Search
  • Francisco Mahfuz

E88. The Essentials of Data Storytelling with Brent Dykes



Below is an AI-generated transcript and therefore it may contain errors.


Francisco Mahfuz 0:00

Hi everyone, Francisco here. Just before we get started, I wanted to share something I'm really excited about. I recently launched the story powers bootcamp, a course that teaches you everything you need to know about how to find craft and tell stories that work. But it's not just an online course, because you get personalised feedback from me for all the practical activities in three hours of life coaching to work through any challenges, or focus on specific projects. So it's like if you bought a cookbook, but the chef came along with it. So go to story powers.com and click on Course, all the information you need will be there. So please check it out. And if you love the show, and would like to support us, you can go to buy me a coffee.com forward slash story powers. I drink about five coffees a day, so any support would be much appreciated. All right on with the show.


Welcome to the story powers podcast, the show about the power of stories, the people who tell them and why you should be doing it too. I'm your host, keynote speaker and storytelling coach Francisco mahfuz. My guest today is Brent dikes. Brent is the Founder and Chief Data storyteller at analytics hero. He's also the author of effective data storytelling, how to drive change with data narrative and visuals. His Forbes article on data storytelling has been viewed more than 300,000 times. And it's often the top search results on Google on that subject. Now, I spend most of my life considering myself on Node, but I have to give it to brands. He absolutely wears it on his sleeve. His first blog was called Power Point ninja. His first book was web analytics action hero. And he ends his latest one called the Marvel superhero legend Stanley. And now that I can see what his home office looks like, just behind him. There are two huge blown up comic book covers showing Spider Man, the Hulk and Wolverine if being a node is wrong, brand doesn't want to be right. Ladies and gentlemen, Brent dikes Brent, welcome to the show


Brent Dykes 2:11

now that you've outed me as a nerd, okay, well,


Francisco Mahfuz 2:15

you what I'm trying to get very relevant. I was reading through your book. And there was a couple of things that I thought were interesting coincidences. One was that when you talk about a story that I know you've you shared, or you share often when you on podcast about I forget the guy's first name but Semmelweis you talk about, there was essentially his discovery that germs killed people. And then later on that story, you mentioned the guy who, who essentially discovered it, who was Lord Lister. Now, I don't know if you've if you've noticed this, but the one of the first five guests of this podcast was an actor called Ralph Lister, who just happens to be the great grandson of Lord Lister.


Brent Dykes 3:01

Oh, wow. Wow. Wow, that's impressive.


Francisco Mahfuz 3:04

Yeah. So that was the first instance. And the second one was that somewhat ironically, you screwed up my data. Because I have been quoting that Made to Stick experiment. And in some speeches are very, you know, data driven and fact based and statistics based, and some some people use stories. And then, and remember the, the, you know, the factual information. And I, I realised after reading your book that I had been quoting it by 1%. Wrong, I had to go back and redo my slides. But you know, it's, it's fucking awesome.


Brent Dykes 3:42

Oh, no. Oh, good. You know, we want to get that accurate. So let's good. I'm happy to hear that I helped you in some way.


Francisco Mahfuz 3:50

Yes, very much. So the first thing I wanted to, to just get out there is, in this podcast, we don't need to cover at all, why stories are important. So that part we it's covered. We have done that when I've done that one to death. I don't think anyone would be listening to this, if they didn't buy into that concept. What I actually want to start with is something is almost the opposite of why why data is important. And the reason I'm asking that I'm not, you know, it's not a softball question is because I'm thinking of, of what I started thinking of as the rokeya effect. So do can you just share the the Rokia story that you mentioned in your book?


Brent Dykes 4:37

Yeah, so there is a it was also taken from Made to Stick that great book by Chip and Dan Heath, probably many of your listeners have already read that book. But in that book, they share this example from the Carnegie Mellon University where they had a bunch of students take a technology survey and and then they basically for taking that survey, they gave them 151 dollar bills, and then that's When the experiment began, because then they gave them one of two pamphlets or brochures on a charity, a real charity called Save the Children. And, and one version of that brochure had all data had all statistics about the suffering of children in Africa due to poverty, you know, famine, disease, you know, and hundreds of 1000s millions of kids being affected and all these statistics. So there's that version of the of the pamphlet. And then there's another version, where they just talked about a seven year old girl from Mali named Rokia and talked about the suffering of her family and how they were struggling to make ends meet. And and then they looked at what were the average donations for both versions, and they found that the rokeya version had double the number of donations of compared to the statistical version. So I think in terms of like, we might say, well, in that case, you know, all we need is the rokeya version, we just need to tell the story. But I think a lot of the stories that we need to tell, especially in the business world, will be about data. And so it's not that we just drop all of the facts drop all of the, you know, we don't include those statistics, I think that's where the data storytelling comes in. That's my specialty, where I say, you have data, you have statistics that you need to share, why don't we weave in a story? Why don't we weave it together with a story in a narrative format, and deliver it in a way that's more like a story. And so, you know, often, you know, as I said, in business, we are constantly working with data, we have insights into the business insights into processes insights into customers. And rather than just taking that pure data approach, and just doc dumping it on people's laps, let's let's construct a story. Let's craft a story with it. And so let's bring in the narrative components that can make it really powerful. I think those two ghosts go together.


