142: Meet The Woman Who Changed Data Storytelling Forever (Cole Knafflic)

You can have the most beautiful graph in the world, and if you can't subsequently talk about that in a way that makes other people want to listen and pay attention and do something with it, the beautiful graph fails.

Okay. Cole, welcome to the Data Career Podcast. So glad to have you. Hi, Avery. Great

to be here. Thank you.

Yes. So if you guys haven't heard of Cole before. Uh, she is the author of the book Storytelling with Data. It is one of the, uh, best books on storytelling with data, but specifically like data visualization and how to present and convince people at your workplace, uh, of your findings. She's also the, the author of the new book, Daphne Draws Data, which we'll talk about in this episode as well, which is, which is more for kids, right, Cole?

It is, yeah. Younger audience, but interestingly, it's a lot of the same lessons that apply.

Okay. And let's, let's get into some of those, those lessons. Um, I want to start off with actually a little bit about, about your career because you studied mathematics in college, right?

Yeah. Math. I have an undergrad in math, uh, or applied math and, uh, an MBA.

Okay. And when you graduated, did you ever see yourself becoming like the author of a storytelling with data book and, and kind of this whole career that you have now? Yeah.

No, it didn't exist as as a career. I don't think at that point I, as I mentioned, I majored in math and I, I remember getting into my senior year in college and still trying to figure out what do I want to be when I grow up? And I remember going to a series of sessions that were, you know, like, What profession to have as a math major. And so I listened to the actuaries and the, the finance people, and I had this moment of, or longer than a moment, you know, the, the crisis of like, Ooh, none of these careers sound like what I want to do. Uh, and I remember then getting some of the best advice that I have received, I think, as I look back from my mother, which was finish the degree. And so, so. Finished my math degree and then got a job in banking. Uh, not in finance though, in credit risk management, where I was building statistical models, uh, forecasting loss, try to understand how we should reserve for the bank. And I loved, I loved the technical side of it, but also being able to Bringing creativity in and where I brought creativity and was in how I was visualizing the data, simple things like colors and some inadvisable things. As I look back like shadows or cram as many graphs on a slide as you can get on there. But interestingly, what I found over time was when I spent. Time and thought on the design of the visuals, people ended up spending more time with my work, and so that became a self reinforcing thing where other people would come to me, and I became the sort of internal expert when it comes to how do you show data fast forward through a few career changes, and I. Was it Google still using a lot of the same statistical methods, but now in an analytics role in HR. So people analytics forecasting things like who's likely to leave the organization and when, and what sort of things can we test out to change that? And I still spent a lot of time on the visuals and the team I was on, we were doing a lot of really complicated things that we needed to communicate to the engineers at the organization and the sales people at the organization and everybody in between. So audiences with widely varying. Needs, technical skills, familiarity with data. And so that was really interesting to see how do you change how you show things depending on who you're showing it to and where, where is that? How can that be more successful when you think about it from that standpoint? So also, while I was at Google, I part of a training program or part of developing a training program where I was creating coursework on data visualization, which was fantastic because it gave me a chance to pause and research and read everything I could get my hands on at that point, which was not a lot. It was like, Tufti, Stephen Few, I think his first book was out at that point, but really start to get an understanding of why some of the things I'd arrived at through trial and error over time, you know, why they work and why some things work better or worse, and really turn that around to be able to teach others. And so I did that at Google, uh, taught courses across the organization for a number of years and around the world. And then realized that it's not just. People in technical roles or at a technology company who need to learn how to communicate effectively with data. These aren't skills that we naturally have, even though a lot of the things and we can get into this, a lot of the lessons are really Practical and maybe even obvious once you say them, but until somebody points them out and shares them, we are sometimes our own worst enemy when it comes to trying to communicate effectively. Uh, and so it was, let's see, back in 2012 when I left Google and started storytelling with data, uh, which is what I've. Poured the last decade plus into really with the goal of helping people create graphs that make sense, but also going beyond the graph to, you know, you don't want to just show data. We want to take the data that we work with and learn something new from it and help communicate that new thing to other people so that we can help drive smarter decisions. Uh, reinforce that we're doing things the right way or that we should change how we're doing things and really have smarter conversations, not about the data, but using the data to have smarter conversations about the business. And so we do that mainly through workshops. Uh, there's the book that you mentioned, um, a couple more after that as well. One focused on practicing another on you as the person who is creating and communicating the data. And then the latest one for kids, as you mentioned,

