Data Visualization | Effective Data Storytelling - Brent Dykes

Okay, we're gonna get going. Uh, welcome everyone. My name is Adam Greco, and I work for Search Discovery and am the ringleader here for the SDEC. For those of you who haven't been to an SDEC session before, the SDEC is a free educational initiative that we're doing just to kind of help people who want to either grow in their skills or especially for people who have been affected by the COVID pandemic and might be looking for new roles or or out of work. We're providing education in a bunch of topics, um, related to analytics, Adobe Analytics, Google Analytics. And today, we're going to talk about data visualization.

If you have any questions that you would like to ask our presenter, Brent, today, please use the Q&A area found within Zoom. Uh, if you put your questions in chat, it doesn't show up in the right place for us, so please use the Q&A area. If you have any technical problems, uh, you can use the chat for that, and I will be there to try to help out.

So, this session is our second session in the data visualization area. I'm excited to have Brent Dykes join us. Brent works at Blast Analytics. I just actually recently joined there. Brent and I work together at Omniture and I want to say it's back in 2005 is when we started working together, which I can't believe is 15 years ago. But Brent is a leader in the analytics space. He writes for magazines and he recently published a book on data storytelling, and he'll tell you a little bit more about that. Um, one last thing, um, we're gonna try something new today. We've had a couple folks, um, who have said that English isn't their first language and they wondered if we could try closed captions. So, we're gonna try that today. Uh, Brent is our guinea pig. So, um, thank you, Brent, for doing that. And, uh, hopefully, it's not too distracting. Uh, we'll get some feedback afterwards and see what you guys thought of it. Uh, we're kind of a beta testers here in the analytics space. So, with that, um, I will pass it off to Brent. Uh, take it away.

All right, thanks, Adam. Excited to be here and to talk about effective data storytelling. So, about 10 years ago, uh, McKinsey interviewed Hal Varian, who's the chief economist at Google, and they asked about different kind of trends and different things that he was seeing. And he added this great quote, uh, where he talks about a skill that we all need. And I broke it down to a couple of key areas. So, one is the ability to take data, to be able to understand it, process it, and extract value. I kind of see that as the ability to find insights. And then further into this, he talks about to visualize and communicate those insights. And that that's going to be a hugely important skill in the coming decades. And so, the ability to find and explain insights, I really equate that to data storytelling.

And if we fast forward to today, you know, about 10 years later from when he shared this quote, basically, we're seeing that the ability to share insights is really important, and the ability to communicate them effectively. Now, insights bring about change. If we look at what's the definition for an insight, it's kind of, I, I like Gary Klein, who's a psychologist, and he said it's an unexpected shift in the way in our understanding. And as we analyze the data, as we go in all the time, we're finding these insights that shape and change our understanding of how our customers operate, how our business works. And what that means is, if we share this insight, it's going to drive a change. We're going to have to make a change. We're going to have to do something to either address a problem or seize an opportunity.

Now, the interesting thing about small insights is that they can be acted on by individuals. If I see, if I'm managing a paid search campaign, for example, and I'm, that's my responsibility, I go into my dashboard, I see some kind of insight around an ad group, ad word group, and then I make a change. And I have, I've empowered to make that change myself. I don't have to check with anybody else. I can adjust my budget, and there I go. Then I have that insight and I move forward.

Now, when we have larger insights, they require communication. Why? Because I may need to get approval from my boss. I may need to get a budget to help fund whatever I'm doing. I may need to get resources to to execute on something. I may need to coordinate with other teams. I may need to get their buy-in. Even if they don't need to coordinate with them, I don't want them to be a roadblock or or whatever. And so, effective communication of these bigger insights is really essential. Because what happens is, when we have poor communication, that's when we see these insights, they don't go anywhere. We don't get the approval from a manager. We don't get the budget or resources that we need. We don't get the coordination or the buy-in that we need to really drive something forward. And so, that really sets up the need for why we need data storytelling.