Francisco Mahfuz 6:48

For anyone who didn't catch exactly why I'm raising this as a potential issue is that people have publicised what you mentioned, you know, when people got just the story they donated over doubled the amount of money compared to a negative data. But what I think was the really interesting part of that which I wanted to raise with you is that when they were given the story, and the data, they donated more than it was just the data. But still less than just the story, which would suggest, if you were trying to make a broad generalisation of that one experiment, that you shouldn't give people the data, you should just give people the story. And of course, in certain cases, like a lot of the stuff you cover in the book, you what the story is about data, there is no way not to talk about the data. If when you give people the data in the story, the story still outperforms, at least when it comes to charity, giving an in that one experiment we mentioned, then why why give people the data, if the data is not the story, if you're just using the support your argument? Yeah,


Brent Dykes 8:05

the interesting thing about what they did is they they took both brochures, they took the data and the story and gave them both together. But what they didn't do is they didn't integrate them together. So they weren't weaved together into a cohesive story using those data points using, you know, using the story of Rukia together. And so I think that's the opportunity that we have, as data storytellers to not just say, here's all the data. Oh, and by the way, here's a story. Let's merge those two together, make them really integrated. And, again, I don't have the data to support this. But my hypothesis is that we can make it even more powerful than maybe just Rukia. Alone. We can take those data points, and it's evidence that support you know, it shows why a family like Ricky's family is suffering. And it's not just, you know, not just Rokia and her family, it's 1000, hundreds of 1000s millions of other families in Africa that were struggling at this time as well. So that's my hypothesis. I believe that you know, when the when you put those two together, you're going to have a much better more powerful communication than just if you did either or, or even just, you know, not integrating them to say, here's the data. Here's the story.


Francisco Mahfuz 9:17

Yeah, I don't think there's any question that if you if you keep the story in the data separate that he might, he might actually hurt your case. But where I think, I think perhaps the concern for your your hypothesis, some other research that I need to figure out where it was from because I keep quoting either can't remember what the sources, but this is something I heard on the Sam Harris podcast, and he talks about they were looking to charity giving, and it was something like the Rokia scenario were they talking about one little girl in Africa, and then they added a sibling. I think it was a brother. And then they added the rest of the people in the village in the form of data. And what they found is the moment you You go from one person, to a few people, even if it's, you know, still children that are hungry in the same village, people's donations went down. So I think it was, then I don't think the issue there was you gave me data and you pulled me out of my, you know, emotional analysis of emotional connection with the person into something analytical. I think the issue there was just, we seem to be moved more, when it's one person and not when it's a few people, only so much you can you can do to avoid that, I guess. Yeah. I


Brent Dykes 10:34

mean, I think it's, you know, we can relate to individuals, we can relate to smaller groups of people, when you say, oh, there's 1000 people, you know, or 10,000 or 100,000 million, you know, we just it then it's hard for us to grasp, you know, so it's like, it was Mother Teresa, who, you know, said, You know, I wish I had the quote at my fingertips now. Now, I'm like, you I wish I had, but she she had a quote,


Francisco Mahfuz 10:57

I think I have it, I have it. I think it's when I look at the mass of humanity. I feel, I don't know, I feel like overwhelmed. When I look at when I look at one person, I feel moved to act or something along those lines.


Brent Dykes 11:14

I think that kind of encapsulates what, what's going on with that and why you know, maybe, and what I tried to do in my data storytelling is I always try to humanise the numbers, right? Can we humanise the numbers as much as possible, because if we just keep it in generic terms of this is how many people converted this is how many people did this or that, you know, it's harder for us to relate. But what I what I really encourage in my book is, let's humanise this, let's, let's let's build a persona for one of these customers and show their experience going through our process or, you know, whatever we're doing and let's give them a name, you know, hey, it's it's, it's it's Mary and Mary struggling to return her product after Black Friday, or whatever it is, and show her journey show her experience and going through the struggles that she had, using data to support everything that we we share about her journey, you know, and so, I do think that that's, that's an important thing to to, you know, not get so focused on the weight of the numbers and the, you know, here's all the evidence and just, you know, oh, it's just overwhelmingly clear that we need to do something. No, let's, let's take it and let's humanise that data, and show it from the perspective of the customer from the you know, whatever, whatever analysing if it's employees, or customers or voters or whatever, you know, let's show it from their perspective and really humanise the data as much as possible. I think that plays into storytelling.


Francisco Mahfuz 12:43

Our one I'm going to want you to define what you mean by humanise the data. But I think there's one other definition that we probably should should give first, which, what is a data story? In what isn't a data story?


Brent Dykes 12:58

Yeah, I mean, I think there's been a lot of misconceptions out there, you know, that was part of the reason why I felt like I needed to write my book to clarify, you know, I felt like storytelling or data storytelling is something that's very powerful. And data stories are very powerful, but they were kind of being used by different vendors and other people in a very kind of careless way. So there's some people who think like data visualisation is data storytelling, a data story is just a chart or a, you know, a visualisation of some kind. And that's telling every, every chart tells a data story, or tells a story, you know, some people would say, and I'm like, no, no, a chart often doesn't tell the complete story, you know, afterwards. In many cases, in order to tell a story, we need to have multiple charts to really walk people through something and so for me, a data story is really where we're taking some kind of insight that we've we've gathered, we want to communicate it with other people. And we use narrative elements and visualisations often because, not not because we need to use visualisations necessarily to have a data story. It's more because a lot of times the data is complex and hard to kind of appreciate. And so, visualisations can be very powerful. And so I talk about those being the three pillars of my data, storytelling book, data, narrative and visuals. And you know, and since I published the book, I've come to realise that, you know, obviously, visuals I'd probably say are the of the three elements, the one that's maybe not needed as much, it often is needed, because we again, like they said, we have we're dealing with complex data. But if you've listened to a really good podcast, and they're sharing data, and they're weaving in narrative, there's no visuals. There's no visuals in that podcast, and I would say that doesn't preclude it from being a data story. So I really think it it really does come down to the you know that your story is based on data insights. observations, and then you're crafting it using the story arc, you know, in using narrative elements like setting and characters and the plot or the story arc becomes part of how you share the details or the findings from your analysis. Now visuals are a, an incredibly valuable tool. And then I would say today, you know, most of the time we are going to tell data stories with visuals or charts, because it's, they are very effective. But, you know, if I was to boil it down to just what the two raw elements are definitely data and narrative being the key pillars and then visuals to kind of support those two.


Francisco Mahfuz 15:40

So how do you differentiate between, between say, for example, a data story that has has data has the narrative but no visuals from one of the one of the three types of data forgery, you call in the book so that the main one to meet the main ones appeals to me, because it's probably what most people do, and maybe I do, which is the data cameo?


Brent Dykes 16:03

Yeah, so in the book I, I talk about three data forgeries and so, you know, on the surface, we look at them, we say, oh, there's there's charts. So there's data, right? There's visualisations it looks like they're telling a story. It's a data story. And really, I look at it, how they're formed. And so if we were to look, if we start with a data story, how should a data story be constructed? Well, it's based on a starting with the data, looking at the data, exploring the data, trying to find you know, something about our business, you know, and we find an insight and we're like, Okay, now we want to share that. And then the next step is then to take, you know, those visuals that we've used, and a key thing is to edit them, basically thinking of our audience, okay, they haven't spent as much time in the data as we have, they may not have full context into the problem, or the issue that we're looking at, you know, there may be different things, and maybe their data literacy level is lower than ours. And so we have to adjust our story and also, you know, when we're exploring the data, we're looking at all kinds of charts of data, but then when we go to tell her story, maybe not all of that data and all of those charts are as important to telling our story and so we remove the noise from our story so that then we use different visualisations we editor visualisations to then tell a data story. So, that's kind of the, the process that you go through you go into the data you find an insight, using visualisations as part of that process and then editing or tailoring the visualisations to then, and crafting the story that to really speak to your audience. Now, the three data forgeries that I talked about, the first one is that I that I cover is the I call it the data cut. And it's kind of like a director's cut version of something. Because what happens is the analysts go in they they start the right way they go into the data, they find an insight, but then they assume what spoke to them, the visuals that spoke to them will speak equally well, to the to the audience. And if that's a business audience that hasn't spent a lot of time in the data, maybe isn't as data savvy as them. That's where the, you know, that becomes, in my mind a data forgery, because it's not going to communicate effectively. Basically, it's, I compare it to a director's cut of a movie, there's no editing involved, nobody went through and edit it to really make it a palatable for the, for the audience. Now, the one that you talked about is more comes, I see that coming more from the business side, where it's like, okay, I need to show that my campaign, my programme, you know, whatever business decision I made was successful. And so then we're gonna go cherry pick the data that supports my point that I'm trying to make. And that's the danger, though, you know, that's why what I call that one, a data forgery, because we're really not listening to the data, we're telling the data, you need to tell me this. And I need to show that I was successful, I need to show that this programme was successful. And I'm not going to listen to it, you know, if I come across a conflicting data point that says, Oh, actually, this aspect of the programme, wasn't that successful, okay? Either. Unconsciously, we'll just ignore it. Because it's like, oh, that's negative information. I don't want to I don't want to tell a negative story, I want to sell positive story. Or we might say, Oh, I'm going to bury that. I don't want anybody to hear about that, you know, that ugly detail of the programme that wasn't successful. And so essentially, we're cherry picking the data points we want. And so I call that when the data cameo that data is there for the ride, you know, but is it essential to the story? No, it's like a cameo role in a movie where you have you know, these guest actors come in and they'll play some little bit and everybody kind of laugh You know, like the Stanley in the Marvel movies, you know, he's not integral to the plot. He's not integral to the plot. He's just there for, you know, whatever, just to just for fun, and that's feels that's that's what it feels like with the with the, the data in those cases and then the third data forgery is one where you may have somebody who's very good at visualising the data. So they may be very talented at Tableau, or one of these tools, and they can create these really amazing visuals. But what they fail to do is really do any real analysis before creating this visualisation. And so I call that the data decoration. It looks pretty, there's, you know, oh, wow, this is interesting. And then as you begin to scratch the surface, you start to see, hold on, wait a second, there's really no insight that they're sharing. They're just sharing a bunch of data and almost hoping that the audience will do the analysis to find something of meaning for themselves. But there's really no point to the visualisation, it's kind of just left for open interpretation. And with storytelling, that's that that's not how we work we really want we have an insight in mind. And we're then you know, guiding the audience down a path to also understanding that insight, and hopefully taking action whenever we find an insight that is meaningful to the business involved to our organisation that that demands that decision. You know, that's what we're trying to do with our data storytelling. We're trying to support decision making, helping to optimise you know, what we're doing as a company,


Francisco Mahfuz 21:09

I guess the the big changing mindset is that the way most people, perhaps in the business world, they shouldn't do this. But when you take for example, the world of speakers or the world of storytellers, or anyone who communicates in more of a massive way, we usually what happens is we we see things in the world, right? So you know, I've something happened, I'm now this is a story, I'm going to share it with people, when there's an insight, there's something I learned from the story. And if I'm trying to build those insights into something bigger, maybe into a keynote, or a presentation of some kind, the normal approach is you go and look for data, you know, you look for data that backs up your claim, now, you won't, you might not necessarily be cherry picking it. But what I think will often happen is you look for you look for data, you find data that supports your claim, and you're done, you know, you're not going and spending hours and hours and hours and hours trying to disprove what you feel you've just proved. And and so that's one issue of it. And but the other the main thing for me from from what I got from your book was that a data story starts on the data and not on the hypothesis. So it's not you've seen something out in the world, you're going to the data to prove it. But you might have seen a problem in the world. And you're going to the data to find a hypothesis and not the other way around. Right?


