how are we providing value to the data?

So value. That was it. And the second takeaway was I didn't hear much about technology. I did not hear about people talking about snowflake or your data lake or your data warehouse or the modern data stack. No, the folks would bring it up would actually be the technologist, the vendors in the room, but the CDOs in the room they were focused on on, on what is the, where is the value?

So that was actually so refreshing. I was so happy to go through this. And I was so happy to not hear about technology and as a technologist and as a vendor, I'm like, it's weird, but this is the right thing to go do. This is the honest, no BS conversations that we need to be having as the CDOs.

And also as the vendors at the end of the day, you wanna make sure you're providing value, not just selling technology. That's gonna go be shelf later.

**Loris Marini:** Absolutely. I think that's success. And I think the winds are turning and I'm so excited that cuz finally we can have those, serious conversations about what it takes to to manage data and use it to actually drive business value. But I wonder like what do you think CDOs are focusing on at the moment?

What are the top? Cause I heard words like infrastructure visibility obviously is a big one. The semantic layer. What is it that you think is gonna be the core focus and Tim perhaps? You jump in anytime.

**Juan and Tim:** Yeah, sure. No, it, I think that's a great question. And I think a lot of CDOs are trying to figure that out right now. They're like what? I think that, that is a lot of where the attention is. Place is like, Hey, as a CDO, there's this sort of stat right now that CDOs only end up staying in their role on average, like two to three years.

So I've got two to three years, what can I do? That's gonna actually make an impact and thinking about oh wow. I ha I need to figure out how we get more value out of our data. I gotta figure out how we, can comply with various compliance initiative. I gotta figure out how we can create new products and use data as a way to do that.

What's the fastest way. And so they're think this is why things I think like data mesh and all these different sort of trends are becoming very popular because CDOs are trying to figure out what's the framework that I can implement that in this two or three years that I have here, I can roll it out in a year and I can make that difference in the next couple of years.

And to add to that, I. Part of the conference and also just conversations that we get to have with so many people is the whole data mesh and we put those buzzwords of size and everything, the treating data as a product. really makes you focus on who is consuming that, why are they consuming that?

So what does, what happens if this doesn't exist? And that's how you are able to go center around the value. So I think the treating data is a product is a very crucial aspect right there that people are trying to go that that's where the focus is. And yeah, that's where the analytics come in.

But at the end of the day, it's treating it as a product is the avenue that is forcing people to. This is a product. It must be providing value and also a lot of our conversations having in the hallways. Like how do I know what that value is? Can I put a price on it? [00:05:00] Yes, no, maybe. I don't know.

Let's go discuss that because if this is actually contributing to some value for you, you would probably be able to go put a price on it. So those are interesting conversations that I had at the conference. And actually I have them with other customers with prospects on LinkedIn, on our podcast.

Actually thinking about like the, when we talk about monetization of data, it's not just oh, what is physically a money amount that we can put on it, but maybe something within organization that's an indicator. What value is? Yeah. Not just selling data to other companies. Like a lot of times people fall into this trap oh, data monetization must mean how do I sell my data?

It's what is the value of the data internally that we're creating and providing to each other?

**Loris Marini:** That's right. Absolutely. The benefits of the business, what the impact on operations on risk management and on so many things, which is which I believe is the hard part, right? That's what makes measuring hard because now, we created these abstractions. We put people in.

Box is quote unquote, you've got domains, you've got lines on business and that helps managing and making sense of it helps hiring. It helps with budgeting because you've got different accounts and different KPIs. It all works fine. If you assume that there is a, I think a centralized brain and the noise at all, or.

And that only works if the organization is small. And if it, we could back go back in time, 30, 40, 50 years. Now, the word is changing so fast. And when the organization scales, it's just impossible to achieve that awareness at the level of the organization, what is happening and what you should focus on.

So definitely. But I wanna dive because this data as a product to me resonates so much, we had a, on a podcast, Brian O'. He's been doing design and data focused design and human, actually people focused design, but in the context of data for a very long time. And we had a fantastic conversation.

So I, what I feel is that there's so many pockets of knowledge. The experts are all there. We've got excellence, software developers, we've got excellent designers. We've got excellent. Infrastructure, architecture, people it seems the right time to come all together and try to work out what is a data product.

So I'm gonna turn the question to you. How do you see the word data product? What does it mean to you?

**Juan and Tim:** So I want you, you kick us off with our data products, ABCs, and then I'll add some commentary based on what. It from the conference, but go ahead. Yeah, sure. No, that sounds great. So we've been thinking and struggling, but excited about what is a data product? Because I think a lot of people are a asking the question that you're asking Loris and what we've found is essentially there's a trend here that we've turned into what we call the data product.

ABCs and so a stands for accountability, a data product means that somebody takes ownership of that product. And when somebody when that product has an issue, right? Who do you call when you're trying to improve that data product, who is in charge of deciding what improvements to happen?

There's accountability around data products and around products in general, B is boundaries, right? A product. Is a thing and it has its surface area and it has its volume and you can point to it and say, oh, that's a product. And so a data product has boundaries. It had has inputs and outputs C stands for contracts, right?

There are certain things that you expect for that product. Is it so if it's a data product, do you expect it to be updated every day? Or is it something that was created once and you can use it or lose it and it's never being updated. Do you expect it to be reliable or is it okay if it has problems, d stands for downstream consumer. Who uses it? What use cases is it supposed to be used for? And then E stands for explicit knowledge. So rather than just allowing all the knowledge around that product to be implicit, it should be documented. It should be clear. It should be semantically meaningful.

