In my last post, I spoke about the importance of actionable data to a business intelligence dashboard for your company. In this post, I’ll address another key aspect that’s needed for a successful dashboard – a context for the displayed data.
In order for your dashboard metrics to make sense to your users, they must have a context. A context for the piece of data can be a date range, a year, a department, as well as additional items like what the goal number should be, what last year’s number was, etc.
For instance, in the dial displayed at the left, you can see that the number being represented is around 124%. Suppose that the title of this gauge is “Revenue”. So, we can ascertain that revenues are now at 124%. But, 124% of what? Compared to what? Even if the title of the gauge was “2010 vs. 2009 Revenues”, we still don’t know what the goal was – was our goal to be at 140% of revenues versus last year, in which case we’re falling short? Is this gauge showing revenues on an annualized basis, or total revenues (if so, can we sit around and do nothing for the rest of the year)?
Also, what context or information is gained by this gauge being so large? Well over 50% of the area of this graphic is whitespace, and having the large areas of color, not to mention the fancy border gain us nothing in terms of knowledge discovery (part of the problem with this graph is that it’s trying to follow the broken metaphor of a car dashboard too closely – a problem I’ll discuss in a later post).
We could make this graph much better by doing the following things to adjust the context of the data:
1. Add the relevant time period to the graph name
2. Add a marker that shows last year’s number (technically 100% for this graph, but let’s just say it’s in dollars instead of percent, since it lets people know more easily how many $$$ they need to find)
3. Add a marker that shows the goal
4. Get rid of wasted space so that more information can be displayed on one screen
One type of graph that does an excellent job at all of these is the bullet chart, which was developed by Stephen Few. This graph type condenses all of this relevant information into the smallest possible space, and is very easy for users to understand. For instance, in the graph to the right, it is clear to the user that the current revenue amount is around 275, and that the goal was around 245. The user can also see from the shading on the background of the graph that there are great, good, and bad ranges of numbers (0-100 = bad, 100-200 = good, 200-300 = great). While it may take a small amount of training to understand what those ranges mean, it is relatively intuitive, and takes up much less space than the gauge above, and displays more information – making it a very effective dashboard tool.