Brent Dykes 22:37

Yeah. I mean, I think it's okay to start with hypothesis. But if you're not open to invalidating that hypothesis, then that's the problem, right? Because then you're not open to conflicting data points that maybe disagree with your hypothesis. It's like, at that point, it's like, no, no, I just want to hear things that support my hypothesis, I don't want to hear things that invalidate my hypothesis, or, you know, sometimes when you go in, you come in with, like an idea, you come up with a hunch or hypothesis. You know, I've had situations where I've had a hypothesis invalidated. I've had situations where it's validated my hypothesis, or I've gone into the data and found something more interesting than my original hypothesis. So it's like, you know, I think it's really about having an open mind. And really, you know, are we going to listen to the data? Are we going to listen, you know, are we going to listen to what it has to tell us. And if we are, I think we'll learn much more. Otherwise, you know, then we get into the cognitive biases that we have, you know, we're always gonna, you know, confirmation bias is rampant in the business world. You know, there's other biases that we have sunk cost, you know, all of these things that we, that that guide, how we make decisions, but it also guides, you know, where we go to get data to support our decisions, whether May, we've already made a decision, and it's like, okay, I need to support my decision with some data to show that I made the right decision, or I want to make a decision, and I need data to support my argument for making that decision. So I mean, I again, I think it's a mindset. And just because we have data now out in the business world, and we talk about being data driven, right, that's a big thing, you know, from the analytics world that I come from, you know, we want to be data driven. But often, it's almost like we have this cloak of being data driven. But it's, it's really not true data driven, because we're just cherry picking what we want from the data to support, you know, what our agenda is, and not really listening to it. And so, I think that's the key thing, if we can listen to the data, one of the things that was interesting, as I was writing my book, there are a number of statistics that people used and other books and other speeches that I had heard. And I was like, Oh, that would be great for my book. And so as I grabbed those data points, you know, I was I was ready to include them in my book and So I just did a kind of a check on where the sources of these data points were. And there were about five or six of these that I found that I just at the end of the day, I couldn't use them. I was like, these are not credible sources. You know, yeah, I even had, in some cases, there are experts in the industry, who used them. And I could have said, Oh, so and so said this and her book or his book. But again, once I went a little bit further looked into where those numbers are coming from they they were very highly suspect, very, you know, you can look on the internet in Google searches. And you can find them used 1000s of times, millions of times, but


Francisco Mahfuz 25:40

was one of them that how stories are 22 times more memorable than anything else?


Brent Dykes 25:45

That was one that I did not use, but I also did not call that one out.


Francisco Mahfuz 25:49

There was no, I've tried validating that one there is just sit the six to seven times is the one the one I've found.


Brent Dykes 25:56

Yeah. I reached out to somebody about that one and asked for her sources and and then I checked the source I couldn't find in those sources. So yeah, that was what I just I didn't call it out, but actually did a blog post on my website on effective data, storytelling, calm and I, I talked about six myths about, you know, how we process information and visualise data. And you know, I call them out as these are, these are ones that are people are using all the time, but I cannot use it with a good conscious in my presentations, or in my book. So I kind of flagged them for everybody. Yeah, I


Francisco Mahfuz 26:30

mean, it's it's the challenge sometimes that although there is more, there is more science and storytelling now than there ever was. Some stuff is not the findings are not necessarily that that strong. I mean, the the Rokia story that we mentioned earlier, I mean, that was an experiment. And by Carnegie Mellon, one could argue that you might have to run the experiment with 20,000 people to get really solid results. The one that we also talked about that I quote, when I do present a presentation on storytelling, which is most of my presentations from from Chip and Dan Heath, I mean, it's a very interesting study, the one the one that turns out that only 5% of people remember facts, or statistics after 10 minutes, but they ran that particular one that on the book, I think it was only with 10 people. Now if you blow that up to 1000 people, I wouldn't necessarily think that the the numbers would come out terribly different. But if you're trying to bank on this is really super solid science. Those are not necessarily the things that you would quote, you can quote me Paul Zacks work, or you rehab essence work, but but even those, I mean, I know Paul Zak has quantified his work a lot. But a lot of that is done on his private practice, he can predict with a lot of accuracy, how TV commercials do based on his work, but that's not necessarily the work that you find easily out there on on research papers. So anyway, moving on to data storytelling and how people do it. Well, we're poorly. There's one thing that you describe in the book that I think this is what a lot of people think data storytelling is, and you don't call it is, but I think it should the cake mistake.


Brent Dykes 28:21

Oh, the cake mistake. Yeah, it's probably very similar. I've, I've started to, I call it the analysis journey. But I've also built a slide now with a cake. And so I think what you're referring to is, you know, when we do, especially analysts will do this where they, they'll do a bunch of analysis, right? And then they'll go to tell their data story. And the story isn't about their findings. It's about their journey to finding the insights. And so they say, oh, first I looked over here, and I aggregated these two datasets. And then I, I did some factor analysis, and I was able to find this, but then I had to segment it and filter it this way. And then oh, and then I also came across this other data set that was really helpful. And so what are they doing? They're talking about how they bake the cake. And I think I kind of mentioned that briefly in the book, but


Francisco Mahfuz 29:11

you do you talk you talk about today, the ingredients and how you measure them and all this stuff that nobody really cares.


Brent Dykes 29:18

Nobody cares. At the end of the day, people just want to eat a slice of the cake. And I think that's that's something that I tell analysts, you know, I think they're the ones who make that mistake the most is that they, I wonder why they do this. And it could be because they they're worried that people won't trust their numbers, or they're trying to show how much work they put into this to show you know, you can you can rely on these numbers because I spent hours and hours and these are all the steps that I went through. But at the end of the day, you know, it's really about so what did you find? You know, that's what business decision maker what is it? You know, is there anything? Okay, that's great. You went through all those steps. That's awesome. I see that you're diligent. I see that you're, you know, spent a lot of time on this, but what did you find? You know, and that's, and I talk to different companies, you know, that I consult with in a in I see this preamble, you know, especially in pharmaceutical clients that I've worked with, where they have all these clinical studies and all of the the details and everything that's needed to kind of back up the data. But in the situation that I, that I found them, they were spending so much time on the setup, to kind of get to their insights, that they're running at a time with their audiences to actually cover the insights. And so I had to tell them look, your, your, your build up, you know, all of this information is, I guess, especially with pharmaceutical, you know, kind of products, it is it, you know, you could be sued, there's, you know, there's all kinds of things that you have to be aware of, but you can put that in the appendix, you can, you know, you can how can we streamline this process to get to the real insights where the, the pharmacists or the doctors need, you know, to know, something that can help their patients or can help their, their organisations to be more effective or efficient with how they're using these products. So, you know, it's, it really is important to not focus on the on the cake building, but actually, let's eat the cake together,