And I like to look at the, the three different elements of data storytelling, which are data, narratives, and visuals. And obviously, this has been kind of communicated as a, as I'm going to talk about data visualization. Well, data visualization is one aspect of data storytelling, and I'll kind of talk about these three elements here. So, I think to understand the power of data storytelling, it's really about looking at the intersections between these different bubbles. And so, in this case here, we have data. You found an insight. You could, you could give all of that raw data to, to your co-worker or to a manager, uh, to an executive, and and hope that they can interpret that information the right way. Or what we typically do is we add some narrative. We, we walk the audience through the data points. We explain it to them. We help them to interpret it the right way and understand it. So, it's clear and and and concrete in their mind what the, what these insights need. So, that's the first advantage of combining narrative and data.

Now, we also combine visuals with data because again, if we go back to that raw data table that we drop on somebody's lap, it may be very difficult to see anomalies in the data, patterns, trends. And so, what do we do? We visualize. Is that use charts to help enlighten people to see things in the data that they would miss if it was just in rows and columns of data? And then the reason why you were up late last night watching your favorite Netflix show is because as human beings, we like that combination of narrative and visuals. It engages us at a deep level as human beings. And so, that combination there is really critical as well to help us engage an audience. And then what we do here is we actually, if we could take the right data, if we can combine it with the right narrative and the right visuals, we have something very powerful that can really shape, uh, people's opinions. It can, it can influence their behaviors. It can really drive change with our, with our data stories.

And so, one of the poll questions that we had for this session today is, if you look at the different, the three different elements, where is it that you struggle the most? Is it with data? Is it with narrative? Is it with visuals? So, take a moment to to answer the the question. We'll provide those results, uh, in a little bit.

Okay, so yeah, so Brent, we got, uh, we're almost done. Giving you another five seconds. We've got about 65% done, but we're starting to see a clear winner here. So, I'll, I'll stop. And it's, I think, I think it's, uh, here, here are the results. So, we have, I don't know if everyone can see that, but we have narrative is the the winner here with 67%, visuals with 22%, and data with 11%. Awesome. Okay, well, that's good because in my book, I took a lot, I talk a lot about narrative, and and it's really powerful. We'll share a little bit today.

So, one thing I want to talk about, the power of narrative and and stories. And and this comes out of a book, uh, that I read called "Made to Stick" by Chip and Dan Heath. Highly recommend it. It talks about communication. But in in one of their chapters, they talk about the power of stories and how how they actually beat statistics. And so, the first way that stories are more powerful than statistics is they're more memorable. And this actually comes from an experience, uh, Chip Heath, he's a professor at Stanford University, and he teaches an undergrad class on communication. And in that class, he has, he has an exercise that he puts his students through where he gives them a bunch of data points about around an issue. He has them take a position, and then he asks them to do a short presentation to each other on on that topic using the data points that he's given them. And so, they form up into groups of four or five students. They each take a turn presenting to each other, and then they give feedback and rate each other on how they did. Now, a little note, a side note on this is that one in five, or sorry, one in ten of these students will actually incorporate an anecdote or a short story into their presentation. So, the students think the exercise is done. And then about 10 minutes later, uh, the professor, Professor Heath, comes back to them and says, "Okay, how many of you remember any of the statistics that were shared?" And the interesting thing is that only about 5% of the students could remember any of the statistics, but 63% of the students could remember the stories. And so, we see here that stories, the way we process information, the way our brain looks to kind of process and understand information, is through stories. And so, if we can package them up that way, they're going to be more memorable.

Now, the next way that stories are more, uh, powerful than statistics is they're more persuasive. And so, this is actually from a study that was done at Carnegie Mellon University where they had students fill out a technology survey. And for completing that survey, they gave the students five one-dollar bills. And then they asked them, "Hey, thank you for completing this survey. Would you like to donate any of that money to this charity?" And it was a real charity called Save the Children. And what they did is they had two versions of a brochure or pamphlet that they that they gave to the students. One version had a lot of data around the suffering of children in Africa due to famine, due to illness, due to war, and talked about hundreds and millions of children's lives that were being impacted by these different challenges. And so, that was in one version, very data-driven version. The other version talked about one girl, she was a seven-year-old girl from Mali, and talked about the struggles that her family went through and told, you know, told it from her perspective. And so, the interesting thing is, when they looked at which version generated more donations, the story version focusing on the one girl, Rukia, and her family, outperformed that of the, the one that was packed with data and showed lots of, you know, hundreds of thousands or millions of of children being impacted. And so, this is important that if we look at stories, they really are more persuasive than just raw data.

