<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=142641566396183&amp;ev=PageView&amp;noscript=1">

Got Vizzies? 10 tips for nonprofit chief executives considering the dynamic visualization corner of the “Big Data” world.

Blog-0216.jpgLike many nonprofits, BoardSource strives to keep up with the rapidly-evolving world of Big Data, both to increase our own data-based decision-making for continuous improvement internally and to deliver more value to our external community. Recently, this ongoing process has resulted in me becoming immersed in the world of dynamic visualizations — the rapid assembly, display, and interrogation of data in on-screen graphical form for business use — or, as the infant industry calls them, “vizzies.”

Based on my research on platforms, demos from vendors, and simply educating myself, I am confident in saying that many nonprofits could benefit from using vizzies to engage their board members, stakeholders, staff, or funders and better showcase the effects of their work. In an effort to help other nonprofits embarking on a journey similar to ours, I have distilled my “Top 10” lessons learned: 

  1. Know your own niche and capabilities. The market for Big Data tools, platforms, software, hardware, and systems, is huge. Giants like SAS, IBM, Oracle, Deloitte, etc., are out to win billions of procurement dollars from federal Departments of Defense, Health and Human Services, and Agriculture, alone. The new Big Data world is still largely vendor-driven with much hype. You, in contrast, are likely a small fry with limited knowledge and a teeny IT budget. You are likely to be ignored by gargantuan marketing efforts, so you have to do your own research and testing. But don’t worry: you can, and will be better for it. 
  1. Believe in the all-seeing Magic Quadrant. The best side-by-side comparison of vizzie software is the annual “Gartner’s Magic Quadrant” of business intelligence and analytics platforms. Gartner has already done the research for you on how the different OTS packages stack up. Make sure you also read their detailed findings about any individual packages you are considering. 
  1. Know your budget going in. Some of the Magic Quadrant products cost less than $10,000, while others can cost a few hundred thousand. Make sure you get the complete picture on costs — not only for the initial purchase, but for any additional tech support charges. Will there be later version update fees? How many users can be logged on for that price? Will your viewers also require new software to interact with your product, and how much will that cost? Which fees are one-time and which recurring? How long a contract are you being asked to commit to? Are the skills required beyond your current staff? Should you hire another employee, or buy in contractor hours to do the vizzie work? Can you split work between design that might require an expert, and routine data stuffing of templates that support staff could easily do? Can you start with a small investment and scale up capabilities once initial value has been demonstrated? Be prepared to ask for a deal; most sales personnel are empowered to dangle offers. 
  1. Believe in the power of Homebrew speed-dating. Find Meetup groups in your area devoted to analytics and visualization where you can ask your peers their opinions. We have 26 such Meetup groups devoted to data science and visualization in the Greater Washington DC metro area alone. Participating can get you up to speed for free, and most attendees at these gatherings are only too pleased to share advice. 
  1. Beware of the “SPASIDE” fallacy (“Sales People Always Say It Does Everything”). Like buying the family car, you are in control until you hand over the check because they want your money. So before you buy, ask for a test drive. Send them one of your own data files (with any personally identifying information stripped out, of course) — just a simple rectangular Excel file with cases in rows and variables in columns. Ask them to prepare some free vizzies that can be done on your data with their product. And if you have particularly difficult operations that you currently do manually, like calculating degrees of consensus, or making compound indicators, etc., then make sure these can also be handled. Some vendors will let you have a 14-day free trial so your own staff can really get under the hood. The proof of the pudding is in the vizzie, not in the sales spin. (Just be prepared to change your phone number afterwards.) 
  1. Some old adages still ring true. As we used to say in the early days of computing, “garbage in, garbage out.” If your raw data poorly measures the real phenomenon you are interested in, then putting it in a vizzie, no matter how glitzy, is still not going to help with understanding. People can be sucked in by the very artistic nature of these products, but it shouldn’t be all about the vizzie. Instead, everything should start with the business or research question, and let that drive which product you need to see. 
  1. It ain’t over with the installation. There are going to be staff learning, implementation, and embedding curves to the eventual effective and successful use of vizzies. Be prepared to educate your board members and senior executives on how to utilize vizzies in their strategic thinking. Lead them away from “gut” calls and into a more analytic mindset for decision-making. Tell them what your vizzies are showing and what they are NOT showing. Explain the scope, trustworthiness, and value of the underlying data being portrayed in this way. Add some simple sliders or selectors, and then demo those dynamically in the context of a business question. Be the trainer MC! “Before you make that decision, let me show you that it’s notable how this picture changes if we are only looking at small organizations, versus large. I can just move this slider here to display results from only organizations with a budget of less than half a million a year, and you can see how this other outcomes chart really changes”. What other kinds of interesting hypotheses or explanations can you suggest, and then follow-up on with instant drill-downs? The goal is to get them to eagerly look forward to next quarter’s charts to see what has changed since their last meeting and to generate discussions. 
  1. Make your purchase within the context of an enterprise-wide modernization strategy. Extract maximum value for your organization by expanding this capability beyond your research team. The real long-term payoff comes from embedding it in work across the enterprise. What can your membership, marketing, accounting, content, and web folks use it for? Bring them along with you early, and open their eyes to think for themselves about use-opportunities in their own realms. Look for gains in even internal routine processing. For example, in many research shops, probably 75 percent of analyst time goes on “data-wrangling,” that is, everything you need to do to the data before actually being able to operate on it the way you want. If the vizzie package reduces that to 50 percent by finding your errors and outliers graphically, that’s quite a saving. 
  1. Leverage your BoardSource organizational membership. Did you know we have an Ask-The-Expert service for members on the BoardSource website? While we’re far from being experts on vizzies, we’re happy to weigh in on any questions you may be grappling with, just as we have here, or refer you to one of our contacts. 
  1. Remember, at the end of the day, it’s just… data. As Einstein cautioned, not everything important is measurable, and not all that can be measured is important. The data are not the “picture.” The map is not the “terrain.” The numbers aren’t the “story.” It’s hard to make good, sound, reliable pictures and stories that nonprofits need in order to communicate to audiences and stakeholders without having data first, but the data are really only one piece.

 So, Happy Visualizing, Everyone!


 Chris Thompson, Ph.D., is BoardSource’s director of research and evaluation. He is a Big Data addict, but it has been almost a year since his last regression and his statistically-significant other thinks he is recovering well.


Topics: Data Visualization, Dashboard, Mission Impact

Subscribe Here!

Recent Posts