Francisco Mahfuz 31:11

I guess they are making a very classic mistake with let's be generous and call it storytelling, which and the advice for to avoid this comes from Kurt Vonnegut, I think, which is that you should start your story as close to the end as possible. So perhaps the concern there is they think, Okay, well, I've been told not to just give people the facts, or just give people the data, I'm supposed to tell a story here. And because they have no idea how to do that, what they think is, okay, I'm going to tell the story of my finding the data, which I think in some cases can be interesting, but can be interesting as a very small anecdote to just give some human colour to it. But but it definitely shouldn't be, I'm going to now spend 10 minutes walking through every single step in the process for getting the data that doesn't, that doesn't add anything. And it says you said it, it's it's setting that you're putting in place or context that you're giving them that doesn't necessarily help now, where I think that might, that could work, not as that approach. But a similar approach is, if you clearly had a path that you believed was going to be productive, you know, your hypothesis was this, you're 100%, you're very sure that this was the way to send a pharmaceutical you're trying to treat a particular disease, you were convinced that this specific approach was going to work. And you went down that approach. Any turned out that that doesn't work at all, or it turned out that that was a complete completely wrong hypothesis. And that is not that that could be what you call as the aha moment, it could be that you went on this right, and this is completely wrong.


Brent Dykes 33:03

It's part of you know, that's a key thing. For me, I think a lot of people use that definition of insight very loosely, oh, you know, like, I get a bunch of insights from this dashboard, I get a bunch of insights from this report, or I'm sharing these insights. And, and I felt like people are using that term too loosely. They're confusing observations with insights, observations are interesting or unusual things that we see in the data might be where we have a metric that is shot up by 20% 200% 2,000%, whatever it is, it's, it caught our attention, you know, but we don't know why at this point. All we know, is what we know something happened, something interesting, unusual happened. And that's an observation, that's not an insight. An insight is when we dig deeper, and we actually learn something that shifts our understanding on something. So it basically, you know, to your point, we have this hypothesis that this drug was gonna treat this in this timeframe. And that's kind of the accepted hype. And this might even be the hypothesis that the audience has, like, oh, yeah, like, that's generally what we believe is going to happen. And yet we did the study, or this couldn't, you know, we found we did some research, and we found no, it's actually not, you know, we found information that invalidated that hypothesis. It's, it actually happens much shorter, it, the treatment only takes a week. And then you know, patients see full recovery, it doesn't take eight weeks, it doesn't take 12 weeks. It's only a week, you know, and so that goes back to that as an insight. That's a shift in our understanding, everybody thought that it would take eight to 12 weeks to recover from this illness using this drug, but we saw massive, massively successful results just for one week of doses, you know, kind of making this up. But that's really what that's it's an unexpected shift in our understanding of something and that's What an insight is. And the key thing is when we have a shift like that, it almost demands that we take action on it, it demands that we share it with others, because we know oh my gosh, I thought this, I thought our customers love feature a of our product. But actually, their favourite feature is feature B, something we haven't invested in, we haven't been investing engineering time, and we haven't been promoting it to the market, we've been, we've always thought that feature a was our selling feature, we thought that that was the one that we go to the market with this new custom research we did, we found out they actually prefer feature B. And what that changes our product development strategy, that change that changes our go to market strategy. This is something I can't just sit on this myself, I need to tell the rest of the teams, you know within our organisation to to capitalise on this this insight that I have. So you know, if it really is a true insight, it is something that we naturally will want to share. And that's where data story comes in. Because then we're going to help to give them the context to fully appreciate what this what this insight means. And also, you know, if we are building out a full data story it's going to talk about what do we do about it? What are our options, okay, there's three options here, you know, we're recommending option B, because we can execute it in a in a shorter timeframe. And it's going to have a maximum impact on our on our company.


Francisco Mahfuz 36:27

So one of the one of the thing that I was quite interested in, in reading your book to find out your take on was, how much of actual stories goes into data storytelling. So for for anyone who, who has no clue what I'm talking about, a lot of people that work on storytelling have this big pet peeve, which is the idea that every everything has become a story nowadays. But the vast majority of things are not stories, at least by our definition. So that to me a story. And this I pick this up from the guys from anecdote, which is a big story consultancy, and they describe it as a story has a time and place a sequence of connected events, a character something unexpected, or surprising. And if it's a business storytelling, it has a point. Whereas a lot of people and there's a lot of very good work done on this from people like Donald Miller and story brands, they will use elements of a story, or the structure of a story to make messages more palatable, because we we instinctively understand those structures. But the the arguments from a lot of people, and I'm not completely gung ho on this, but I tend to be of the opinion that if you can tell an actual story, instead of just using elements or a structure tell when an actual story. Now a lot of the examples you've used in the book, by that definition, wouldn't be actual stories, they will be using story elements or a story structure. So my question here to you is, do you think that you should, as often as possible, tell actual stories with characters, or the vast majority of data stories in your experience won't go down that route? I would say,