that's such a wild and cool story. Congratulations on all the success. I actually attended a, uh, storytelling with data workshop at my company at ExxonMobil in 2020. And it was, it was awesome. And, and obviously I've, I've read the book and, uh, I actually have multiple copies, one of all the success in this, this really cool career that you've had. If you go back to that first job, you know, one of the things that you said was that if you designed your charts. Well, and you use best practices for data visualization, your boss and your boss's boss would care about them more and pay more attention to your work. And actually I was, I was rereading your book and I pulled this quote and you said, I quickly learned that spending time on the aesthetic piece, something my colleagues didn't typically do met my work garnered more attention from my boss and my boss's boss. And I just want to kind of talk about that for a second, because. It's not necessarily that you were doing better work or that your analysis was better or it was more meaningful. It was just easier for them to understand. And because it was easier for them to understand, they valued it more and they valued you more. Is that true in your career?

I think, yeah, I think it's exactly that, that it became When the graphs made sense and the messages made sense, it was more of a direct line into the value that the work was having. Whereas, if you imagine the same work being done, but being communicated in a really complicated way, or, you know, really going deep into the statistical methods instead of pulling back to say, What does this mean? What does this mean for you, the audience, or the person, the people to whom I'm communicating? What does it mean for our people? Business, how do we put that complicated stuff into words that makes sense to somebody who wasn't intimately involved in the process that when you don't take the time to do that, it can really easily become a barrier to the good work that's being done actually having the impact that it otherwise could. And that's what I think when we spend time thinking about how do we make this make sense to someone else? How do I look at something and say, all right, this might be what made sense to me, or it's the view that helped me reach that aha Eureka moment, but it doesn't mean that that's the same view or the same path that's going to serve my audience best. And so it really is this paradigm shift because I think often and I think Especially people in technical roles, we, we get so used to seeing things a certain way. And I think for me, at least as I look back, there was joy in figuring out the puzzle, right? Figuring out how the pieces fit together when it wasn't obvious. And so I think there's part of something in us that wants us to then be able to kind of show that puzzle to someone else, but have it not be clear so that we can have them experience some of what we did, but that does a total disservice because what that does is basically take the value that we could have added and obfuscate it instead of saying, all right, I did this work. I've, I've found, you know, the, the interesting thing now, rather than me take my audience through all the details and the work I went to to get to the interesting thing, it's actually just lead with that. And we may, in some cases, not even have to get into any of the detail. I think sometimes that. Feels bad when it shouldn't, that is success. That means your audience trusts you. It means they trust your finding because I can remember times I can remember times at Google, I can remember times at banking back prior to that in private equity, where I worked, where my team and I would spend a ton of time on an analysis or on a study. And then putting together a really dense recount of what we did and what we found in all of the methodology and. When it didn't get presented after at the end of all of that work, that would feel bad when really that was a success scenario. It didn't not get presented because we didn't talk about it. We talked about it and actually didn't even need to go into all of that detail because of the trust over time that was established to our stakeholders were able to go in with the story and then have the conversation focus on really understanding that and it. Understanding how we apply that to the business going forward. And it doesn't mean we didn't need to spend all that time putting together the document. We needed to have that. We needed to do that work in order to get to the, the answer or the finding or the interesting thing to communicate. And there will be times where you do need to take your audience through a lot of that detail. And so you need to have it there, but. The dense communication is not the, the, the goal, right? Going through that is not the goal. It's having the impact through the work.