Do I like to say, we hear statistics, but we feel stories. And this is actually backed up by neuroscience. When neuroscientists look at how we react to data and facts as it's being shared with us, there's a couple of regions of the brain that light up that are associated with processing language. However, when somebody's being told a story, they notice that other regions of the brain would light up. And this is because the listeners are actually experiencing the story like the, the storyteller, uh, experienced it. And so, there's almost like a neural coupling that that goes on between the audience and the storyteller.

Now, another interesting thing happens is that if we approach somebody with data and facts, there's a natural reaction for people to put up their guard, their shields come up, and they don't want to be tricked or deceived. They're crit, they're a little bit more critical and skeptical of the information, and they want to make sure that, you know, it, they're on guard, essentially. Now, the opposite effect happens when you approach somebody with a story. Their guard actually comes down, and they're not going to nitpick on the details of the story. They want to hear where it goes. And, and so, this opens a door for us to share insights in a way that's going to really be received more readily by our audiences, and and they're going to have a much more open mind to them. And so, what we want to do is we want to take those insights, the data that we have, and we want to connect it to that emotional center of the brain through a data story that's going to speak to that emotional center of the of the brain and help us to get our, if our information and our insights through to an audience in a much easier and better way.

Now, today's presentation, I'm going to take three tips from my book and I'm going to share them with you of how to, how to turn your insights into data stories. Now, one of the challenges that we have is obviously, we use data visualizations to find insights, right? So, we, we use the data visualizations to look at charts and trend out different patterns and and find insights. Now, when we then go to communicate those insights with other people, we may need to adjust our visualizations. We may need to change our charts so that they communicate more effectively to other people. And I'll give you some ways in which you can do that.

So, here we've, we've had, we've done some analysis. We've, we've compared the marketing costs, and this is, uh, field marketing here. We're looking at the different costs we have for different events, uh, for two different years. And so, we're doing some kind of analysis here. Now, what we could do, if I go back here, you'll look, how do we analyze this data? Well, we have to look at the bar chart links and then compare the links. Obviously, we have the the labels there to kind of help us. But we could use a different chart, a slope graph chart, for example. So, the slope graph basically, you look at the slopes or the angles of the of the lines to kind of see which ones are are going up or which ones are going down. So, that can give us maybe an easier way to kind of see which of these events increased in cost from year to year. Uh, the other thing we can do is we can control the colors that we use. And so, what effective manner is to push the contextual information to grayscale, so it's in the background. And then we bring out using a bold color to the foreground, the key point that we want people to to focus on, while we still have the context of the other numbers there, but they're not the focus of our messaging. And then another thing we can do, often when we're doing our exploratory analysis, we give kind of, uh, titles that are more just descriptive of of the data, uh, not really that insightful. And so, what we can do is we can actually change the titles so they actually convey the key points and and emphasize the key points that we're trying to make with our graph. And so, the, the title here, if we use it effectively, it can actually reinforce the key point that we're making in our charts. Oh.

Now, the second tip is, we want to enable comparisons. And often, when we think about data storytelling or data visualization, you know, maybe in our mind, we start to think of these really complex charts. And really, when we're, when we're shifting to communicating our insights, it's really about facilitating comparisons. And I like this quote from Edward Tufte. He said, "We're always trying to to set up that compared with what, right?" So, it comes down to making and showing smart comparisons. And so, how can we do that with our data as we go to communicate our insights with other people?

So, in this example here, there's a couple of things that we can see clearly in this chart. We can see the total revenue in which region, in this case. So, we have four different regions here, and the, the sales of different products, four different products. And we can see which region had the most sales. We can see that very clearly. We can also clearly see which of the product A, which regions bought the most of product A. Now, what's not as clear from this chart is, okay, if I'm looking at product B or product C or product D, it's harder for me to kind of clearly see which, which of those had the most sales in each region. And so, what we may need to do in this case is modify the chart to a panel bar chart. And so, now, what we can see here very clearly, I mean, I could take away the labels, and you could actually assess and you could, you could clearly see which, which region had the most sales of product B, C, or D. And so, by having a, a baseline, an axis that we can then refer to, it makes it easier for us to make these comparisons.

