About this episode
Technical folks miss the boat and are boring when they talk about the features of data catalogs – such as glossaries and data lineage – to business people. In this episode Krystin Kim will share how a data catalog should be presented to the business: the ultimate place to share ideas across big companies, a treasure trove of use cases for others to discover and make even better, recipes to make amazing things that returns millions in business value.
Tim Gasper [00:00:01] Hello, everyone. Welcome. It's time once again for Catalog& Cocktails. It's your honest, no- BS, non- salesy conversation about enterprise data management with tasty beverages in hand, presented by Data. world. I'm Tim Gasper, longtime data nerd, product guy, customer guy, joined by co- host Juan Sequeda.
Juan Sequeda [00:00:20] Hello, Tim. How are you doing?
Tim Gasper [00:00:22] I'm good.
Juan Sequeda [00:00:23] We're global today, so hello-
Tim Gasper [00:00:26] Where are you?
Juan Sequeda [00:00:28] Well, first of all, it's Wednesday. It is end of the day, 5: 00 PM, here where I am. I'm the East Coast. I'm in Orlando. Where are you, Tim?
Tim Gasper [00:00:36] It's 10: 00 where I am. I'm in London, in the UK. This is a global podcast episode today.
Juan Sequeda [00:00:42] I'm here live from Gartner, and I'm together with our fantastic guest, who I met actually at a Gartner last year or I forget, but we just sat down and we had a phenomenal conversation. I'm super happy that you are a listener, and now your part is a guest. It's Krystin Kim, who is a senior director of Decision Science at Post Holdings. How are you doing?
Krystin Kim [00:01:00] Great. Thanks for having me here. I'm really excited to be part of the show. I've loved the show to listen to both of you and all your great speakers, and I'm excited to be part of it.
Juan Sequeda [00:01:10] Awesome.
Krystin Kim [00:01:11] I do have to say, for my legalese at Post Holdings, that all of the opinions I'm expressing today are entirely my own. I'm not representing Post Holdings today.
Juan Sequeda [00:01:21] Thank you for that. I acknowledge. Let's kick this off. What are we drinking? What are we toasting for today? Tim, you're on the other side of the pond. What are you drinking tonight?
Tim Gasper [00:01:31] I am drinking a special cocktail, a special old fashioned. It is an Italian old fashioned provided by the bar that I'm in, so the bar is called Ruby Lucy. It's a very fun, kind of rock themed bar. They've got these Marshall Amp kind of things that are here for listening to music and things like that. This Italian old fashioned's got some amaro in it, so good stuff.
Juan Sequeda [00:01:56] Oh.
Krystin Kim [00:01:56] Nice.
Juan Sequeda [00:01:57] How about you, Krystin?
Krystin Kim [00:01:58] I am drinking a hibiscus cooler, which has some hibiscus liqueur, vodka, and for hydration, coconut water.
Juan Sequeda [00:02:07] There you go. You see getting all the-
Krystin Kim [00:02:10] All in one.
Tim Gasper [00:02:11] That sounds refreshing.
Juan Sequeda [00:02:13] I'm by the pool, and I'm like, it's actually nice and sunny here in Orlando. It's finally not humid, because every time I come here it's hot and humid.
Krystin Kim [00:02:20] Beautiful.
Juan Sequeda [00:02:20] It's beautiful right now, and I'm like, I just want a nice, cool beer. They have one of my favorite IPAs, the Jai Alai, I think it is. Anyways, delicious, that. What are we toasting for? What's something special to toast-
Krystin Kim [00:02:33] I had a great day today. Thank you to Mike Sullivan and Gartner, who helped me get a little bit of a wonderful day. I got to take my picture with Magic Johnson. Truly inspirational. The way he makes people feel is amazing.
Juan Sequeda [00:02:49] How about you, Tim? What are you toasting for today?
Tim Gasper [00:02:52] I'm going to toast to being connected to other people. It's great to meet interesting people, be connected, so I'm going to toast to that today.
Juan Sequeda [00:03:02] Cheers to that. Awesome.
Tim Gasper [00:03:03] Cheers, Krystin. Cheers, Juan.
Juan Sequeda [00:03:06] We're traveling and we're at a conference and stuff, so tell me what are the most fun things that you've done with a large group of friends or colleagues when you've been at a conference?
Krystin Kim [00:03:16] Something I'm going to say out loud. No, just kidding. We went to a conference, not this particular time, but usually, whenever I go to a conference with my team, I take them to a basketball game and-
Juan Sequeda [00:03:30] That's already a fun thing right there.
Krystin Kim [00:03:32] Yeah, it's amazing. The team itself is super fun. I'm a huge basketball fan, so that's my enamored day of Magic Johnson. We always go to a basketball game and enjoy the local sports, because it's around the culture and the sports and being a part of the city.
Juan Sequeda [00:03:51] Tim, what have you done, fun, crazy-
Tim Gasper [00:03:54] Well, I also have to make that same comment as Krystin, which is like, hmm, it's okay to say on a podcast. It was really fun going out. There was a conference I went to several years ago where we all went out karaokeing afterwards, and that was so much fun, so karaokeing is definitely up there.
Juan Sequeda [00:04:13] I think that was what I was going to say. I've been from academic conferences to industry conferences, the karaoke part. I did a conference once in Japan. Oh, that was wild. Actually, another crazy karaoke one was here at this Gartner one, because there's that Japanese restaurant in here and we ended up karaokeing. That was a fun night. Anyways, let's kick it off.
Krystin Kim [00:04:33] That's a whole other show.
Juan Sequeda [00:04:34] Honest, no- BS, Krystin, what do you mean by the power of collaboration, which is the title of our episode today?
