The Art and Science of Impactful Data Visualizations in Adobe Analytics

Transcript

I want to welcome you to our Adobe Analytics webinar, the Art and Science of Impactful Data Visualizations. My name is Justin Swanson. I’m a Senior Customer Marketing Manager here at Adobe, and I’ll be your host. Today, you’ll hear from David Geist, a business consultant here at Adobe and formerly an Adobe Analytics champion. Before we jump into our presentation, I want to go over a bit of housekeeping. While I do that, go ahead and say hello and drop where you’re joining from in the chat. To start, this webinar is being recorded, and after the webinar, we’ll be sending out a copy of the recording. The console you’re seeing in front of you is completely customizable, so you can resize or minimize any of the widgets on your screen. Feel free to make the slides or videos larger on your screen or smaller based on your preference. We’ve shared several resources related to today’s webinar. You can find those in the Related Content panel on the top right side of your screen. Throughout our session, there are two ways for you to interact with us. First, you can ask a question directly to David in the Ask the Presenter box in the bottom center of your screen. We’ll do our best to get to all the questions during Q&A, but if we don’t, we’ll follow up in a discussion thread on the Adobe Analytics community. And second, if you’d like to chat with other attendees and ask a group a question, you can use the Attendee Chat panel. This is a great way to hear from others or share your own thoughts and experiences during our session. And finally, at the bottom of the screen, you’ll find your webinar console. Here, you’ll find additional information about our speaker and a survey. Please be sure to take that before you leave. That’s how we pick topics and presenters for future sessions. There’s also a Reaction button, which can be used to give our presenters a bit of love. Go ahead and click on that emoji, and feel free to try that out now. OK, now that those are out of the way, let’s get to our agenda. We’ll have a 40-minute presentation with David, and then make sure you stick around for a 15-minute Q&A, where David will take your questions from the session. So without further ado, I’ll pass it over to you, David. All right, thanks, Justin. So as Justin said, my name is David Geist. I’m a business consultant here at Adobe. And I haven’t been here too, too long, so I joined in April of this year. Prior to that, I was at SunTrust Bank and Truist Bank for about 12 years, 10 of those years, working directly with Adobe Analytics. So I do have a little bit of experience. I think at about six months after I got here, I think at about six months after I joined, Analysis Workspace was introduced somewhere around there. So I feel like I’ve grown up in Adobe Analytics with Analysis Workspace and getting to use all the new visualizations and features that they’ve introduced over the years. So hopefully, we can walk through a lot of those today. And here is a picture of my family. This is us from Halloween a couple of days ago. As you can see, we were Peter Pan-themed for most of us. My oldest son there in the middle decided he was not going to be Peter Pan-themed and was going to be a ninja instead. So anyways, there’s my family. We live in Atlanta, Georgia. Yeah, so go ahead and keep going. This presentation was kind of built on the back of this what’s called the Adobe Analytics Data Visualization Playbook. This is something that I had put together earlier this year, finalized earlier this year. And even as I was building out this playbook, there’s a lot of gray area and a lot of the recommendations I give. And all along, I was thinking, man, this would be a lot better as a conversation or something like this, like a webinar. So you can find a link to that. There’s a link at the bottom here. But I think you can also find a link to it in the resources panel. So check that out if you want more details around some of the stuff we’re going to talk about today. But again, that’s kind of what this was all built on top of. All right. And throughout this presentation, we’re going to have some of these polls. This is what they look like. We’re going to give maybe 10 to 15 seconds, or try to hit about 10 to 15 seconds to let everyone answer these polls. But first question here is, on a scale of 1 to 4, how comfortable do you feel using visualizations? And I would say specifically in Analysis Workspace, since that’s what we’re focusing on today. All right. That should have been about 20 seconds. Let’s see if we can see. We’ve got some results coming in. All right. Just a few ones. A lot of people there in the middle. OK, so that’s good. That gives us a little bit of a baseline. And we can hopefully get these numbers up by the end of the presentation. All right. So first up, we’re going to talk about why we visualize data in the first place. And I think this is really important. We’re going to keep referencing back to this as we go through the recommendations throughout this webinar. So here’s my definition. I think this is probably close to what most of you would have said. And I put this together specifically for this deck, because again, we’re going to keep referencing back to this. But how I define it is, why do we visualize data? It’s to make it easier for our brains to quickly and efficiently absorb something specific about the data. And I’ve highlighted some key terms in here, because those are really the important parts. We want it to be easy, quick, efficient, and in this case, really something specific, because data visualizations as a whole can be a little bit overwhelming. We really want to call something out in the data. That’s really how you make an impact and how you cause change is when you are talking about something specific. And there’s really two, in my mind, two objectives for visualizing data. There’s two reasons why we do it beyond this overall definition. The first one is data analysis. So this is where you’re looking at a data visualization or data to try to learn something about whatever the data can tell us. So data is representative of something that’s happening in the real world or in the digital world, and we’re trying to learn about something that’s happening. And that can take a lot of different forms and that data visualization can help us out with. So again, it’s for us to use it so that we can quickly and efficiently absorb this information that we wouldn’t otherwise be able to even comprehend. So that’s the data analysis. But you