I love that. And I think in today's society, as much as all of us might enjoy working on something we're passionate on, uh, I think people rather be doing their hobbies or spending time with their families. And so if you can just make your results as clear as possible, as quickly as possible, uh, that bodes well for you. Because some, sometimes I think. As technical workers, we want our work to speak for itself. Uh, and we want them to recognize, yes, I did all this work to actually accomplish this, but the sad truth is most businesses don't care. Just give us the results, tell us why it matters. And a lot of the time I even saw this post, um, from Kelly Adams on LinkedIn. She's like a LinkedIn creator. She was like the most of the time my boss doesn't ask me how I, how I even got to. Like doesn't ask to see my code ever. It doesn't ask to like actually figure out how I came to my conclusion. They just trust me to, to do the analysis and come to the right point.

Well, and I think that's part of the, part of the magic magic. It's not quite the right word there, but is really assessing a situation and. Anticipating what is going to be needed and what level of depth you're going to need to be able to walk someone through or show someone, uh, because when you can make that match the situation, that's when when things go really well, because you could easily take that and say, okay, well, so my manager trusts me. And that means. You know, I still need to be buttoned up on my work, but maybe I don't need to show all of my work. But then as soon as you get the question back, or you, you, if you misanticipated that or misread that, and now you have, or you're using that and going in front of another audience who actually is going to want to be convinced of the robustness of the analysis that was done, you need to be able to anticipate that so that you can meet that. Need. I think that is where things most often fail, where we create a report or a presentation for, for ourselves or for our data for the project and not specifically for the person or the people to whom we're communicating. That's that paradigm shift I was referring to before that when we can get out of our own heads and really think about, all right, here's what I did, but now how do I make this work for the people who need to understand it? And take measures to make it work for them, both through the visual design and through how we talk about our work, how we communicate directly, that that's where all of that can work really well.

So I think if, if I understand what you're saying correctly is your presentation, your communication, maybe even your, your graphs should almost dynamically change based off of who you're showing it to.

Yeah, I mean, ideally, so if it's a critical scenario and you have audiences who are, whose needs are sufficiently different, then you may want to think about, there will be times where it would make sense to have different communications for those different audiences. Now, in practice, that rarely happens. In practice, we try to create this one size fits all, but it's easy through doing that to then not exactly meet anyone's needs. So, I think A lot of the time we can get to the good enough scenario where, you know, if we, if we craft the communication and it's 80 percent meets this audience and 80 percent this audience, right, there's some overlap and that's probably okay, but where audiences are caring about really different things. So bring up an example from Google, since we talked about this a little bit earlier, internally, our main audiences were. Engineers on the one hand, highly technical, needed to be convinced that the methodology was sound, wanted very detailed information. We needed to get them on board before we even did the research a lot of the time so that they would eventually buy into the results. And then on the other hand, we had the staff. Sales organization whose general sentiment was leave us alone. We're the ones out here making the company money. And so for them, we needed to be direct and short and concise, focused on what mattered to them and not until they needed to act upon it. And it was like, it was, it was. After trying to communicate to both of those audiences simultaneously at first and just failing for a variety of reasons that are obvious in retrospect, that we decided, you know what, that's not the right approach. We actually do need to communicate to these audiences separately, not only in what we share and how we talk through it or show it, but also even when we communicate to them.

I think there's, there's people listening who, who might be thinking, well, the analysis is the analysis, but it's so funny because. You wouldn't necessarily think this, but the packaging that you put are around your analysis really matters. And oftentimes, like if, if let's just say we're, we're almost in the holidays, let's just say I'm giving you a Christmas present of some, some new headphones, right? Like if, if the headphones just in a cardboard box. They're not going to be as valued as if I put these headphones in like a really nice, like box that has really good, like opening mechanisms and really good wrapping paper and a bow and a nice card. Even

though even the wrapping paper, right. It's going to be different around the holidays than around birthday or something else. So yeah, it's the same contents, but the way you present it will and should be different.

Let's, let's talk about some of the ways that, that we can present well. So we talked about like. Addressing your audience. So if you're, if you're talking to your boss's boss, you're going to present it differently than to like your colleague or a engineer or a programmer or something like that. What are some other things that people should know when they're, when they're making data visualization and presenting?

