Author: Jay Leslie

Tidbits from Stephen Few’s well-revered book, Now You See It Ever since we began including a section on data visualization in our formal training, we’ve only gotten one negative piece of feedback. The particular attendee suggested something like “I could have done without the content on how to make a PowerPoint slide.” We welcome and take all feedback very seriously of course, but the comment came off as ironic given that in the training, I discuss the Dunning-Kruger Effect and its importance in successful data visualization in business today. Early in his book Now You See It: Simple Visualization Techniques for Quantitative Analysis, Stephen Few talks about specific attitudes that a data analyst might have that would help them excel in their field. We actually don’t think about what might be the proper attitude for a certain role or activity in business. At least, I don’t ever recall having been asked about my attitude toward what I might be doing in an interview. Nor have I experienced a company whose culture fosters certain attitudes. But it makes sense, right? Although Few is focusing in on data analysts, my personal feeling is that several of these attitudes are also something that our audience should attempt to call upon in order to best absorb the information being presented: Open-Minded and Flexible Here, he quotes the physicist Alan Sokal who said “science requires two fundamental attitudes: Willingness to accept what you find Willingness to discover that you are wrong” I think we can pretty easily adapt this to business. Our plans are always changing, and often the entire direction of a project might need to be altered to account for significant realizations, as the initiative evolves. Humility helps here. Imaginative How many executives have you encountered that were open to new ideas? “If I had a nickel…” certainly does not apply here, although it does happen. As Few says, “We don’t have to be creative geniuses to blaze new…trails.” Skeptical Although you may be skittish about this one (because we want our audience to trust us) think about what healthy questions and skepticism will inject into the effectiveness of delivery of the information or the project overall? More collective thought usually equals better solutions. Aware of What’s Worthwhile The must-have follow up trait to Skepticism. I once had a Director berate me in a large meeting because some of my bullets were slightly larger than the others. It’s important to be detailed, but was that really worthwhile and relevant? Maybe not the best example, but you get the idea. As Few says, “Not all the questions that we might ask about data are of equal value.” How to foster these attitudes if they don’t currently exist is something different altogether. It’s not your job to coach your leadership or colleagues, and working to impact attitudes falls squarely in that category. Leading by example seems like the only approach that might be within your power.

I love this topic because it elicits a higher level of thought around designing the data visualizations we need in planning, in a way that my simple mind can consume. In her book “Storytelling with Data,” Cole Nussbaumer Knaflic points out very early on that there are really two kinds of data visualizations: exploratory and explanatory. Exploratory visuals are created to help us figure out what the important things are within the data…they have an analytical purpose. Explanatory visuals are meant only to show us the important things…there should be little to no intended analytical value. Makes sense, right?  What if I ask you what type of visuals you are creating now? Your answer likely will be “explanatory,” but based on what I see daily, that may not be entirely true. Apparently, Nussbaumer Knaflic agrees; “…it can be tempting to want to show your audience everything, as evidence of all of the work you did and the robustness of the analysis. Resist the urge. Concentrate on the information your audience needs to know.” My personal experience tells me that there is probably a balance here, in terms of what data you should include to offer context (another topic of the book, but for another time). This also translates into a spectrum of explanatory vs. exploratory. For example, to communicate why your plan might be behind schedule, you may have to include lots of backstory data. But once that story is told, is it necessary to revisit in every visual? After conversing with so many of our end users, I very firmly believe that there is a lack of awareness on the part of both the audience, and the creator of the plan communications, as to what type of visual is necessary. Is your audience expecting you to present them with a report, or do they intend to decide what’s important, for whatever reason? Often, I don’t think this conversation has even occurred. Most of the time we’re simply asked by our audience (someone senior to us) to create a visual, and we just do it…we’re programmed that way. It’s important to know when it’s time to do what your told. But if we can facilitate the ability to ask “why,” we’ll have a much clearer understanding of what it is we’re trying to accomplish…what our audience wants or needs out of our plan communications. After all, your audience has probably had absolutely no training in data visualization, communications, or graphic design. They’re just pulling on what they know, based on experience. Simply put, it comes down to design. If we can initiate a design phase (or redesign) for our data visualizations, we’ll understand early on where our chart will sit on the explanatory vs. exploratory spectrum. Armed with this knowledge prior to creating our visuals, we can more easily groom our data, remove what isn’t needed, and come to more effective plan communications results in OnePager. Related Content Webinars (watch for free now!): Better-than-ever Project Plans: Using Cognitive Psychology to Communicate Effectively Best Practices in Data Visualization Articles: 11 Quick Tips for Project Managers Creating a Custom Report in Project 2013: This Week’s Tasks Storytelling with Your Gantt Chart