also have data communication. And this is where you’re taking that analysis and you’re spreading it out to someone else. You’re communicating it out to another stakeholder that needs to have the same understanding that you do. Maybe not to the same level, but you need to communicate that analysis out to someone. And these are two very distinct purposes in my mind. And the first thing that separates them kind of goes back to the title of the webinar, The Art and the Science of Data Visualization. This may be the only callback to that reference in the title. But really, with data analysis, it is more about the science. So do you understand how your data gets generated? Do you understand what it represents? And do you understand how the visualization is taking that data and visualizing it, like what it’s doing with the data, so that you can accurately interpret that visualization? So that’s really more of a science. There’s not as much art to that. But on the other side, with data communication, we really need to focus on the art. And by that, I mean we need to structure the visualization in a way that the audience, or whoever we’re sharing it with, they can quickly and efficiently absorb the information. And the way that you lay out your visualization and the decisions you make when you’re putting it together are sort of artful decisions. And you’re going to see that as we walk through some of these examples. There’s a lot of gray area. And there’s not like a real right or wrong way to do it. But there are ways that are better. And those are all going to be dependent on what you’re trying to communicate and the audience that you’re trying to communicate it to. So that’s the first two things. Oh, and the other thing, when we try to communicate data, I think a lot of times, as analysts, I do this, I stay in this analysis mode, in this science-focused mode, where I’m trying to clarify absolutely every detail about the data to this audience that may not care. They don’t need the same nuance or the same granularity that I went through in the analysis process. And we just have to trust that. We have to, first of all, establish trust with our stakeholders. They trust that, even though we’re not giving them absolutely every detail about a specific segment that we put together, that we’re representing the data in a way that is accurate and that is just conducive for their understanding. So again, sometimes when we’re communicating data, it is more about the art. And you may have to sacrifice some of that science-y side of it, so some of the accuracy, some of the nuance. But that’s OK. I think we need to do a better job of that, because it’s getting in the way of effective communication. And again, I’m speaking for myself here. This is a huge struggle of mine that I’m trying to get through. So if you’re dealing with it too, totally understand. The other axis here you may have picked up on is the audience. So data analysis really is you are the audience. It’s an internal audience. You’re doing analysis so that you can get a better understanding of the data or whatever the data is trying to tell you. In data communication, the audience is an external audience. It’s, by definition, someone else. And that audience could change. It can be a variety of different people. Even for the same message, you might have different groups of people. And that communication then might need to be different. And the visualizations you choose and the details that you include in those visualizations might also need to be different. So just keep this in mind as we go through all of the following examples. And as you’re building data visualizations in your jobs, we really want to keep in mind when we’re communicating data, we want to structure those visualizations in a way that makes it easier for our audience’s brains. So you have to put yourself in their shoes and make it easy for them to understand and where they’re coming from. And I think that’s the overall message that I’m trying to get across today is that be thoughtful with your communications. Don’t just take whatever analysis you did and pass it along directly. You might want to make some tweaks to it if you’re really trying to cause change in your organization. Because you can be the best data analyst in the world. But if you can’t communicate your findings effectively, and that means getting your audience to understand what you’re saying and your point behind what you’re saying, then you’re not going to cause change in your organization. And you might as well have not done the analysis in the first place. All right. So now that we’ve got that out of the way, we are going to talk through a variety of different use cases using Analysis Workspace visualizations. So my goal here, and I’m keeping an eye on the clock, is to get through all the visualizations we have, all the options we have available right now. The first set of these are going to be questions that you’re trying to answer or things you’re trying to visualize where you have options. And we’re going to look at a few different options that you can have. I’m going to let you all vote on what you think you would pick in that circumstance. And then we’ll look at what I picked and talk about why. And if you disagree with anything I say, please keep note of it or put it in the Q&A, and we’ll come back and discuss it. Because even as I go through this, even this morning when I was going through this deck, I was like, man, I think I disagree with my answer here. Just a general answer. So keep up with those responses. I’d love to have some conversation or at least some discussion here at the end talking through why certain decisions are made. And again, that’s why we’re doing this. It’s not to say this is the right way to do it or this is the wrong way. As long as you’re being thoughtful, just be thoughtful about your visualizations. So that’s the first part. The second half here, we’re going to look at visualizations where there’s not really choices to make as far as which visualization to use. It’s more about how do you use a certain visualization and then maybe some tips about using that one. All right, let’s keep going. So here’s the visualization options. And if you don’t know how you get to your visualizations, if you open up Workspace panel, you can go over here to the left side of the panel and click on this little icon. That’ll show you all the visualizations that are available. And that’s where you would get to this. So just wanted to throw that out there for those that aren’t familiar. All right, first question. Monthly site visits over the last six months. So this is what these are all going to look like. And we’ve got, in this case, three options. And this is maybe the most basic visualization you can put together. It’s probably something that you do daily without even thinking about it. So which one of these, if you’re trying to view monthly site visits over the last six months, which one of these would you pick? We’ve got our quiz here. I’ll give you about 15 seconds. All right. So we got 62% line chart, 25% bar chart, and then area 12%. So most of you all agreed with me. So I picked the line chart here. And I don’t, like I’m going to say with a lot of these, I don’t feel strongly about this. And I think you could certainly make a case for a bar chart or even an area chart. I’ve started using area charts since putting this together because I like the weight of the area chart a little bit more than that line chart. The reason I chose line chart here is really because of the options. So if you click on the gear icon, it’s kind of, I think, outside this screenshot. But there’s a little gear icon above your visualization. If you click on that, you can see the different settings that you can control with your visualization. The line chart has more options here. So you can see with the overlays, we can show a min, a max, and a trend line that neither of the other options have. So that really just the ability to add more context, especially if you are trying to call out a min value or a max value or show one of the trend line options that we have available, that just in and of itself was the reason I chose line chart. But again, depending on what you’re trying to communicate, this one’s pretty simple. There’s probably not a lot of room for confusion. So whichever one you think best, I think is probably OK here as long as it has the options that you need. All right. So the next one here is weekly visits by marketing channel over the last 12 months. And this is showing four different channels. So it’s a selection of the channels. And I’m going to use marketing channel as an example throughout this. I like that as an example. It’s pretty widely used. So it’s just an easy dimension to use in these examples. And this visualization is very similar to the last one, the difference being we’ve got four different lines or bars. And we have double the amount of the x-axis marks. So there’s 12 weeks as opposed to six months. So which one would you use here? All right. That should be about 15 seconds, I think. All right. So we have interesting. OK, so line chart moved down. And we got pretty close to a tie here on the area and bar charts. So I actually went with line chart again here. And for a different reason. And again, I’m not going to argue with either of the other two options for this specific use case, though I’ll explain why I would probably not use them. Anyways, the line chart here, I really picked it not because it’s better than the other two, but because the other two, I have faults with them. So the area chart, my one fault with this, it’s a little bit busy down here where the lines merge together. But you can say the same thing about the line chart. So I don’t know if that’s a good reason. The biggest reason was if you have values that are always ranked in the same rank order, meaning one through four, they’re always one through four. They never cross each other. Your area chart, I think, could start to look like a stacked area chart. And maybe not. Maybe that’s overblown my concern with it. But if it does look like a stacked area chart, like if your audience is confused at all by that, that could become a little bit misleading or, again, confusing. So that’s why I picked, I said, not the area chart. Again, I’m arguing with myself now. I think it’s probably fine. With the bar chart, at some point, and maybe this isn’t the point, but at some point, you get too many values across the x-axis to make this a viable option. Just because you can’t really, like with the line chart and the area chart, it’ll drop some of those values and just assume that they’re there. You can still have the line pass through. With the bar chart, you need a distinct set of four bars for each, in this case, week. So if you expand this at all, it really starts to get busy down there. And it’s hard to follow. This is one, though, that I had put in the playbook. I was back in February. And after that, I saw an example of a bar chart used, trended like this with multiple values, and actually really liked it. I think it was in some article or some publication. But they had taken, I think, the one value and highlighted it. And the rest of them were kind of gray. So it really made that one value stick out. I really liked that view. So I do think that there’s cases where the bar chart is fine, too. Again, it goes back to what are you trying to communicate and who is your audience. With this one, with any of these, none of them are great options. They’re all a little bit confusing. If you give them to your audience without specifying what you’re looking at, they might get lost. So any time you can, I’d recommend either taking a screenshot. Really, you’re going to take a screenshot or export this, and highlighting or circling the change that you’re talking about. So again, from this, it’s really hard to know which change I would even be referencing, depending on the comments. So just make sure the audience is clear about what change you’re talking about in that trend line. All right, number three, total weekly site visits over the last 12 weeks by marketing channel. So this is, again, very similar to the last one. But we want to see the total. And that’s a good call out. And I don’t have a question here for these. But if you’re most interested in changes in individual values, like dimension values in this case, then use the normal line area bar charts. If your biggest takeaway is the change in the total value, so if these add up to a total, if that’s what you’re most interested in, then you should use the stack version. Because obviously, you can see the total value up here at the top. You can’t see the changes in individual values quite as easily. OK, so all that out there, what would you all pick here? Do you like the stacked area chart or the stacked bar chart? This is nice. I can take a drink every time. All right. Cool. Let’s move on and see the answers. OK, so stacked bar got some more votes in here than stacked area, but relatively close. All right, and I’m sticking with the stacked area. As you can tell, I don’t love bar charts in a trended view. Maybe that’s not