Kevin Loudermilk, Director, Auto Data, USAA
Executives, your company’s infrastructure is not the most significant challenge impacting your data visualization program. Of course, you must address the technology, people, processes, and data issues that can, and eventually will, reduce your effectiveness. But once you have done this, you must face the biggest challenge to useful visualizations – you. Your expectations and requirements will ultimately determine how successful your data visualization program is.
The “Easy” Issues
Technology, people, process, and data issues are not always easy to address, but they are the obvious issues that we know about and create plans to resolve. Your company has probably already looked at these, but let’s quickly walk through them, just in case.
Choosing which software best fits into your IT stack and budget can be a difficult task. What business questions do you need to answer, and how will you consume this information? There is no right answer, contrary to what the vendors at a conference will tell you. There are many choices, and all have their strengths and weaknesses. Trust your analytical and IT community to make a good recommendation and stick with it. Changing technology regularly will guarantee failure.
You must provide training for your analysts, not just on the chosen tool(s), but on design, storytelling, data manipulation, and even business ethics. The visualizations they create will drive your business decisions, and you will need clear, unbiased information to make the best choices for your business. Leaving training up to on the job self-development is inadequate. Invest in training your analysts to maximize their effectiveness.
Unless they are addressed early and strategically, data issues can quickly escalate and create rework, or worse, decisions that will negatively impact your business goals. You have to decide what works best for the volume, security, and reliability needs of your company and establish appropriate processes for governing these requirements. Ensure someone is managing the quality of your data, whether in IT or on the business side. Poor data quality leads to poor analysis. Finally, think about having one source of the truth. That means eliminating data replication, defining metrics once for everyone’s use, implementing a version control process, and reconciling reporting across multiple visualizations as appropriate. Data maintenance is complicated, but you cannot afford to neglect it.
The visualizations they create will drive your business decisions, and you will need clear, unbiased information to make the best choices for your business 
The Hard Stuff
I realize that sometimes, technology, people, process, and data requirements are overlooked in the desire to stand up a visualization program quickly. But they are not the most significant challenge out there. You are. Don’t be offended. You didn’t do it on purpose. Examine your expectations and see if you are guilty of any of the following: Instant Information – Do you expect to have instant access to data through a clear, concise, robust, yet straightforward dashboard of your 30 favourite new and yet-to-be-defined metrics? Building useful visualizations take time. Having precise requirements helps, as does an understanding of the business questions you need to answer. Agile development that is incrementally delivering business value as the dashboard is completed and refined is a great way to meet your initial time requirements while providing analysts time to supplement and improve their ultimate deliverable.
Changing Requirements – How often do your key metrics change? Do your executive dashboards require metric swaps every few months? Chasing a tactical metric frustrates your analysts and keeps you from seeing your business from a strategic perspective. Your key metrics should be tied to your executive goals for your business and should not change unless those goals change. Select your parameters, clarify the requirements for them, and don’t change them unless necessary. You need to build history to put the present into context to keep from making knee-jerk reactions to business volatility.
Overly-detailed Data – How deep down the rabbit hole do you want to travel? Do you want to be able to self-serve down to the most granular level of data available in your systems? Do you have that kind of time? Trust your analysts. Let them know the questions you need answered and let them do the massive analysis. It’s what they do and what they love. It will also improve what they can deliver to you. You will get clear, focused visualizations instead of complicated, confusing data dumps. There is a vast difference between visualizations designed for executive reporting and those designed for analytical exploration. You have to look at your business from a higher level, so don’t let yourself get down in the weeds with your expectations for data.
Challenge or Champion?
You will determine the path of your data visualization program through your proactive decisions or lack of direction. Invest in technology, people, processes, and data infrastructure to set your company up for success. After those are in place, your work is not finished. Your expectations can maximize or minimize visualizations. Understand the time requirements necessary to build compelling, sustainable visualizations. Be clear and strategic with your needs. Remember your role and your analysts’ role – don’t switch them. Will you be the biggest challenge or biggest champion to your viz community?
See Also:
