Fiona (00:01)
Welcome to UnDUBBED up to the podcast that's unscripted, uncensored and undeniably data. I'm Fi
Sarah (00:08)
And I'm Sarah. And Fi I've got to say this has been one of those episodes I've been looking forward to since we started our Undubbed podcast.
Fiona (00:16)
Same. We're joined by someone whose work has genuinely shaped how millions of people are communicating with data. Cole Nussbaumer Knaflic is the founder and CEO of Storytelling with Data, a five-time bestselling author and the person behind what has become the go-to resource for anyone trying to make their charts actually mean something.
Sarah (00:48)
more than 500 universities worldwide.
Before building SWD, honed her craft through analytical roles in banking, private equity, and as a manager on the Google People Analytics team, where she built and taught their internal data visualisation course.
Fiona (01:09)
home for us too. Both Sarah and I have been through Cole's storytelling with Data Workshop and it genuinely changed how we approach presenting data.
Sarah (01:18)
today we're diving into Cole's newest workbook, Storytelling with Data Before and After. This one she co-authored with Mike Cisneros. my God, I knew I was going to stumble on that.
Fiona (01:28)
No.
Sarah (01:30)
which she co-authored with Mike Cisneros and Alex Velez It's packed with 20 real world examples showing exactly how to turn unforgettable charts and slides into clear, compelling data stories. We've read it, we've flagged our favorite bits, and we have questions.
Fiona (01:49)
But before we dive in, it's that wonderful message from me. Remember to pump up the algos like and subscribe to Undubbed, and most importantly, share this episode with anyone on your team who builds charts, creates slides, or presents data. They need to hear this.
Sarah (02:04)
Welcome to Undubbed, it's so good to have you here.
Cole Nussbaumer Knaflic (02:07)
Hi Sarah, hi Fi, I'm excited to be here and to chat with you today.
I was trying to think back, so the Tableau user group that we did, that must have been before I lost my voice, because on that trip to New Zealand and Australia, I completely lost my voice, which is problematic when one is meant to speak to a number of audiences.
Fiona (02:29)
Yes, absolutely. But I actually took my whole JLL team from Australia along to your Melbourne session. And, you know, it was really at the start of my journey with the team there. And it helped to solidify that we would be doing things a little bit differently. And it really stuck with them as well, your philosophies around how to really communicate so well with data, even when you lose your voice.
Cole Nussbaumer Knaflic (02:33)
I remember that.
Yes.
Fiona (02:56)
Anyway, so let's get started. We always love starting with the person behind the profession. Could you tell our listeners a bit about yourself and your journey and where you've come from and where you are today?
Cole Nussbaumer Knaflic (03:08)
Yeah, I mean, it depends how far back we want to go. Maybe I'll start in Seattle. I had freshly graduated from the University of Washington. We're going way back, right? This is a couple of decades ago with a math degree and was trying to figure out what do I do? And I have always loved math and numbers and the way you can logically put numbers together and make sense of something.
And so my career started off in banking, in credit risk management. I was building statistical forecasting models and communicating to our leadership and realised pretty quickly that where I liked to play in terms of making a graph, making things visual, making data, something that other people could see, wasn't only a creative outlet, but when I did that well,
other people paid attention to it in ways that they just weren't to my colleagues work. So fast forward, I found myself later at Google. So after realising I wanted to get out of banking, subprime crisis, all of that stuff, I looked to say, what are my skills and where can I put them to good use? And at that point, I had gone back to school, got my MBA, and so really enjoyed this intersection between
highly technical statistical things and the business side and really breaking what we can learn through numbers and data down into something that can be used to help people have a better conversation or make smarter decisions. I ended up at Google in Mountain View in California at headquarters on the People Analytics Team, which is an analytics team that is in Google's People Operations or HR.
And I had the benefit of joining right as that team was being formed. So it was small and scrappy, which was fantastic. That's my favorite way to work because it means you get to dig into lots of different things, wear a ton of different hats. And it really felt for those first few years, like every rock we turned over, every bit of data we looked at, it was like we were the first ones doing it. And so looking back, I got to work on things like
Some of the work we did on managers has been pretty well communicated. And none of it was probably groundbreaking in terms of we didn't find things that aren't probably common sense, but we were able to put data behind them in a way that had never been done. And that meant we could quantify what makes a good manager and actually measure people against it and hold them accountable and test out different things and understand the impact.
And so it was this really interesting way of taking the same sort of statistical methodologies that I'd used in banking on credit risk data and applying them to people and behaviors in ways that were super interesting. Also, while I was at Google, I'd continued to hone my craft for making visualisations out of what we were communicating and helping my colleagues.
present them to all different parts of the business, ranging from engineers who have this like insatiable desire for detail to salespeople on the other end of the spectrum who just want to know the big picture and how it means they need to change what they're doing and why. And through that, I was asked to develop a course on data visualisation which was originally for a program that we were building within people operations.
But there was such interest in the data visualisation course that we ended up rolling it out company wide. And I got to travel to Google offices across the US and across Europe to teach people at the organization how to communicate with data in a way that is easy for people to understand. And it was after speaking at my first couple of conferences that I realised these skills aren't needed by Google alone. Rather,
everybody and even increasingly since then who works with data needs to know how to do this. And they aren't things that were typically taught in the way that we're taught statistics or business communications. And so I left Google, this is back in 2012, which is a long time ago now, and started storytelling with data where at first it was just me going and talking to anybody who would listen about
the things that I had learned when it comes to not just showing a graph or designing a good graph, but going beyond the graph and thinking about your audience and what you need to motivate them to do and how we can weave the communication into a story that gets their attention and builds understanding in a new way and hopefully drives them to act. Over time,
My team has grown, so we are now nine people strong. And we still spend a lot of time teaching and doing workshops where we'll go into a client organization and spend half a day or a day really getting into the tactics and the strategies for communicating effectively with data. We also have a wealth of other resources that we are constantly putting energy into.
We still have a blog, our podcast, our videos, we have an online community, and of course our books, which I think we're going to talk about more today.