validated, and someone can tell me why that’s wrong in our Q&A. But OK, so I went with the stacked area chart. I think it’s cleaner. I think you can see the overall movements in the total traffic trend easier in this one. And it’s easier, again, when you start adding even more of those values. Obviously, if you have a daily view for over two months or something, a bar chart would get really cluttered there on the x-axis. That said, when would I use the stacked bar chart? I do think it’s a little easier to see those individual changes with the stacked bar chart than it is the area chart because it is a cutoff. If you look at this in the middle between, I think, that’s March 19th and 26th, there’s a big spike up there. I think you can see those individual changes a little easier than you can in the line chart just because that change is gradual. It’s like a slope. So it’s a gradual change. So if your call out is like this value change, traffic from text messages changed significantly between these two weeks when our overall traffic spiked up, I think that would be a valid use case to use the stacked bar chart. So you can see, as I go through this, very subjective. All right. So for this one, we have site visits and bounce rate trended weekly. And this is not going to be a quiz because this should be a no-brainer. But the first thing I want to call out is anytime you have two metrics with significantly different axis, use a dual axis. So the obvious example is anytime you have a metric that’s a whole number and you want to trend it with a metric that’s a rate, so in this case, bounce rate, those two aren’t on the same scale. It would be invisible. Obviously, on the left chart you see here, you can’t see bounce rate. So you would never do that. So just calling that out. You always want to use a dual axis. If you don’t know how to do that, you can click on that gear icon again. Click on the gear icon and select your option for display dual axis. And it will split those up. Again, it has to be two metrics. It can’t be more than two metrics. So that’s kind of a limitation. But you do have that ability. So that’s nice. OK. So the real quiz question here is between line charts, area charts, bar charts, and combo charts. So these are going to disappear when I go to the quiz. So figure out which one you like. Here we go. Remember, it’s visits and bounce rate compared. All right. That’s probably enough time. Combo chart, nice. And no one liked the bar chart. I gave that hint on the last one, I guess. Great. And that is also what I picked. And so it’s probably obvious, but I’ll call out. In this case, we’re talking about two different metrics. Site visits and bounce rate are two different metrics that we’re trending together. You want to do as much as you can to let the audience know that these are two different metrics. And the combo chart is the only one in here that really differentiates between the two visually. And I think that’s really important. And that’s really a benefit of using the combo chart. If you were showing two of the same metric trended over time, but they were on different scales. So maybe it’s site traffic from direct marketing channel versus billboards, if you have some billboard marketing channel. Those are going to be very, very different, right? Direct traffic is going to vastly exceed billboard traffic. And so they’re going to be on different scales. So even view them together, you’re going to probably have to put them on two different axes. And in that case, I think using the line chart or the area chart is probably fine and maybe preferable because it is the same metric. In this case, though, it’s two different metrics. You really want to differentiate and emphasize the fact that they are two different metrics. All right, number five. We’ve got total visits by marketing channel last month. So we’re moving away from our trended view, which is kind of nice. We have a little bit more flexibility here. We have control over the x-axis now, whereas before that was always kind of predefined as the date granularity or dimension. So yeah, bar chart or horizontal chart showing total visits by marketing channel last month. A couple more seconds. All right. OK, so bar chart got about 20%, 80% to horizontal bar. And you all agree with me. So that’s good. Horizontal bar chart, in this case, I like better. And there’s two reasons. So first of all, we’ve got 12 different values in here. And they just fit better if you put them in a column together, as opposed to trying to spread them across the x-axis over on the bar chart. Obviously, it’s having to drop them down. It’s hard to read. So I really like that they’re all sort of uniformly aligned in this column on the horizontal bar chart. The other thing is you have more real estate or pixels across this x-axis than you do vertically in the bar chart. So on the horizontal bar chart, you can see those differences between channels much, much more clearly than you can over on the bar chart. So I think this is another one where, if there were maybe fewer values, you could use a bar chart just fine. Like if there were four different dimension values, that would be fine. But in this case, I like the horizontal bar chart. All right. So next one here, what is wrong with this visualization? And this is the only time I’m going to open up the attendee chat and see, can anyone tell me in the chat what’s wrong with this visualization we’re seeing right now? And there is a little bit delay. So sorry if I’m having to wait for a while. I’ll give it a couple more seconds. There you go. Yep. So first one, Mark got it. Yeah. There are, so you can see here, this is the top 20 pages last month. There are 10 values and there are 20 bars. So this is one thing that I’ve run into. It’s not obvious. There’s no like error indication that we’re dropping those labels. But it does. It just, if there’s not enough real estate here and the font is as small as they can make it, they drop all the values in between each of these dimension titles. So in this case, page names. And it’s not obvious. This is, so you really want to avoid this, especially if you have like an automated report or dashboard or something that’s going out to an audience. They’re going to get this. And if they actually try to read it, they’re going to be completely confused. What on earth am I looking at? I’m going to hide this chat again. I’m not distracted. And yeah, so that’s just look out for that. The solution here is pretty simple. There was a video that was supposed to play to show this that doesn’t work with this