Krystin Kim [00:04:44] The power of collaboration. Working through everything from self- service to our data catalog implementation, and working with the communities around data, what you really find is what I call the undervalued value, and that's in collaboration. Everybody is focused on the ROI of a use case, and I absolutely believe that you need to mine for your gold and make sure that you're doing things that make sense, but at the same time, the undervalued value is what actually helps you get there. In a data catalog, what happens is you end up with a treasure trove of all of those individual use cases, and now you have somewhere to go to get that use case and exactly how to build it. You can stumble upon an idea, you can amplify your own idea, you can amplify others, and each one of those, that stumble upon moment and that enlightenment, is something that you really can't put a value on, but is really the value of the data catalog.
Juan Sequeda [00:06:01] I think one of the things that we were talking earlier was having these, I like that, you stumbled upon moments, but then also, you have the whole technical conversation that we don't get bogged down into all these features. It's like, yeah, you need these technical features, but sometimes you're like, that seems to be like the ultimate goal, is to get these features implemented, and it's like, but wait, wait. These are just means to an end.
Krystin Kim [00:06:23] Exactly.
Juan Sequeda [00:06:23] You had a rant on this.
Krystin Kim [00:06:27] Yes. From a rant perspective, when people are talking about value, people talk about what I call the what, the why, and the wow, so what are we doing, why are we doing it, so oh my God, we're going to miss this opportunity if we don't do it, that's why you're doing what you're doing. We're in business for a reason. The rant is, is where you start to go sideways is that when you talk about how do you do it, people get caught up in canonical models and taxonomies and glossaries and lineage, and actually, nobody cares about that part. My data architects are going to say, " Yes, Krystin, I care," but really, nobody cares. At the end of the day, you're focused on what are you building or what are you bringing to the table with it? Really, being able to show, being able to see what other people's are creating, and the fact that, oh, somebody already created this. I don't have to start from scratch. I can do this and make my idea better, and start so much quicker, accelerate so much faster, be able to react to something and say, " Oh," and we have that all the time. An example is with our self- service program, with our data fluency program, somebody created a set of dashboards that actually ended up, with the dashboards and process, they ended up driving almost$ 4 million in a quarter in lift to our revenue. Nothing to sneeze at. What happens if another customer team takes that and amplifies it because they found it in the catalog, and another customer team and another customer team? Those are the things that you don't really add up, but that's that untapped value that is created by collaboration, and it's created by the catalog itself.
Tim Gasper [00:08:29] I love that. How do you bring more visibility to that, though? Because I think sometimes when we talk about collaboration and things like that, it feels like the soft stuff, whereas even though I think you're right, there is stuff that is kind of the how and the what that people obsess over a lot, but they can really see it. They're like, " Well, I have lineage. I turned it on. It's working now." It's very visible when it's on and when it's working and when it's not, but the collaboration piece is a little more intuitive and it's a little softer. How do we bring visibility to that so people do really value it and see it?
Krystin Kim [00:09:05] I think it's storytelling, just like it is to show data value. When you're showing the value of the data, you're not saying, " Empirically, I'm looking at this math equation to use this algorithm to get it." You're telling the story of the data, and to just tell the story of collaborations value and the softer skills, I think is telling the same story. One is making it more relatable. Kind of a little bit of a left turn to your question for a second is being a food company, I used an analogy to help make the data catalog a little bit more relatable. I had a really difficult time explaining it and what it can do, but if you turn it into an analogy, so for example, I think data is an ingredient, and one of the ways that a data catalog shows up is it's the grocery store. We sell a lot of cereal, Honey Bunches of Oats, but think about walking into a grocery store and being able to pick up a piece of data ingredient. You take a little customers, you're going to throw it together with some sales, and a little bit of inventory, you've got yourself a forecast. You don't have to worry about how fresh it is, because the data catalog tells you exactly how fresh it is. You don't have to worry about it being there or not, because you just found it just on the shelf. It's just there for you, waiting for you. You take these ingredients, you go to your kitchen, which is your enterprise platform, it's your database, it's your visualization tool, it's your ML tools, whatever, and then you go and you make your model, which is your recipe. It tells you exactly how to put it all together so that you can make your meal. And then the data catalog has lots of models, lots of assets, so it's really your cookbook. At the end of the day, it's as priceless as Grandma's cookbook, and you're sharing those recipes of exactly how you do it. You get a, I'm going to add a little bit more sugar, I'm going to add a little bit more salt. You're going to add your own secret sauce, you're going to amplify the idea, and make something even better.
Tim Gasper [00:11:15] I love that analogy.
Juan Sequeda [00:11:18] I think many people use this analogy about the kitchen, but one thing that is a bit subtle, but I think I want to just really highlight this, is that you're cataloging all the ingredients and you're cataloging the recipes, you're cataloging the result of the recipe, the meals, and those go back in, and then maybe somebody else is going to say, " I'm looking for this recipe and this ingredient that I want to mix, or an existing meal and add." It's like you are able to now make all these possible combinations around that, and it's not just about only ingredients, it's not just about... Because I think the other point is sometimes we create these meals and we create the meal, and that's it. What if somebody else wants that same meal? Well, then, you go and figure out how to go create it again. I'm like, we're reinventing the wheel over and over again, and so I think part of taking this analogy is like, everything needs to get in your catalog, not just the ingredients, but also, all the meals that are being prepared. It goes back to the whole circle of your catalog, what I always bring, your catalog data, your catalog knowledge, all this stuff needs to get connected.
Krystin Kim [00:12:24] It needs to all get connected. You just made me think of something else, Juan, like what do you do with your leftovers? No, just kidding. That's a whole different-
Juan Sequeda [00:12:33] Well, you do... This is another thing, Tim, that we do on the podcast, is that we come up with these analogies and then we keep them going, or these metaphors, and see how long we keep them going, so let's see how long this one is.
Tim Gasper [00:12:44] I'm already thinking, you need to put it in the data compost.
Juan Sequeda [00:12:49] We do need all of that, but before we get there, we need to have the leftovers catalog, because you can only keep them for some amount of time.