Sarah (08:36)
Cole, I really love hearing about your journey and just how it's evolved and congratulations on having a nine strong team as well. That's amazing. And you talked about your ecosystem being books and workshops and a community and also a podcast. What does a typical day look like for you now that you're running all of this?
Cole Nussbaumer Knaflic (09:00)
So I think that is one of the reasons that I love storytelling with data so much is because there are so many different outlets we have for sharing knowledge. And just going back, that has always been a foundational thing for me. When I first started at Google and I mentioned I was going to speak at some of my first conferences, I realised if I'm going to go talk to a room full of people, I need to have something to point them to at the end.
And so that was when I started my blog. This was back in, what would it have been, like late 2010 when you could start a blog and actually get some traction, right? I don't know that that's the case today. And I remember I quickly put up two or three blog posts so that by the time I spoke at this conference, I had somewhere to direct people.
But that's always been a core facet of really sharing as much knowledge as we can, as widely as we can, and for free so that people around the world can learn how to be effective when it comes to the sort of things that we teach. And we've been fortunate to have lots of paying clients along the way so that we can continue to provide a ton of resources for free.
Fiona (10:13)
Yeah, I mean, I love the concept. You don't see a lot of people not putting things behind paywalls. know, a lot more people are saying, okay, get my subscription and keep, you know, keep them on that retainer and keep the things coming through. And, you know, that's great, but it's also important to be able to lay that foundation and share some of the free content as well. You've had amazing success with your books. Your original book, Storytelling with Data, I...
inhaled literally. So I read it cover to cover, like laying down one day on my couch and running through it. And one of the things that I really admire about you, and people would even hear it in the way that you're speaking right now, is you're so eloquent and you make things really accessible for people who are trying to perhaps understand concepts that they're not familiar with, or even if they are familiar, like I was.
You make it super easy to read and engaging. It's not boring. So many data books are really, really dry. How have you approached that storytelling or what's the secret sauce that you have?
Cole Nussbaumer Knaflic (11:19)
Well, first off, thank you. You're making me blush. Yeah, by the time, so where to start with this? Well, I'll share an anecdote, which is my husband likes to take credit for the book because he says the book was his idea. He said, no, no, the book wasn't your idea. And he says that because he was telling me to write a book long before I ever wrote the first one. But I wasn't ready yet. In my experience, you have to have an idea and
content and like the book has to happen so much that it is like pouring out of you because if you don't start with that I don't or I should say for myself at least if I didn't start with that I don't know how I would sustain myself through the process because it is a lot of work and focus and intensity and So for me, I had to wait until I reached that point of boiling over if I didn't get the words out
And I think one benefit of that is by the time I wrote the first book, I'd stood in front of audiences and facilitated hundreds at that point of workshops. And so I'd said a lot of those words or versions of those things before. And what that enabled me to do is just write in a really conversational style, as if I'm teaching you and taking you through my thought process, because that's exactly what I was doing.
I think also having written the blog for a number of years before that, that it really helped me get clear on my voice and what that sounds like. And I was able to try out teaching different things in person in a way that I circle around and refine how I explained things and what language I used and collect all of the examples to use to illustrate the different tips.
And so it really evolved organically. And the benefit came together in, as you said, a book that is straightforward and fast to read. think actually, so I recorded the audio for the book, and I think it's like four hours, or give or take. So it's not a very long read, which I think is one of the things that people enjoy about it, is that it doesn't feel like work, I guess, to consume.
Sarah (13:25)
And I think when you were saying about how you the style in which you wrote it, I feel like I'm kind of sitting in the room with you and it's like a calm environment. And I'm just absorbing all this information.
Cole Nussbaumer Knaflic (13:35)
Everything's blue and gray.
Fiona (13:37)
Yes!
Sarah (13:39)
Exactly. Now, storytelling the data before and after is your first co-authored book. How did you find that process?
Cole Nussbaumer Knaflic (13:45)
Yes.
Well, all I will say yes and no, did have before that Daphne Draw's data, I worked with an illustrator, which was another sort of, so we didn't write together, but it was collaborative in some ways that I think actually probably ended up being helpful for having co-authors on Before and After. I think when it comes to Before and After, this was a book that we had conceptualised years ago with the idea that,
Anytime we work with a client, we solicit examples from them ahead of time, use that to understand how they communicate with data currently, what challenges they might be facing, use it to customise content where it makes sense, and also as the basis of a number of the hands-on exercises that we take people through during a workshop. And so what that means is we have all of these great examples.
And now of course we have to change them to be able to share them more broadly. And so we've developed this really interesting skill set of completely swapping out like industries and specifics so that the each example is still true to the original scenario. But basically all of the details have been changed so that there aren't any confidentiality issues. But for us in workshops, you know, it's one thing I think when you teach a lesson and you show it through an example that you bring along with you, it's like.
hey, OK, that makes sense. But then it leaves room for the argument, but my case, I'm a special snowflake. My case is different. That's not going to work here because it just leaves things open for people to agree but maybe not change behavior. Whereas when you see those same lessons applied in the example we bring, but then also to your own work or your colleagues work,
That is where the magic happens. And when we're in person with a group, you see it and you hear it in participants. I was at a workshop late last year and I went through, we have kind of a capstone example that we'll do for a client at the end after we've gone through all of the lessons that's meant to be comprehensive and show the totality of what's possible.
You know, not as the solution, but as one way to approach things based on the lessons that we've covered and the assumptions that we're making about their scenario, because we never know as much as they do about the specifics. But I got through this one, and at the end, there was silence. And then somebody said, wow, if we had presented it that way, we probably would have gotten our budget approved. I was like, well, that feels great. And I'm sorry, but now you can do it better next time. And so.
The book is really about trying to bring that magic to people more broadly. And of course, it won't have quite the same effect because it's not your examples that we are remaking. But in every one of them, you can see something that you can connect with, something familiar that you've done or seen done or a challenge that you've grappled with. And I think that's the really fun thing.
how much you can learn from every example, even if at first blush, it seems totally like it's from a different world or different industry. That's one of the things that we see across clients from all sorts of places is that fundamentally, people are struggling with the same sort of challenges. And there are a myriad of ways that we can go about easing those challenges.