platform. But all you do is drag. You just drag the, make it taller, make the visualization box taller, and it’ll pop those values in once it’s big enough. So just watch out for that. All right, our last one, we’re skipping to seven because that was six. We have percent of overall traffic made up by each marketing channel. So in this case, we’re showing a percent of total broken down by all of these different dimension values. So we’ve got a donut chart and a tree map. And I’ll let you answer that. And you could argue that you could use like a stacked bar chart in this case, which is fine. Usually you’d probably do that when you are trending or have some other dimension across like an X axis. You wouldn’t really do a stacked bar chart on its own. Like a single bar chart. I guess you could. All right. I didn’t look at my timer, but that should be enough time. All right, donut chart wins over tree map. Okay, and this is the one that I feel like I changed my mind this morning. And I think I want to go donut, but let’s talk about it. So in this case, I do not like doing donut charts with more than like four or five values. It gets really cramped in there. You usually don’t have an even distribution across the values. This case, it’s all of this data is fake data. So we do have pretty even distribution across these marketing channels. In my experience, you’ve got like four, well, two marketing channels with really high traffic. And then the rest of them have increasingly smaller traffic. So it’s really hard to view all of your marketing channels or any dimension that has a lot of values in a donut chart. You can see here these two values are overlapping each other. You can’t read that. You also have this legend over here on the side that you have to scroll. Not a fan of that, especially if you’re sending this out as either a screenshot or a PDF. You know, I just don’t like it. The TreeMap solves for a little bit of this. So at least you can see your dimension names. And in this screenshot, it does show the total number here, but you can change that to a percentage. But it also drops these smaller values. Like you can’t read those, right? Whatever those are, those little 2% and 3%. You can’t read those down here. Yeah, so in some ways, it’s a little bit better. It’s definitely not as pretty as the donut chart. And what I think what finally made me change my mind was really thinking back to who the audience is. And I would say 100 times out of 100, your audience is more familiar with a donut chart or a pie chart than they are with a TreeMap. So really at the end of the day, if that’s what your audience is familiar with, then yeah, I changed my mind. It’s not TreeMap. You should go with the donut chart because that’s just going to be more consumable by your audience. In almost every case. I can’t imagine a case where someone knows about a TreeMap, not a donut or a pie chart. So all that to say, again, this is all very subjective. Think about your audience, think about your message, and make your decisions based on that. OK, so we’re moving on to the second section here, where we’ll talk about single-purpose visualizations or things that they don’t really, they’re not really single-purpose, but we don’t really have options for visualizations to use. I think I’ve mislabeled this section because they’re not really single-purpose. And the first one here is, we have 10 minutes. I think I’m going to have to go pretty quickly, but if there’s any of these you want to look back at and talk about, please leave those questions in the Q&A. We’ll try to get to them. But all of these have lots of documentation on how to use them in our Experience League. And so you can go there if you need more details. So flow visualization. This is really good for when you don’t know how people are navigating through your site, or you start to see something weird and you need to confirm, like, how are people making it from page to page? And that’s how, I would say, 90% of the time this is used with the page dimension. You can use any dimension here and see how people are flowing through. But yeah, this is where you would go in and explore and say, OK, starting with this one page, how are people going from that one to where are they going to after that and then after that? So it can be really good from an analysis perspective. It gets a little bit clunky as you’re clicking around to use this in any sort of data communication. But one thing that might help is a tip I learned, actually, from another session similar to this one. I can’t remember when it was or who did it. But right-click. Right? So product team is always talking about how, when in doubt, right-click. This is a case where that holds true. So right-click in the white space above the visualization, and you’ll get these options. One of them is to copy the data to a clipboard, and the other is to download that data. And this is what that looks like. So if we’re just looking at the home forward data, because it really is two different data sets, it’s one for previous and then one for the next steps. So those are two separate. If you downloaded this whole thing, it would be two different tables. But anyways, if you’re looking at the home node, you can see how many path views that had. And this was taken in these were different date ranges. That’s why the numbers don’t line up. But then you can see home and search results. That’s how many people went through the search results from the home page. And then home search results home is 1303. So you can get that data in a tabular form and then play around with it and put it in a format that might be a little bit easier to consume or to narrow in on the specific thing that you’re looking for. So that’s just a little tip with the flow visualization. And we actually used that with a client recently. All right, fallout reports. So whenever we talk about flow reports, usually the next one is fallout reports. But this is where you have a more predefined path that users are going through. And it should be very intuitive. In this case, we’re looking at a simple shopping cart. So you have product views, cart additions, and checkouts. And how are people getting through that? And what I love about this, my favorite thing about the fallout report is the ability to compare different segments. So in this second screenshot, we’ve got all traffic compared to desktop in blue and then mobile phone in the orange color. And this is really, I feel like, intuitive and easy to understand. You can easily see very quickly that mobile phone has a higher conversion rate through