And so ABCDE. These are the things that we're finding are consistent about what a data product is and how you create value around it and connecting it back to the conference, the MIT C D O I Q conference. I presented. Going through this, a the ABCs of data of the data product. And that was very explicit with the audience saying, Hey, what Tim just said, this is our framework of what we're thinking.

We're not saying it is the framework. We're not saying it is a complete framework. I'm, we're trying to gather from the audience, from the community what do you think, do you agree with us? Do you disagree? There's something you would change to something that's missing or something completely like this is completely false.

No. And what has been surprised? I have been surprised that nobody has come up and say, I. nobody has said that I'm wrong like this. No. And the majority of conversations has just been in agreement. It's yes, this is right. And actually a couple of people have come up with saying, oh, you [00:10:00] should think about liability.

Some kind more, some nuances about it, but B, C, D E. Are there other things that are out there or anything else falls into these different five things that we just said, but talking about the data products and actually at the. there was this one person who I there's one. I was going to all the data mesh talks.

There's another person who was also going to the data mesh talks. And he was always asking this question. It was like, what is a data product? and everybody was answering things. It's oh, it's a thing that provides value, right? Oh, it can be a table like this, but it was, there was no consistent answer.

This person is mark Nochi, who is the global CDO of EY and a very senior experienced gentle. And I just had so an awesome conversation with them throughout the conference, a very philosophical and very practical at the same time about what is a data product. And at some point, I was I'll acknowledge.

I was getting annoyed myself. I was like, I we're just giving you a bunch of definitions. Why aren't you happy it? the thing like. There is no consistency about it. And that's already an indication we're not agreeing on what a, we can't have a definition about it. What is a database?

A database is a collection of data. Very simple thing. That's textbook definition around things. But when we talk about data product, we are all these things and then, oh, we can provide value. So a Tableau dashboard is the data product. Yes. Somebody would say it is a creative product, but somebody would say the word go off.

And then at some point he made the arguments like then everything's a data product. And if everything's a data product, then that word is useless. Meaning it doesn't provide that much meaning. So the takeaway here is that yes, words matter. And we start talking about data products, but I think we also need to be careful about this hype, that we're talking about data products and as a community ourselves right now, we're trying to go figure this.

I believe that there is some sort of a spectrum and the whole framework of the ABCs is informing what that is, but there's still a lot for us to go do. And I think we are in this paradigm shift as a data management community industry for the last 30 plus years, dealing with data that we're now trying to go elevate this thing that we call data that goes into a data warehouse that it's something else.

And what is that new thing? We're just figuring out ourselves right now, and this is fine. We just need to have these conversations and I appreciate the opportunity to have this conversation on your podcast.

**Loris Marini:** Oh man, absolutely. You most welcome. I can tell you that I've been trying to explore that for some time and I don't think this is gonna ever end, one, one of the points that resonated to someone at the conference, you said that data is never done. And I really believe that because it, it, data is an instrument.

Yeah. Yeah. It, it's a raw ingredient. We can use it to solve problems and the word is never gonna run out of problems. So the there's no there's no point where we say, okay, done. We're done and dusted enough, but I think boundaries in contracts really resonate with me.

So when I think about a product, I think. Something that someone wants wants to use, wants to buy whether the exchange involves cash or time or whatever, cognitive resources, emotional resources, whatever it is, there's always a cost associated with consuming the thing, whether it's. Data information, knowledge, concepts, semantics, whatever, it could be a box.

It could be a pen. Yes it's broad, but I don't think it's gonna lose meaning if anything, it really forces or, encourages people to think about the user. Behind the product at the end of that segment, along the very long value value chain, from ingesting all the way to consumption and decision it's a long way.

Brent Dykes has a beautiful graph that it publishes often on, on LinkedIn of, how. The sort of return on investment or the value, if you want to use that, those two words is meaning the same thing where you see the curve going down for the very, for half of that length of the data value chain, and that's all the data management stuff, ingestion, manage transformation and cleaning and.

VI visibility and testing and all that. And then you start getting some sort of semblance of an ROI when you start plotting it and visualizing it. But until you get to, the actual decision and it, you can't measure the impact on the business. So it's gonna be really hard to We need to come up, I think with a new framework to measure value of products and break it down into segments and say, Hey, when you ingest data, that is also a product.

It doesn't mean if you didn't make any cash out of it, it doesn't mean it's not a product because someone's gonna use it. How do you see this? I.

**Juan and Tim:** I, I think that's an interesting way to think about it is. I think of it, like in [00:15:00] terms of like a manufacturing analogy, right? And like when you're building actual products like this laptop, there was a supply chain and a set of steps and factories and things like that had to put stuff together is a circuit board, a product.

Yeah, it's a product, but it's not the same kind of product as the laptop, as a product. And the kinds of people that are consuming and working with the circuit board are different than the people that are consuming and working with the laptop. Are they both products? Yeah, they are. What about the, when the circuit board is half built, so it's halfway down the conveyor belt, but it's not fully built yet.

It's just half a circuit board. It's not really useful yet. It's not finished. Is that a product? Eh, maybe that's not a product, what about when you combine the laptop now and the laptop comes with a TV and a desk? Is that a product or is that a bundle of products? Obviously we're talking about semantics here, but it makes you think about what is a data product?