Fiona (17:09)
For sure, it's like a bit of magic, I think, when you have a before and an after, and it's like, it's like you've had an interior decorator come through your house, you know, and you're preparing it for sale or something, and you've got the before, and then you've got the staging afterwards, and that uplift, it's like a completely new house or a new build.
and you feel really fresh and you go, this is really inviting and that's somewhere that I would like to belong as well.
Cole Nussbaumer Knaflic (17:37)
I almost think more powerful than the before and after is when you're allowed to see the progression in between those two states, right? We think of like home improvement shows or makeover things like that's the fun part is seeing like everything that gets done over the course of it that brings it from what it was to what it could be. And I think, yeah, that's one of the reasons we had a ton of fun with this book. And interestingly, we...
Fiona (17:43)
Yeah, right.
Cole Nussbaumer Knaflic (18:02)
when we were putting together marketing images after the book was done, one of the, we did a lot of things with just the before and afters from different case studies. And in a lot of cases, the before and after on their own maybe aren't as, I don't know, they weren't maybe as striking as we thought they might be. But it's because that process piece is where you get to unveil what happened behind the scenes. And as you see.
said kind of show the magic there.
Fiona (18:30)
The messy middle, suppose. We call what you do a glow up. So we take something that's original and we add a bit of fairy dust on it and go through that messy process and then make it over so that it really gives a different positioning perhaps and a different communication style to really land the data. One of the pieces, one of our...
Cole Nussbaumer Knaflic (18:32)
Yes, exactly.
Ha, I like that.
Fiona (18:53)
like secret sauces that we have that might help our listeners as we use a tool called Tonic Fabricate where it creates synthetic data and you can really prompt it. You get a free, I think it's $10 per month to burn on credits and it really helps you to build the synthetic data, push it back and forth a little bit better than a Claude or a chat GPT because it's specifically for data. So that might help you but also our listeners as well.
Cole Nussbaumer Knaflic (19:23)
Yeah, that's great.
Fiona (19:23)
⁓
One of the themes that does run through your entire book is really understanding your audience. And what do you think that looks like in practice when someone brings you a chart and says, can you help me make this land? It's really not getting, it's not sticking with my audience.
Cole Nussbaumer Knaflic (19:40)
Well, and I think that is the first question that I would put back to somebody in that scenario is, who is it for? Who is your audience? What are you trying to make happen with it? Because this is often when we see data gone wrong is when the audience is anyone who might be interested in this particular data or stakeholders, these generic large
audiences who will have different needs, they'll have different things they're looking for or wanting to get out of it. And it is difficult to craft a single communication that is meant to meet the needs of multiple different audiences if those audiences are motivated by different things and care about different aspects. And so the more specific you can be about who the graph or the slide you're creating is for,
the easier time you're going to have making decisions about the design. Because once you know who you're communicating to, you can be really intentional and thoughtful about how do you make it work for them. So I find often when we're putting together a graph or we're building a presentation deck, it's for ourselves. It's for our data or our project. And when we shift that paradigm into thinking about creating graphs and slides that are not for ourselves, but rather,
for the people on the receiving end of it, it totally changes the game. Because now we're thinking about not what do I need out of this, but what do they need and how do I make that happen? How do I make this make sense to them? How do I make it relatable, understandable? How do I make it resonate so it's something they actually pay attention to and want to do something with? How do I direct their attention within my communication?
So you can then get to all of these deeper level questions that'll help you with the actual design.
Sarah (21:34)
Sometimes it's really hard to get to that answer, right? Particularly when you're maybe talking to not the decision makers. When that happens, how do try and get in front of the right people when you've come into a project and maybe you're dealing with the analysts that don't know those answers as well?
Cole Nussbaumer Knaflic (21:55)
Yeah, so certainly the more information you can get, the better. And so even if it's not you in front of the executive, if you know eventually it's going to the executive, you can design things for that. I mean, there are some instances where it's going to take some trial and error and understanding when you try something and it doesn't work and why that is. I think one question you can come back to then is,
what are we trying to make happen with this? Because sometimes if you can answer the what, then the who becomes more obvious. And it doesn't mean you're always gonna get to be in the room with that person. But if you know who the eventual audience is and any that are in between you and them, then it becomes more complicated because now you have to design with both of those in mind. But you can still do some things to make that work.
For example, let's imagine a scenario where I'm the analyst, I'm putting together something that was directed to me. It's my manager who asked for it. And my manager is the one who's eventually going to go and present it to the person who cares. So the more I can understand that that's the case, one, getting that understanding is helpful. But then second, if I can think about that ultimate audience,
Sarah (23:02)
you
Cole Nussbaumer Knaflic (23:19)
and my manager, I can think about how do I frame things so that it works for the ultimate audience, but make it so my manager is inclined to focus on the places where I want them to. This is little hand wavy in assuming that I have enough context to be able to do that. But in that instance, I can say, well, I'll design the data. Where do I want people to look? Let's use colour or other means of contrast sparingly there.
let's title the graph or the slide with the takeaway so that that's obvious. If there are clear next steps or a clear recommendation, write that in text on the slide or on the graph. And that means now you've designed something that your manager is gonna take and present, but you know what they're going to present because you've kind of not given them other options in the way that you've designed it. And now certainly better than getting the request, going away, coming back with the final thing would be
having interaction with your stakeholders, whether it's the manager or somebody else along the way. And when it comes to that, I'm a huge fan of starting low tech where you're sketching or you're storyboarding with pen and paper or Post-it notes because you can put together something rough and take that to, if we keep the scenario going, I take that to my manager and I say, here's what I'm thinking. This is rough, but I want us to get alignment here before I go start creating things.
the roughness of that is useful because it forces people to focus on the general direction rather than the specifics. As soon as you put something into a graph or a slide and it starts to look like it's been designed, it begs for feedback on the design. And sometimes when that happens, often when that happens, you actually miss them the conversation on, is this the right thing to be looking at in the first place? And so I'm a big fan of starting low tech.
having there be feedback loops between you and the person you are designing for or somebody like them so that you can get your plan set first. And then by the time you turn to your tools, you already have a sense of what things need to be. It just helps with alignment and helps with that frustrating scenario where somebody asks for something, you go away, you put it together, you bring it back and they're like, that wasn't
quite the thing that I wanted. And you're like, this was just what we talked about. I can remember, it's funny, I had thought it was Steve Wexler, but I brought it up with him one time and he was like, no, that wasn't me. But I heard this analogy at some point from somebody that wasn't Steve Wexler, that was, I had a client and they asked me for a rock. So I went and I got the rock. I give them the rock and they're like, well, this rock's flat. I wanted a bumpy rock.
