SELECT video_content FROM LIBRARY WHERE SLUG='003-she-trained-20-000-data-analysts-here-s-what-they-got-wrong-w-mico-yuk'
003: She Trained 20,000 Data Analysts: Here's What They Got Wrong w/ Mico Yuk
Loading video player...
Hey, everyone. Welcome to More Than Numbers Live, a data show focused on turning data teams from support functions into engines of growth. I'm Ollie Hughes, one of the cofounders of Canva co Canvas based BI tool. I will have privilege every week of interviewing a new mind in the industry to learn how they're turning their organizations into value engines. And this week, I'm really pleased to say we have Mico Yook. Mico is a multi time founder. She's a keynote speaker. And most importantly, she's a trainer of twenty thousand analysts in how to improve their reporting and how to drive more value to their organizations. She's also a dear friend. And, also, she has the same birthday as me. So hello, Mikko. Amazing. Amazing. Hey, everybody. And you forgot that I was also an analyst and data scientist when I started. Sure. Yeah. Yeah. You it's coming from roots. It's got a lot of roots. In fact, that's a good first question I love to ask, guests when they first join. Like, what was their first BI tool they started loving or using other than Excel? So maybe you can give us a bit of a flavor where you started. Alright. So this is gonna age me a little bit, but my complete obsession when I started was Excel, obviously. But then I bumped on a tool called Excelsius. That tool was then acquired by Business Objects, which then evolved into SAP Analytics Cloud. That's what people know today, but it was one of the best visual tools on the market. That does age you slightly. I well, actually, I haven't heard of it. It sounds like one of those tools that you find, like, deep in hedge funds still, but it doesn't it's like like, it sounds like it's a thing you downloaded maybe and installed on your desktop when you first started out. Is that aging you too much? Well, what happened is when they kill when Steve Jobs killed Adobe Flash, it killed the tool, but it was a fantastic tool. Really? Okay. Cool. Right. That's awesome. Well, thank you so much for joining us. We're gonna get more into, like, the what you've learned from twenty thousand interviews and training sessions on analysts. But first, again, we also have to ask a bit more about the context, like what you're who you're dealing with, what data tools you're dealing with. So we call it, like, what's your data stack? I know that you're slightly different because you're training a variety of different people and different tooling. But maybe you can have a sense of, like, what are the common tools that you're seeing that you're working with people on regularly? And that'll give a flavor of, like, the tools that there may be people dealing with as well. They can see how you're thinking. Yeah. Absolutely. And so, you know, we did the when we did the initial recording, I put high level, but I went back and did some research. And what we're seeing a lot of when it comes to data warehouses, as you could imagine, is Snowflake. Right? We definitely are seeing some Amazon Redshift. We definitely are seeing still some BigQuery, right, from Google. When it comes to more orchestration, we see a good bit now of DBT and Airflow. So I think everybody in this audience is probably shaking their head. Data ingestion, lots of Fivetran. And then from a BI perspective, we have the three stooges. Right? Power BI, Tableau, and Looker. The three stooges. I love that. I might use that internally at Cowder. I'll come back to my major incumbent competitors. I'm sure they'll love that phrasing. That's awesome. Thank you so much. Okay. That's wonderful. Let's dive in. Like, I want to make sure that we're talking about value. And as you know, Mikael, I've shown this this kind of framework to you many times over the last year or so, the idea of value for data teams being the four tenants, that there's four ways a data team can drive value, operational clarity, signal from noise. That means becoming problem solvers, time for decision, how to collaborate better with the business, and then measuring yourself to make sure you're driving as much ROI as possible. This is the way we like to frame our interviews. I I wonder if we could just talk about what back to, like, the core lessons that you've seen. You've you've trained thousands and thousands of people in analysts who are coming to you to ask for help. Maybe to start off with, like and we can frame it in the tenants. Like, what are the problems they're coming to you for? Why are they getting trained by you? What is it they're looking to solve? Yeah. So I see three core problems, Oli. And I and by the way, you know, I love the tenants, for count. One of the main problems we see is that, typically, we have data leaders that are coming in, and they have teams that are asking the wrong questions to the users, which then leads to the wrong answers, which then leads to KPI metric overload, which then leads to producing stuff