What, isn't a data product and at what points of the pipeline, are there products coming from products? And I think that goes into what the consumer is. I think I was having a conversation with somebody on LinkedIn the other day. A brick. It's a breakup product. For the con for the contractor, it is I bought this house. Brick is not a product. This isn't the wall or whatever. I don't care about the brick. I don't, it's not a part for you. It's not a product for me, for my consumer, but somebody did decide to go purchase and they purchase one brick instead of the other brick around that stuff.

So I think so. A, as a computer scientist, I always like to the way my framework of thinking about problems is to always break things into smaller pieces, such that the output of one is the input of something else. So when you have that type of of let's call it that, that box right there.

The output of one product, the output of one box, that's a product that can be an input for another box and all these things get changed and connected to what you just said, Tim, like that's the supply chain around things. So I think that's the way. So when you zoom out. Somebody who's buying the result of these of 4, 5, 6 different products.

They don't care what happened in between. So this is really why the within our ABC frameworks that we discuss is it's crucial to understand your downstream consumers. Now, I think there's also different types of kind of these manufacturing. Even if we look at many manufacturing, this is something I learned talking to my buddy at Mohamed OER mean he's a former CEO of Mackenzie talking about what is a continuous shop in like in job.

so job shops in manufacturing is like I have, I am going to go manufacture a very specific product. That was that I have a customer who says, I need you to go to create this very specific type of box for some reason, whatever, how many I gonna go produce? Not that many, because I'm only, I'm presuming you very specifically for.

But then I realize, Hey, the market's going for this. More people want this. This is not such a specific thing. More people want. Eventually this goes into a continuous

And then suddenly, oh, wait, I, this is, I'm just doing the same thing over and over again. Let's make sure this goes in. One thing is to go have a Toyota Corolla versus one thing is to have a very specialized rolls Royce.

But at some point, like these things start crossing. And so if you start creating a product that's very specific, but then you start realizing, oh, more people are consum. Thing. Why am I spending so much re I'm trying to, oh, optimize how manufacturing is. So I think there's a lot of, I'm not saying that the way we define the way we should do product data products or how we manufactured cars or bricks or whatever, but there's a lot of learning around these things.

I think also a message that I want to give out is let's read history, let's go out of our bubble to go see how other. Industries go do things. The data world is so invested in their bubble and we think that we're doing the latest, greatest stuff and we're the best. And we live on this high horse and I'm like, no, get off your high horse.

Be more humble. Go read.

**Loris Marini:** so much to unpack there. Absolutely. I think. When start in reverse order. Yes, absolutely. We need to hit read history books and we need to look look around, and example is the modern data stack. Just having a conversation a few hours ago with the chat Sanderson, we're gonna have a Pakistan.

And the comment was like, you see this new trends, the MDs is great. There's real benefits that this new type of storage and transformation and management, the semantically control that DBT brought. Something completely new. None of these things along by themselves are sort of the solution, but it's very easy to get really caught up with the new thing that we, the focus, the range of our focus of our attention span, narrows down to the thing.

And now everything is MDs. Guess what people in the [00:20:00] enterprises have been doing semantic control taxonomy ontologies for a very long time. The word is full of experts and we need to bring those experts in because we can learn from them. So I'm gonna, I'm getting ahead of myself, cuz I wanna touch on ontology and semantics.

We're gonna drill down, but I wanna make a comment in the reverse order, on the concept of the production line and Tim, I think this is a brilliant analogy. The. Circuit board and the individual component. Think about a resistor or a transistor, that you can use it as part of something much bigger and more useful, like a laptop, because first of all, you've got specs, right?

This what we call metadata in our field metadata, the metadata tells you what that piece of data actually is. And with the transistor, you've got a product ID. You can look it up, it's. It's not just a piece of plastic with some metal soldered, FSED together. It's a thing that has a clear boundary, right?

To site your framework has a clear contract. You know what the transit, the specs of that thing, a current is gonna get, come in. The voltage is gonna go up, you've got curves and that's how you reuse it, but you gotta be able to find it first. So it's everything is coming together to something is U useful.

If, you know the context around this, if you know what it actually does and you can find it. And so the catalog, the idea of the catalog, and now, if you look at Gartner catalogs already dead, everything is already dead these days, but the catalog is there to facilitate that that process of this discovery.

So how do we just to drill down on this semantics? What is, how do you see the process of. Discovery and labeling stuff so we can find them again when we need them.

**Juan and Tim:** I just wanna pause here quickly of saying, I love how you guys super excited. You have a PhD in physics, right? Like I love how you're just making this connection between like great. So that's awesome. Yeah. That's perfect. I love it. Like you, you talked about you, so you mentioned about like discovery and Cal, like how does that kind of stuff all tie together with products and I guess the quick comment I'll make is that I think it goes back to this analogy of the supply chain kind of to build the product.

But then on the other side of that is the marketplace. Like where are we selling this product? Where do people go to get and procure this product? And I think that like catalogs and, discovery and governance and things like that more generally are around looking at both of those lenses.

Because for example, the marketplace that's easy, right? People think like amazon.com, I want to go to Amazon and I wanna buy stuff and I wanna. I wanna see the books and I wanna see the music. And so it should be organized in a way that makes sense to me as the consumer, whether it's a broad consumer or a consumer within that specific enterprise.