And so you go back, or you find a bumpy rock, you give it to the client. Well, but this bumpy rock is gray. I wanted a bumpy rock that had some red through it. And you can imagine this goes on and on. It looks very much like a client interaction when you're providing some sort of data, where oftentimes people don't know what they want until they see what it is that they don't want. And so the faster you can make that without putting in a ton of effort, the less
tense all of that is, right? So if I can do that in rough form where we're drawing things on a whiteboard or I'm sharing a storyboard and walking through it and we're moving things around, that speeds that process up. And so that you're giving people something to react to versus going away and creating what you think is the final thing only to be frustrated when they tell you no, it was something else they were looking for. So I've taken this.
in a bit of a different direction than we started with, but hopefully that's still helpful.
Fiona (26:55)
It totally is. And I'm just reflecting on the work that Sarah and I do with our clients. I would say we don't go so well. We probably go in a different direction. We don't go lo-fi because we find that if we're using fabricated synthetic data, we can drag and drop and probably do things faster. But it challenges my thinking definitely around
I know they do get into the weeds sometimes on design, like on colours or whatever else. And so we do actually do a lot of prep with colours to make it psychologically the same as what they've got already. And so, but I think you're right. And it makes me wonder, you know, how can we be better to have things, for instance, in Miro, where we could drag and drop some types of charts on or Figma or whatever it is, still be a little bit because my drawings are shocking. This is what I'm trying to avoid.
Cole Nussbaumer Knaflic (27:20)
Yep.
Well, and
that's the thing. When you've got a lot of data and you want to look at it different ways, drawing can be cumbersome. So if you have a way to iterate quickly through tools, you just want to be careful. you almost want it to look low fidelity there so that, again, so that people are focusing on the bigger picture of, this what we want to be looking at? Is this the right way to look at it, not colours and fonts and things of that nature? Which that feedback is also useful, but at a much later stage.
Fiona (28:07)
So I might use something like, use Tableau to wireframe and then shove it back into trace table which makes it look like it's being drawn with a pen.
Cole Nussbaumer Knaflic (28:11)
Mm-hmm.
you could do, you totally could do.
Sarah (28:18)
I think the whole point is we're supposed to be more throwaway at the beginning.
Cole Nussbaumer Knaflic (28:21)
Yeah,
Fiona (28:22)
have real struggle
Cole Nussbaumer Knaflic (28:23)
totally.
Fiona (28:23)
with MVP. I have a real struggle. I'm trying to be better this year. Like, what's the skinny version is my mantra this year. What's the skinny version of delivery? You touched on some amazing pieces of magic as you were walking through and winding through that journey with us. So the little elements like colour and using that to highlight, you know, being really intentional with things as well.
These are all things that I believe that people can learn so quickly from you and I've never seen anyone explain it in such an elegant, accessible way. So I highly recommend if you're starting out or even if you've been in this environment for a long time, it may actually help you to train and teach other people how to talk about it effectively.
not all of us in the analytics business are perhaps as elegant and articulate as you Cole. So just hearing it and being surrounded by somebody that gets it on point can really help to navigate through some of those conversations as well.
one of the chapters in the book that we see play out with our clients all the time is explaining with data stories rather than dashboards. And Sarah and I are big believers in knowing when each one is the right tool.
How do you help people recognise when they need a story rather than a dashboard?
Cole Nussbaumer Knaflic (29:46)
Great question. for me, the scenario in which a story is going to serve you well when it comes to data communication specifically is when you have a specific audience and you need them to understand something in a different way. And probably as part of that, you need them to take some sort of action.
when you don't need a story is when it's a status update and there's nothing in particular that people need to pay attention to. And I think that's where confusion can come in sometimes. So when we're teaching about this and in the books, we often make the distinction between exploring data and explaining data, where both of these are important parts of the process, but you use different tools for them. So as you're exploring data, you might be, you
Digging through it, looking at things in different ways, graphing things, pulling together different data sets, try to understand what's interesting here that someone else might care about. And I actually lump any sort of regular reporting or standardized dashboards into this exploratory space. Because these are tools that are meant to be diagnostic. You want to look at it, or somebody who knows what they're looking at looks at it to say, where are things in line with our expectations?
And probably more interestingly, where are they not in line with our expectations? Where is something unexpected or interesting happening? And it's when you find those interesting things and there's a reason for somebody else to care about it and do something with it that then you move into explanatory space. And this is where you have someone specific, you need to do something specific. And this is where data storytelling comes into play.
And so my guidance is at that point, maybe you use the dashboard to find the interesting thing. Now let's take that interesting thing out of the dashboard and tell a story. It doesn't even mean we take it out of the tool. Like you could use a dashboarding tool to tell a story. Tableau has, story points or did at one point anyway where you could still use the same tool, but now you actually think about it in the context of a story. So we often, when we're teaching on this, we use the narrative arc.
which is a simple story structure that maps really well to a lot of different business scenarios where you start off, there's the plot, tension is introduced, the tension builds in the form of a rising action, reaches a peak at a point of climax, there's a falling action and a resolution. And I think one...
beautiful thing about the narrative arc and thinking about our communications in terms of structuring them like a story rather than what we typically do. So what we typically do is much more of a linear path. Often it's chronological, it follows a similar path as we took through the analysis or the project where maybe we start off with hypothesis.
Then there's the problem statement, what question we set out to answer. There's the data, where did we get it? What assumptions did we have to make? What did we do to clean it? Right? All the nitty gritty. what's the analysis we did? What were the statistical methodologies that we employed? Then the recommendations or the findings, what did we learn from that? And then the recommendations, what do we now do? So this is the path that a lot of data presentations take. And I've been known to characterise it as a really selfish path.