Krystin Kim [00:12:54] Well, there's definitely the leftovers in doing it, but what I was actually going for was what I call sweating your data assets, because you actually are buying data now, too, and it is expensive, not expensive, but you have to curate your data, you've got to keep your data. There's a cost associated to data. I call it the hidden cost of data, but there's a cost associated to it. If you're going to spend money on something, you want to sweat that for as much as you can. You want to use it in as many different ways as you can, so that you get the most out of it. If you don't make it accessible to everyone or make it available and just kind of market it, really, that it's available, then that's how you get the maximum value out of it again.
Juan Sequeda [00:13:39] More juice for your squeeze.
Tim Gasper [00:13:41] More juice for the squeeze. Oh, go ahead Krystin.
Krystin Kim [00:13:46] I was just going to say, in the consumer products business, consumer behavior is really important data. What are people buying at the grocery store? What else are they buying at the grocery store? That information is for more than just marketing. The demand planners want to know how much is being sold in the stores to help us plan our inventory. Finance people want to know. Everybody wants to use it, but without knowing it and seeing what they've done with it.
Tim Gasper [00:14:15] The storytelling, the marketing, the communication around all this, obviously, is very important. One of the things that you told Juan and I as we were prepping for this session is around that people kind of miss the boat and can get really boring if they're just thinking of this as some kind of a Dewey Decimal system for data. Why is that the wrong thing to think about, and how do you change your frame around that?
Krystin Kim [00:14:39] Because name a businessperson that knows what a canonical model is. Nobody cares. Nobody cares how it was really made. In fact, Juan and I were just talking about how do you measure the ROI? I would say the exact wrong way is get really hung up on exactly how to calculate the precision exact right answer for ROI. It's really around what productivity, revenue lift, cost avoidance, did the business achieve? Really, not the true ROI of one particular technology, so taking that analogy a little bit further. I heard somebody doing A/ B testing to be able to truly value the cost or the value of AI, and I'm like, if you want to know the value of the project and what the business value gave to you, you don't really care if it was AI, in particular.
Juan Sequeda [00:15:45] That's an honest, no- BS take, because you would think, oh, as data people, we want have all the data. I want to be able to show all the evidence that this is the right... I'm going to give you a number that I can back it up, but at the same time, you've got to be practical about this stuff. The practicalities is how much is this actually increasing productivity, how much is it making us money, how much did we reduce costs and so forth? But for that, we need to then really have, it's not just the baselines, but we need to understand how the business is working. I think this is the other thing, is that we have all these disconnects between data teams that are not really understanding how the business works, which I want to connect it back to the storytelling, is how do we become better storytellers, then?
Krystin Kim [00:16:28] I'm going to go back to the way that I tell my team. First, you focus on the what, the why, and the wow. It's not just what we're doing, but why we're doing it, but really, it's that last piece, is the wow. We say it in a funny way within my team, and I do it with a Paris Hilton meme and I say, " I don't get out of bed for less than a million dollars," because really, it's about how are you going to spend your time? You want to spend your time on the most needle- moving, impactful ways. You don't want to spend your time saving one person an hour a week or something like that, so you really want to spend your time on the most impactful ways.
Juan Sequeda [00:17:18] Tim, we've got two t- shirts there, the what, the why, and the wow, and I don't get out of bed for at least a million dollars.
Krystin Kim [00:17:28] That's how you focus on value.
Juan Sequeda [00:17:30] I was looking at our notes here. One other thing we were talking about, the army and the gold miners. You were talking about this Levi Strauss model. Expand on this.
Krystin Kim [00:17:49] Data teams do an excellent job. Getting on a project, you figure out what the project is, what is the goal? You're going to do a demand forecast or something, use a predictive model to do a demand forecast, customer segmentation, whatever. Each use case is hunting for gold. You're looking for value. Again, we always measure our projects and the value that we're realizing either in revenue lift, cost saved, cost avoidance, or efficiencies gained. My team returns tens of millions every year back to the bottom line. But one thing that we started to do as we matured is taking the army and the gold miners approach, and it's really, I call it the Levi Strauss model. I'm from California, and in the Gold Rush times, Levi Strauss didn't go mine for gold himself. He put the tools, and the jeans, in the hands of the people that were going to mine the gold, and who ended up the richer one? At the end of the day, Levi Strauss is the one that really created the empire. Maybe Magic Johnson said it even better today. He said, " It's not just yourself being successful, it's that it's helping everyone else be successful," and for us, helping everyone else in the business be enabled to be successful using data to drive value for the company.
Tim Gasper [00:19:21] I love that. One thing that I think is unique about what you just said there is when you said, just from the beginning, when you said cost avoidance, revenue created, efficiencies gained. Your team contributes tens of millions of dollars back to the bottom line. I think that folks don't always think about that. I can tell that you're naturally thinking about the ROI, and I feel like data teams are not thinking enough about that and working that into how they're talking about the work that they're doing. I guess something's different. Is it a mindset shift that folks need to embrace?
Krystin Kim [00:19:58] It's probably a mindset shift, but who better to do the analysis and tell the story than data people about the value that was just created on what they just did?
Juan Sequeda [00:20:12] But I do feel that there is a large portion, I think it's large, that just focus back on these features, on the tech stuff, and then it's like, the data engineers have to go deal with all these pipeline issues. It's like yeah, but we're not seeing the bigger picture. You're being very operational about these things. We were talking earlier like, well, this is the big Gartner IT part and there's all this infrastructure and security, which you need to have these things, but that's like the cost of just doing the business, but what we're discussing here is being strategic for the business, and I still see a disconnect. This is my hot take here. I would argue that there are teams in different organizations that you're like, you're just an operational versus you're actually being strategic. You are actually thinking about, well, you just bringing up this Levi Strauss model, you're thinking about the storytelling, while other folks, they're not. They are waking up, and not for a million dollars.
