Five Rules of Data Clarity
Updated: Nov 3, 2022
As early as 2006 Clive Humby, a UK-based Mathematician, preached that “Data is the new oil.” It seeps into every aspect of daily life, in work it’s a turnover calculation for candidate outreach, interpersonally it determines who’s on the front page of dating apps, and it even intersects with relaxation as streaming services build out more and more complex recommendation systems.
Nearly two decades later, there is still a foundational disconnect between data practitioners and business leadership. Be it culture, data literacy, or sheer logistic complexity, an organization that doesn’t capitalize on its data is like a ship without a rudder. Because of this disconnect, it becomes the role of data workers to take steps to decrease the barrier to entry to data-based decision-making. What follows are five simple rules to make data accessible whatever the audience.
1st - Like most communication, there's a narrative arch to displaying data. It's far too common to display every drop of information in one place. Instinctually this is efficient, saves storage space, and makes analysis easier. However, it can also overwhelm non-technical audiences resulting in information overload and disinterest due to complexity. What sales manager wants to spend 30 minutes digging through a dashboard to find five clients when they could simply be making calls already? Instead, the role of the analyst/architect is to streamline and simplify that information flow. In this case, let’s borrow the first rule of public speaking: know your audience. Presenting information to sales, giving them top prospects in a dynamic text table. Presenting to accounting, display outliers and averages first. Presenting to executives, a selection of KPI cards with month-over-month analysis will keep them engaged far longer and far more meaningfully than any spreadsheet. Simply put, whatever you show, make it relevant to the people seeing it.
2nd - Continuing with the idea of storytelling. Make your message clear, most business intelligence software allows you to create dynamic titles so you can make a title that tells the story. For a quick example, what gives more information "Monthly sales report" or "Monthly sales have increased by X% since Y." People remember stories so analysts should prioritize narrative flow and simplicity over showing everything. Taking this a step further can result in sentence-based KPI cards, more fluid tooltips, and a more conversational feel. The result of these seemingly meaningless edits is a more humanized experience that is much more likely to get people involved.
3rd - Moving into more practical suggestions charts and graphs should be simple. For clarity, this doesn’t mean content, but a more visual focus. Remove gridlines, only mark important points (min, max, etc), and remove pointless titles. An incredibly useful practice is asking about the purpose of each element. This could be written in a spreadsheet or with a colleague, but make sure there is a justification for each element present no matter how minor. This decluttering has a few immediate effects, it naturally generates more white space for a cleaner look, creates nature prioritization for users by giving them less to get lost in, and reinforces the idea of the narrative discussed above.
4th - In a very literal sense contrast creates focus. If most of the screen is white, a brightly colored section will immediately draw attention. This allows analysts to control what is prioritized be that KPI cards or a text table. Alternatively, this is even more useful on the micro level. Decreasing opacity for all but the most impact trends on a line cart will make the story clearer. Avoiding too much of the theoretical, most people will scan the screen left to right so you can either go with that flow and put important information on the left so it’s processed first, or you can make important information stand out from the rest of the presentation. This same trick applies to most communication, but for crystallization try bolding the most important words on a slide.
5th - For those building several dashboards (which should be everyone considering rule one), establish a templated approach to building dashboards. Keep titles, fonts, spacing, etc. as consistent as possible across the organization or brand. There are two primary purposes for this. For users who might access multiple dashboards, it immediately establishes a sense of familiarity and eases exploration because they are used to processing your style, and for creators, it reduces build time by reusing resources from previous builds. By templating there's a natural consistency that's established and that will eventually evolve into a personal style that users come to expect.
1. Make all analysis contextual. Who are you talking to?
2. Keep it conversational. No one wants to parse a spreadsheet.
3. Keep it clean. Don’t let users get lost in grid lines.
4. Where should their eyes go? Establish visual priority.
5. Standardize what you can. Familiarity is far more appealing than innovation.
While this sounds simple, complexity tends to be inevitable in a world dominated by exponentially growing technology, so it becomes the role of data stewards to prioritize accessibility. If nothing else, these two themes can improve data coms considerably. Keep it simple and make it intentional. These should be the guiding light during dashboard building, slide crafting, or presentations.