And then I want a very, I wanna be able to see what is it right? What are its boundaries? What can I expect from it? What's the contract, right? Oh, if it's messed up, is there a warranty? That's the accountability, so all these things matter, and then I wanna get it. I wanna buy it. I want to have it right.

And so that's me extracting the value from that. But then on the flip side of that's the supply chain side. And that is well, how a little bit different from your amazon.com with data we usually care about where did it come from? And can I depend on it? And how often does it come and all these different aspects.

So that's the supply chain aspect. And that's just as important, if not more important than the marketplace aspect, cuz you need to know that you can trust the data. You need to know where it comes from. You need to know whether it's a product or if it's not really a product. So it all fits together into that sort of picture of discovery trust.

And access around data. And one thing I wanna add is I've had this conversation with a buddy of mine Gary, George from indeed he really opened this, opened it to me and realized it's not really two catalogs. It should be, you have to have two lenses of a catalog and you understand who those personas are.

One lens is for those data producers or for those folks who are interested in the bits and the bites and the transistors of things. They're the ones who are interested in the most detailed things, right? They're the ones who need to know the technical metadata, the quality, the lineage, and all that stuff.

All this stuff happens later on within that supply chain, which by himself will continue to use the catalog. But at the end you have the marketplace, which is the other lens of the catalog. It is the lens for the consumer. they want a catalog to go discover and search for the product they're buying that computer, that phone, which by the way, has all that transition to the chips and that.

They don't really care about the details of it. They don't, the consumer doesn't care. What the quality what type they care about. So much of the quality. They don't care about the details of the quality. They don't care about the [00:25:00] details of the lineage. They don't care about the details of the technical.

They don't care how often the factory checks the floor every night and things. No, exactly. They just want to know, is this yes or no? Great. And guess what, how do they know it? Reviews comments. If I have two different products, I always do give this talk. I always have water bottle with me. Look for water bottles on Amazon.

How many water bottles are you gonna get? Find Amazon, not one a thousand. Which ones are you gonna look at? How is it ranked? The ones that have most best comments. So even when people say oh, but how can we have multiple products about customers? The data will be duplicated. We go, yes, this is the complexity of the world.

There is not one manufacturer of a water product, a water bottle. There's gonna be many of those things. We embrace them. Which one are you gonna go use? The ones that you feel more satisfied that you feel that answers your questions, that you see, that your colleagues, that your, that the people that you go trust, go use that.

That's one you're gonna go use and the ones that are exist and they don't, and they don't have good quality of stuff. They have a very bad description of their metadata. You're not gonna go use that stuff. And that is gonna slowly die. Is this sort of data capitalism here, maybe? At the end of the day, it's like the strongest will survive and there's motivation there's competition.

So even if you think about it, the data products, when you think about the, a for accountability, and we talk about ownership and I know people like, oh, ownership is the bad word, then call it trustee or call it an ambassador. You know what also you should have within account. marketing. You're creating a product.

You're investing this product. You want more people to go use it, go market it,

**Loris Marini:** Yeah.

**Juan and Tim:** we need, this is bringing product. Thinking into data.


We are product marketers.

**Loris Marini:** I think, yeah, I think you're into something really big here there, which kind of opens the conversation towards the idea of decent translation and not worrying too much about one topologically speaking, not one node that knows it all, but letting teams with their own expertise and subject matter expertise and domain knowledge to add that knowledge into the the data layer, the semantically to many layers that compose this thing.

The problem I see with that one is that. There is going to be, we need someone that will make a call and say this model here. Piece of the data ecosystem is now obsolete. We need to retire. We need to get it out of the way. Cause that, governance and in general compliance is gonna be really hard to do if we don't do that.

So it's gonna be, it's super interesting. How do you do that? I have no idea.

**Juan and Tim:** Yeah, let me go start. Cause I know you have more to add here and I actually don't know how much we agree on this one. We'll see. But I think there's always this balance between centralization decentralization. There are some things that you need to go centralize because that's the industry you live in it's regulation.

You need to do GDPR. You work in finance or in pharma, you need to go regulate because if you don't do this stuff, you're gonna get fined. So you are obligated code. Do obligated to go do that, right? It's a carrot to stick. You gotta stick to go do that period. I don't care how you feel. You gotta go do that.

That's centralized. Now on the decentralized parts goes back to like you think Amazon is gonna go say we're only gonna sell a hundred different water bottle. No go whatever you want and who's gonna win the strongest, the strong. And what does the strongest mean? The people who rate the most write more comments, I think that's how we should treat data within large organizations.

Now this balance between centralization, the centralization is going to evolve. Depends. The size of your organization, the culture in your organization the industry that you're in, that's going to, and it's going to evolve. And I just said something very bold right there. That's saying, Hey, let this stuff die.

And you're gonna go invest. I don't know what are your thoughts, Tim? To go with what you were saying, Laura. So like sometimes you want, you don't wanna wait for it to die. You wanna kill it? and so there's I, what I go to is a little less of what you said at the beginning, but more what you said at the end.

Okay. Which is that, I think this is. Designed like we get to choose as organizations, how we approach this. And so I in college I studied economics. So I come from an economics background. And so when I look at data and I look at data supply chains, I look at data products and things like that.