Another challenge sometimes is when we have time series data and we're trending lots of different values. In this case, we've got campaigns and and in each week, how many leads do they generate? And so, it can be, you know, sometimes we can have what they call a spaghetti chart, right? So, there's lots of inter, inter, uh, interwoven, uh, data, time series data. And so, what we might want to do, again, going with a panel chart approach, in this case, a line chart approach, we want to break apart the, the data series into their own individual charts with their, with the axes are, or the scales are the same. And so, it's very easy for us to go in and kind of then analyze these data points and and see, uh, the trends of these campaigns without that overlapping data, which can sometimes create noise.

Another thing too is, you know, as we go through exploring the data, in this case here, I've kind of looked at employees by location, and it's, it's done by alphabetical, and I've got a column chart here. Now, one of the challenges, you know, we want to reduce the friction that our audience has in analyzing data. In this case here, it's harder for us to look at a label on an angle like that. And so, what we might want to do is change the rotation of the of the chart entirely to a horizontal view, so that we're not looking at the, at the, the names of those cities at an angle. And another little thing that we can do to kind of again, make this easier to consume, is by sorting the values. And so, here we can see side by side, kind of the similar values as it goes from largest to smallest. And so, these subtle effects, these subtle treatments that we can do, can really have a huge impact on our presentation and and on how we communicate our insights to an audience.

The third tip that I want to talk about is narrative structure. And that was obviously something that came up in the poll as a key area of focus. And often, when you hear about stories as related to data storytelling, uh, and narrative, often you'll hear, a story has a beginning, middle, and end. And and this comes to us from Aristotle more than 2300 years ago, who looked at Greek tragedies and and found that they all kind of had this arc to them or this format. Now, I, I don't really find that, um, that, you know, stories having a beginning, middle, and then that helpful because you could look at a textbook or a report and say it has a beginning, middle, and end. That's not really what the key thing is here. And so, I kept looking for other kind of models. And I discovered a guy by the name of Gustav Freytag, who's a German playwright, uh, and he studied Shakespearean plays. He looked at Greek tragedies as well, and he found that they all had a very similar story arc, and sometimes called the Freytag pyramid. And I've taken that approach and I've applied it to data storytelling and how we can display or present our data using this arc.

And so, the first thing that you're going to do is you're going to have a setting and a hook. And at this point, what we're doing is we're establishing context. We're establishing the background information so that people can then understand the insights that we're going to share with them. And so, we're, we're establishing, okay, here's some of the the status quo, you know, this is how this thing is performing. They were analyzing in the past. This is what we typically see. Um, the characters will be the maybe the customer segment that you're analyzing or, uh, mobile users or whatever the people are that's behind the data. And then you have this hook. And what's the hook? Well, the hook is the observation that kind of caught your attention, that starts this, uh, this whole journey that we're going to take the audience on. And so, this could be an observation, uh, where we saw a spike in a particular metric or a drop in a particular metric that's of interest and of concern to the audience. And then from there, what we do is we start to reveal other insights, other supporting details that that dive, you know, almost like peeling an onion, where we're going deeper and deeper until we build up to our aha moment, which is our major finding. So, often sometimes when you see data being presented, it's kind of a data dump. And there's, here's something interesting. Oh, here's something interesting. Here's something interesting. But where are you taking me? Where's the destination? And really, with the aha moment, that becomes the crowning, uh, uh, ex, you know, insight that we want to share with the inside or with the audience. And if they only remember one thing, it's going to be the aha moment. That's going to be the major finding that we want them to remember and act on.

Now, we're not done at this point because again, going back to, we want to drive change, we want to drive action, we need to also provide maybe a solution or potential options that they could consider. And what are the next steps and where do we go from here? And so, that might mean additional analysis to see, okay, what's the impact of this, uh, this opportunity or this problem, and how do we address it? And through this process, hopefully, by sharing this data story using this approach, we're going to increase the audience's knowledge into the business. We're going to hopefully position them in a manner to drive action and make a decision. And that's really why we want to have that narrative structured.