Because at no point along that linear or chronological path do I, as the person communicating, have to give any thought to my audience. And for me, that's the big difference when you start thinking of how you communicate in terms of story, in terms of that arc. Because to have that shape, you have to have tension. That is what gives the rise and the fall.
And it's not the tension that matters to you, it's the tension that matters to your audience. What is at stake for them? Because if you can frame your communication around that, they are naturally going to care because you've made it about them. And it's a way of taking basically the same information you probably would have gone through, but kind of turning it on its head and thinking about it from someone else's perspective.
which means there's probably a totally different level of detail that you'll go into in different components of that than you might otherwise have done. Because at the end of it, what your audience needs is not a book report recounting everything that you did. It's a guide on what they should do next or what they need to understand from that so that they can be informed in what they do next. So it's just a different way of thinking about how we communicate. And again, it comes back to audience.
Sarah (34:15)
Yeah, great. And I love that distinction. It's like, are you exploring data or are you explaining data? It's such an important distinction. One thing that I reflect back on rereading, or sorry, reading your book is around the power of annotating and really like calling out and honing into something and having those, you know, like saying it and saying it again and saying it in such a clear way that it's so directive and at the same time.
very, very calm.
Cole Nussbaumer Knaflic (34:44)
Well, what you want to think about, one of the phrases that kills me is a picture is worth 1,000 words. hear people say this sometimes. And I think the way it gets misconstrued and used in data communications is to mean that we should just put the graph up. We don't need words to help people understand this data visualisation, which anybody who's seen a graph without words knows is totally false.
is a picture might be worth a thousand words, but if it says different words to everybody who's looking at it, that's a failure of communication when we have something specific we want to say. So I'm a huge fan of annotation, Sarah, as you bring up, takeaway titles, right? If there is something you want your audience to know from your data, title it with that. It sets the stage for then how they process the rest of the information. There was actually some really interesting research, this was done a few years ago, I believe out of Northeastern University.
Michelle Borkin's group, I want to say I may have that wrong, but it was that if you title the graph with the thing you want people to see, they are more likely to remember that thing, which just is hugely important when it comes to how we think about how we use words together with data.
So I'm a fan of, if you're taking a graph and you're putting it on a slide, make the title of that slide the takeaway that you want people to see in that data. Use colour or other means of contrast sparingly to draw people's attention to where you want them to look for evidence of that. Add annotations so that they can help understand context or what this means. If there is a recommendation, which I...
there should be when you're at the point of explanatory communication, then make sure that is verbalised on the slide directly. So use your words wisely.
Fiona (36:30)
So on that note around using your words wisely, one of the things that strikes me, and I phrased it a couple of times already, is how...
you can communicate effectively through that story. And you talked about the narrative arc that you use and the process that you follow. But me personally, I think my brain thinks in patterns and gets so excited about everything. I'm really verbose. I over contextualise. I give everything thinking that that's gonna help people.
I wonder whether or not is there a way that we can get Cole to write a GPT agent that would help to take the information from people like me and push it through your process so that we sound more articulate and more on point and really communicate really well.
Cole Nussbaumer Knaflic (37:26)
So I'm going to challenge that, Fi. I don't think you want a bot to do that for you. We want to use our own brains and the people around us. So one thing that we often do in a workshop setting, because you are not alone in this desire to want to hold on to all the detail and take your audience through all the detail. I mean, it took time and energy.
It might feel like it helps build credibility. But the challenge is when we take our audience through everything we've done to end up at this one point or maybe a couple of interesting points, we're basically asking them to shadow us along the analytical process that we undertook when really that's the value that we're meant to be adding. They don't want to sit through all of that for the most part. You'll have some audiences who do care about that detail. But when you know that they don't,
Fiona (37:53)
Mm.
Cole Nussbaumer Knaflic (38:18)
the more succinct you can get, the better chance that you're not going to lose them in the details that may not matter or not matter as much. And one way to get good at this is just to talk about the thing with other people. And it doesn't have to be anybody from your eventual audience. It could be a partner or a friend or a colleague.
So we go through an exercise in workshops where we actually, have people work through a worksheet that where they start with their project in mind. And then we're assuming, you you've already analyzed the data, you know what you want to say. This is just how do you get good at saying it. And at the end, they're down to a single sentence that is meant to encapsulate the gist of what they want to get across. Not that it'll ever be that single sentence that you
articulate, right? This is not a you go and say the sentence, drop the mic sort of situation. But if you think about if you can't clearly articulate your point in a sentence, how in the world are you going to put together a graph or a slide or a presentation that gets it across? And just the amount of wordsmithing and thought that has to go into getting from detail, lots of detail down to a sentence or two is hard. But there's
interesting clarity of thought that happens during that process. And if you can then take that sentence and say it to somebody else and have them ask you questions, it's this really interesting process of getting ideas out of your head, out into the open, where you can consider them from a different viewpoint and where you can start to see like,
OK, I thought we needed to go through all of this, and maybe we need to go through some of it, but this one thing is the key thing. So that's what I can't have get lost. And that might mean when it comes to your ultimate presentation, maybe you do keep a lot of the other details. Perhaps it's an audience whom you've not interacted with before, so you're not sure how much detail they want. And you need to establish credibility. So you might want more of that there in that scenario.
But if you can encase it with starting out by saying, here is the main thing we're going to talk about today, and this is why it's important, and then you back up into some of those details and you keep coming back to this north star that's guiding you through, now by the time you get to the end, you've reinforced it several times. You've made sure every bit of detail you're going through actually helps support it.
or drives the right sort of conversations about where you need to get people aligned. It just allows you to frame things in a different way and allows you, I like storyboarding as part of this too, where I am a big fan of little sticky notes. I like actually even the size smaller than this. I'm holding up a three by three. I like the like one and a half by one and a half. I don't know offhand what that is in metric.
The little sticky notes where you can write down ideas and just, you if you have something you're going to be communicating, first you start by brainstorming, write down an idea per sticky note of things that might eventually be data you include or a slide or an example, get it all out of your head for a few minutes by brainstorming, then start arranging. And as you arrange, you also have a discard pile because I think there's something really powerful about writing down an idea, considering it.
and deciding not to include it versus if we just go straight to our tools, feels like they need to answer, you know, what we put together needs to answer every question that might come up. And that's one of the reasons we end up with these massive decks that don't actually get people to understanding in an easy fashion. And it doesn't mean we don't still need to know the answer to that question if it comes up, but it allows you to hone the content that you create in a very different manner.