I think one thing to be clear on is that you likely know the situation, you know, the data better than anyone else. And what happens through that Is when you look at the graph you made or the slide you made, it's super obvious to you where to look and what to see. But to make those things as obvious to someone else, it means you have to do things to make that happen. And so when it comes to the design of the graphs and the slides, you can think about how you might employ visual contrast, for example, sparing use of color to show your audience. where you want them to look and then using words either through your spoken narrative or written directly with the graph or on the slide or a combination of those two things that tell your audience why you want them to look there. And a lot of the time, just those two simple things. So making it clear where to look and what to see, even if it's maybe not the perfect graph type for what you're using, or there are some, you know, there's some clutter or, or something else, uh, You can still get your message across and it gets the job done.

That's something that I think you, you cover really well in storytelling with data. Um, just like the idea of how do we, how do we declutter our graphs? Because you know, it's funny, you're, you're, you're big enough that, um, maybe, maybe, you know, the answer to this. Um, but, but in this book, like you do all of this, I'll call it pretty ization of, of data visualization in Excel. All of the graphs that you do in the book are, are done using Excel. And what I mean by, by you're big enough, like your brand and your, uh, recognition has gotten to the point where it's like, can't Excel start? Like, It's actually a lot of work to make a graph look pretty in Excel. Can we talk to someone at Microsoft and have it like he defaulted better? Cause one of the things that Microsoft defaults does is if you have like eight different lines on your chart, they're the all different colors. And one of the things that, you know, you talk about is like, okay, let's only use color on one or two of these lines. Like why, why does Excel make it so hard?

Well, I don't think so. No tools trying to make your life miserable, right? Um, that, uh, any tool is trying to meet the needs of so many different situations, all at once that it's never going to exactly meet any of those, right? Take the example. You say like, why, why is everything colorful? Well, because if, The legend is, you know, off to the side or at the bottom, which is how that charts going to be at the beginning, then you have to have color as a differentiator. So you have some way to tie those back. The way that you can get around that when you are intentionally designing is you figure out, well, where could I label those lines where proximity is the thing that ties them instead of the similarity of color? But yeah. You have to make that decision in light of the data because it depends on how it lays out on the graph to say, well, can I label it within the graph? Or is that going to make it hard to read? Or there simply isn't space to do so. And so there are all these decisions that we make every time we're working with data. And you're even, you're implicitly making decisions when you're not changing these default things, because then you're letting the tool make the decisions for you. And. It's funny because I, I had thought for a long time, like, Oh, I should make myself my own template in Excel and make, make it just really easy. So I can have the starting point that I want. And I made several of these years ago and found that I never used them because for me, part of the process was looking at the thing that was never going to be quite right. And then figuring out how to intentionally make it work for what I need. And I think there's value in that and in the time and thought that it takes to do that. But we have to be intentional about doing it because otherwise we can just plug data into any tool and it will spit out something and it's never going to be what we need. You know, we pick on Excel, but this is not unique to Excel. Uh, it's, it's anything you're working with. And so I think there's an important part of the process that comes into play when we are taking the time to make those decisions and change the default settings to make them work for our given situation. I guess it takes

time. It takes human brain and it's just the laziness inside of me that wants it done automatically, but it's also, it's also probably. Something to look forward to for me and our listeners, because it also keeps us employed, right? Because if it was done out of the box automatically, perfectly, then maybe we wouldn't have jobs, but it requires a human brain. So that's good. I want to, I want to transition into talking about, uh, you know, a lot of people who listen to this podcast are trying to land their first day at a job. They're transitioning into data careers. Um, maybe they're teachers or physical therapists, or they're in sales. Do you think there's room for them? To, to stand out using data visualization and ultimately pivot into analytics.