Krystin Kim [00:21:25] I call that when data scientists get lost in walking in their random forest. A little data science humor there.
Tim Gasper [00:21:36] That's another t- shirt right there.
Krystin Kim [00:21:41] I guess, at the end of the day, I never asked to be the leader of the team, but I'm really grateful for the opportunity to be a part of a shift. Everything that we're doing in data right now and technology is a game- changing shift. I think that our job as leaders is to focus on what matters. At the end of the day, I'm in a Fortune 500 company, and we're here to increase shareholding value, improve people's lives. What is the goal that you're trying to accomplish? The how is super fun and how is super important, but you've always got to remember the why and the wow.
Tim Gasper [00:22:30] I like the what, the why, the wow.
Krystin Kim [00:22:34] The what, the why, and the wow, and then you can talk about the how, maybe the how and the who.
Tim Gasper [00:22:40] Get the three Ws in first. You talked about the Levi Strauss model, and I love that kind of analogy or that sort of example of how it's not just about the gold itself, it's like how do you enable. There's even more value, in some cases, in the enablement that there is even sometimes in the folks you're enabling. How do you enact that in an organization? What's an example of something around the Levi Strauss model that we could point to as like, for those who are listening, either they're part of a data team, they're managing a data team, what's a way that they could be leveraging that concept?
Krystin Kim [00:23:20] Well, then, you're going to take a turn into the culture conversation. Everything around that is around the culture, and it's not just a data culture. Back to your value question, it has to be really the company culture, and that you have a place to innovate, to collaborate, and inaudible. I think about culture, culture is really three things. It's enablement, having the right tools to do your job, to hunt for the gold. Number two is really learning and knowledge, like teaching the smart people how to use the tools. It's also about teaching them how to know more about the business process itself. If you don't know anything about the business process itself, the tools aren't going to help you. Also, it's about learning to tell the story of the business process with the tools that you have. The last piece is really the collaboration piece. Collaboration is the sustainability of the first two. Collaboration is what makes it stick. Collaboration makes it be a part of the DNA, part of the culture, as opposed to just being one and done. It's those three working together that actually create the army and the gold miner. You asked for an example, I'll tell you about a story. It's kind of a case study, where we took someone into our data fluency program. It's the Vis Academy. I call it data fluency over data literacy, because I think it's so much more empowering and positive to say, " I am fluent in the language of data," as opposed to data literacy, which makes me sound kind of stupid, but-
Juan Sequeda [00:25:27] Just a quick pause. Thank you. Thank you. People hear me saying, I don't like... We should stop using this word data literacy because it implies that you're illiterate, and that's not true. I really, really like the whole fluency, where it's like, I am fluent in a language, or I'm not fluent yet, I'm learning to be. I love this, data fluency.
Krystin Kim [00:25:47] It's just positive and inspiring, I think, but anyways. People don't always know the tools, especially some of the data visualization tools are kind of difficult to learn. It takes a little bit of grit to learn a data visualization tool. But taking the time, and they go along a journey. We do a fantastic class. I'm really, really proud of the class that the team has built. It's maybe a 10- week course, and you start at the basics. You really just learn around the language of data, and you level up into actually how to build your dashboards and how to tell your data stories. You're learning about the business process. You learn about the certain data sets that you're interested in. You're not just learning about the tools, you're learning about the process, you're learning about the data, you're learning how to tell that story. Taking someone up that journey and seeing that light- bulb moment, just they've got to eat their meal, and they're like, " Oh, mamma mia, best pasta I've ever had." That light- bulb moment like, " This is amazing." And then they go, " You know what? I want to make my own." Then we've had people in our classes go on to create the most either basic dashboards that have changed business. Who cares how pretty it is, as long as you got the message across and we're able to use your dashboard or to tell a story to change something for the better. Like I said earlier, somebody was able to change for the better, and found these micro opportunities to increase our sales that translated into millions of dollars. Another one is able to, they come up with these dashboards that we hadn't created, and it's like putting those tools and that knowledge in the businessperson's hands who really knows it, and they're able to do it. Now, what we do in our collaboration, so we also have these community meetings, we tell the stories of what other people have done. The gold miners that we have armed, we celebrate them. We do a case study on them, we let everybody applaud them, and share what they do to inspire other people. At the end of the day, another one of those undervalued value is, at the end of the day, everybody wants to do a good job and they want to create value. People want to do a good job, they want to do something that helps the company, they want to see what they have contributed, and they want to be seen for it, and that's priceless, as well.
Juan Sequeda [00:28:33] That's part of the culture, too, the celebration of the... There's another topic that comes up about, is celebrate the wins.
Krystin Kim [00:28:41] Celebrate the wins.
Juan Sequeda [00:28:44] We've been talking very positively about things. In your experience, what happens when things, they don't go well, and why is it that they're going wrong and how do you fix that? Going back to a couple of your things, the what, the why, and the how. It sounds easy. Is it easy? I'm listening to this right now, I was like, it should be pretty good. The what, the why, and the how, let me go do this, the what, the why, and the wow, but when do you struggle to get to that? If you're struggling, what does that mean? I'm just curious, let's talk about when things don't go very well.