this funnel than desktop. And you can do whatever segment you can think of as a comparison. So you can do date-based segments. That’s a really good use case where you have, if you’re looking at last month compared to the month before, you can see how your fallout has shifted month to month. But really, any segment that you can think of, I think, is valid. And again, this one especially, I think, is really easy to understand from an external audience. You don’t want them to get bogged down in the actual numbers because those aren’t really going to align with your, like the product views here is not going to align with product views if you just pull the metric, same thing with all these. But overall, it’s a really approachable visualization that you can use. And again, I like the comparison feature. Next, we’ve got the bullet chart. And this one is very single use case. I don’t know, you could probably think of another use case. But really, here, we’re trying to measure progress to a goal or an objective. You’ve got your options in here. And the teal bar in here is referencing something in a freeform table. So really, single use case, I don’t see this a lot. But I have seen this where teams have specific static objectives for a certain metric in a certain time frame. You can use this to keep track of that. The next one here is histogram. I think what I end up using this the most is when we think that we’re not getting enough detail from either a mean average or median average. So most of the time, you’re looking at a mean if you’re looking at maybe, in this case, like product views per visit. You look at the mean and you’d say, OK, how many product views per visit? But you’re like, well, that’s probably different across different users. So maybe let’s look at a median and get rid of some of those outliers or see what the average is without those outliers. But then even beyond that, if you want a really full view of how many product views there are per visit and see what that distribution curve looks like, you can pull a histogram. And that’s really when I use it the most. There’s probably other use cases. And one tip I have here is a little bit of a hack. And it’s something I like to do. When you have some of these visualizations, you can do this. Usually, most of the visualizations, you build a freeform table and then build the visualization off of that. Some of these, though, you go through a little, you make a couple selections, and Workspace will build that visualization for you and build the freeform table in the background. So if you click on this little circle next to the visualization title, you can click Show Data Source. And that’ll show you the data that was generated. In this case, this is what was generated. So it’s kind of on the left side. I changed a little bit. But it creates all these segments. And then you’ve got this one segment showing your histogram for your bar. You can add a segment to that. In this case, I changed it and said, I want to show visits with a card edition compared to visits without a card edition. And it works similarly to the fallout where you have this comparison bar. So you can see the histogram for one set of visits versus another set of visits. So that might be helpful if you’re trying to compare things. Again, it’s just like a little bit of a hack that you can do with your histogram. All right. The next one here is pretty basic. But they are important. The summary number, summary change, and key metric summary. These are all really important. It’s really the best way to call out a specific number change, which is sometimes hard to do with individualization. The one thing I’ll say is you do want to make sure and lock down your metric. Because you don’t want someone, so especially if you hide the legend down here, if I had named this like visits, someone could go into the free form table and click a different column and say, I want to see this. And it would update here and you wouldn’t really know it. So make sure to lock those down. I think in most cases, you’d want to do that so that people aren’t confused. I’ve had that happen before. Someone clicked another cell and it changed it. It wasn’t visits anymore. It was something else. And people got really confused. All right. We’ll probably speed it up a little bit. Maps. I have historically kind of not been on board with maps because I’m like, really, what value am I going to get? But I’ve come around a little bit. And the main reason is maps are really approachable. So if you have any reason to throw a map in and you have an audience that’s not super data savvy, they’re going to latch onto that map and be like, I understand that. This makes sense to me. And they’re going to feel better about themselves, which makes them more likely to engage with the rest of the information that you’re presenting. So this one kind of goes back. I mean, I’m sure there are use cases for maps. But this kind of goes back to really aligning with your audience and making sure that what you’re sharing with them is presentable and understandable for them. And I think this can help. So use those maps. Then diagrams. I think most people probably understand these, but it’s going to show you the intersection of up to three segments. In this case, we’re looking at intersection of marketing channel traffic by visitor. So how do you get that? So how do visitors use multiple marketing channels or marketing channel groups? There’s usually some overlap in here. So these can be, I’ve used these for sort of dissecting or trying to figure out different problems as well as just showing overlap in traffic for certain things. So I did at one point try to figure out how to add a third segment using that view data source. I couldn’t get it to work in the visualization, but I did. I couldn’t get it to work in the visualization, but if someone can figure that one out, I’d be really impressed. All right, scatter plots. So these are maybe my favorite visualization and one that I don’t use nearly enough. But it’s going to show you the relationship between two different, so with a given dimension and the various dimension values, how do those stack up against each other based on two different metrics? So here we’re looking at bounce rate and checkouts per visit. And then on the bottom here, the diameter is dictated by the amount of traffic, the amount of visits per marketing channel. And this can be, if you can figure out two different non-correlated values to put on your X and Y axis, this can be really insightful and show you values that are sticking out. Obviously, over