Brent Dykes 38:18

you know, whenever we can, like, first of all, I would say more stories is good. Okay, if we can share a personal experience, you know, something that represents what we saw, I mean, I can, I can see that incredibly powerful, especially in sales, you know, like when you're when you're talking about your product or different things. And, oh, by the way, here's what another customer did. And you tell a story about, hey, they're a retailer just like you, and they were having problems with this area, their, their company, and, and we were able to help them to do XYZ. And then this is what they had, you know, I think sharing any kind of story is really powerful. But yeah, in some cases, I also wouldn't force it, I wouldn't, I wouldn't try and tell a real story for every set of data or insights that we have, because it would just come across as kind of like contrived, maybe forced, it may backfire on you, because it just feels like, you know, you're you're forcing it too much. So I would say you know, I probably more in the camp of, you know, many of the things that you said about, you know, what makes a story, a story based on anecdotes kind of things, those are things that we you know, something unexpected, you know, like, if you have an insight, you automatically have something unexpected that you want to share, you know, and you should have a main point. That's another thing that I talked about in the book, but the six, the six elements of a data story, I talked about you you should have a main point. It's not about taking a bunch of insights, and then doing a data dump, as we call it in the analytics world. We call it where you just, oh, here's something interesting. Here's something interesting. Oh, and you thought that was interesting. Here's something else and we just pile on. But there's no theme There's no main point. And that's one thing that I emphasise in, in the data storytelling elements is you've got to have a main point, you've got to have a theme to your each data story, you may have two insights, well, if they're not related, if they don't support each other, you have two stories that you need to tell you, if you didn't tell, build a story around the first insight needed build a separate story around the second insight. So you know, will they perfectly measure up? Will they be like JK Rowling novel or a Steven Spielberg movie? No, they're, you know, data stories are going to, we're going to try and borrow as much as we can. Because I do think that, as you said, instinctively, we as human beings like to process information in a narrative format. And there's lots of power to that, that we gain. And so we, I think we should aspire to incorporate as many of those elements as we can and make our insights and findings and our data stories as story like as possible without, you know, artificially bending things to make it almost hokey or silly, you know, where it's just like, Okay, you're, you're trying too hard to, to have a villain, you're trying, you know, you're adding intention when, you know, we're just trying to make a decision on something that, you know, we should be making a decision on this week. So I wouldn't go overboard and try and convert everything into a true literary kind of story that, you know, that's not what we're striving for. We're just trying to borrow elements that help us to communicate our information better. Yeah, I


Francisco Mahfuz 41:34

wasn't even thinking of literary literary stories. I think what I was thinking more of was, because when I talk to you about storytelling, I tend to describe stories as a real life example that make a point, which is, which is some somewhat imperfect definition of stories, but for most people, it gets them out of the literary Hollywood definition of a story, or the children's definition of a story. So what I what I was thinking of there is, for example, something which might be a process you called zooming out, in one of the ways of one of your story points. And is this if you have a bunch of customer feedback, and that is part of your data. And from that customer feedback, that was maybe the beginning of your data exploration that led to this data story, you're now trying to tell. If you have genuine customer feedback, and that customer feedback has come in the form of something that happened to a real customer, then I listened the way I would normally approach things, I think you could go in, there could be worse ways to start the data story than actually just telling a very small short story about that customer. And once you establish this problem, or this insert this thing that's happening to the customer, then you blow it up to a broader picture that involves more data that involves everything else. Whereas most people's go to approach won't be that they will get people completely out of the story. And it will just be numbers and graphs. And no points there will be an individual that may be we can relate to and that that I think tends to be the issue with a lot of presentations is that there are no human beings for you to care about is just a whole bunch of numbers.


Brent Dykes 43:21

Yeah. And that goes back to you know, that when I talk about after I introduce kind of like the narrative structure that I introduce my book to kind of, you know, how do you organise your findings into a narrative? Then I have I talked about, can you have heroes? Or can you have characters in your data story? And I say, Yes, you know, and that's part of that humanising the numbers, you know, wherever we can, kind of one of the steps in that process is put your audience in the shoes of the of the hero, meaning, you know, in most cases in business scenarios, that's probably a customer experience. Right? So we're, we're sharing okay, you know, and I'll give a couple of examples here, it could be somebody calling into the call centre, you know, and there's a number of steps and information we collect, and questions that we ask and steps that the, you know, the person calling into the system has to go through. Now, people in the organisation may not be familiar with that process, and all the steps that, you know, that are required, they all they know is that, you know, maybe we have an abandonment rate within the call centre, we have, you know, a satisfaction score that we're collecting, but they don't know the actual steps that people go through on a successful or unsuccessful call. And so one approach would be to actually, let's walk you through the steps. Here's, you know, almost like a flowchart where you're going through, here's the first question we ask, you know, and maybe you're inserting data 80% of the people make it past that question. Then they get to this question, and then and then that's where we're getting a lot of people having kind of getting circular, they're clicking on different places looking for in they're getting lost or whatever. So it could be really very high level kind of process kind of experience, helping them to understand Oh, okay, this is what our customers go through. The Another example would be actually to show, you know, what they're experiencing, whether that's on a website or an app, or in store it actually taking screenshots or pictures of that experience. And in helping again, people in the organisation understand Oh, so once people click on this button in the app, they go here, and whoa, I didn't, I thought we provided this information as well. No, it's not there. There's no information on this, this and this. So that's why they're having these issues with linking their accounts or whatever. And I think, as we kind of help people to kind of experience that now, in some cases, also, I talked about having, I come from the marketing world. And in marketing, we we use a lot of Persona building, you know, as we're doing redesign to websites, or campaigns or different things, we usually create different personas to kind of represent who we're targeting with these campaigns, or these experiences, I think the same thing can go for how we analyse their data, we can show hey, here's Roger Rogers trying to renew his subscription with us. He's gone into the app, he's clicked here, you know, to pull up his profile, he goes to the profile, and then and then we, for whatever reason, we lose his his profile information. And this is what he sees. And so Roger is frustrated. And he calls our call centre, you know, blah, blah, blah. But, you know, we can do that we can kind of craft these kind of story experiences, again, to your point, focusing on the individual making this data that yeah, we might have data for hundreds of 1000s, millions of customers who've gone through this experience, but then by sharing the example of Roger, and walking through the visual experience, or the flowchart experience, to kind of help the internal audience understand the pain and frustration that customers like Roger are experiencing, all of a sudden, you know, it, it hopefully, sparks interest in addressing the problem.


