10 Lessons About Using Data to Spark Action

A few years ago, I was peripherally involved in work that a small county in California was doing to raise awareness about the large number of falls among seniors across the county. To help make a compelling case, the key staffer there wrote a data report – in this instance, in the form of a story that integrated photos, quotes, and charts. Most of the time, such reports sit on the proverbial shelf, too seldom leading to impact. But in this case, county supervisors took notice, and increased funding for the fall prevention program by adding an extra staffer. 

Such examples where data leads to impact can be all too rare. It’s hard work, after all, when you think of all the steps busy nonprofits and government agencies, or others, need to follow to activate change from data: first finding relevant datasets, then analyzing results, then visualizing the findings and putting together the resulting brief or report, and, finally, implementing a communication campaign to reach constituents who can take action. It’s a tall order, and all too often we’re so focused on the initial steps that we have little time or energy for the all-important communication work.

"In my travels, I’ve learned that there are some concrete steps we can take to increase the odds of transforming data into action – such as more awareness raised of an issue, policy or program improvements, bolstering funding for an initiative, or empowering communities with data they need for decision-making."

Here are some recommendations you can apply in your work to achieve such last-mile impact: 

1. As one wise data person once told me: “If you can’t be with the data you love, love the data you’re with.”
Unfortunately, that dataset you yearn for that has that exact demographic breakdowns you want simply may not exist, even in this age of data, data everywhere. Rather than wait for that magical dataset to appear, utilize what’s available to effectively support the case you’re making. You may find that what you resort to using works quite well as a proxy.

2. On a contradictory note, try to drill down as much as you can when presenting data findings. If you have the option of communicating, say, just county data vs. county data plus numbers for all school districts within that county, err on the side of reporting out the more finely tuned data. More granular data will be more relevant and relatable to your audiences, who are more likely to identify with the town they live in and its school district, not the county where they reside. Take, for example, this simple, unadorned map of the CDC BRFSS 500 data (it was put together as a very quick prototype). Choose your city, and you’ll see breakdowns by Census tract. You may not even notice how basic the display is, given that you can focus on such granular findings.

3. Create data stories, not just data visualizations. While it’s true that a picture (or map or chart) may be worth a thousand words, that picture + words can be worth infinitely more. Find ways to integrate the visualization you’re creating into a broader canvas where you can tell a story to help make your case. Your story needn’t be long – the digital equivalent of a one-page brief that includes a few graphs or maps and contextual photos or a video may be all that you need to engage your readers. The data website, FiveThirtyEight, for example, used a visualization of just dots to help readers understand the 33,000 annual gun deaths in America; it’s the supporting narrative, presented as a slideshow, that really makes this visualization compelling.

4. Appeal to both the heart and the head. We may understandably assume that presenting data means our arguments should all be logical and fact-based. But remember that we’re all persuaded in varying ways. Building on the data story recommendation above, it may be advantageous for you to appeal to emotion as well. For some people, the logic of data speaks to our head. Others might be persuaded by a quote and/or poignant photo that tugs at our heart strings. And some of us may be persuaded by both. Why should we need to choose? Through data stories, we can appeal to both the heart and the head.

Here, for example, are some findings published by the Sonoma County Department of Health Services sharing results from a survey of local farmworkers. Sure, it’s a data report, but you have to scroll down the page to even see the first graph. What’s front and center? Poignant photos that draw you right in.

5. Enlist writers and designers to help you tell your story. Another wise data person observed to me: “The hardest thing for a data person to do is to speak English. The second hardest thing for that data person to do is to write English.” If you’re a data person who lacks such communication skills, don’t try to harness your inner Hemmingway. Your time may be better spent searching for contract writers and designers who can help you package findings for maximum engagement. 

When I was a program officer at the California Health Care Foundation, we took this approach, recognizing that we didn’t have the bandwidth internally. We hired a writer to help us come up with a compelling way to describe hospital-level disparities in birth outcomes. That writer’s resulting “Tale of Two Births” concept helped the design firm develop an effective display format, and the resulting infographic became one of the most widely visited publications in 2015 for the foundation.

6. While it’s important to be accurate, don’t shy away from being bold and provocative when presenting data findings. I've found that when you ask a data wonk – such as the owner and publisher of a dataset – to review what you’re planning to disseminate, their edits may soften and over-explain the findings, often rendering the messaging as, well, mushy and far less meaningful to an external audience. There's likely to be tension between these two camps (communicators and data wonks) as they collaborate to publish a data visualization, and it’s important to at least make the push for presenting findings in a way that can help your rise above the noise. 

7. Always talk to your users. Before you build a data story or some other data product, ask your audiences what messages most resonate with them. Find out, too, how you can best package information in ways that will encourage them to engage with your content and activate them to take action. Think through, for example, whether your users will be most likely to read your findings on a computer, a tablet device, or a phone. Or, ask them if they plan to use this material offline instead – e.g. via a live presentation at a community gathering or in a one-on-one meeting.

Especially with cross-sector coalition work, this methodology of talking to users serves another related benefit: You surely know your data quite well – arguably too well from a communications vantage point – and individuals from other sectors may be able to help you find messages that will relate to a broad audience. This website summarizing zip code-level asthma data for California, for example, uses a format – a slideshow – that’s easily accessible to all by presenting findings in bite-size chunks. Moreover, the slideshow doesn’t assume the reader is fully steeped in the topic; the first slides, after all, answer some basic questions – what is asthma, how is it measured, and how prevalent is it?

By walking in your users’ shoes, you’ll know how best to present data findings in ways that compel your readers to do something. And the most effective way to get these answers is to go out and meet with your audiences and learn about their information needs – before you build, when you have prototypes ready, and after it’s launched (so that you can iterate).

8. Don’t start with the full story: Especially with data, we think in terms of larger reports with lots of charts. But we live in an age when we’re consuming information on the go, often from our cell phones – e.g. while we wait in line for the bus or make breakfast and get the kids ready for school. In such situations, we can’t expect our readers to digest long reports. It’s much better to offer increasingly deeper levels of details. For example, lead with a graph on Twitter (just one graph to entice the reader may suffice). That graph can link to an online fact sheet with the main points, which, in turn, can link to the full report/story.

9. Involve local data ambassadors. In this modern information age, we can’t be reliant on just one organization to both create and disseminate content. For both the creation and dissemination, we need to involve our partners across the local data ecosystem: county and city agencies, school districts, community foundations, advocacy groups and nonprofits, health care providers, elected officials and their staff, and coalitions and collaboratives. They all can pitch in to help ensure a message is delivered effectively to the communities who can take action.  

10. Create a call to action. Data findings often are presented in an FYI way – that is, it’s interesting, perhaps even worthy of forwarding to colleagues, but there isn’t an overt call to action in the data visualization. Think about whether you can include such a call to action in what you’re communicating. For example, could your readers be persuaded to sign up to join a campaign? Could they register to get alerts when these data are updated? Can they send a note to an elected official with targeted data for their region?  Before you publish, think about the ways in which you can create calls to action from the data you’re presenting, then make sure to incorporate these in what you build.  

So communicating data effectively is about far more than designing and building the visualization. In the next blog post, I’ll share how Velir has expanded its data practice to focus on some services described above – audience development, community organizing, communication strategy, etc. Our collective ability to build eye-popping data visuals is certainly on the upswing, and that’s an important development. but Velir is focusing, too, on building our clients’ capacity to communicate data once you’ve created your data da Vinci. To that end, do you have lessons you’ve learned from the data engagement frontier? If so, I’d love to hear what comes to mind for you. Just send me a note at [email protected].

A version of this article appeared on All In: Data for Community Health's blog.