Everyone wants to get more out of their data, but how exactly to do that can leave you scratching your head. Our BI Best Practices demystify the analytics world and empower you with actionable how-to guidance.

Preserving insights

One of the biggest pitfalls in data is the preservation of insights when analysis is handed off from the data team to a business professional. Often, the data experts have been exploring the data for a while, developing a clear sense of its structure, assumptions, and conclusions. The data analyst has had a great opportunity to pinpoint an insight, but when it comes to sharing that work with the business person who will ultimately make a decision, they fail to fully communicate that information. Here are some tips to help you improve your data visualization so that you can add the most value to your teams via a dashboard.

The first step to making the most out of data collaborations is to set up a meeting where you discuss the relevant business questions. I outlined some important guidelines for this meeting in 6 Tips for Data Professionals to Improve Collaboration, so if you haven’t already read that, it’s a good place to start. Data experts should expect to come out of that meeting with a list of questions that they can translate into queries to build the initial dashboard. 

In this post, we’ll cover the next steps in the dashboarding process.  

BI & Analytics for Business Analysts

Work from a blueprint

The first step in building a great dashboard is to review that list of questions and group them into larger buckets. When you read the entire list, think about which themes emerge. When you’re identifying those buckets, you want to look for general topics that are only answered by combining a few of the individual questions. In the ideal scenario, all of your individual questions can be grouped into a handful of broader themes.

Once the questions are grouped into buckets, it’s time to build a blueprint of the dashboard using those general themes as the headers and the individual question as the charts that fit underneath. My personal blueprinting process uses a lot of sticky notes for this. I’ll write down each of the questions on my list, then arrange them into groups on paper until I’m happy with the design. It also helps me to make sketches of the individual charts to get the most complete idea of the final dashboard. 

At this stage, I’m making decisions about what is included in each chart. For example, if there are three line charts next to each other that can be combined into one chart with three lines, this is the step to decide to do that. If I decide to combine charts, then I need to make sure the title and scale of that chart fit all the information appropriately. If I decide to keep three separate line charts, I think how can I format them to show the unique information clearly.

Getting your charts in order

Once all the organizational work is done, I start arranging the charts to match my blueprint. Every question from that original list gets turned into a chart that’s placed on my dashboard according to the blueprint I made in the first part of the process. 

I like to build a dashboard in horizontal layers, with the very top layer being the most important high-level KPIs and then each of the layers below tackling one of the buckets I identified in blueprinting. To help guide users to understand the purpose and context for each section of the dashboard, I often use text as signposts. Additionally, within each chart, I use titles, colors, and other visual cues to help the charts explain themselves. Finally, when deciding on the arrangement of the charts within each section, I start with the chart that will be most frequently referenced on the left and then work my way to the right in order of decreasing frequency of views.

For the additional layers, I like to provide context for those high-level charts. For example, if the most critical chart in the dashboard is a revenue tracker, the layer directly underneath has more charts that answer questions related to revenue. The next layer down contains more detailed information about the second high-level chart, and so on. This design strategy lets that first row of the dashboard act not only as a quick summary of the most important information, it also turns it into a table of contents for the rest of the layers.

In my ideal dashboard, the top layer has one high-level KPI chart to summarize each of the buckets I identified in my initial conversation with the business professional who requested this data. Then, each section underneath that contains answers to all of the related questions we listed that belong to that topic, starting with the information that will need to be referenced the most frequently and working down to the information that will be referenced the least. Data is messy and it doesn’t always fit neatly into those layers, but this mindset makes it easy to compartmentalize and organize a long list of charts.

Readability means insight preservation

The goal of your dashboard isn’t to allow business professionals to easily find answers, it’s to help them find the right answers easily.

An important thing to remember when creating a dashboard is that most of your consumers are busy professionals with their own long list of work priorities and deliverables. When they read your dashboards, they are most likely looking for quick answers to a particular problem. They need to take away the important learnings that you’ve found in the data, but they won’t have the same amount of time to spend studying the data as you did. This attention gap is a place where the insight can erode significantly, so you need to make sure that it’s easy to get the right insights quickly.

One of the best ways to focus a reader’s eye is through the use of color. Using the same color for all the charts related to one topic of your dashboard is a shortcut to making sure all of that information is digested together. The goal of your dashboard isn’t to allow business professionals to easily find answers, it’s to help them find the right answers easily.

It’s always good practice to title the charts as specifically as possible to minimize confusion about the insights. Translating the blueprint you designed to an actual live dashboard will always result in a few unexpected hiccups, so it’s crucial that you review the dashboard as a whole and the topical layers for places where insights could get lost in the handoff back to a business professional. Through iteration, your dashboard will not only have all the right data, but it will also have a form that resonates with the person taking action, leading to higher adoption and higher impact.

BI & Analytics for Business Analysts

Christine Quan is a seasoned data and analytics veteran, focused on data visualization theory and building tools to empower data teams. She’s an expert at constructing SQL queries and building visualizations in R, Python, or Javascript.

Tags: | | |