Krystin Kim [00:29:30] Well, good question. What I would say, well, first of all, we do have... I'm a, again, big sports fan, so I think that you either win or you learn. I truly believe that. Things might not always go perfect, but you learn from it and you get better every time. If you're going to fail, you can fail, but let's try to fail fast and let's try to not fail too big. If you do it fast enough, you're not going to fail too big. I think that for us, we are so focused on value, and part of being focused on value is speed- to- value. Sometimes you go so fast you might trip over yourself, and where we tend to trip out ourselves is project management and seeing all the angles. Project management is truly an art as much as it is a skill, and you have to really balance not overdoing it and slowing things down, but really having the path paved. That is, I think, difficult to do, that art and blend of how much is too much. That's probably a little thing, but a much bigger thing, I would bet that the number one cause of people for failure is really focusing on what grownups call change management. It really goes back to, again, is focusing on value, and focusing on your audience, is really the other thing. Who is going to benefit for the value, and what do they need to be able to ingest that change, to realize that change? Because you actually have to make the person who's going to benefit, you have to change their day. In order to change someone's day, you have to get in their mind and you have to really understand their pain points, you've got to understand what lights them up and what's going to drive them and be motivated to change their behavior, and then you have to make it stick. You can't be one and done. It's not Field of Dreams, where you just build it and people come. You have to go, you have to understand their mindset, you have to follow them along, you have to show how what they're doing is creating a value, you have to make them excited, and that's an art that not necessarily all data people are good at. A lot of data people want to focus on their models and their algorithms and whatnot, but change management is where you might actually have the hardest time.
Tim Gasper [00:32:31] I like that you said the grownups call this change management.
Juan Sequeda [00:32:34] What do the non- grownups call it?
Krystin Kim [00:32:40] I don't know. I don't have a good word for it, actually. I know.
Tim Gasper [00:32:45] The youngsters are like, " Man, it's politics," or something like that.
Krystin Kim [00:32:50] It really isn't. It's just paying attention. It's paying attention, and it's really focusing on people. People often just overlook the... And I hate it when people call other people users. I hate that term. I'm like, are we doing drugs?
Tim Gasper [00:33:15] I'll be honest, we even struggle with this at Data. world. Are they called end- users? Are they consumers? What are they?
Juan Sequeda [00:33:20] Do they call the business users, the businesspeople? You bookend everybody as this group.
Krystin Kim [00:33:26] Yeah. I think that it's the people that we're trying to help or the process that we're trying to change, but-
Tim Gasper [00:33:36] Do you have a phrase that you call those people? Is that like your community, your business community, or what do you call it?
Krystin Kim [00:33:41] I just call them people. I just call them people.
Juan Sequeda [00:33:48] You said something, there's the people we're trying to help and the processes-
Krystin Kim [00:33:55] That we're trying to change or improve, or something like that.
Juan Sequeda [00:34:02] There's all these little things that, for me, it's like these words, these small words that we happen that it generates a culture that we're all just used to this that we go back to the features and technology. Then I'm like, so we have to be data literate about this stuff and we have to do change management, and then I go, we have the users we need to go... But we focus on all these little things, and we really forget to zoom out on the big pictures about it. This is why it's such a refreshing conversation right now to know that there is this big picture out there, and we need to really understand that. But I'm now thinking, you said it earlier that you work for a very, very large organization, and then how much of what we're saying here is easier or possible in a large organization, and is it also applicable for small organizations? Do you think there are other things that are easier in small organizations that are harder in large organizations? How does this compare in different types of organizations by size?
Krystin Kim [00:35:11] Throughout my career, I've gotten to work at all sizes, from companies with 30,000, 40,000 people down to smaller startups and whatnot, and to be honest with you, the bigger the companies are, the harder it is, but that's another place where the technology and the way that business models have changed or organizational models have changed make it easier now than ever. What by that is, and I'm going to get really specific around data organizations, so data organizations, they always debate where should the CDO sit, which C, whatever? And it really, that stupid answer of it depends. But at the end of the day, the problem is with the way that typical corporations or corporations have always been structured functionally, the salespeople sit here, the finance people sit here, the IT people sit over here, how many football games are won by teams of all quarterbacks? You need discipline from every single area, all different disciplines coming together to take their skills together. What you can do with the data team is create these smaller teams that have the discipline that you need for the focus that you have. The sales team will have salespeople, but they might also have a data integration person, they might have a data scientist, they have a data visualization person, as an augmentation to their team. These smaller, I call them either focus teams or a tagger team, I think there's all sorts of names right now in the industry, but I think that that is the way you are able to bring smaller sets of teams, because it's impossible to hire data people right now, you need to spread them across the table, but keep the culture, disseminate that mindset, that thinking, throughout the organization, so now, you can still have that same thinking at every level. Does that make sense?
Juan Sequeda [00:37:36] Let me rephrase this, is that in large organizations, let's go back to your analogy, a football game is not won by just only having quarterbacks. That means that we actually need is the diversity folks, and that diversity is going to change with respect to what focus you're going to have. First of all, on the sales side, you're going to have more salespeople than on the marketing side, but maybe for some particular focus, you're going to have, I don't know, the same split of SMEs on sales or with the data side, but for marketing, maybe more marketing, whatever. It really depends on that focus, so I think that's an important point right there. The second one is you're talking about a very distributed, decentralized, kind of federated model when it comes to all these team structures. The question is, is there the central data team that actually manages and moves the data people, or do the data people live in one org and then they get moved to another one and they get restructured all the time, because that can also change within the culture, because the cultures of sales may be different from the cultures of marketing? Anyways, there's a lot that I just said there-
Krystin Kim [00:38:47] My honest, no- bullshit answer is that I lean into what they call, there's so many different terms for it, and I hate this term for it, but that center of excellence model, because you have fragmented or you have centralized or you have center of excellence. What you really want is as many data people as you can afford closer to the business, but at the end of the day, a data scientist in a sales organization is never going to be the head of sales. That's not a good career path for him. He is going to be maybe a head of data. While they would maybe technically report through a arm where they can have good mentorship and somebody to lead them to become a even better data scientist than they are today, but be out in the business, so that's where they can add value, and get their life skills, get their business skills, they can understand that process. But I think that's what good looks like, is having these teams of specialized skills sitting with the business, focused on the value and the problems that the business is focused on, but still having a reporting relationship to something central, so that their careers can be groomed and lifted.
Tim Gasper [00:40:12] That makes sense.
Krystin Kim [00:40:12] Not everybody agrees with me on that point, but that's what I think.