Now, let me walk through an example of how this would work. And this is actually taken from a real, uh, example that a data scientist team, they were working with a retailer and they're going to do some analysis of their customer orders. And so, they, they expected to see a chart like this, where there's a normal distribution of orders of different sizes, uh, over, you know, this kind of distribution. Now, what they discovered when they, when they started looking at the, the actual data, they found that there's this, the second peak. And that really got their curiosity, and and they're wondering what's driving this second peak. And so, as they dived into the data, they found that this retailer only, uh, had stores in the US and so they didn't have any international presence. And what was happening is they had a lot of, uh, international, uh, store owners come into their local stores, purchasing bulk, large orders of of clothing, and then take it back to their home countries to sell in their own stores. And so, this was the big aha moment, that there's these large contingent of international national buyers coming into their stores and making these large purchases. And so, from that, then, okay, what do we do about this? Well, there's different options that they had. They could look at, do we expand into these, these countries and start to open up stores globally? Uh, do we market to these, uh, these individuals and kind of help them to make larger purchases? Or maybe do we partner with them in these other countries and set up relationships that way? So, again, looking at those different options, making recommendations, and then driving forward with an action plan. So, as you can see here, all of this data that the data scientist team could then be put into a narrative structure that then can be communicated to the decision makers and help them to make decisions.

Now, what I want to do in the last part here is just share an example of a data story. And and this actually comes from a historical perspective. Now, I have never had surgery before in my life, other than my wisdom teeth. But I know that if I had a doctor operating on me, a surgeon, I would want him or her to wash their hands and make sure that they are, uh, have, you know, no germs on their hands. Now, if we go back in time to the 1800s, it wasn't until Louis Pasteur discovered germ theory that that the world started to learn about these bacteria and germs that were actually making people sick. Now, Ignaz Semmelweis, he was actually a doctor that worked at a, a university hospital in Vienna. And and he was actually the administrator over two clinics that they had for training doctors and midwives. And he inherited a problem when he first came on board. The previous six years, they had had some real problems with something called childbed fever. And we can see here that in the two clinics, the doctors' clinic actually had a higher rate of mortality from this childbed fever, which mothers would come in, they, they'd deliver their babies, and then they would get sick shortly thereafter, and then many cases would die. And it would also affect their, their babies as well. And so, this is a huge problem with the doctor's clinic in particular, and they didn't know why. Uh, there's this difference between the two clinics. And and they looked at different things. They looked at maybe it's the, uh, the bad air. At this time, they believed in miasma, which was like foul air, smells in the air would, would you know, they believed that contributed to making people sick. And then they looked at maybe, maybe one of the clinics was overcrowded and that was contributing, or maybe it was the temperature of one of the rooms. You know, that the temperatures were different. And and they kept looking at all these different factors to see what could be influencing this significant difference in the mortality rate.

And it wasn't until one day when a fellow doctor of, of some of Semmelweis's was actually performing an autopsy on a woman who had died, uh, from childbed fever, and he was performing this autopsy with a bunch of students. And one of the students accidentally cut the hand of this doctor as he was performing this autopsy. And sure enough, uh, that wound got infected, and the doctor died a few days later. And Ignaz was actually out of town when this happened. I think he was on vacation, and he got called back and had the difficult task of actually performing the autopsy on this doctor. And that's when the aha moment came for him. That he, as he's performing the autopsy, he noticed that the pathology of how this doctor died actually mirrored the way that these women were dying. And then he started to, the dots started to connect. And he was like, "Oh my gosh, what are the standard practices at this hospital? Many other hospitals, they were training young doctors was to have them come in in the morning, perform autopsies, and then for the rest of the day, the rest of their shifts, to do examinations, and then to make deliveries." And so, then he's like, "Oh my gosh, we may be passing some kind of particles or something." Again, he didn't know about germ theory, but he was maybe, there's some particles from these, uh, dead bodies that were passing to these women and making them sick. And so, he instituted a, a new, uh, policy of of hand washing. And so, he said, "Okay, we're going to introduce this hand washing policy. We're gonna have a chlorine lime solution that we're going to have the the doctors, uh, wash their hands in after they've done these autopsies, before they do their rounds around the hospital." And so, immediately, it had a huge impact where 82% of the of the mortality rate went down, 82%. And then for several months, then it was, it was he was able to to maintain a relatively low, uh, mortality rate. Although, you will see that it is going up a little bit. And that's because a lot of the doctors at the time couldn't believe or wouldn't accept that they were contributing to the death of these women. They, they couldn't believe that they were the ones responsible for these women dying. And so, he had some of the young student doctors rebelling against them and not really adhering to his standards. So, he introduced strict controls to really enforce the students, they followed his policy. And for a couple of years there, you'll see that there is actually zero deaths for a couple of months. But you'll notice at the end that it actually goes back up and goes up to 4.9. And you may be wondering, what's going on? How could Ignaz let this happen? Well, Ignaz was actually let go. He butted heads with his superiors, and and they could not accept that doctors were contributing to the death of these women, even though he had a lot of data to show that they were and showing that the hand washing policy was effective. And so, he was basically ostracized, uh, basically in that hospital and and the rest of the Vienna medical community, and had to move back to Budapest to his home country to continue practicing. And he waited 10 years for his, uh, research and his his policy and his approach to take to catch hold within the medical community. And eventually, he published his findings to kind of talk about what he, what he had accomplished and was rejected and and never saw his ideas embraced or adopted by the medical community in his, in his lifetime.

So, why wasn't he successful? Uh, if we, if we stand back and kind of look at what he had, he had accurate data, of course. He didn't know, he couldn't prove germ theory like Louis Pasteur, but he did have 18 months of data on hand washing that showed that it was extremely effective at reducing the mortality rate. Incredibly valuable, if you think about not only if he was able to save lives at his hospital, but also around Europe, around the world, because this is a standard practice of doing these these autopsies. And so, women were dying around the world needlessly because the doctors were not washing their hands. And then how actionable is this? Well, you know, having a chlorine lime solution to wash your hands as a doctor, a small, uh, task to ask, and yet it wasn't embraced and wasn't adopted. And if you actually look at the, the journal that he published on it with his data, he didn't visualize any of the data. He had 60 plus data tables, none of them were visualized. He didn't tell the story and didn't capture the hearts and minds of the doctors to really compel them to embrace his insights. And so, ultimately, he was not able to see success in his lifetime.

So, I just want to conclude with this quote from Stephen Few. You know, I really do believe that we all have insights that we want to share. We all want to see those insights acted on, and we want to see them drive value for the businesses that we work for. But it's going to be up to us. We have a responsibility to be a clear and convincing voice, to give, you know, a voice to the to those numbers, to tell their, to paint the picture of what's happening, and really explain it to people in a way that's memorable, engaging, and persuasive. And so, I'm kind of a data geek, a comic book geek, and I really like that the quote from Stan Lee, who said, "With great power comes great responsibility." And I would add, and also great opportunity. You know, if we can tell these data stories in an effective manner to the organizations, we're going to drive more action, we're going to drive more value, and it's going to impact not only our organizations but also our careers. And this, and with that, now turn it over to Adam and whatever questions the audience has. And, and I guess here's, here's a copy of of my book. You can actually download a free chapter, uh, from my website, my book website, at affectedatastorytelling.com, or you can go to Amazon or other retailers to pick up a copy of of my new book.

Awesome. Thank you so much, Brent. And I did put a link, uh, in the chat to your book and also to the one you mentioned, "Made to Stick," as well. So, uh, everyone can click there and go check it out. But thank you so much for sharing so much information in such a short amount of time. So, if you all have any questions for Brent, now is a great time to go ahead and post them in the Q&A area. Um, and I will, uh, start reading off some as we already have a few. So, uh, one person had said, um, how do you see narratives playing a role in a dashboard? I can imagine that even when the chart explains the data on its own, there's still people wondering what they can do with the data they see.