in production that no one uses. So that entire I call it the epidemic. The entire epidemic is the reason why we have we trained so many analysts. The second core problem we see, which is mind blowing two thousand twenty four, but I guess the field we are just, you know, technical people, is that to focus on tools. You know? Again, the Three Stooges, Tableau, Power BI, Looker, so fascinated with the tool that that's the focus and that doesn't work with users. And then I think the last one, which really, really has been an issue but hit a real hedge, And I will say this, it hit a real hedge with Chargebee T coming out is critical thinking and not asking at least seven times why why why. And the reason I bring up Chargebee T with that is because you and I know tools like AI and Chatubiqui and Cloud or Gemini are all great, but the answer is only as good as the question. And the thing is, as and I used to tell my user this, that you can get AI to lie. Okay? You can literally ask a bad question and get it to light. And I think if you think about the way that AI works because you can manipulate AI. You can ask a question a certain way, and I've gotten it to say and apologize and said, I'm sorry. That was incorrect. I, you know, I I wasn't I wasn't truthful. And so I think that what we see with users now is we need very, very good critical thinking. I love that. I agree. Those three points are really good. Let's go into more detail. I love the way he described the epidemic of, like, the wrong questions, the wrong answers. At Cali, we talk about it as the vicious cycle rather than the virtuous cycle. It's like the vicious cycle that the more dashboards you have in the business, the more confused everyone is, the more they ask for another, question to answer what's going on. That just creates this feedback loop of just charts and metrics and dashboards everywhere. And it just means there's massive clutter, and it just leads to that kind of feedback loop of negative well, of production of stuff. Is that kinda what you're meaning by wrong question, like, Lyric? Yeah. Absolutely. You know, data leaders come and they're like, we are, first of all, shorthanded team wise, and you notice from, you know, call in the canvas. So we don't have enough people to keep up with the queue. But on top of it, they suspect this queue isn't as valuable as we think. Right? Everybody is coming to ask for everything, and my team doesn't understand what questions to ask them to really validate what makes sense. And so not to plug count, but that's why I like the canvas. Like, the one good thing about cone canvas is that it gives it forces you to put your thoughts on paper and on a on a in a visual way. Right? And it forces you to let people know, hey. What are we trying to do before we get to the visuals? Because you can also do that in cone. What are we trying to achieve, and let's have the conversation? Yeah. I I love that. I I think people often say it's not our tooling, and that's correct. But it is about workflow, and the key is you have the like, getting humans to work the right way together is a thing worth getting a tool for. It's It's not like getting more technical features. In fact, in fact, actually, I only just wanted to ask you about your training. Like, how much time you spent teaching them, like, visualization techniques and how to format charts well? How much of it actually is mindset? I bet it's more So you know So you know what's interesting? When I first started this about a a little a little less than a decade ago, people would come in and it's like, we have to do a full day of visuals. Right? Like, it was just a visual to visual. Fast forward to where we are today, literally in a three day training, two and a half days is spent on storytelling, mindset, asking questions. You talk about workflow. I have a storyboard that they use, which you've seen because we have I think we put one in the Canvas and count, and they storyboard out it. The visuals are less. We probably spend less than half a day in visuals, and you want me to be honest, we work with enterprise clients like Costco, like Shell. A lot of them tell us, don't even do the visual. I gotta get my analyst mindset correct. It's it's it's one of those really difficult things because there is a huge value in visualizing well and making sure that there's no block interpretation. But, actually, if you're optimizing your time there overthinking through the the story arc of, like, how to make the clear argument and to make someone understand the point you're making. It's the you can emphasize that at the detriment of more valuable levers you can use to drive drive impact and solve problems. So I think that's you've just told that story very well. Actually, the emphasis is such a good one. Can we talk about the critical thinking? I think the whole point of critical thinking is key. Right? That's the that's the the thing when you have a analyst AI agent that's spewing out charts at someone's whim. The the real value of an analyst is gonna start becoming more