Yeah, I, so I would say for the person who is trying to make that pivot and is in a role that is not working with data on a regular basis, currently, first thing is to look for opportunities where you. Where is their data and what you're doing today that you could work with? Because that almost always exists. If it really doesn't, then you can look elsewhere in the community for ways of practicing and honing those skills. For example, we have our online storytelling with data community where we host a monthly challenge. That's always something very, um, specific in theme, but open ended in term of how you address it, where typically you're finding data. Data that's of interest to you and doing something with it. I think the one we have going on currently, uh, so November, 2024 is just finding a graph in the wild that isn't perfect. And then taking steps to improve it. Uh, we also have an exercise bank that has hundreds, probably at this point of exercises that are more focused on developing a specific skill where the data, the instructions, it's all about. All provided. And so all you need is, you know, five minutes, 30 minutes and something you want to work on, uh, in terms of practicing, whether it's, you know, like we talked about, maybe it's taking a graph and figuring out how to change the color of just one line and make everything else green. Gray or, uh, designing a slide. Um, and there's a variety of other things as well. So looking for ways to practice to hone your skills, which I would say again, first look within your role to see if there's anything you could be doing there or more broadly at your organization. Some will allow there to be moonlighting or, you know, shy of an internal transfer, but still getting some exposure to skills that you would want to be using. So look for those, if not in your current role, then look to the community to see where you might do that. And then I think for anyone who is not currently in a data role, but wanting to get to where they're working with data, visualizing data, communicating data, the thing to not overlook is how you communicate, how you communicate verbally, and how you talk about yourself in terms of, you know, how do you introduce yourself, or how do you portray Your work history and your skills when you are interviewing or doing things like that and spending time working on that, uh, and also how you engage your audience through the way that you speak. Um, because this is one of the things that over the years, and I think again, as I look back, it's not surprising and seems obvious, but it wasn't until. Fairly long into things that it really became clear to me that the graph or the data visualization is really just one part of the puzzle because you can have the most beautiful graph in the world. And if you can't subsequently talk about that in a way that makes other people want to listen and pay attention and do something with it, the beautiful graph fails. And so I think both for those who are wanting to transition into data roles. Also, I would say for those who are currently in a role working with data and communicating data work on yourself because you can be just as strategic when it comes to how you speak about your work, how you portray yourself, how you communicate as you can with, you know, what graph you're choosing and how you're choosing to portray things visually. And when those two go together, you've made a good graph. And you can get other people's attention through how you speak and through the passion you show for the work that you've done. That becomes a really powerful combination.

It's, it's a great point. Um, and whether we like it or not, we live in a world, uh, where your appearance really matters. You know, it's not, if you're trying to land the data job right now, it's not the. The smartest person or the person who's best at at sequel that lands the job. It's the person who's able to best portray their skills that they'd be, you know, able to help the company. And the same is true. Once you land a job, it's not necessarily the best employee that gets the promotion. It's the employee that appears the best or gets portrayed as the best. And they, you know, it really doesn't stop until. You become like the CEO. And then even then like appearances still really matter. So it's, it's maybe unfortunate and you'd want maybe just pure talents and skill to win, but the way that I think this is

part of the talent as well, being able to being able to communicate adeptly and one resource that I'll point people to in case like, okay, I get this, but how do I actually do that in the yellow book storytelling with you? This is the one that goes back to. There's data visualization in it, but it goes beyond the data into how can you develop yourself to be able to plan, create and deliver content? Uh, the penultimate chapter is crafting the story of you, and it's basically taking people step by step through how you can be really thoughtful and robust in how you plan and how you talk about the story of yourself, which can be useful in a variety of scenarios. And it's actually, it really, it becomes an interesting case study and way to practice a lot of the other things that are introduced that are grounded more in how you would communicate data. But things like, you know, brainstorming on sticky notes and really considering your audience and making all of that work together with. Using a subject that people know really well themselves. Um, but then after going through that chapter, you can come out of it with a really clear plan and ways to practice when it comes to talking about yourself that you can then translate into talking about other things as well.

That sounds like a superpower to master that. I don't have the yellow book, so maybe I'll, I'll have to look it up. Look into that one. Let's talk about your, your brand new book. Daphne draws data. Uh, tell us a little bit about what it is and why you decided to do this. Yeah. Look at it.

