Eric's space

With a little practice, it's quite easy to create professional-looking charts and graphs using the BI tool of your choice.  It's another matter entirely, however, to create a chart that utilizes proper design principles and answers the business questions relevant to the information consumer.  In this post I'll walk you through a chart-design scenario, highlighting various decision points and considerations at each step.

For this example, let's say we have a request to create one or more visualizations around monthly sales, including a breakdown by salesperson. 

The first visualization we might use isn't a chart at all, rather it is tabular data with graphical indicators.  This is better known as a scorecard, and in this case, we'll create a scorecard with a Sales ($) KPI.  KPI stands for Key Performance Indicator, and within the context of a BI scorecard it is simply a measure or metric with a defined target or goal value.  (A more formal definition of a KPI exists in the business strategy domain, in which a KPI represents the measurement of particular strategic business objective or initiative.)  Our very basic scorecard with the single Sales ($) KPI might look like this:

Figure 1: Sales Scorecard, Sales KPI

So what is this scorecard telling us?  First, let's examine the layout and content of the scorecard to better understand that data we're presenting.  We have the Sales ($) KPI with six numeric values.  The first column is "Actual."  This is the actual sales ($) value for the month of Dec-09. That is followed by a target (remember our definition of KPI above), in this case, "Plan."  This is our sales target for the month.  Next we have the variance of actual and plan as a percentage.  Here is where we introduce the visual stoplight indicator which communicates KPI performance at-a-glance.  We've established thresholds of performance for this target (>100% is on target (green), >90% is slightly off target (yellow) and <90% is off target (green)).  Next we have the measure value for "Last Yr" which represents performance for this period last year.  Finally, we have "Last Yr Var %" which represents current performance compared to performance last year as a percentage.

What insight can be gained from the data in the scorecard?  The up arrow indicator draws our attention to our performance versus last year, which is good (up 6%), and the yellow indicator draws our attention to performance versus plan, where we're missing (only 90% of target).  What's the next question that our information consumer might ask?  Probably something like, "which salesperson had a particularly good/bad month?" or "which products were strong/week this month?"  We'll answer that question with one or more charts...

Let's focus on the salesperson perspective.  So we know that our total sales for the month were $90,000.  How was that distributed across our team of five salespeople? Here's a chart that answers that question:

Figure 2: Sales by Salesperson

Displaying the data graphically, sorted from highest to lowest, illustrates very quickly our top and bottom performers as well as a sense of the variation across the team.  Can we enhance this chart to to provide additional insight?  Looking back at our scorecard example, we know that adding context often enhances insight.  Look at this next chart. 

Figure 3: Sales and Plan by Salesperson

We've added the sales plan measure to enable visual comparison of each salesperson's actual to plan performance.  We've preserved the insight around top/bottom performers, the variation/range of data, while adding context with the plan measure to facilitate yet another level of insight.  It's useful to know that Mary was the top salesperson for the month, but even more useful to know that she exceeded her plan.  On the other hand, Rosa was the bottom salesperson, and she missed her plan.

With this example we've illustrated how basic charts, simply designed, reveal multiple insights in a common business scenario.  In the next post, we'll look at charting some calculated measures to further enhance the message in the data....

This week I'm attending the TDWI World Conference in Las Vegas, NV.  In this post I'll share some of my key takeaways from the Tuesday morning session I attended on the topic of developing a DW/BI Strategy.  The course was led by instructor, Michael L. Gonzales, from Claraview.

Here's the course abstract:

I found an interesting article this morning on User-Centric Dashboard Design. The author describes his approach this way:

"The user-centric approach requires a keen understanding of the user’s needs and the ability to design around them. This can be reduced to two fundamental questions:

• What decisions or insights does the user hope to make by interacting with the dashboard?
• What data is needed and how does it need to be presented to answer those questions?"

I like it.  Simple, but effective.  Focusing on user needs helps to ensure that we build the right solution.  Too often, we (those of us in the business of designing and building Business Intelligence solutions) focus on the details of the data at the expense of user needs.  It's easier to build a solution by ignoring users and instead focusing on the data.  "Show me your data and I'll build you a solution."  After all, those pesky users don't really know what they want, and when they finally tell us, they'll just change their mind later, right?  To mitigate that risk, we often include everything we can think of in the solution and hope something sticks.  Not very efficient, huh?  Taking the approach to focus on user needs takes some additional investment up front, but the payoff is in a more focused data development effort, and more satisfied users who are engaged throughout the development process and beyond.

At BlueGranite, we take the user-centric dashboard design concept one step further.  We talk with our customers about the process of moving "from insight to action."  It starts with an effective dashboard design that facilitates understanding, and continues with the support structures of a collaboration tool which allows individuals create and implement a plan of action.  Once you've gained valuable insight into the root cause of your business performance, what are you going to do about it?  What are the possible courses of action?  Once you've made a decision, how do you implement the change?  Who needs to be informed?  Who should be on the team to implement the change?  How will the change initiative be measured?  These are just a few of the many considerations one encounters when "completing the loop" in moving from insight to action.  The recipe for success is information plus collaboration; your BI solution should do both.

There are a number of vendors and toolsets available to the IT professional tasked with creating BI applications for organizations.  At BlueGranite, we've developed expertise in delivering solutions with the Microsoft BI platform or "stack."  We find the Microsoft tools to be both powerful and easy-to-use, and the price to value equation is very compelling for most organizations.  The guys over at Microsoft BI Blog linked to the latest Gartner report which positions Microsoft in the leaders quadrant, ahead of many of the other major vendors such as IBM/Cognos, and SAP/Business Objects, specifically in the "ability to execute" area.

The graphic below illustrates the primary components in the Microsoft stack.  As you can see, it includes three widely-adopted Microsoft Server technologies (SharePoint, Office, and SQL Server), with the addition of the new flagship BI product, PerformancePoint.  The Microsoft BI public site explains the capabilities of each of these products in more detail.

The purpose of presenting this overview is to provide a frame of reference for subsequent posts where I'll explore the capabilities of these products in producing effective data visualizations.