Tim Gasper [00:40:15] Well, I know some people argue like, well, should that person be a part of that group like sales, but then kind of dotted line into the central organization, or are they part of the central organization, and they're sort of dotted lining into that, or are assigned to that particular subject matter X area? Do you have opinion on one model that you like more than the other, or is that-
Krystin Kim [00:40:38] In my opinion, it's the latter.
Tim Gasper [00:40:39] You prefer the latter.
Krystin Kim [00:40:40] Yeah. But they need to be in every weekly status meeting, treat as part of the same team, because they have to know how you think, they have to know what you care about. They have to be able to finish your sentence, just like the other sales guy.
Tim Gasper [00:40:54] Really immerse yourself, kind of become part of the team, even though your reporting structure is also aligned to your career path.
Krystin Kim [00:41:01] Exactly.
Juan Sequeda [00:41:03] Who else would follow this type of approach, not just in data?
Krystin Kim [00:41:09] Finance people. The finance people don't just sit in their little accounting world. The finance people are actually the ones that are going out to the salespeople and truly understanding maybe the cost of sales, what gets you to net sales. In manufacturing, they really understand the cost of every ingredient, the amount of labor that goes into every widget that you produce. They're doing those analyses, they're doing pricing strategies. The finance function is the same thing. They're going out to that particular arm of the business and applying their secret sauce, their own discipline, on that business function in order to help that leader make decisions. It's the same thing.
Tim Gasper [00:41:59] That's a really good point. I don't think that's ever come up on this show before, how similar the finance organization is actually to the data organization.
Krystin Kim [00:42:08] Smaller, but yeah.
Juan Sequeda [00:42:10] This is actually a fascinating point, because sometimes we talk about this as like, oh, we're going to go do this thing, that the data is going to move there. Your point is like, we already do this for finance. This is nothing new. Why do we make such a big deal about it?
Krystin Kim [00:42:26] Well, I think data is at bigger scale. Finance is probably one person, and data is probably a bigger scale, so that's why it's probably a little bit it has a pucker factor.
Tim Gasper [00:42:35] That's a fair point.
Krystin Kim [00:42:40] I think, I don't know for everyone, but a lot of data organizations have started off in IT, and so they're thought of with the same lens as being IT. There's some things that are very close. We use a lot of technology. I have my whole platform team to support the technology. There's no finance person that has their own platform team. And then at the same time, we're a service, we're a support function. We're not monetizing our data right now. Some people do, but that's not the core business function. It's an add- on.
Juan Sequeda [00:43:28] We're going to start wrapping up here, but again, there's so much more you need to talk to. Let's go back to the question that you said, " Well, it depends." Where should data report to? Where should the CDO or the subs report to? What are the different options that you see?
Krystin Kim [00:43:50]I don't remember what the percentage is, but I'm sure there's percentages out there. What I have heard is is that in an infantile stage, they report through an IT organization. As a company matures with data, it often goes to the marketing organization, because of the power of data science in consumer behavior, so they tend to go there. But then what happens is that, I'm in a manufacturing environment, but the operations group gets a little bit jealous, because there's really a lot of opportunities for data science and predictive maintenance and line optimization. There's some really great use cases out there for data science, so they want their team, too. Ultimately, where do I think it needs to fit in? My opinion would be either a chief operating officer or a chief financial officer, or just being a seat at the table.
Juan Sequeda [00:44:52] Well, I'll give you the final word, Tim, but one thing is that this is a trend that I've been seeing, that moving to operations, at least that's the desire of it. It makes sense because if you're in the IT, you're like, well, you're the cost center. No, data's not just a cost center, just making sure that the reports are generated. There's much more than that. An ultimate goal is you have a seat at the table, you're reporting to the CEO, you're at the board meetings, but something in the middle is like, no, you are actually there to make sure that the business is operating effectively. It's operational excellence, so you should be reporting all the way to the COO.
Krystin Kim [00:45:32] You should be part of the revenue strategy, really.
Juan Sequeda [00:45:33] Yes. This is a shout- out for an upcoming episode that we'll have with, I was going to say Phil Williams, but Phil, you get the joke, it's Pete Williams from Penguin Random House UK, because actually, he just announced it yesterday here on his LinkedIn, he runs data for Penguin Random House, and he moved out of the CIO and he's now reporting to the COO. He's gone through that whole entire journey for years to make that happen. It's going to be an interesting conversation we're going to have with him-
Krystin Kim [00:46:03] I can't wait to listen.
Juan Sequeda [00:46:05] Tim, any last, we'll pass it on to you before we hit our lightning round... Well, we actually have the AI minute.
Tim Gasper [00:46:11] Oh, we have the AI minute, too, so let's not forget that. Coincidentally, just amusingly, I was actually having dinner with our friend, Pete, not Phil, from PRH, so funny coincidence. But the one last thing I'll say is that I think from a collaboration standpoint, you've really done a great job illustrating how to really focus on the people, and the what and the why and the wow, I think that's really helpful. Just as a final question before we move into our final bits here, how do you create some good carrots and sticks to really motivate what you're trying to motivate here? Do you have any tips on that before we move forward?
Krystin Kim [00:46:56] Again, I'm a sports person, so there's no better driver and motivator than a scoreboard, the scoreboard of how many dollars are you putting up for revenue? What did you do to lift revenue, or how many dollars did you save in costs? Really, and then again, celebrating those wins, so scoreboard.
Juan Sequeda [00:47:18] I love that, scoreboard, and then it forces you to actually start quantifying this. Let's hit our AI minute. You have have one minute to rant, say anything you want about AI. Go.
Krystin Kim [00:47:32] You know what bothers me about AI? One is people are getting fatigued about it. It is awesome. The power is game- changing of what we can do with AI. Two, people think that they're all trying to say, " This is AI and this isn't AI," and trying to maybe oversell it or undersell it. At the end of the day, who cares? Nobody cares what it's called. What can we do together? What can we build, what can we create, and what can we solve that we didn't get to solve before? I think that's interesting about it. I don't know. Not a full minute.