“Storytelling” is something marketing experts have been talking about for years, as a better way to communicate brands, products and what sets businesses apart. It’s now a recognized, tried and true approach, and one that has taken a foothold due to its effectiveness. Our customer experiences online are, in large part, shaped by “stories” that marketers have set up for us to more easily get to know what they’re selling and eventually buy it. Because of this success, storytelling is now quickly making its way into business vernacular, specifically with respect to data. With so much data being collected over the course of doing normal business, we need better ways to communicate that data (the stuff we’re “selling”), in such a way that it can be easily consumed (“the stuff they’re buying”). Technology is providing us gigantic leaps in data visualization capability for business writ large, but how best to leverage these new capabilities in a way that allows us to easily grasp the information is still — and probably always will be — evolving. The same is true in the planning world. In project management our stories tend to be question-driven and are similar between initiatives yet different depending on where we’re at in the process or at what level we want to view the information. Here are some examples. Project-level Conception/Initiation/Planning/Design What is our plan and are there options to choose from? Who is responsible for what? How long do we think it will take? What will it cost? How will we measure ourselves? Project-level Execution/Performance/Monitoring: How are we progressing? Where are we falling behind? What can we do to fix the issues popping up? How are we tracking against estimates? Are there any new decisions to be made based on what has been learned so far? Project-level Close: How did we measure against our estimates? What are the occurrences that we can take forward in our business to learn from, and improve? Portfolio-level: Where are my risks and financial exposures the largest? What can be fast-tracked, if necessary? When can I expect revenue, and at what scale? Do I have any patterns of success or delays across my portfolio? These questions — or stories — are high-level and not at all-inclusive. They’re also where a manager might start when he or she begins to design the communications plan (otherwise known as “data visualizations”). Often, we jump straight into attempting to squish many of these things into a single visual, either because we’re asked to or because we’re driven by an inherent attempt to “wow” our audience. By lumping lots of varying data into a chart, however, we end up making two big mistakes: First, we try to deliver too many options of focus, which slows viewers down when they’re reading. What we do is deliberately camouflage each story. To absorb each dimension of data, we have to bring forth each one separately in our mind’s eye to allow people to consume them together. This is difficult to do, so why not isolate each one from the get-go into its own visual? If we do this, the information our audience needs to absorb will become simpler and clearer, which is what we want! Second, Scott Berinato quotes the psychologist Robyn Dawes in his book, Good Charts: “…Cognitive capacity shuts down in the absence of a story.” If we cram a bunch of different dimensions of data into one visual, we’re diluting the narrative of each. If, instead, we extract those dimensions into separate pages/images, our readers will be better able to quickly grasp the information they need to read the stories. As communicators, we first need to uncover what it is that our audience should know — what story they need to be told (using the data) and then create a visual for each (at least). I love the story-time image above because it’s not too far off when it comes to business. Although our audiences aren’t children, they’re also not intimately familiar with the project data that we’re often presenting. We, as the project managers and schedulers, eat, sleep, and breathe this data, but our sponsors and actors often must learn what it is they are looking at, to some extent, each time you present to them. On top of that, our audiences often suffer from the Dunning-Kruger effect, when it comes to plan communications. They like to tell us what they want to see, even though they have no expertise in the fields of communications, graphic design, data visualization or planning. If you’re working on gaining knowledge in those areas, don’t hide it. At some point, when you feel confident enough that you can make positive changes, propose a concerted effort for improvement in your plan communications. How do you like to tell stories with your data? Share your ideas with the MPUG community in the comments below. A version of this article originally appeared on the OnePager blog here.