about how they wrestle with the business, actually understand the problems, and form rational arguments. The technical skills are gonna go away. So, like, what do you how are you helping people come to that? Like, how do you diagnose their level? Any sort of suggestions around that? Because I think that's such an important it's a good way to simplify the problem down. Like, if you can critical think and you have technical skills and you can still tell the story, you're gonna drive value. So it's interesting that you asked that because a show that I did last year, I answered this, and I think it's a little bit controversial. Someone asked me recently, and then I also advise startups. Right? So I I had a founder in France ask me this recently and said, what's the future of all this? What is data storytellers and data analysts gonna be doing? You know what I said to them? I said, in our data field, we need to prepare for the fact that users are not gonna come to us anymore to say, hey. Could you build my dashboard or fix it? They're gonna come to us and say, hey. I put in these prompts into the AI in our tool, and I can't get what I'm looking for. Could you help me fix my prompts? I think that that's where we're heading, and that requires a lot of critical thinking. Because as someone who uses chat, I feel that I don't know about you, Ollie, but I write, like, twenty prompts, and then I try to go back and condense them to try to get a single answer to understand. So I personally think that data analysts, one core component is gonna be with their critical thinking, being able to condense out proper prompts, well taught through prompts to get the outputs that the users want. The the reason I'm laughing is I I should tell you on this that Mikko is probably one of the most biggest early adopters I met. She is a complete fiend for new software, so I'm not surprised to hear it. They're already, like, optimizing prompts in whatever AI tool you picked up. You have so many every day. Single one. I have every I got mid dirt. I have all of them. I have quad, Gemini, nice, you know, sister. I got everything on the cocoa. Oh, yeah. But the point still stands that, like, the the key to framing logic thinking is the key thing here. And and, also, it's communication. Right? The human layer of this is so important that Trust. Trust, like, the ability to help someone. That's all I think people really struggle with is like and it doesn't matter if using AI or using a dashboarding tool like the, or the kind of three stages you mentioned earlier. Like, ultimately, there's a human who's interpreting information that's being given to them, and the better that can be translated, the better. And and that human should do that best always. It's just about how you get there. And that's why I think critical thinking and thinking through empathy is also really key. I love it. Well, another thing and, again, you didn't mention that I used to work for Compe, but I'm gonna bring this up again. One of the reasons that I actually was fascinated with the Canvas, right, BMI because I was considered to be a super analyst. Right? I worked for the feds and all this stuff. You know my history is that it forces you to work with the user and visualize your thinking. And I I don't know. I'm like a a fiend for that. Right? I really like the fact because if I go into Power BI, I can't really get, like, text and workflows on a screen. Right? I can get the visuals. And again, I and I said this earlier, what I really like with the critical thinking is it it forces that thinking and it makes it visual. So you're doing it live. Yeah. I I thank you. Your good plug. I love it. Wink wink wink. I mean, it's a true True. I know. But, right, like, the thing which I often say when we meet customers is they, like, the biggest barrier between your ability to as a data team to better help the business is the fact you've got a dashboard as a wall between you and them. You're building the and they can't see any of your working, any of your critical thinking, any of your structuring of a problem. They just see some charts, and you have to try and that's all you've got to influence, help them, is a read only read only dashboard wall that they can't see any workflow. So the biggest one of the many benefits of the Canvas is the ability to let your workflow become open and visible. And that's an incredibly powerful way of doing it. It's the transparency, and it becomes this agile feedback loop of, like, them helping you as you help them, and it's really, really cool. So And to explain it to my data analysts because I've had them ask me about count. In the traditional dashboard world of Tableau and Power BI, when people ask you why, you typically have to either create an additional drill down report or a pop up with an additional chart. I feel like what I like about this new way of working, and that's what attracted me to the canvas the canvas where the notebook world is that you have the why upfront, and that's built up first in a workflow, and then you can have