And I would say in the scenario that I mentioned where if you don't have credibility, you're not sure how much detail to include, I still don't recommend cramming it all into the communication. And one approach there can be you keep the story upfront. You apply best practices to your graphs and your slides. You keep that succinct and to the point. And then if there are
Or is other data where you anticipate you might get questions or you just want to have it there just in case stick that in an appendix. Don't worry about beautifying it. Just have it there in case you need to turn to it because that's actually an incredible way of building trust with your audience is if they ask you a question and it's clear and how you respond to that that you've anticipated that question and you can eloquently address it.
Sarah (42:58)
Yeah, and I would say we are getting better at putting things in the appendices.
Fiona (43:03)
Yes, we are. And one of the things that works really well with the two of us, because there's two of us and we've got a lot of experience to share, we can bounce that between one another and make sure how does this read through? And then the other one will come through, readjust the story, you know, and think about editing it down or adding things that don't make sense. So I definitely recommend having that kind of body double to help you out.
Cole Nussbaumer Knaflic (43:29)
Definitely helpful. And you can look at things through different lenses. So for example, if you've put together a deck, one way to review it is where you just flip through and look at the slide titles. And those titles, and this could be applied in different scenarios as well. It doesn't have to be a deck specifically. But those titles alone should tell the overarching story. And if they don't, then you want to figure out, I missing something? Or do we need to pull some things apart? If you are working in
PowerPoint or similar, can be in slide sorter view where you can actually move things around after you have them developed and vet different flows. And then yeah, talking through things with somebody else. These are all great ways to be able to figure out if what you've created is going to serve its intended purpose, or if it isn't, where you might focus your iterations.
Sarah (44:18)
And just staying in the storytelling space as well. I really like the way you, what I call, layer the information. So you kind of build on the story and that may be using something like PowerPoint where you bring up maybe the bold headline to start with and then comes the chart and maybe the annotation and the takeaway kind of all just, and you can kind of tell how you would talk that through and you're not overwhelming your audience with, here's everything that I know.
but layering the information.
Cole Nussbaumer Knaflic (44:48)
And that's really, again, is coming back to designing things for our audiences. Because any slide you've built or graph you've built, you know it because you put it together, which means when you look at it, you already have this curse of knowledge where you've already seen the thing there is to be seen. So you can't any longer not see it. And so sometimes we forget that for our audience, they're encountering it.
for the first time. And even if they've seen it before, they've seen a zillion other things since, and so it's not going to be top of mind for them in the same way that it is for you. So when it makes sense.
having a progression like you described, Sarah, can be so powerful. And it doesn't take long, but where you can say, all right, today I'm going to walk you through something that we've done. Here's the title that sets the stage. I'll put up a graph, the skeleton of the graph, before I even add any data. Here's the y-axis and what it means. Here's the x-axis and what it shows. Now let's layer on some data. Let's highlight one thing so that we can talk about some of the context there. And what that enables you to do is you can actually
build to something dense and complicated that no longer feels complex because of the way you've broken it into pieces through your communication. So it can be an excellent way to bring people through a communication when you have the benefit of being live with them.
And actually, it works beautifully in virtual land as well, because you can tap into people's fear of missing something. If you put up a dense slide, you are asking people to turn their attention to their inbox. Whereas if they know, if they tune out, they're going to miss something, now you've created a bit of FOMO that might work for holding their attention that much longer, which can be useful.
Fiona (46:30)
all of these great techniques. I love building that FOMO and also the curse of knowledge. That's really powerful.
Cole Nussbaumer Knaflic (46:38)
Yeah. It's actually a thing, right?
Sarah (46:39)
You
Cole Nussbaumer Knaflic (46:41)
It's a bias. Mike talks about it in one of the chapters he wrote in before and after.
Fiona (46:46)
Yeah. The book also has a framework called the Big Idea. Can you explain to our audience what it is and why it's important to get right, even before you open your tools?
Cole Nussbaumer Knaflic (46:49)
Yes.
Yeah, that's that single sentence that I was talking about earlier. So the big idea is a concept I initially was introduced to from Nancy Duarte in one of her books. think resonate is the one where she goes into detail on the big idea. So big idea has three components. It should articulate your unique point of view. It should convey what's at stake and it should be a single sentence. And so over the years, we developed a, what we used to,
do an exercise in workshops where we would just say, okay, take your project and form the big idea. But that was hard for people because it is very hard to go from all the details down to a single sentence. But there's such value in doing it that we iterated over time and came up with a worksheet to help with this, where it just breaks things down into first, let's think about your audience. Who are all of the various audiences who might care about the thing that you're going to communicate? Now let's get more specific. Who are you communicating to?
at this point in time or who is the decision maker, how can we narrow? And it doesn't mean that there aren't still other audiences who will care or will be listening, but when you can narrow, you can really design things with your highest priority audience in mind, which improves the chances of it working better. So after you have audience comes what's at stake.
And what's at stake, you can frame either in the positive. Here, audience is what you stand to gain or the benefits if you act in the way we're recommending or the negative. Here is what you stand to lose or the risks if you don't act accordingly. And then finally, let's put it together in a single sentence.
And so when you break things down piecemeal, it makes it harder for extraneous details to work their way in and easier to get to that single sentence. Oftentimes, it does take writing out more and then cutting and refining. And again, having a conversation with somebody else can be so key for refining because it helps us understand when we're using.
language that might not be commonly known, acronyms have a way of creeping their way in big time in lot of places, or where we might be making assumptions that we aren't articulating or that may not even hold true. And you can have a lot of really good conversations as you work to refine that big idea. And I'm a big fan, and actually, Fi, Sarah, this would work awesome for the work that the two of you do. But if you're working on something as part of a team,
have people first individually work their way through the worksheet, come up with a big idea, and then share those. And from the individual big ideas, come up with a single one that everybody agrees on. It helps ensure that there's alignment.
always people approach this in slightly different ways. And so you may find like, Sarah, use this one phrase, that's key. We've got to have that there. know, Fi, yours had this word that really makes it work, or you approach it from the negative, not the positive. You know, what will work best in this scenario, or where are we more likely to have success? So the conversations that ensue from that can be super useful in terms of getting people aligned and coming up with a really good strategy.