…Well, in plan communications and Gantt charts, it does. Size, when used as an attribute to denote meaning in data visualization, will likely force our brain to look at the largest items first. In her book, Storytelling with Data, Cole Nussbaumer Knaflic tells us, “Relative size denotes relative importance.” Size is just one of the “preattentive attributes,” or visual clues, that tell our brain where to look. When employed correctly, as Nussbaumer Knaflic suggests, these attributes can be used not only to tell our audience where to look first, but also to create a “visual hierarchy” that will allow us to communicate layers of data to our audience in a natural flow. The hierarchy allows us to better tell the story of the data, even in a single visual. The preattentive attributes in visuals are: Image credit: Colin Ware, Information Visualization, Third Edition: Perception for Design Motion was also noted, but not shown above. Nussbaumer Knaflic, smartly, also shows us preattentive attributes in text, which are more obvious, relatable and make sense to note here, given how often we have text in our Gantt charts: Let’s look at a couple of applications using preattentive attributes in a plan communication. For some context, assume that we’re working to communicate a portfolio of potential drug candidates in development. Our executives want to know where the major milestones lie, but the final milestone is the most important one. In the above example, we’ve used several preattentive attributes: Size Color Enclosure (swimlane borders) Bold text Shape My eyes are first pulled to the large red milestones and from there to the legend to figure out what those mean, supported by the other colors/shapes. I then begin to look at the large bold font within the swimlanes, emphasized by the swimlane borders. The actual bars that connect the milestones are thin and therefore come across as unimportant. This is a hierarchy. With some more work and feedback from others, we could probably get this hierarchical design as close to perfect as possible, but for now it works. Here’s another example using the same plan; but in this presentation we’ve been asked to point out only the final milestones as well as anything that is “critical.” Our audience also wants to understand how firm certain estimates are. Here we’ve pulled out all the stops, using: Enclosure (in two ways: swimlane borders, and the blue highlighted box) Size Color Italicized text Underlined text Bold text While the preattentive attributes overlap a bit between the two images, they’re quite different from each other. In this second example, the first three milestone shapes in the rows become almost non-existent. They’re relegated to the background due to their color when compared to the final milestone, like the bar/line on which they sit. In the second example, we would probably pick up on a couple of those bars being black (critical), which draws us to those top two rows more easily. Last, we now have a light blue box highlighting milestones that are considered “directional” with regard to their estimates. You have probably used preattentive attributes before. But now that you’re more conscious of them, we hope you’ll do two things: reflect on how they can improve a chart you’ve created in the past and play with different uses of the preattentive attributes in order to make your future visuals even better. Image Source

Nine. Well, so says Scott Berinato, in his book Good Charts. He bases this number on a conversation he had with Tamara Munzner, a data visualization expert and professor of computer science at the University of British Columbia. Here’s an example Gantt chart with more than eight colors. Munzner is quoted (in the footnotes) as saying “There are fewer distinguishable categorical colors than you’d like. You don’t get more than eight.” This is an opinion — a subjective view from someone who has studied data visualization for over a decade, but an opinion nevertheless. Humans are all different and are going to have a different threshold for where that too-many-colors-line is, but the number eight feels right… and maybe even a little on the high-side for me (my brain is feebler than most). Have you ever heard the expression “Can’t see the forest for the trees?” This is perfect for explaining what we experience when we’re presented with a chart that has too much complexity. We have a hard time visually and mentally extracting what the most important information is. Instead, our brain wants to look at everything at once. Based on his own research and Munzner’s thoughts, Berinato adds an accentuating comment that, “The threshold at which individual data points melts into aggregate trends is surprisingly low.” This makes sense, right? For that reason, we, as the ones creating these plan communications, need to be very conscious of over-using color to illustrate attributes in our data. The principle can be applied beyond color, as well. In his book Scott Berinato also wrote, “Simplicity is courageous.” Any sort of complexity will take away from the important information — the “story” — that we’re trying to get across, and we should do everything we can to simplify for the sake of clarity. Is Munzner right? Again, it’s subjective. You should ask your audience. When you’re building your communications plan, use your judgment. Don’t accept what’s being asked of you as the right way if it doesn’t make sense. Ask questions to find out what it is your audience wants and needs to see. Go through a process of iteration and drafts. From that point you can work to find better solutions within the designs of your data visualizations. A version of this article originally appeared on the OnePager blog.