here we can see this value has a really low bounce rate and a really high checkout per visit rate. And again, it’s all fake data, but a lot of times when I’ve done this really thoughtfully, I’ve come away with things that I never would have thought. So use those scatter plots. Finally, maybe the most intimidating of all visualizations is the cohort table. And I say that just personal experience as well as what people have, I’ve heard people talk about, but cohort tables can be intimidating. And it’s probably not your first option for what to use. But I do think similar to the scatter plots, if you make yourself use this, you will come away with something that you didn’t know before. It does things that no other visualization can do. So we’ve got a couple of examples in here. How many weeks after a visit does a checkout occur? So this is a pretty standard example. And what you’re looking for in here most of the time, I think, is what are these cells that pop out? So it has this formatting where we’ve got darker cells and lighter cells. Where do they pop out? And what does that mean? So you got to think through, OK, what is this telling me that this value is a little bit darker green for February 5th through 11th than the rest of the values? So yeah, play around with it. See what you can figure out. This next one is what desktop browser is driving the highest rate of return visits. So in this case, I’m showing that you can use not only time-based cohorts, but also any dimension value that can be your cohort. So in this case, it’s the browser that you can use. And I’m speeding up a little bit faster than I wanted, but we’re running out of time. And then the last one here is just showing that you can look at the number of visits showing that you can look prior to an inclusion event and then after an inclusion event. And I didn’t even really touch on inclusion and return events or criteria. But you can look that up. And there are, I think, plenty of videos that you can learn how to use these cohort tables. So play around with it. Try to use it. I do think it will be valuable, especially from the analysis side, coming to new insights that you couldn’t otherwise come to. OK, so we are done with the presentation portion. Our last poll here is on a scale of 1 to 4, how comfortable do you feel using visualizations? Again, the goal here is that you are more comfortable than you were when you started. I hope we have not progressed. Give me a couple more seconds. We’re going to have maybe 12 minutes or maybe 10 minutes for Q&A. I think there’s a closing slide at the end. All right, let’s keep going. Oh, good. Look at that. Man, look at that Instagram. Very nice. OK, so questions. I think Justin’s popping back in here. There you are, Justin. Yeah, that was fantastic. What a great deep dive into visualizations. I personally learned a lot. I will not be using bar charts as much as I have in the past. So thank you for that. There’s going to be a dramatic drop in bar chart using. We’re going to be like, what’s going on? We’ll see that drop. All right, so we’ll be taking a few questions now from our audience. So feel free to ask a question in that Ask a Presenter panel. And as a reminder, if you have to leave early, please don’t forget to take that short survey. It’s just three questions and helps us select topics for future sessions. All right, so we got some good questions here. So this one says, is there a way to separate separate flow chart to only use specific events as for each step? To only use specific events for each step. I know you can use do event based. I don’t know if you can do specific events for each like touch point in the node. If needed, we can say that for a follow up. That seems like it might be a technical one. You might want to play out. It’s something I should know. I honestly don’t use those a ton, at least all the different features that they have available in there now. And they’ve changed slightly over the since it was released originally. Yeah, no worries. Good question. All right. Yep. This next one says what to do in cases where one channel is too high and another is too low and flattens out the line chart. So I think that question is kind of getting to the dual access. There is a normalization option in your visuals. So you can normalize the data. And that should shrink them together a little closer. That would be another way to view data that really has a different scale. So that might be one option is to use the normalization. I wish I could share my screen. I feel like this is more. I know I had the safe thought. You could. Yeah. Probably just like everybody watching some of this. It’s okay. We’ll do a follow up. I have that. See if we can answer a lot of these and show more screenshots. Screen shares. All right. This next one says, is there a way to customize visualizations like add spacing between bar chart items? As far as I know, you cannot add spacing. And that’s honestly like as much as you can do in Workspace and if you can stay in Workspace, like that’s great. But it’s in my mind not a great reporting tool. It is a great, it’s a fantastic analysis tool. But because you don’t have all of those options, like I find myself, if I really need to explain something and control a lot of those different features in the visualization, I a lot of times will export it to Excel and whatever your visualization tool of choice is. But yeah, really the options you have, you can find all of them in either the gear icon. And then some of the visualizations have another, have other settings that you can adjust, but most of them you can find in that gear icon above your, above the visualization. Awesome. Great. Okay. Next one, how to interpret the data on the cohort table. Should we look at the rows first and then the columns? Maybe dive in more into those cohort tables. Yeah. Can we go back? Let’s see. Y’all can still see my screen, right? Yep. They can see my screen. Okay. So how do you interpret this? Let’s look at the basic one here. The, and I did fly through this because we were running out of time. Basically a cohort table is building cohorts. You have different groups of users. So in this example, we’ve got five different groups of users that their inclusion criteria put them into a group based on the day that they came to the site in this case. So when you’re reading across it, you can say the group that came in from January 2nd to 28th they came back. And I think that the checkouts is the return criteria. So they came back and had a checkout after one week and then, or, and then after two weeks, after three weeks and so on. And so what you’re doing with this is you’re comparing these cohorts and you’re saying, okay, the group that came in this first week, they came back and all of these are the same percentages. So I think if you added another digit there, you might be able to see the differences. And that’s what the conditional formatting is showing us. But you really want to compare across these columns. I think it was the original question. If that makes sense. That’s probably how I would read it because that’s, it might change depending on the use case. But in this use case, the important thing is like, how are these people coming back to the site and checking out and how is that different across these different cohorts? Awesome, thank you. Yeah, cohort tables are a great tool. So I’m glad you’re diving into those. They’re not utilized enough, I feel like. So appreciate you sharing that. Yeah. All right, next question. How do you properly trend percent change? I’ve never met to make it work for me. Trend percent change. That is a good question. Because it would have to be like a running metric. I don’t know if you can trend percent change. You can obviously trend percentages like of a total, but percent change, you basically be saying like, this is the metric over the previous month. I think if you could do it, you’d have to, I’d have to play around with it. I think if you could do it, you’d have to build a calculated metric showing percent change from a prior, with like a dynamic date. So you’d pull in this is, yeah, I’d have to like actually show it. But you’d have to pull, you’d have to create a calculated metric where your date was like last, last week and then last, like two weeks ago, right? And then show that, calculate the change based on that dynamic date. But I’m not sure if that would even work. I’d have to play around with it. Man, a lot of these are like tough questions. Yeah, they’re not taking it easy on you for sure. I was hoping you were going to curate some of these out and throw me some softballs. No way. Give you the hard ones. Like I said, we could do some follow-ups. A lot of these are, sound like really hands-on questions where you need to be in the product and kind of play around. So we can do some follow-ups. All right, here’s our next one. It says, is there a way to add a line to compare of the baseline, say average or moving average? Sorry, I missed that first part. Can you say it again? Sorry. Is there a way to add a line to compare the baseline, say average or moving average? Yeah, so if you go into, again, go into the gear icon, you can add a trend line. That was one of the options we have with the line chart that’s not available for the other ones. But the line chart gives you an option to show a trend line. And one of those options, so it’s a little drop down box, one of those options is moving average. And I wish you could show just the trend line because sometimes it gets a little, there gets to be way too many lines. But yeah, you can add those trend lines for any, let’s see, I’m pulling it up right now. And it’ll compare moving averages. That’s I think what you’re looking for. Is again, a line chart, go to the gear icon, click show trend line, and then select moving average. And you can select the number of periods you want included in your moving average. And that’ll reference like, obviously, is it a weekly or daily or monthly chart? Great, thank you. In the fallout chart, how did you get the visualizations on the right? It looks like you broke it down. I haven’t seen this before. Let’s go back. All right, fallout chart. I assume it’s, they’re talking about this. To get the comparisons, and if you’re not, please clarify. But to get these, you just drag your segment into this just right here, basically right in the top above your fallout, the visual. Just drop it in there, and it’ll add that comparison. It should pull up a little drop zone once you are hovering over it. But yeah, it’s just any segment, just pull a segment over. Great. I’ll jump us back to the questions. Okay, it looks like we have time for probably one more. Is it possible to consolidate the bottom variables as others, or horizontal bar can be useful with having more than 20 variables? Yeah, that’s a good question. I don’t think so. Other than, I mean, there’s obviously like workarounds you could do. You could do it through segmentation, which is not ideal. Yeah, I don’t think you can do that in Workspace how you’re wanting to. I did a lot of that in Report Builder when I was using Report Builder. Maybe one of the main reasons I used Report Builder for certain canned ongoing recurring reports was just to lay out the dimension items and show the top ones and bucket everything else into another bucket. You got to be careful with how you aggregate things. But yeah, I don’t think there’s a way to do that in Workspace. Thank you. All right, we have a couple other questions we didn’t get to. So like I said, we’ll be doing some follow up after. Maybe we can get into more of these technical questions as well. But I just want to thank everyone for joining today. Before we go, I wanted to remind everyone that this webinar was recorded, and we’ll be sending out a link out with the on-demand recording. We’re also excited to share with you some of the questions that you’ve asked us. We’re also excited to share our first in-person experience maker skill exchange, which will be live in New York City for Adobe Analytics coming up next week. So if you’re in the New York City area, join us on November 9 to learn analysis, workspace best practices from Adobe experts and customers just like y’all. You can find the registration link up in the related content section on your console. And as always, you can find more information about the upcoming analytics webinars, similar to the one you attended today and register on the events page for the experience league, which is also linked in that resource section. Lastly, we would love to get your feedback on Adobe Analytics. If you have thoughts, please share them with us on Trust Radius. We will actually give you a $25 gift card as a thank you for your time. You can go ahead and scan that QR code or click the link in that resource section, the top right part of your screen. As long as you do it by November 20th, you’ll qualify for that gift card. So feel free to share your thoughts. And again, just want to thank you, David, for your time sharing such great information. And thanks everyone for joining us. I hope you all have a great rest of your week. Thanks, Justin.