Sarah (50:02)
Yes, so much to take away from there. And I think, you know, we spend a lot of time curating and perfecting and making everything look as clean and clear and logical as possible. We also run the track sometimes of when we present something that's like, that looks easy or, you know, that's obvious. And sometimes it can get like, you know, it's like the devil is in the detail, but also
What's that Richard Branson comment or quote of like, it's any idiot can make something complicated. It's a genius that can make it look simple. I probably totally misquoted that, but that's where I'm going.
Fiona (50:42)
Sounds good.
Cole Nussbaumer Knaflic (50:44)
No, it's true. And it's about simplifying, but never oversimplifying, and finding the right balance of detail for your audience. Because I think that's where we run the risk of simple not being the right thing. Because it really is going to be in the context of who you're communicating to. And that's why I think a lot of times people like to have hard and fast rules in this space, particularly as you're learning. And I understand that.
but it also is misguided because any sort of rule is going to have its exception and its understanding why that really, I think, helps people make decisions in a smart way, not only in a given scenario, but more broadly. can think of a time, for example,
I was at a public workshop. This was in Seattle, where we had a mix of people coming from different organisations. And we're on the topic of colour. And I made some offhanded comment like, you never use Hot Pink in your board presentation. Guy in the back of the room raises his hand. He's like, I work at T-Mobile, which, if you don't know, is a US telephone company. Hot Pink. OK, you use. OK, so that's an exception. Or people talk about, well, never show a p-value on a graph.
Sarah (51:50)
Hot pink.
Cole Nussbaumer Knaflic (51:59)
I'm sorry, but if you're communicating to engineers or statisticians, if you don't have that there, you're going to lose credibility immediately. So it depends, is the only reasonable answer to any question like this. And a lot of time, what it depends on is your audience. I know we keep coming back to that, but it really is such a critical component of all of this. Because ultimately, what we're trying to do when we're communicating, whether with data or with anything, is
to get somebody to understand something or to do something. And so by keeping those people in mind throughout our process, we actually are more likely to create something that's going to work for them, which will make us more successful.
Fiona (52:41)
Spot on. like, I think one of the things with that industry changing at the moment so quickly, you know, we're seeing the introduction of model context protocol, which means that people are able to ask questions through Claude, for instance, of their existing data, generating charts, you know, everything's changing so quickly.
it's really difficult for us to even keep up on the basics, let alone thinking about where we need to evolve our skills. You challenge it? Because I think that with generated charts, talk like, am I like, am I off on this? You're not a fan of it? Or how do you see the role of human judgment and storytelling evolving?
Cole Nussbaumer Knaflic (53:21)
Yeah, it's
such a great question. actually, I've been preparing for a panel that I'll be doing in a couple of weeks. It will probably be already have passed by the time people are listening to this. But the topic is data storytelling in the age of AI. And really where we're going to end up focusing with it is on the uniquely human pieces of that process. The push button chart.
We will eventually get there, but we're not there yet. And the fact that people are using things that way makes me so nervous. I think most people who have played with GenAI for creating graphs has run into a scenario where there's data that just gets made up. one of the things that we know about AI is its success, quote unquote, is pleasing us, which means it's serving up things and
what with such confidence that that can be really dangerous. so I am really curious and am constantly testing how can we use the speed because I think that is what it has that humans will not be able to match is just such speed of.
Being able to take so much information and do something with that so incredibly quickly. I think I've commented on this before, but I think back to one of my first projects when I first started working at Google, we have a massive employee survey that we did every year. And I was reading through verbatim comments, like thousands of them, trying to figure out...
How might I categorize these? What are the categories? OK, now that I have the categories, how do I attribute? Does this one fall into career development, or is this more of a manager issue? Or is it both? How do we count this? And I spent months on this. And this is something prone to error and inconsistencies and human judgment and all the things in all the wrong ways. And this is something that natural language models like
can do instantaneously now and consistently. So that is amazing. And we should absolutely be thinking about where and how we can use the speed of AI. But there is a dark side that I am still figuring out where. Where do we use it? And where are the uniquely human things that we need to have stay that way?
I think I end up using it mainly today as a brainstorming partner. If you think about the different parts of the process of going from data to telling a story with it, where can things go wrong? Well, they go wrong when I have a gnarly audience member who's going to challenge me in ways that I didn't anticipate. Well, we no longer have to have that because I can now tell AI to be my gnarly audience member and help me poke holes in what I've done. And it will do that to the nth degree, right? It has an incredible amount of
patience So if I have questions that somebody else might not have the patience to go through with me, like brainstorming and doing back and forth in that manner, I think can be really useful. And so one of the things I've been thinking about is just mapping out the process that we teach in the books and in our workshops and really being thoughtful about where could you use AI here? Where might it be dangerous? And it's maybe not there yet or
Fiona (56:05)
you
Sarah (56:25)
you
Cole Nussbaumer Knaflic (56:30)
there is this idea that if we outsource our brain to AI, are our brains not going to work as well any longer? Like think that's a real risk and I want to use my brain. Maybe I'm naive, but I plan to continue to and I might be able to get out of this stuff early enough that I don't have to outsource it to AI, we'll see.
Fiona (56:49)
Absolutely. mean, I think some of the some of the ways that the brain degrades, you know, we don't want to end up all being sitting there mute in the retirement villages because we haven't had the opportunity to really strengthen things. we do use it as a sparring partner day in and day out. But even this week, I've had
huge successes and huge failures in the AI space with the hallucinations, know, the huge success. I connected Tableau MCP, used the Superstore data set, and there's a really interesting insight, it's very familiar for the people in the community where
tables in the East are not profitable and causing some issues. And it really brought that up to the fore. And I just wonder if there's a way that we can iterate through this, but still have the human element where we're verifying what's happening and then also presenting it in a way that I still don't think that AI is there yet in terms of the visualisations.