Juan Sequeda [00:48:17] No, it was good. I love how practical you are about... It's like, the goal is to solve things together. It's kind of obvious, and it is obvious, but sometimes we forget about it.
Krystin Kim [00:48:28] Isn't this the no-bullshit talk? So, practical.
Tim Gasper [00:48:31] How do you not get pulled into the pedantic stuff? That's the fun stuff to argue about. How do you avoid that?
Krystin Kim [00:48:39] Oh, I don't. I can talk data at the deep and, I don't know, technical as anyone else can, but at the end of the day, the referee is what are we doing, what are we trying to do?
Tim Gasper [00:49:01]You're saying in our next episode, we could make it about the importance of canonical models.
Krystin Kim [00:49:07] Maybe you should get someone else for that, though, or a lot more alcohol.
Juan Sequeda [00:49:16] Oh, there are so many follow- ups on this one. Lightning round. I'll kick it off. You mentioned a 10- week course when you're developing your data fluency training. Do you target the broader business for this, or is it more focused?
Krystin Kim [00:49:35]It's the broader business. What we found is, first of all, we created it because there was a demand, and then so we let people in, but what we found was that when people had a purpose and they had an intention of what they wanted to do, or they had an expectation, their managers expect them to do a function, it stuck a whole lot better. Now, what's happening is there's too much demand for how many classes I actually have people to support, so what we think about is really intentional now, like what area of the business has either the most pent- up demand or has a real opportunity to capitalize on a project that they're doing to, to append a class onto it, as well, to make that initiative, project, what have you, even better?
Tim Gasper [00:50:33]That makes sense, really focus it. Second question. Today, we talked about the power of collaboration. A lot of collaboration happens very organically, perhaps haphazardly, in places like Slack and Teams. Is this a good thing?
Krystin Kim [00:50:50]I think it's a great thing. I think that Teams and Slack are great. We use Teams quite a bit. I love it. But again, I think that the distribution channel for data and the culture of data is the data catalog. I think no matter where you are, remote, hybrid, whatever, nobody cares. You're distributed, anyways. Very few companies are all only in one spot, so you have to be able to create a place where you can find what you want at the right place, right time, right way, and also be able to contribute, so that's key.
Juan Sequeda [00:51:34]Next question. Is AI going to help data collaboration or harm it?
Krystin Kim [00:51:41]I am an optimist, so I'm going to say help. It's going to level everything, level it up. I'm really excited for some of the pedantic stuff, just the code generation, like being able... I feel like I'm going to be superhuman, with what you're able to accomplish now that would take you a long time to do before, so I'm looking forward to my new superpowers.
Tim Gasper [00:52:09]They're already here.
Juan Sequeda [00:52:09]I agree with that. Tim, take us away.
Tim Gasper [00:52:14] No worries. We have some solutions people on our team that are using code generation, and they're doing stuff in 10 minutes what took three days. It's crazy.
Krystin Kim [00:52:22]It's insane. I can't wait.
Tim Gasper [00:52:24] Final question. Self- service was also a theme today. Should everyone in the organization be empowered to create reports in dashboards?
Krystin Kim [00:52:36]Everyone who wants to. I guess I don't know about everyone, because that's a really, really broad word, but I don't think everyone has to. I think that, let's face it, no matter what AI or what tools we put out there, it's going to take some amount of grit. You're going to have to want to do it to be able to use it. It's a power to be used for good or evil. When you go to the everyone, then I don't know if we're going to take a term on data ethics, but you should be using it to create value, to make the world a better place, so being able to use it for good.
Juan Sequeda [00:53:24]Well, Tim, it's takeaway time. Take us away with your takeaways.
Tim Gasper [00:53:30] We started with what is the power of collaboration? Krystin, you really started us off with it's really the undervalued value that matters the most, and it's that when you've got a catalog, for example, you've got this treasure trove of all these individual use cases, you can stumble on things, but the true value is really the collaboration, it's the connection, it's the moments, it's the stories that come out of leveraging these tools, these data processes, doing these data activities. Things like lineage or specific features, these are really a means to the end. It's the undervalued piece, it's those moments, it's that collaboration, that matters the most. When people talk about the value, it really has to be about the what, the why, and the wow. I like that, because I think a lot of people say the what, the why, and the how, and I like how you kind of flipped that last piece, like before you even get to the how, let's start with the wow. I know we've got a really fun saying, which I'll mention a little later. I won't jump the gun on it, but you really have to focus on the business value. You went on a little bit of a rant where you said you can go sideways, where people get really stuck on the canonical models, the taxonomies, the lineage, all these more technical or more specific or the pieces that we can argue the technical aspects around, but ultimately, are those the pieces that are creating the value? We have to focus on the value. You mentioned that you have programs that are creating multiple millions of dollars worth of value for the organization and dashboards that are having multiple millions of dollars of value. That's where the focus has to be. That's where the impact has to be. How do you show the value around collaboration? You said it's around storytelling. Make it relatable. Use analogies. You used a great analogy, which is around the kitchen, like data is the ingredient, the catalog is like the grocery store. You've got to get the data ingredients, you've got to mix them together, you've got recipes, and then you created a forecast. You baked a forecast, good job. Go to the kitchen, go create the model, create recipes, and try to create repeatability. I think this is really awesome. Find those analogies that resonate with your organization, because every organization's going to be different. Don't sweat the cost of data, get more juice for your squeeze, focus on measuring ROI, and try to help your organization, the data team and the broader organization, become better storytellers. I think that you had a great quote, which is that Paris Hilton said, " I don't get out of bed for less than a million dollars." That's the mentality that we need to have across our organization, don't get out of bed for a million dollars. Focus on the stuff that makes a difference, focus on the wow, focus on the cost of avoidance, the revenue created, the efficiencies gained, that's going to be the difference. The last thing I'll say before I'll pass pass to you, Juan, is that, Krystin, you mentioned the Levi Strauss model. He didn't go mine the gold for himself. He put the tools and the jeans in the hands of the people that wanted to go mine the gold. If you have that kind of a mentality, an enablement mentality, you're going to be even richer than those gold miners, and so that's a great way to think. Juan, what about you?