Francisco Mahfuz 47:05

And I'd be remiss if I didn't ask, but can we just go quickly through the structure that you've, you've landed on, for most people storage, so if I if I, if I, I let me just see if I if I still remember it if it stuck with me or not. So I understand is that you have the the setting, which shouldn't be too long, the hook, which is to get to get people's attention, which I guess, more often than not, that's going to be a problem that you're trying to figure out, then you have rising insights. So this is when you support that problem. And you should, those should lead to your aha moment, which is really worth your trip. The presentation is really about an after that, you should have solutions. And next steps. Perfect. You remembered it. Yeah. It's helpful that I've only finished reading it earlier today. Yeah. Fresh, fresh. Okay. So that structure is pretty clear. I know you base that originally on three tags, structure, which is fairly well known as rising action, falling action, all that stuff. Now, there's two questions that come from there. The first one is, as simple as that is, I think a lot of people would think this is all way too much work. Do I have to do this? Can I just do a like an you know, quote, unquote, normal presentation, and you have this this thing you call the story zone, which defines when you should, when you should bother with this. And when you can just like, you know, phone it in, like most people do with most presentations. So what are the types of what's the type of information or insight that calls for a data story, and you know, you're doing you're doing your insight, a disservice if you don't use a data story, which what falls on that story zone for you? Yeah, so


Brent Dykes 48:52

there's basically it's a matrix for people that are listening in, in, there's basically two axes that I look at. So the first x is pretty basic, is just Is this a low or high value insight? You know, so we're talking something that may only impact one or two people, or, you know, it's a fairly quick, easy fix, low impact on the business, that's going to be a small insight that maybe doesn't require a data story. If it's a medium to high, you know, where it could have a significant impact, there could be high value to the organisation, then I think that's where we'd want to focus on an insert or date data story. Now on the other axes, then I look at what type of insight are we talking about here? And I look at it kind of it's kind of a look at it from two perspectives on this. It's Is it a hard or an easy insight? And what I mean by that a hard insight would be something that's either hard to understand, or something that's hard to accept in. So let me give you some examples. If if I find an insight that is counter intuitive to the essential kind of way that we do things that are coming here on our team, that's probably a harder insight to accept, because it's counterintuitive. And so that's going to, in most cases have its high value or medium value. And it's counterintuitive, then that's going to be something where I'd probably need to tell a data story. There are other other things that I say, you know, if it's an unexpected result, that's quite jarring. If it's if it's kind of disruptive to our business, and how we do things, again, that's probably something that will require a data story. If it's costly to implement, again, that may be something so basically, if there's barriers or resistance to the insight, then that's when we need to employ a data story. Now, on the flip side of that, if we've got something that's expected, you know, like, we ran this campaign, the results come back and looks, you know, is a smashing success. And oh, and we're gonna be asking for some more money. How hard is it really, for people to kind of embrace them? Like, sure, I mean, wow, that was so successful, or, you know, if it's, if it's intuitive, or if it's, you know, any of these things that kind of, or it's a standard practice, it's not, you know, something disruptive that we're doing. So if it's, if it's anything that's just easy for us to accept as an insight, then that's where, you know, do we really need to invest the time to build and craft a data story, which can take time can take a lot of effort? And so I would say, No, you don't need to do a tele data story in those situations where, you know, it's an easy to accept insight. And even if it's high value, if it's easy to accept, again, I don't think you have to do a big production of a data story. It's only in those situations where it's high to medium value. And it's hard. It's, you know, we got bad news, you know, this programme that we just launched, after three months, now, we've analysed the data, we found it's not connecting with with our employees, hasn't improved their, you know, our retention numbers. And, you know, really, we need to reevaluate whether we should continue pouring money into this, or we need to, you know, maybe explore some other options. And that's a tough conversation. So whenever we have tough conversations that that are based on insights, that's when we need to invest the time to tell a data story.


Francisco Mahfuz 52:23

I guess another way of thinking of this is, do you need to change the story that people have in their minds? Because if you don't, if their story they're playing in their minds is that we ran this campaign, we expect that it was successful for these reasons. If all you're going to do is confirm that, then you don't need a story. But if you if they have one story, you're not going to beat that story or change that story with with facts alone, you're going to need another story.


Brent Dykes 52:50

Right? That's, that's a good summarization. Absolutely.


Francisco Mahfuz 52:54

And one final thing is just this idea of, of the most probably one of the most common objections you get, and I get, and anyone that works with storytelling gets when it comes to presentations, which is, you know, can you just, I just want the facts. I don't want a story. I don't want anything else. So you, you talked about the data trailer, instead of a data story. And I have recently done some work with a company that does branding and packaging, but mostly packaging, right? So they get a commission, they're trying to change, change the packaging of a product, and they have to figure out what the design is going to be. And they say that their biggest challenge is that the client, if the client had their way, they would just look at the design. Like they don't want to be told anything, they barely want a presentation, they just want to say, Okay, we went for these designs, here they are. And now let me tell you a little bit about them. And I asked them, Do you ever do it that way? I was like, No, we never showed the design first. Because you want to set the stage. You want people to understand what led to those things. And, and when they were pushing back a little when they said to tell stories. What I said to them is you just need to make the story good. And quick. Like you don't want to take 10 minutes telling it but if it if you jump into it, and within the first 30 seconds, it's interesting. They're not going to be hurrying you out of the of the stage, or saying Can you just stop whatever you telling us and just show us this thing. And so that's my, that's my view, I find it the day of the trailer that actually reveals the big insight to be pretty much saying, it's just not going to work. But if this is really what you want, here you go.