Like it activates the economics part of my brain. And I'm like, oh, we're designing a system, an economic system here within our organizations and sometimes across organizations and we get to make [00:30:00] choices. And so there's this marketplace that's happening here. And we can choose to let it be a free for all, and that's gonna entail certain things, we can choose to have a federated governing body where it's more of a consortium and the different decentralized parts have to meet together. They gotta coordinate and there's a tax that they pay, but it gives them a lot of autonomy. Or you have more of a centralized governing body, right? And you give them, this much responsibility a little bit, or you give them a lot of it, of responsibility.

And depending on those choices that you make and what responsibilities you give them different pros and cons happen. And so I go to what you said at the end there, Juan, depending on what industry you're in, depending on what you're trying to optimize for. You have choice and you get to choose what you want in your economic outcome to be around your data.

I, I would argue, I agree. We're on the same page and I would argue. That historically we've been in this one central place and that's it. And we've certainly tended to gravitate towards that. Yeah. And I think we just realized that we can't, that's not sustainable. We can't do that anymore.

It's sustainable. If you have a small company, you have a small team, but once you start growing, that's not sustainable. And if you're treat, if you keep trying to keep that up and going. While your company's going. You're not, that's why you have multiple silos and all another thing, blah, blah, blah. We do all the time.

This is part of the interest around things like data me, is that okay. The country has gotten too big. Our data country has gotten too big when we need. We need states. We need counties. We need cities. We need some other way to start splitting this up to, to scale and we have shadow it.

I know I'm probably pushing this analogy more, but that's why you have rebels in some area, whatever. We could get some interesting, this analogy has legs. This analogy could go weird.

**Loris Marini:** yeah.

**Juan and Tim:** pass to you before we get into some weird conversations here.

**Loris Marini:** No, definitely. I think I totally agree. It, there has to be some sort of, making sense of this fragmentation at some point because you can't, the one thing that really concerns me is what happens when. You do self serve or achieve some sort of self-service kind of functionality into any product.

Doesn't have to be a data product. And then people that don't have really the understanding of the implications downstream, they don't experts in the field, but now they can, they feel they can come in, can play and bang some things together and get some functionality and get really excited. It's great.

But. It can also backfire big time and in data in particular, because we know, I've worked as data scientists come from that background of, trying to extract meaning from data. And then I got passionate into building trust, and allowing the data management side. But my, because of my background, I know that you can get the Datatel, whatever you want, and if you don't have that literacy, so it's you have to have all these elements in place because if you focus on only on one thing I feel it's gonna be a recipe for disasters.

So there has to be some sort of bigger picture there's someone at the very top. And it should be probably if you ask me the chief executive or, definitely the chief financial to get up with the chief executive to sit down and say, Hey, this matters to us as an organization. We need to have.

The systems such that they know what our people know and vice versa, there's this knowledge exchange between our databases and the heads, the brains, the flesh, that is working around or working remotely, there's still a human machine interface and we need to work how that interface.

It's gonna look like to enable scale. And I think I'm so fascinating about, this topic because it's brings together so many different fields of science, from psychology to technology. It's just, it's awesome.

**Juan and Tim:** So I, one of the things that I've really been push pushing right now, why, in addition to data literacy, we should talk about business literacy. And this is something that I think at the executive level. Now everybody says data literate. That is what we call an ordinary truth, right? Neil's bore you have an ordinary truth and profound truth.

An ordinary truth is one that, who that the opposite is obviously a falsehood, a profound truth is one who's obvious is also profound truth. It's negation, right? So we obviously need to be data L. Yes, because not being data literate is stupid. I think what we need to go do, which we're not doing is to be business literate.

And what I mean by that is that everybody within the organization let's actually focus on the data teams and their, they need to understand how the business works. Ask yourself, do you know how the flow of information goes across your entire organization? Do you know how the flow of. Goes through the organization.

And do you know how that flow of money is aligned with the data systems that [00:35:00] exist? Let's start. I have, I spend money on ad campaigns on Google and LinkedIn money goes in. There's a system, Google analytics, LinkedIn sent me some stuff. Those marketing campaigns generate what leads. Those leads flow into what, into a CRM system, HubSpot, Salesforce, whatever you want to go do.

Those leads are different types of leads. Your organizations have different types of leads, and they're gonna go through a sales process. What is that sales process? Is it a qualified lead? Is it a sales lead? Is it a marketing lead, blah, keep doing of these things?

**Loris Marini:** a, do you wanna, B do they want

**Juan and Tim:** Exactly go do things, right? You have a lead, a person as a company, they become a customer.

What is the customer? You have a cus another type of system that manages customers and the customers need to make sure that they keep you need up. So what is that entire process and what are the systems involvement? So when you go talk to somebody and they say, oh, I need this data about. You're like, okay.

Which department are you in? What part of the flow of the data and the money you're interested in? What data do you need data from this system, from this other system? Do you live on the border of these systems? Like the problem right now is that the data people don't understand this entire flow. They live in their own little tower, right?

Their high horse or whatever they need to not just be exposed. I think they need to understand, and the ones who will succeed. the ones who will be the leaders are the ones who truly understand the business. And and if you don't wanna spend your time learning the business, I'm sorry, you're gonna be second class.

**Loris Marini:** Yeah, absolutely. This resonates. Me when it's unbelievable, I think Very that it has to be a synergy like the business folks need to feel definitely less threatened by data people that tend to be very technical inclined. They have their own language. And so this, we have to literally come together.

So that means. Opening up and not closing down when it needs. We need to stop using divisive language as in, oh, you don't understand cuz you're mark. No. No. Okay. I don't under you. Don't nobody understand. Nobody understand. Can we get together and then understand it together. That's curiosity, which really resonates with me that unlocks.