Yeah, so I, I actually see, uh, dashboards as not really data storytelling. And that may jar for some people, but I like to look at it as a funnel. And I look at dashboards as the top of the funnel. They're, they're for monitoring data as it's coming in. And obviously, that data can change, and and it can change from week to week, month to month. And and so, we use those dashboards to to monitor the, the key, you know, key performance indicators that we have, and then potentially dive in, do more analysis, or whatever. And from a dashboard, when we have an insight that we need to communicate with others, we then need to create what's probably more like a data story in a, in a presentation format or maybe in a report or something like that. But there, it's a different, uh, vehicle. And so, what we're doing with the data stories, we're taking snapshots of data, right? And then we're adding commentary, we're adding, we're explaining the numbers, and we're pulling out insights. And then when we build that data story, it kind of serves a purpose, and then it's, its usefulness is done, and we move on. Whereas a dashboard kind of lives on forever. And I call that kind of story framing. The role of a dashboard is to frame the right information, the that we would then find data stories from. So, you know, obviously, we have a ton of data that we can look at. And so, a dashboard is really great at kind of framing our attention, focusing our attention on key, key metrics and dimensions that we should be looking at. And then from that story framing, we can then identify these insights that then need to be told, uh, with, with, you know, snapshots of the data to different audiences. So.

Okay, cool. Uh, next one from our very own Tim Wilson. What tips do you have for actually capturing or working out a narrative with a document outline, Post-it notes, a shell set of slides, or put another way, tips for figuring out what the narrative actually is?

Right. Yeah, so one of the things I, I talk about in my book is kind of a storyboarding approach. And and so within that, you may have, you know, you'll, you kind of lay out all of your insights, all of your findings in front of you. You might put them on Post-it notes or a whiteboard. Um, it's good to kind of be able to move them around. So, I kind of like the, the Post-it note approach. And so, then you start to say, you know, and this is something I cover in my book, uh, but we say, okay, the first thing I would start off with is, what is your aha moment? You know, that's the first thing. No, you don't have a data story if you don't have an aha moment. And so, that's really where, and that's the destination, right? That's the star of your data story. And so, that's the first thing I would start off with. And then the next thing that I would want to look at is, what's the hook? You know, what's, where, where does this whole story start? Because one of the, one of the challenges that we have sometimes when we, as analysts, we, and I've seen this done many times, as people will do what I call the analyst journaling or analysis journey, where they go through, "Okay, first I looked at this, and I didn't find anything. And then I looked over here, and I didn't find anything. And then I did a segmentation, and then I found this, which led to this, this, and so they walked them through their analysis process." And honestly, a lot of stakeholders and business people don't care. They don't care how you found your insight. They don't care about all the steps that you took to bake your cake. What they want to do is eat the cake. They want to unders, you know, they want a slice of your cake that they can enjoy. And so, it's really about understanding, okay, if we're going to have this data story, um, where does it start? What is going to be that hook that's going to intrigue them and get them interested in in the data story? And then, and then for backing up from that, how much context do I need to provide to then understand that hook? So, I might need to show, okay, we saw the spike in this metric, but here's, you know, here's what's happened for the last 12 months preceding that spike. You can see it's relatively flat until this spike. And then people are like, "Oh, okay, got it." And then now that we have our, our hook and our aha moment, then we have to look at all of our data and say, okay, which, which findings and which data points are going to get us from the hook to the aha moment in the most concise and clear manner? And obviously, we're going to have lots of data points that we did, and we dove into the data different, uh, degrees of detail. And what we can then evaluate is, okay, this one builds the story. This one is tangential. I'm going to put that in the appendix. This one here, this one adds to the story. It kind of connects to that first point. Okay, we'll add it in. This one is kind of duplicative. We're going to put that in the appendix. We're going to remove it. And so, you start to connect the hook to your aha moment. And then the last step is obviously to think, okay, how am I going to drive people to act? What's the information they need to make a decision to evaluate this in the most, you know, clear way? And, and so, then we might have to do some additional, we might find that we don't have, uh, we haven't done all the analysis to really drive action. And so, we might have to go through and, and do some additional, uh, discovery to kind of build up these options or these potential paths that the audience can take. And so, all of this can be done by just whiteboarding, going through our, our analysis, pulling out the key points, and also in thinking about the audience as well, right? That's another key factor here. What, what questions will they have? Can we anticipate what those questions are and address them with, with findings in our analysis? You know, and different things like that. But it's, hopefully, that that that's helpful to kind of understand how we can take our findings and, and then, uh, you know, build into that narrative structure.