Brent Dykes 54:39

I really struggled with this because I had people coming to me after I talk about data storytelling and they'd say, Well, I love you know, I love this approach and stuff but I'm not sure if I can I'll be able to tell data stories in my organisation because I think I kind of feel like it's where executives and organisations are. They've been bombarded with data so often and efficient so often that they're just like cut to the chase, you know, give me give me the what is it that you know, you're sharing. And so people are programmed to kind of, they want executive summaries, they don't want to hear the long data dumps, they don't want to hear all this information. So I was like, Well, how can I? How can I adapt to this? How can I make something where, you know, in these situations where you have leaders or organisations that just just tell me the numbers? How can we adapt to that? And so I came up with this concept of a data trailer where basically we take the hook and the aha moment. And so those are really the 10 poles of any kind of data story in my mind. Because if we have that, that hook that shows, oh, there's something going on. That's interesting. And then and then what does that? How does that what does that mean to us as an organisation? What, what kind of shift of our understanding do we need to have? So combining those two together into the world's worst movie trailer, essentially, because you're, you're giving away the climax of the movie? Why would anybody go and watch the movie, but I, but I anticipate that what we're doing by sharing a data trailer is were peeking their interest, because at the end of the day, they're going to want to know, well, how did you get, you know, from the hook to the to the AHA, and what do we do about it? So yeah, you said, it's going to cost us $2.3 million and extra cost this year? What do we do to fix that, and then, by sharing that data trailer, our goal is to then pique their interest into, here's the rest of the story, you know, we now I can now they're giving me permission to tell them the rest of the story. So that's why I came up with that concept of the data trailer. As a way, you know, I'm hopeful as people get familiar with data stories and data storytelling, nobody today is saying, Francisco Tell me a data story. Nobody's going out there and saying that, but what they are interested in his insights. And if we package it up in a way that is like a data story, it's going to connect with them in a much, you know, it's gonna be much more engaging, it's going to drive more action, you know, and it's going to be a better experience for for both sides. So, you know, that was one workaround that I came up with. And we'll see, you know how that goes with that, that option. But I think people needed it,


Francisco Mahfuz 57:11

I guess my approach to that with most people is, you can fail in a way that everybody fails, which is just do a presentation the way everybody does, and it's boring, it doesn't generate the interest you want those engaged people doesn't actually lead to the action, you want to to move people towards. But nobody's going to say anything about it, nobody's going to criticise you over that. Or you can do something that might be considered a little riskier, but has a much higher chance of success if you do it half as well. And again, maybe this is this analogy is going to be lost that anyone who's not a football or a soccer fan. But you take penalty kicks, for example, the vast majority of penalty kicks that are saved, are saved when they're taken at low to mid height to the side, because that's the perfect place for the keeper to jump to, in most keepers gas the corner when someone takes a penalty kick. So they will jump one of the sides. If it comes around that height, they will often get it. If you just spank it hard in the middle, then very few keepers will ever get that because they've picked a corner and jumped. But if the keeper decides for whatever reason not to move in, you kicked it in the middle when they just grab it without having to move, you look terrible. You and people go, why are you doing that? Why you keep it right in the middle. But like the so easy for the keepers like but doesn't matter if even 100 penalties you kick that way, the keeper is gonna stay in place five times. Whereas if you kicked it the way everybody kicks it, then they will save 20 times. One way gives you more chance of success. But if it fails, it makes you look worse, because you weren't conventional. So I think my my workaround to most people is you just need to, if you're going to tell a story right at the beginning of a presentation, for example, which I think is one of the best ways to get people wind to the to the subject, just pick a short one in practice a few times. So you can tell it well, within a minute or a minute and a half. Nobody's going to be cutting you off that early on. And more often than not, they're going to they're going to find it refreshing that you haven't started the way everybody starts by introducing yourself or, or doing something equally. Yeah, here's the gender something equally boring. So alright, so your your book just to give people the name, again is effective data storytelling, how to drive change with data narrative and visuals. I'm gonna I'm going to endorse it in a way here. Unfortunately, I can't endorse every person that works with storytelling. You actually tell a lot of stories, which is refreshing. It's one of the most painful things in my life, that I'm occasionally researching people for this podcast and I cannot find them telling any stories or show Good stories. So your book has plenty of that. And even though the subject can be dry at times, I find in the found it in the main to be a pretty easy read. So other than the book, which I know is available everywhere, where should people go if they want to see more of your stuff? Yeah, I


Brent Dykes 1:00:16

think a good thing would be if you're into storytelling and data storytelling in particular, definitely connect with me on LinkedIn. So I'm on LinkedIn there. And I often post different LinkedIn posts there. So that's definitely a great place to connect with me. I also have a website effective data, storytelling comm. And also on Forbes, I write for Forbes more about data and analytics, sometimes about storytelling. But you know, I'm also a Forbes contributor. So I think those are the three main places I'm also on Twitter analytics hero, and you can follow me there as well. So those are the probably the main areas to connect with


Francisco Mahfuz 1:00:50

me. Perfect. Brent, thank you for your time. This. This was great, man. Thank you, Francisco for the opportunity is great to talk to you. Alright, everyone. Thanks for tuning in. Take care of yourselves. And until next time.


I hope you enjoy the show. And if you did, I'd love for you to subscribe and leave us a review or a rating on the Apple podcasts app. It's very easy. You open the app and find this show and scroll down a little and when you see the stars tap. I'd really appreciate it and it does help other people find this. And if you'd like to get in touch or find out more about what I do, reach out to me on LinkedIn or visit my website story powers.com



Recent Posts

See All

After 100 episodes, what storytelling lessons have I learned? Well, a few, so here are 23 for you, and they cover: why stories matter, what do you use stories for, where do you find them, how do you t