the visuals right on the side or below it. So when a user comes back and asks questions like, how was this defined and how did you measure this and, you know, well, why did this happen? It's literally all sitting there. It is context rich. Agreed. Yeah. It's actually it lets you work a different way where you can you are building backwards. You scope the report or the problem, bringing the analytics into the same space, and then you bring you can start modeling to match the any modeling needs to happen to be done off the back of what you know is important to the business. And it's reverse and it's additive rather than being, oh, I do some scoping in a random place, and then I'm gonna go do some modeling, and it's it's completely broken workflow and breaks the workflow. So let that's also I I love that. Let's let's talk more about, like, what so when people come to you and they've got a all of the three stooges and they've got that vicious cycle, that epidemic as you call it, which is the kind of they've got to the point they've got all these all all these the wrong questions and the wrong answers you say. How do you unpick it for them? Right? Where do they start? Is it about burning to the ground and rebuilding, or is it about can they can they because most people are gonna be somewhere in that spectrum. Right? Do you help them Yes. To iterate there, or do you just come in and say, delete first, then we'll start again? So the first question I find out about is the relationship with the business. Right? Some people have burnt through the business so badly that, like, if they don't show up with a rabbit, it's over. There's no time to go back and act. Like, I have users tell me, we can't go back and ask more questions. Okay? So that's one scenario I'll get into. The second one is, oh, yeah. I work with Ollie. He's great. He's flexible. He'll talk again, so on and so forth. So in scenario one where you have burnt the bridge down and if you don't have a rabbit commodity, a hat like Las Vegas, you're out of the user doesn't want to see it, I recommend that you burn it down, you savage what you could get, and you focus on value and sexiness. Dashboard at a time. Right. Yeah. Yeah. Okay. One dashboard at a time. Just burn it down. Don't try to come back and ask only a clar there's no clarifying questions. But most users are amicable. So So in a second scenario where they're more amicable, I actually have them pull out our storyboard, and I I tell them, let's extract what we do have here. And what our storyboard does, Oli, is it shows all the gaps. Right? The like, you call it a workflow. It's my storyboard. It shows you all the gaps in your story, and then you can go back to them and fill it out and continue from there then to gather the data and build the visualization or report. Yeah. Yeah. It's based on it goes back to that trust. Right? What have you done before you got to me? How bad is it? How bad is it? Does Breeze make a good now the big guns or she gonna help you with the scalpel? There's a range of different tooling, different situations. Well, you know, I get so many scenarios where it's like the leader comes and they're like, they've hired a consulting firm, but they haven't closed a contract. We need help. So, you know, when they've hired a consulting firm, you know, only they're done. You guys are like, they're just done. The other thing which makes me laugh about this is, like, you've, you know, you've done twenty I I can't I don't wanna age you again, but you have done this for a number of, like, a decade, maybe more than a decade now. You've, like Yeah. A long time. And what I think is I think if you've been in the industry for a decent amount of time, one of the things that, you know is that and what's frustrating to many people I meet who have been in it for a decade or so more is these are the same problems that we had. We've had these problems as an industry for decades, and I think that's that makes it difficult. And I think we I think, it's a joy to see things like the tenants to help people actually understand what really matters because there's so many rabbit holes you can go down and so many ways people are trying to artificial impact by dashboard views or number of reports I'm building per week in all these terrible ways. And it's just wonderful have a community of color that's around the the tenants. But I also think it's really important to recognize that AI isn't gonna solve the problems. No. It make them worse, as well. I agree. I again, I'd like and then I wanna bring up one thing to what you're saying that, you know, one of the things that I really appreciate haven't been this industry for a while is remember when I started, I was hashing together a Google doc with this, a spreadsheet with with with this, another you know, ways of working have changed. Yeah. Like, you know, again, why was it talking to the Canvas? Ways of working have changed. Right? And so we I was in that scenario of having four different documents to where you scope here, you do the data validation