Yeah. So Daphne is a delightful pink dragon who has a unique talent. She enjoys drawing. That's not so unique. Well, maybe for a dragon it is, but the thing that she likes to draw the most is. Data. She likes to draw graphs. And so the story is a really fun, I mean, it's a picture book, really fun, brightly illustrated, uh, about Daphne's adventure. She decides, well, if she's not being appreciated at home, she's going to go off and find a place where she can fit in. And so she goes to the jungle and outer space and underwater and all sorts of places. And in each location, she encounters some creatures. Uh, and a problem they're facing, and then helps them solve their problem by drawing data. So she collects it, she draws it in very pictorial forms of graphs. Uh, the word graph I don't think is used once in the book though. It's really introducing the concepts through story and through pictures. And then, uh, I won't give away the ending, uh, other than to say it's a happy one. And the story ends, but then the book continues into a graph glossary that goes more into what the graphs were that Daphne used over the course of her adventure. So there's a page each devoted to bar charts, line graphs. pie charts and scatter plots, uh, showing examples from her adventures, helping kids understand how to read them when they work, and then introducing activities that kids can undertake using data that's of interest to them because one great Parallel that we can make across adults communicating with data and kids and the use of data and graphs is to make it about something that's meaningful and something that can be acted upon because when I see my kids come home with graphs from school, so far, I've been pretty disappointed because they're graphing things like the weather. The weather in September, okay, it was sunny. You experienced that. It's not so interesting now to draw it in a graph or they'll do things like roll a die, uh, you know, a bunch of times to see that, you know, and then graph it to see, okay, I rolled all the numbers about the same amount of times. This isn't anything that they can then Use to understand things better. Uh, and so I really would like to make the data that we're having kids work with be something that they're interested in, because I think this is such a, it could be such an amazing way into mathematics in a way that isn't portrayed as boring or complicated or completely abstract when it comes to kids day to day. Uh, so, you know, let's have them track. How many hours they're spending on a screen every day and how they feel plot that, or, uh, you know, where's their favorite place to read and, you know, how might we then emulate some of those things in the classroom to promote more reading, like things that we can actually, uh, help kids learn about themselves and about the world around them in ways that is fun and engaging because what I've seen through my kids. And their friends is that kids are fantastic and love doing a couple of things. One, asking questions, particularly like, I don't know, kindergarten, first, second grade. There's no filter yet and kids are so curious and they ask questions about everything. And if we could teach kids how to hone and get really good at asking questions that can subsequently be answered with data, that is going to be an amazing foundation for everyone. Any sort of problem solving, critical thinking, analytical career, and they also love drawing. And so if we can let them take some of that creativity and do it with a graph and with numbers and let kids approach that creatively, I think it's a very refreshing change from math being something that's either right or wrong because graphs, there's more leeway. Uh, there can be creativity. People can approach things. differently, and we can celebrate that and learn from that rather than say, no, don't do it that way. Do it this way. And so for me, I think it was a combination of just going back to the impetus for writing the book, a combination of, you know, seeing the adults who we teach and so many saying, I wish I had learned this sooner or earlier, and then seeing my kids and how, just how they learn about the world around them, how they develop. Language and logic and realizing we could take the visual language of numbers and introduce that a lot earlier than we do, sort of those two things coming together. I think there's an opportunity to really help our kids recognize this superpower of comfort with numbers and asking questions and answering those questions and drawing and plotting things that, um, It'll be a great foundation for them for so many things going forward.

You're building the next generation of data analysts and a data viz specialist, a ripe young age. So, uh, that is very cool. Where can people find this book?

Oh, anywhere books are sold. So yeah, favorite independent bookseller. You can order it. It's on Amazon. Uh, and, uh, yeah, is around the world.

Okay. Awesome. Well, I haven't checked it out yet. I'll have to check it out. I'll have to check out the yellow book. Um, but I'm also a huge fan of, of the storytelling with data original book. So if you guys haven't checked those out, be sure to check them out. We'll have links to all of them in the show notes down below. Uh, Cole, thank you so much for coming on our show. We appreciate it.

Thanks for having me, Avery.