Juan Sequeda [00:56:42] Well, I'm going to start with another good quote for a t- shirt that says, " Don't be the data scientist that gets lost walking in the random forest." We got like four T- shirts out of this conversation today. When talking about culture, three things you said. One, you need the place to innovate, and make sure you have the right tools to do your job. Second, it's about learning and knowledge. Not knowing about the business processes, the tools aren't going to go help. You need to know things. You need to learn how the business works. Collaboration is what makes it stick. It makes it part of the DNA, instead of just a one and done. Second, I'm so glad we're connected on let's not use data literacy, let's use data fluency. It's more positive. It's inspiring. In your case, you talked about a 10- week course that you have where people can learn about the data sets, the process, the tools, learn how to tell that story. They cook their own meal and they eat it, and they're like, " Oh my god, this is great. I want to keep doing this." And then people will come up with dashboards and they will find micro opportunities that will save millions of dollars around this stuff, and then those are the wins. Those are the gold miners out there, you're going to go celebrate them. That's actually part of the culture. I brought up like, these all look great, but what about when it's not great? I like how you said, " It's either you win or you learn. " That's another one that I really like, another good quote there. Fail fast, not too big, and if you are failing fast, then it's not going to be that big, so I really like that point. What else is hard there? Project management is hard. It's an art and a skill, so you've got to be careful about overdoing it, but you still need it. Other thing that's hard is what grownups called change management. I think that's another T- shirt right there, grownups call this change management. Change management, effectively, is just paying attention and it's about the people part. What are we trying to do? Who are the people we're trying to help? What are the processes we're trying to improve? Who are you helping today, what processes are you improving today? That should be on everybody's, like you wake up or you go to the office or whatever, you turn on your computer, that should be right there in front of you every day. Ask yourself. Then we also talked about what are differences between big companies and smaller companies, and how is this being set up? You said, " Bigger companies, it's harder, but the way that organizational models have changed make it easier now than ever." I think another thing that you brought up is how many football teams do you see that win that are just full of quarterbacks? That doesn't happen, and that's why you need that diversity. You need data teams that have rich experiences that are going to push with the sub- teams, and that diversity is what brings it together. Specialized data skills are going to be sitting in the business, but you also want to have some reporting relationship to some center of excellence, so they can keep having that mentorship and growth. This isn't just unique to data. This happens already in finance. The finance folks, they need to understand the cost of everything. But one thing to acknowledge is that for finance, it's smaller, and I think with data, it's going to be in a bigger scale there. Eventually, you just want to apply your secret sauce to every business unit. Wrapping up with, even though we had the question about, oh, who do you report to? It depends, I like how you gave this kind of growth pattern. From an infant state, you're in the IT. Then you start growing, and then you probably end up going in to marketing because you focused a lot about consumer behavior. Then operations get jealous and you go there, and eventually, you probably get the seat at the table. That's where that needs to go. We wrapped up with how do we motivate the change to tap into this power of collaboration that we started off with? Following the sports analogy, just implement a scoreboard, quantify it, put it there, and just celebrate those wins. How did we do? What did we miss?
Krystin Kim [01:00:15]You guys have quite that talent. I don't know. We were paying attention.
Juan Sequeda [01:00:19] Well, this is you. We're going to wrap it up very quickly. Three questions.
Krystin Kim [01:00:25] Uh- oh.
Juan Sequeda [01:00:26] What's your advice about data, about life, who should invite next, and what resources do you follow?
Krystin Kim [01:00:37] My advice is stay curious. I'm sure probably everyone has said that, but my favorite people are lifelong learners and people that are constantly curious, never feel like they're finished, so stay curious is my advice. Who should you invite next? My friend, Laszlo, at Data Storytellers, I think he does a great job about connecting people and bringing ideas to life, so you should meet my buddy, Laszlo. He's an amazing, amazing thought leader in the data space. What was the last question?
Juan Sequeda [01:01:12] What resources do you follow, people, conferences, books, magazines?
Krystin Kim [01:01:17] Well, here at Gartner, so that one's easy, that's a layup.
Tim Gasper [01:01:21] A layup. You're awesome.
Krystin Kim [01:01:24]I follow the IBF, which is forecasting. I think it's like, I forgot, Institute of Business Forecasting, or I forgot what it stands for, but it's really around demand planning and business function. They really do a great job of putting up... They have a podcast for forecasting, that space. I also listen to The CPG Guys, which is for the consumer product goods area, and of course, I listen to Catalog& Cocktails.
Tim Gasper [01:01:52] I'm glad that you could be on it.
Juan Sequeda [01:01:58]Krystin, this was such a pleasure.
Tim Gasper [01:02:00] So fun.
Juan Sequeda [01:02:00]I'm so glad we got to do this in- person here. Tim, enjoy your evening in London. Next week, we have Dr. Mike Dillinger, who formerly set up all the knowledge graphs work at LinkedIn, and he's been posting so much about knowledge graphs and large language models. You should definitely be following him. He's going to be on the podcast next week, and that will be fantastic. I will actually be in DC next week having another Honest, No- BS Dinner, because Krystin and I, we're going to our Honest, No- BS Dinners that we have later tonight. With that, Tim, see you next week.
Krystin Kim [01:02:36] Goodnight.
Juan Sequeda [01:02:36]Bye, everybody. Thank you so much.
Tim Gasper [01:02:38]Cheers, everyone.
Juan Sequeda [01:02:40] Cheers.