here, and then you build over here. And then you do a report in the corner. You know what I mean? And what I like now with the tools, the way of working is changing. I think that also data analysts, data teams, they have to change their way of thinking and work issues. Yeah. Yeah. They need new tools. I think that that's so it is the big secret, I think, that we we believe account is that it's about the workflow, less about the function. The function is these You go back to Excelsius. You'll see that that had a lot of the functionality that most of the modern BI tools or three stages have, if I'm honest. The big difference of a modern tool, like like like CALT, I guess, what we're focusing on is what's the how to change with human to human interaction and make the workflow better, not just give more buttons to press. And that's really the but the core of value generation is the it allows a different conversation. It allows a better more a a better value generation. And, yes, we have all the buttons too, but it's just putting it into different workflow and a different encouraging different frequency and contact points between humans, and that's what really matters. Well, Holly. Right? The thing about con Canvas is that remember, I use whiteboards all my career. Count Canvas is like having a whiteboard with everything you need installed. It's like having it with your Power BI, access to your data, and access to, like, a Visio diagram. Right? So those are all things that I use in my career in separate ways that I have to invite you into the cloud, and then you could see the diagram here. You have everything in one. Like I said, the ways of working have to change. You're you're you're a grab. Good for us. I love it. I love it. That's great. We we need to move on, actually. We because I wanna make sure we cover other questions. We one of the things I love to fit oh, by the way, before I finish off, I should say, Mikko is kindly, giving us some of the training material she uses with her Yes. Gonna give to us to use in the show notes. So if you head to the show notes at the end of this, you'll see I think it's the analytics data dictionary, and we'll also include the storyboard on the cameras. I think, Mikko, is that right? Or I give away to that? Yes. I have to do the work. But, yes, I would definitely give you an analytics data dictionary downloaded over ten thousand times used by probably every organization that makes sense on the planet, and then I would definitely give them access to the storyboard. That's amazing. Thank you so much. Yeah. I will make sure that's in a Canvas as well. Okay. So, Mika, one of the things I love to ask people, particularly if they're listening to us and they're just in the middle of their early in their daily career, they're hearing people say how dated it's been. Like, tell us the one thing that you wish you knew at the start or maybe the thing that you tell your, like, people coming to your training course, the one thing you wanna take away. What is the what's that lesson that you wish you had as an analyst maybe? Man, Ollie, and I know you're not gonna believe this. I'm an introvert, and I know because you see me on keynote stages, you don't think I am, but I am an introvert. I didn't realize that this freaking data business is hospitality. Like, I tell all I there's two things that I tell my students. I say, look. One, data is TLC. We are in the hospitality business, and empathy is the currency. Number two is sales. Learn how to sell. You're gonna be selling everything. You're selling everything from the ninety minute workshop where you're asking for additional time people don't have, to the dashboard or the report that you produce, to the tools that you want them to learn, you're selling yourself. Right? Trust me. I'm an analyst. Here are my credentials. So learn sales, not just communication. A course in sales and then understand and accept we are in a hospitality business. I think if you get those two correct, you're gonna have a long, good career in data. That's really helpful. I think that's the idea of being of TLC tender loving care that your hospitality business is quite a big it's not the way the industry's been hiring, I would say, in the last five, ten years, I think. It's been more focused on technical skill, but I think that is the people who want to stay in it and make value from it. I think recognizing that means human human contact and better influence is really important to help. Yeah. Yeah. But when it comes to data, TLC is tenderness, listening, and clarity. Oh, I see. I wanna be very clear on what that means to me. I assume that. Actually, I got it completely wrong. There you go. I'm glad I didn't acronym this. This is good. That's awesome. Okay. Final thing. Well, the final thing is a big we we often ask the community of people watching like that who we're speaking to, what topic is, and we get an I we call it, like, dated anonymous letters, like, an agony art letter. I've I've got one for you here. I wanna just read it to you, then we get your