I wish I had a dollar for every time I’ve been approached by a client who said, “I need a dashboard.” (I’d probably have enough for a nice round of golf. ) The impetus for his/her request is not really a need for a dashboard, rather it’s a plea for a more business friendly approach to data analysis and thus an easier way of making informed decisions.
What is a Dashboard?
Dashboards are visualizations of data meant to give the user a snapshot analysis. They are not meant to require any mental aggregations, additional analysis, or even math! Imagine if your car’s dashboard required you to complete an algebraic formula in order to calculate your speed, there would be a lot more speeding tickets! You just want a simple, intuitive, and quick method of assessing your speed. That’s why the gauge is the most prevalent speedometer component in your car’s dashboard.
Where do I start?
My first response, after getting the plea for a dashboard, is to ask for whom the dashboard is intended. Ultimately, the determination for success or failure of a dashboard lays in their hands/minds. If we fail to understand who “they” are, then we are doomed to fail.
Another way to think about it is that a dashboard should provide you the information you need to make a decision in less than 10 seconds, between 11 and 60 seconds is a scorecard, and greater than 60 seconds is a report. I often play the use case game as the company’s CEO. My goal is to obtain a snapshot view in 10 seconds (or less) that allows me to decide whether it’s coffee or phone calls as the day’s first agenda item.
Once we’ve identified our target audience, we next need to understand how they measure their business; meaning what KPI’s and/or metrics are used to benchmark success. If these are too high-level (Net Sales, Revenue, Profit, etc.), then we need to work with the business to list out all the contributing factors or sub-measures. You know you’re getting close when multiple whiteboards have been filled and you’re asking for more paper.
After compiling the list of KPI’s and/or Metrics, we need to identify the top 5-7 for the Dashboard. (That is right. We are NOT putting the full list of 200+ KPI’s and Metrics on the screen at once!) This will and should be the most painful step in the entire process. There is a tendency to fall into the “All of these are important” or “We really need to see all of that at once” mentality, but the reality is that no one will or can look at all 200+ KPI’s and/or metrics. The goal here is to create a set of tiered KPI’s and/or metrics that can be used to relay how the business is operating. The top tier KPI’s and/or metrics show on the main dashboard, second tier to second tier dashboard(s)/scorecard(s), and third tier+ on drill-down, detailed reporting.
Step 4: User Interface Best Practices
The last step is the most important, because if done poorly it has the power to render all of your work up to this point moot. The successful delivery of your information’s message is the most critical element of all dashboards. Showing too much information will only obfuscate the information, providing too much noise that will muddle your message. Additionally, avoid over highlighting, the Pie Chart from Hell, Eye Charts, and the Million Gauge March. They don’t communicate anything meaningful to business stakeholders. The goal of your dashboard UI is to deliver your thoughtful and targeted message in 10 seconds. Less is more!
When good graphs go bad… Don’t try this at home. (or the office!)
Where can I go wrong?
Real-Time isn’t always Right-Time
A big movement in the marketplace is towards real-time data feeds. While I see data agility as a differentiator over legacy technology, it’s not always the right answer. It needs to be a clear and high priority to refresh data on a specific interval, but as prioritized by your business against the other priorities. Anything can be accomplished with the right amount of time and money!
The primary goal of the dashboard is to deliver concise information (not data) in a short period of time (<10 seconds). In order to accomplish this, there are some common mistakes to avoid. Avoid using your dashboard as the ETL layer, aggregating, modeling, or joining data. Avoid building dashboards against transactional systems (aka direct-connect dashboards). However, if you’re building an operational dashboard, then we get into a bit of a semantics argument as I would consider that a scorecard. Slim to win!
Aside from sweat pants, I’m not sure if there’s anything that’s one-size-fits-all anymore. In a time where you can order running shoes personalized to your size, color specifications, and monogrammed with your initials, it’s even more important for your BI to be customized to the appropriate audience/person. Don’t fall into the trap of attempting to design something to fit all needs, because if you attempt to please everyone then you’ll end up pleasing none.
Nowadays, each BI platform vendor has multiple tools for reports, scorecards, dashboards, visual analytics, etc. The most important technical decision can often be the tool selection. Just because the vendor’s marketing material says Product X is the dashboard tool, doesn’t make it the appropriate selection for your project. Review the functionality of each tool and make sure that you’re not just a hammer seeking your nail.
Let Axian come to the rescue and help define your BI strategy, develop a roadmap, work with your business community to identify the next project, and provide clarity and direction to a daunting task. For more details about Axian, Inc. and the Business Intelligence practice click here to view our portfolio or email us directly to set up a meeting.