Empathy unlocks communication and ultimately problem solving because you don't come in the room knowing, or looking like the one that knows or stop pretending a lot of people have this thing, why is it I it's part of the culture maybe, but. It just bothers me so much. When you see that there's a clear problem.

Everybody's complaining. There's a bunch of people finally in the room after months and months of people complaining someone from the top say, Hey, we need to get together a brainstorm session of six hours. It's exhausting. You drink water, you don't eat. Coffee is after coffee's in the end. There's not a lot progress.

We've got these masks on our face and, we want to look like the ones that know what if for once we are knowledge that we don't know and it's fine, that's I think it's gonna be massive. Yeah. I don't know. I don't have a question, but just wanna share that.

**Juan and Tim:** This is good. I love what you're saying there, Laura and maybe to add one additional thing is. And beyond just the, not just the business becoming more data literate, the B the data, people becoming more business literate is how do we take what does it mean to be literate?

It means that we have some shared understanding, right? We have some explicit meaningful that's happening. And I'm gonna, I'm gonna travel a little bit into Juan's territory here and talk just a little bit about semantics and say, like, how do we get more. You implicit and inherent with the knowledge that we have in our organizations, because to your point, like six hours in a workshop we're drinking copies, we're talking at the end of it.

We end up with a bunch of boxes and arrows and we think maybe we're on the same page. Maybe we just tell ourselves that because we wanna feel like it was a good meeting before we all go out to dinner. But like in the end, like how do we capture that knowledge? How do we put it somewhere? And actually.

In some cases agree this is what a customer is, but in some cases disagree, we could, we argued for six hours that we couldn't come to an agreement. Okay. It's because there's more than one truth. There's a marketing customer and there's a finance customer and there's an accounting customer and there are three different.

Notions of customer and we're gonna approach them a little bit differently and that's okay. We're getting explicit. We're getting embedded, right? The explicit knowledge. It goes back to the framework of ABC, the ease, explicit knowledge, right? So if we're talking about customer, we know what customer, so you know what business [00:40:00] literacy means to me, if somebody says I want customer data, The other person's saying what type of customer data and the other per the other person saying, oh, I'm apologies.

What I mean is that I need the customer data about finance about oh, that's not the customer data I have a way marketing. I'm so glad we had this conversation because I'm glad we realize that we're talking about different customers and I'm not gonna waste your time because your data's not what.

That is business literacy, right there. That's one thing. Second, something that I, one of the very hard learnings I did in my PhD. And probably you did you learned through that too? Is guess what I don't know is a perfectly valid answer. we need to be comfortable with the complexity of the world and realize we don't know all these things because a lot of people like to blah, blah, blah make up a lot of things because if, oh, if I apparently don't know this stuff, I'm gonna look bad.

No, we need would, the world is complex. Organizations are complex around these things. And I think this is the shift about when I talk about business literacy, what I talk about. The knowledge first world, the knowledge first world is where I'm like, we're gonna be explicit about what these things mean.

We're gonna understand what the context is. People, what I call, we call knowledge versus people. First context. First relationships. First, you talk about something. What context are you coming from? Who you said the word customer? Who are you? Oh, you're Tim. You work in product. Oh, so you're talking about the CU.

So Tim is the person who's talking about customer in the context of a product, very different from Haley who works in finance and her definition of customer may be different. That context, but guess what? There is a relationship between the finance version of a customer and the product version of the company.

What is that relationship? Let's go understand that. That's a knowledge first world around that. I always tell people what is, what does knowledge first world mean? Go catalog your question. Go catalog the business questions that people have, who's asking them and what department, the context that they're in and when they start asking those questions, guess what?

There's important words that show up. Oh, customer user customer activity. Oh, those mean things. Let's go then go deep. Let's go dig in. So I think this is the knowledge first world of. Focused on people asking a lot of why we need new roles within organizations. What I call the data therapist, the knowledge scientists, the analytics engineer, the data product manager, blah blah.

At the end, these are the people who are trying to bridge the gap and focus on the meaning, the knowledge, the semantics, that's the knowledge first world, which is different from what we've been doing the last 30 plus years. That's a paradigm shift we go into.

**Loris Marini:** Absolutely. This resonates so much. One, I think the psychological stance has to be, that there is a, there is the right mental attitude to that and the wrong one, clearly wrong one to me. And it's rare that I find, black and white, so that can delineate very clearly concepts.

But I, I really believe that to do that. We need to be really comfortable with looking stupid and asking the why of something that looks like an obvious thing. But. It's never obvious. There's always stuff that you can learn by asking why, so that resonates so much with me. But so that's the psychology, but there's also like a actual processes and technological challenge, how do we cause a lot of those conversations happen in the water cooler and when you're going for a coffee after lunch with a mate discussing, oh, I'm really stuck with this.

And so by having that harness. Non threatening, safe conversation. Cause that's what happens when you walk to get a coffee, there's no someone in a room measuring or listening or ranking. It's literally just you with your mates talking about this and what happens then to me at least, I don't know, maybe you guys experience this the same, but.

You get to not only solve the problem because you get to explain it, but also you learn different angles to the same thing that you weren't even conceiving because inevitably that person has a different view, a different experience, a different understanding of the business. So relationships here, and of course there's only so many coffees we can get during the day.

