1 dimension + n metrics
Compares each values of a dimension on several metric. Effective to discover if metrics are corelated or for detecting exceptions namely if a dimension value is remarkably high or low because it outperforms or underperforms on a certain metric.
A simple example
You want to compare quality of visits per browser.
Try it with your own data on the Visualization Playground
Let's analyze a bit
Clearly, Chrome is by far the most popular browser used by your visitors.
But looking at the most used browsers lines, it appears that more visits leads to a higher bounce rate, less time spent and fewer pages viewed per visit.
On the contrary, fewer visits apparently leads to a better bounce rate. Look at how all these lines have the same shape, low on the visits axis and high on the bounce rate axis! (higher meaning better)
The Android Browser for example, has only been used 62 times but has the best bounce rate, almost the best time spent, and a high number of pages viewed per visit.