thoughts on it. It's a it's a it's a few it's a few lines. Let me read it to you. It they go, hey. I'm a senior analyst at a growth stage start up. I enjoy my work. I've got a good relationship with a number of different executives and their wider theme themes, teams, and I feel I can make suggestions which get actioned. Recently, I've been thinking about my wider career option. I feel a bit stuck. There's no promotion opportunities in the analyst path I'm on, and so I'm currently thinking about either moving into a domain specialist role, possibly product manager, or even becoming a data engineer, analytics engineer because I think the pay bands are higher. I'm gonna have a chat with my manager about this, but wondered how if you had any any ideas on how I should frame the chat. Wow. Yeah. Absolutely. And you know what? It's interesting. You have a lot of data analysts that are facing that wave. We had a big wave during the pandemic, and now they're trying to figure out, hey. I got in the door as a data analyst. Now what? So I think the first thing that I would do, and I'm, again, going back, I started as a data analyst and a data scientist, is that you first need to kind of figure out what motivates you and what you enjoy doing. So a good example is if you enjoy tackling business strategy, working closely with users, building technical solutions and pipelines, maybe being a product manager is a better fit for you. Right? And your your analytical background will help you define metrics and influence roadmaps and all these fun things. I think that if you are more focused on data modeling, if you enjoy pipelines and automation, I think the data analytics engineering is gonna be a better fit. So I think it's more about sitting on and having a conversation around, hey. What works for me? Either way, when it comes to your credibility, being a data analyst and being exposed to data is definitely gonna get you there. But you have to redetermine. Do you wanna be in pipeline, or do you wanna be doing road maps? What do you enjoy the most? I I, so that that's good advice. I think that's obviously there are different fundamentally different skills there, which you've gotta go deep on. I think my I've gotta say my answer is kinda triggering me this question because, it lands on my, like, pet peeve that I have for most state organizations and companies, which is that the problem path. Isn't yeah. Yeah. It's not that, problem isn't where should I go next. It's the fact there's no career path for the analyst at all. But the thing That's right. There's one thing I wish that there was there was. It's like there was a genuine analyst career path where you could go junior analyst, analyst, senior analyst, principal analyst, staff analyst. You could go all the other bands and be valued as you get better at the skills that make analysts great, that data communication piece, that problem solving function. It shouldn't be that you go off and do something more technical or more domain. Like, I wish that there was there's one thing I wish. It'd be that there was a genuine Data storytelling, Ollie. That's why I I'm a date chief data storyteller. Data storytellers are invaluable. So I do have to plug in here and say, this is why I transformed my academy into data storytelling a few years ago, and it's been highly successful, is that I always tell data analysts. The difference between you and the next module is gonna be the stories that you tell. And so if you learn to tell the right stories, people are never gonna forget you, and you're gonna continue to move up because communication is a key factor. So I actually think that there is, and a lot of my students have evolved into leaders, and some of them have evolved into data journalists. But data storytelling to me is an evolution from a regular, quote, unquote, data analyst or a data scientist. I, agreed. Agreed. I just wish that, also, like, HR departments could see that progression was it was about there's a valuable progression there. It wasn't just Yeah. Analyst the problem. Into more technical domains. It's like, actually, there's a role here. We need to invest in training, send people some Mika's training courses to make them better, how we recognize that, how we therefore increase the pay, increase the prestige of those roles as we go. But this is my pet peeve. I'm gonna change the industry in that way if I possibly can. Thank you, Mika. That was amazing. I think we're running out of time. I wanted to say, if you wanna know more about Mika, you can find Mika. I think Mika York dot com. Is that right, Mika? Yeah. You can find me and also just on LinkedIn. LinkedIn is probably where yep. You can find me on LinkedIn as well, Mika York. If you wanna learn more about Cal or wanna learn more about the tenants and also find those resources that, Mikko's offering us very kindly, you can see in the bottom of the show notes, the outbound. Otherwise, thank you so much for being with us. Thank you, Micho. Have a great time. Thank you. Thank you. You guys are awesome. Bye.