But if we were to create an environment where. We can have those sort of conversations. The safe space looks I'm talking as a psychologist, but literally, like it's yeah, it's an opportunity. It's an, it's a space to exchange ideas where no one is measuring or judging or trying to get promoted necessarily.

We're just sharing knowledge because we know that together we can achieve way. Then the sum of the individual parts, it's that getting together, being that. So if you think about the skills that we need to develop and look at the future, what do you see? What should we invest on? What are the people that are going to actually make a [00:45:00] difference?

**Juan and Tim:** Oh, my goodness. I think I'll start. And then I'm curious where you go here. I think a lot of this has to do. Empathy. And I know that's what a lot of people would think of as a softer skill, but I think that's a lot of what we need to do. If we're gonna really understand each other, we're gonna, if the business people are gonna be, become more data literate, if the data people are gonna become more business literate, we're gonna connect together.

We're gonna have empathy for each other, the problems that we need to solve and the path that we need to take in order to get there. And so there's a lot of communication, a lot of documentation, a lot of collaborating. The, the combination of people, process and technology together, that's gonna help us to create an environment where we can be empathetic and not just in the data realm.

But across the entire organization. Like we talk about oh, it should be, there should be honest, no BS. There should be safe places for you to have these convers. That doesn't happen unless you create this culture of curiosity, not knowing, asking hard questions, being polite and empathetic while you do it.

But like that, that these are skills that might be considered soft skills, but this whole hard, soft thing that needs, we need to throw that out the window. Yeah. That's not useful, that the, these, so so-called soft skills are the things that are. Organizations and people actually successful and others not.

We had a guest on our podcast. Ergas JLA you should have him. He's been writing a lot about data modeling and we had him and we asked him the exact same question. What are the skills that should transcend time, that you should be learning to go do data modeling? We're talking about data modeling and his was.

In curiosity. And this was an episode we had a couple months ago and it has landed with me that I'll never forget about this. And now I start every day in my work and with my customers and in sales, whatever, I am trying my best to be empathetic, to put myself in the other person's shoes to say, what are they trying to go do?

And, but at the same. challenge them because that's part of it. Part it's part of it. It's asked the, it's asked. why are you asking for that and why keep going why you really want to get deep down? And I think what are the things that, again, I'm gonna say I talk about this as soft skill soft skills, but I'm like, to stop saying this, these are just the skills needed gonna have.

Communication is key. You need to have. And part of that empathy is like what I call like the data therapist is actually, let me go, tell me about your problem. Let me go understand that. And let me go ask you why it's almost like a sec, like a Socratic method. Keep asking questions to go understand that type of stuff.

So those are the types of things you wanna go in. Understand why do you want a data catalog? Because we need to be data driven. of course be data different. Why didn't be data driven because we need to democratize data of course you need to democratize. Why do you need them opportunity? Why and what happens when you keep going to, why you realize that there's some group of people you're talking to that they don't know why they're just following orders.

Okay. They need to challenge themselves and figure out, okay. I receive this order from somebody else all the way to go let's understand what, where this is really coming from. Why did they tell me to go find a cat? Exactly. Why do you need, why do you need the, I need high quality data. Why? Because we can't make great decisions.

I know we need high quality data, but why what is behind all these things, right? So ask yourself, challenge yourself. That's one thing. Second culture. I truly believe that this is, this needs to be an example, coming from the executive level to be comfortable. The executive not say, gimme the customer, oh, I don't know what customer is.

So we link. No, you need to be able to go say we have multiple definitions of the customer, Mr. Or Mrs. CEO, which customer do you think? Because it's these five different. Your talent, go challenge your CEO too, to go do that stuff. Then for more of the, like the more of the technological skills, I think we need to go focus on the things we don't do today.

Data modeling, understand knowledge, logic. These are crucial things that today we're not sunny. People can get a master's in data science and they don't take no data. Mono node logic. It's ridiculous. It is actually ridiculous. How the heck are you supposed to generate insights from data? If you are not studying what logic is and how to go do data modeling,

**Loris Marini:** Or if you don't like talking to people, if

**Juan and Tim:** if you don't like talking to people.

**Loris Marini:** I met a lot of those folks, right? The people that love to be behind a laptop in their Jupyter notebook, doing models. And they, it's not so that relationship and I think we should start using, we should start calling them human skills.

Cause that's really, it's not soft. It's not soft at all. What do you think

**Juan and Tim:** a great moment I will from now on, [00:50:00] remember this moment, stop saying soft skills. Yeah, it is human skills. So you have your technical skills and your human.

**Loris Marini:** Not even people, right? Because people doesn't have that, what's unique about human beings is that we've got feelings and we've got irrational and an emotional brain, and we need to stop ignoring that the emotional brain exists. Can we please do that? Can we please stop ignoring that? There is it's there evolution speaking.

It evolved first. So if anything is the emotional brain is the first class seated in anyways, I'm now going

**Juan and Tim:** I actually tell people, you know what if you go to a therapist, Just to experience it in case you don't you. I go to the gym so I can exercise my body to feel stronger and lose weight and so forth. I go to a therapist so I can exercise my brain to feel stronger and lose the weight of things that I'm thinking about.

Go, if you here's a technique here. Here's an example. Go. If you have it go to a therapist, just experience it. And you will, you'll be exposed to a different world. Of just going to the hurt at the beginning. You've never done. It's uncomfortable itself. And you're just gonna Hey, I should, I'm learning a bunch of techniques.

