This post kicks off a partnership with Cole Nussbaumer Knaflic from Storytelling with Data in which we are interested in exploring how organizations improve the way they communicate data. This is one of the biggest questions Cole and I both hear at workshops and conferences, so we thought we would start a bigger conversation by posting some initial thoughts here and at Cole’s site. To kick this off, Cole is writing about steps an individual can take and I’m talking about steps an organization can take (you’ll notice some interesting overlaps). We’re then asking you to provide us with your thoughts and experiences. You can get in touch with us using the comment boxes on either site, Twitter (me/Cole), or sending us an email. We’re not sure where these conversations will take us, but we’re looking forward to the journey.

Let’s say you’re really into data visualization. You’ve attended some Meet-ups, read some blogs, keep up with things on Twitter. Maybe you’ve even taken a course or two. And now you’re implementing your new-found strategies in your ever day job as an analyst, researcher, or marketer. You believe your new approach will improve the way your organization communicates with your audience and help that audience adopt your organization’s ideas and messages. How do you get others at your organization to buy in? How do you convince them that thinking differently—more strategically—about communicating data, can help your organization improve? Here, I’ve outlined 4 steps I think any organization can take to improve the way they communicate their data.

1. Build the Team. I believe there are three core skills needed to create great, effective visualizations. First, you must have some understanding of statistics—how to work with data and understand its pitfalls, errors, and limitations. Second, some understanding of design—how color, font, layout, and user experience, for example, work together to bring a person into the visualization and keep them there.

Finally, to create interactive visualizations, some understanding of programming languages. It’s certainly not the case that every visualization needs to be interactive (in fact, I think far too many visualizations are interactive), but in the cases where drop-and-drag tools won’t suffice, some programming may be needed.

I don’t think People who are great at all three of those skills—statistics, design, and programming—exist, so I call them Unicorns. In The Idea Factory: Bell Labs and the Great Age of American Innovation, Jon Gertner writes that teams at IBM “required researchers with material processing skills, chemical skills, electrical measurement skills, theoretical physics skills, and so forth. It was exceedingly unlikely to find all those talents in a single person.” As at IBM, what I think we need to do is to not rely on or try to find individuals who can accomplish all three tasks, but instead, to create teams within our organizations who, together, become the unicorn.

2. Embed the Team. Now that we have teams of people, let’s place them close to the content creators and decision-makers. In too many organizations data folks sit in one silo, analysts in another, and managers in yet another. Then the designers sit at the end of the hall in their own silo wondering what’s going on. What we need to do is bring those groups together so that we have a complete whole.

In their article on creating a data culture, DJ Patil and Hilary Mason note that, “If the data scientists are isolated in a group that has no real contact with the decision makers, your organization’s leadership will suffer from a lack of context and expertise.” An organization that successfully communicates its data is broader than just where the teams sit. What’s as important is that those collections of people are empowered to work with one another and communicate their work to decision-makers and the audience. If a part of the team is made to feel like their input is secondary to the primary output, then the whole thing fails.

3. Change Habits. This may be the hardest part of the whole thing. To get more people in an organization to buy into a spirit of better data communication, you need to change the habits and the culture. In Charles Duhigg’s book, The Power of Habit, he suggests identifying the parts of the “habit loop” from which you can then work to break or change the habit. The three parts of that loop—Routine, Reward, and Cue—lend themselves well to the data workflow of an organization.

Say that every week an analyst is given a data set and asked to build a report (Cue). She works with the data and builds a report or table (Routine). That report is handed off to the manager, who is pleased because she can go to the exact same spot of the table each week to find the familiar number that matters (Reward). But what if something else is going on that keeps getting missed because of this standard, well-worn routine?

Instead, let’s identify the parts of the habit loop that can be changed. The analyst still gets the data every week (Cue)—maybe the format of those data can be changed, maybe not. Instead of building a table, however, the analyst uses some visualizations, adds color, highlights, or other attributes to highlight important trends, statistics, or distributions. The analyst uses (hopefully good) data visualization techniques and principles to reveal important trends or other findings. It’s not so much about providing the weekly report because that’s the task at hand, but using data visualization principles and strategies to help the managers (and others) make discoveries and find insights in the data (Routine is now changed). Now, not only is the manager pleased (Reward), but so is the analyst’s colleagues and other managers because the new report has more depth and enables users to make better decisions.

4. Build Success. Start with small projects and build on your success. In my experience, a lot of organizations want to start by building the latest, greatest, biggest interactive, exploratory data tool. Yet more often than not, those organizations really just need a short report, blog post, or a simple bar chart—their audience just needs the bottom line and important insights.

If you look at the Urban Institute, for example, we built smaller, interactive pieces and demonstrated success with those projects (measured by webpage views and feedback from clients, funders, partners, and policymakers). Then, as those successes grew, more and more researchers (and grantors and partners) recognized the value of such tools and products so that the demand grew. Just look at this progression of visualizations from 2010 to 2016.

Once you’ve demonstrated small successes, don’t stop there. Now is the time to branch out, both in terms of the types of visualizations you can produce, but also the types of people in the organization who may want to be involved in this new push to improve data communication. Here it comes to pushing boundaries, trying new things, and experimenting with new tools, technologies, and means of communication.

Improving the way we communicate data to an audience—be it a colleague, the public, the press, decisionmaker or policymaker—is how we can affect change. Relying on the same old ways of publishing PDF documents with a standard line chart surrounded by text may not be the most effective ways to help someone embrace your ideas or make discoveries. These four steps are by no means an exhaustive list of how you get buy-in from colleagues, managers, and your audience, but I believe they are a good way to get things started.