Okay, so we just have a couple minutes left. We have a bunch of questions. Let's see if we could do some kind of quick hit ones. Um, how does the age of Twitter with short attention spans impact the ability to tell a story in a quick, quick, uh, crisp way?

Yeah, so one of the challenges I get often is, what do I do with an executive who's only got 10 minutes or 15 minutes or five minutes? And so, there's a modify, I call it a data trailer. And so, you basically take the hook and the aha moment, and that's what you present. So, it's kind of, it's kind of like the, the world's worst movie trailer, because you're giving away the climax of the, of this, of this, of the story. But what you're doing is you're gaining permission to then get their attention, to then invest it in hearing the whole story. So, that's my quick answer for that one.

Okay, next one is, how can we make storytelling more relevant and personalized, uh, maybe for different, uh, reporting teams that hits the right chord with the audience and makes the maximum impact?

Yeah, one of the quick tips that I have is obviously your audience. If you're going to care about your customers, right? In most cases, we're talking about marketing data. They're going to care about the customers. And so, if you can, as you start to tell your data story, if you can focus on, uh, the customers, bring them in, you know, I, some people don't like clip art and stuff, but I think sometimes that can be helpful if you can tell their story. Maybe there's bringing in qualitative data. There may be answers from surveys, from product reviews, from, um, on from social media, different quotes and different things. And so, you can bring alive the, uh, the analysis. You can humanize it, essentially. And I think when we humanize our data and our, our data stories to connect, you know, our data to people that the audience cares about, then that's going to, that's going to resonate even more powerfully. And that's something I talked about in my book.

Yeah, and one thing I'll add to that, Brent, is, um, I've had a lot of success working with tools like Decibel Insight, where you could actually show people, right, using your website. And there's, and I always say that like watching a really good video is worth like a thousand analyses, because, you know, when you see people get really frustrated on your website, like you can't look away, like it's pretty impactful.

Absolutely. Uh, next one, um, can you talk about the distinction between data exploration and data communication that's discussed in data visualization literature? Um, and so, you're using narratives as both useful for data exploration and data communications. I'm not sure if I butchered that, but hopefully that makes more sense to you than me.

Uh, yeah, I would say the narrative comes in on the explanatory side. Um, what you're doing when you're exploring the data, you shouldn't really have a narrative in mind. You might have a hypothesis, you might have a hunch that you're exploring, but you should be open-minded to what the data will tell you. And so, on the exploratory side, you really shouldn't be, I mean, that is one approach. You could say, "Here's the story I want to tell. I want to show this campaign was a success," and now I'm going to go cherry-pick the data that supports that story. But that, in my mind, is not true data storytelling. Um, and so, I think you want to be careful there that narrative really isn't a part of the exploratory until you've found an insight, and then you go through that process of, okay, what is that narrative that I'm then going to tell on the explanatory side?

Okay, I'm gonna squeeze one last one in. Um, how important is the organization's analytics maturity level in the impact of storytelling to drive actual change?

Yeah, obviously, uh, data literacy. If your organization is not as data literate, it's going to be more challenging. But I see data storytelling, if you can, if you are that data literate person that can tell data stories, because you have, you know, you understand the data, you can start to raise the data literacy of the community or the employees internally. Everybody likes a story, and if you can format your insights into a story, that's going to, that's going to go further within your organization. And so, I kind of see data storytellers as people that will raise the data literacy of the internal employees. If you think about with small children, what do we do to kind of make them more literate? Well, we read them stories every bedtime, right? More, more stories, and then they get curious and they want to read the stories themselves. And so, that, I think, is the same thing. And obviously, the maturity of your organization, uh, will impact. But I think if, even if you have a low maturity, I think data storytelling and sharing those stories can start to increase the literacy level of the organization.

Awesome. Well, Brent, thank you so much. We will try to provide a written format, answer to some of these questions as well in the Slack channel. And I will post this recording in the Slack channel later if you want to review it or if there's, um, something if you missed the beginning. So, Brent, again, thank you so much for coming. Um, this was awesome, and thanks, everyone, for joining. And we'll see you at our next SDEC session.

Thanks, Adam. Thanks, everybody.