Like I can apply this at work. I can ask the why I can be empathetic about these things. We, the world needs more empathetic, curious human beings.

**Loris Marini:** Absolutely. Absolutely. I've started my own journey in therapy in January this year. And I think I found that, I should have started 10 years ago. I think there's a lot of stigma around that, but that's just pure BS. It's just what you described.

It's it's. It's a discovery journey into the nature of your own mind, how it works, what are the, the blind spots that we all have. And I think that if we had a little bit more of that, heck we should create a new, see level, office. I know you're gonna hate me because there's so many, there are too many, but it should be like a chief curiosity officer, someone.

It has got one mission is to go and talk to people and get them in, in, encourage curiosity.

**Juan and Tim:** you're not crazy. I actually, I was right about this the other day. I'm like, there should be some sort I call the data therapist. I think you could, should start out having a data, the, not even data, a therapist with the organization that reports to the CEO and this therapist, what is gonna go do is ask start with the VP's leader.

Ask two questions. What keeps you up at night? What metrics would you use to, to keep track of that? And then keep asking why I get more information catalog that put in the spreadsheet. If you have a data catalog, put it in your data catalog, by the way, shouldn't I hate the word data catalog should be a data knowledge Kello, because you're cataloging into questions and the people are asking them and then you'll be surprised what you.

**Loris Marini:** Definitely looking looking forward to that. I think I think this is the time is ripe. I don't know how you felt team about this, but there's been so much, going on in, in the space. And I feel like we are all converging. It's beautiful because we are coming from different. Paths, we are approaching the problem in a completely different way.

You got people that have been selling software that realizing this people that have been, deep involved, deep in the widths of data teams of, in enterprises that are struggling, people that are in startups. And they've got the modern, cool data stuck and still struggling. And then of course there are those that are implementing data emission organization.

50 people, which is to me, is like, what are you doing? But yeah, we all, we are all converging to to the same ground truth we need to have this multidisciplinary kind of thinking it's if you're in data, it doesn't mean. It's not enough for you to know how to code SQL. SQL is definitely an important part of your skill set, but there's so much more to it.

And so that means changing how we hire, how we promote people. It's just fascinating, to see how the industry is slowly is converging and maturing. And hopefully we get, we can have less hype cycles in the future and more productivity plot, always a productivity. So we can actually do stuff.

**Juan and Tim:** I hope so. It, as you're mentioning, it's such an exciting time right now and I think that. We're finally moving into this world where, people were talking about 10 or 15 years ago, maybe it was even longer than that. When software really started to be like something like every company was like, oh, we have to be, we have to be a software company.

Like every good company is a software company and we're gonna hire developers and things like that. And I don't know how successful that was ultimately. Now you've got the SAS movement and things like that. It's oh wait, maybe we don't have to build all our own applications. Maybe that was a bad idea.

But I, I think this time with data, it's actually a much better and healthier movement, right? Oh, like every company is a data company. We should all be data companies. That's, what's gonna make us successful and every person should be a data person. [00:55:00] And I think that's a really healthy thing.

And I think if we're successful and we may be many decades away from this, but it feels like the winds are blowing in this direction. If we're successful there, shouldn't be like, oh, the chief data office and all those are the data people, we all are part of the data office. We are all data people. And if you're not using data for your job, what are you doing?

Like you use Google docs and you know how to type words, right? In important to know how to type words to do your job. It's probably important for you to know how to work with data, to do your job. And you should know the business. You should be business literate too, right?

**Loris Marini:** Guys, I'm gonna steal your one minute thing because I think I love it. So gonna have, I've had my timer right here. So you got 60 seconds. Help me wrap this up in three to one who wants to go

**Juan and Tim:** All right. So takeaways, you gotta be curious. You gotta be empathetic. You gotta understand the data, but you gotta be business literate too. And ultimately you want to be able to balance decentralization and centralization and create data products. And you're gonna be a lot happier and data products, B C, D E accountability.

boundaries, contracts and expectations, downstream consumers and explicit knowledge. And,

**Loris Marini:** Boom.

**Juan and Tim:** that's it.

**Loris Marini:** Awesome job, man. That was a 20 sec, 22nd, 23 seconds. That's impressive. And I'm gonna add the multidisciplinary thinking. Curiosity, get curious, ask why go for it. It's gonna be a rewarding journey. You're gonna learn so much.

**Juan and Tim:** This was fun. I'm excited that you did the minute we do the data mesh minute at at catalog and cocktails. And it's a

**Loris Marini:** This is even.

Yeah, guys, I really wanna thank you. This has been a great conversation. We are approaching really fast at the end of our time, but I'm looking forward to many more in the future, so we should get together and hang more often.

**Juan and Tim:** Absolutely. The world of data is changing and your podcast is an awesome way for people to learn about how to do things better and have the right conversation. So thanks for all you do.

**Loris Marini:** Oh, thanks to you. And that reminds me do catalog and cocktails. We're gonna put links in the description. If you haven't listened to that podcast, you should definitely start downloading em. That's gonna be part of your top three that's for sure. Cool. Guys. Big thank you. And a big shout from Sydney Australia will will catch up soon and see on LinkedIn

**Juan and Tim:** Awesome. Bye. Thank you. Cheers Lauras.

**Loris Marini:** C.

Contact Us

Thanks for your message. You'll hear from us soon!
Oops! Something went wrong while submitting the form.