Sarah (57:49)
Fi, you bring up a really interesting point. We know when it's hallucinating, because we've been in the industry enough, but there's a lot of people that, you know, or if we were looking at a subject that we knew nothing about, we wouldn't particularly know if it was hallucinating. And I think that's when it's really dangerous.
Cole Nussbaumer Knaflic (58:06)
Absolutely. I was laughing because I was just for part of this panel that I'm going to be on. doing a before and after of what an AI generated output would be for a given data set versus what I would do with it. And it's some different distributions from a manager training program and AI puts it in a box plot. And I had used this for a presentation that I did a while ago and was just refreshing it to fit with what I need.
And the amount of prompts it took even to just get it to work in the size and the colours. I noticed only after, like, one of the box plots, the whisker is actually extending off of the graph. So you don't even see little things. But the amount of instruction it needed to get to, like the slick looking bad, was already a lot. yeah, it doesn't yet get to the good when it comes to data storytelling. Maybe it hits it every once in a while.
But think the challenge is it can look pretty slick. So if you're looking at it quickly, you're like, this is great. It's only when you start to dig in deeper that you're like, but it made up that data point. Like this thing extends off the graph. Like, actually, it looks shiny at first, but it's right with errors.
And as the person presenting that, if you're doing that, that's risky because that's your credibility that's on the line. You can't just say, well, AI did that. That's not going to work when it comes to passing off responsibility. So tread carefully.
Sarah (59:34)
For sure. For someone in our community who's read the original storytelling with data book and wants to take their skills to the next level, what's the single biggest shift in thinking that before and after will give them?
Cole Nussbaumer Knaflic (59:50)
I think, so before I knew you were going to before and after with that question, the single biggest thing I was going to say is really about time. I think oftentimes we don't leave enough time for this part of the process, the data storytelling piece, which if you think about it is really the only piece that others see of the whole process. And so it should take probably at least as much time as all of the rest of the process.
so that you can make it count. And so I think the way that before and after illustrates that is you see the iteration. It's not just the before and the after. It is that process of trying out different things, considering different strategies, looking at things different ways, planning out for various scenarios so that you can understand not only
what was the thought process we went through and what were the considerations that time and what can you can learn from that, but also just recognizing that this does take time to do it well. And it almost always takes more time than we think it should. And yet often when faced with a project, you have to go through the other steps. You have to grab the data and analyze the data, but you can almost get away with throwing it in a graph and being done.
But that graph, if that's the only thing that other people see, deserves at least as much time and attention as those other steps. So I think carving out time, recognizing that it takes practice and learning from each time and iterating with new knowledge, that this is one of those spaces where you can be consistently improving. I've always hated the idea of an expert.
when it comes to data storytelling, because I think we can all continue to be more nuanced in how we do it. But that happens through a process of actively paying attention every time to what you're doing and what's working and what's not working and refining and iterating. And so a lot of what has ended up in the books are the things that I've gone through or that the team has gone through so that we can hopefully help speed that process up for others.
but it really still does take getting your hands dirty and going through the steps yourself. And I think when you take the time to do that, though, the benefits are so immense because it means you can shift the conversation where the conversation is no longer about the graph or the colours or the font or whether the data is correct, but rather the conversation is about
what you learned and what now you can influence in terms of a business decision or a conversation that has real impact. And that is a hugely powerful thing.
Sarah (1:04:02)
What's the most common chart crime you still see in the wild?
Cole Nussbaumer Knaflic (1:04:07)
including unnecessary detail.
Fiona (1:04:10)
chart type forever, what would it be?
Cole Nussbaumer Knaflic (1:04:13)
donut chart.
Sarah (1:04:14)
What's the one thing people should stop doing immediately with their data slides?
Cole Nussbaumer Knaflic (1:04:19)
designing for themselves, not their audiences.
Fiona (1:04:21)
Yeah, earlier when you said, you know, it's selfish. was like, that hits, that hits hard. And I think that's absolutely right. You need to be designing this for your, your stakeholders. Well, that was a really quick, quick fire round. Cole, this has been such an awesome conversation. But before we wrap up.
Cole Nussbaumer Knaflic (1:04:37)
Super fun.
Fiona (1:04:39)
that you really want our audience to take away from today.
Cole Nussbaumer Knaflic (1:04:44)
I think I'll underscore the theme that we've been talking about that's come up again and again and again, which is the next time you find yourself needing to communicate data, pause and think about your audience and what you need them specifically to know or to do. And then the rest of the puzzle pieces will more easily fall into place.
Sarah (1:05:08)
I love that. Stop being so selfish and think of your audience. Thank you so much for coming on and being so generous with your time, especially as what I know is a ridiculous hour on your side of the world and yours, Fi. It's been a treat.
Cole Nussbaumer Knaflic (1:05:25)
Thank you for having me. It's been super fun to chat.
Fiona (1:05:27)
Yeah and so if people want to connect with you or find you, get the book, can you just share some of those details for us?
Cole Nussbaumer Knaflic (1:05:35)
Absolutely. All of the resources that we've talked about can be found at storytellingwithdata.com. There you'll find information on our team trainings, our public workshops, videos, books, podcasts, our online community and all the rest. Also just mention briefly that for those who have little ones in their life, daphnedrawsdata.com is another great place to check out.
Sarah (1:06:00)
We'll make sure to drop all those links in the show notes below.
Fiona (1:06:05)
And if you've loved this conversation as much as we did, don't forget to hit like and subscribe, leave us a review and the most important is to share with your fellow DataFam members and colleagues.
Sarah (1:06:16)
What a conversation. If there's one thing I'd like to take away from this episode, it is the goal isn't just to make beautiful charts, it's to make charts that are impactful and help people take action.
Fiona (1:06:28)
and not be selfish. Cole's book, Storytelling with Data Before and After is available now wherever you buy books.
And if you're someone who creates dashboards, creates reports or presents data in this form, make sure that you get the book. It'll change how you think about every chart that you make.
Sarah (1:06:45)
And thanks for listening to UnDUBBED where we're unscripted, uncensored, and undeniably data. See you next time.
