It's All About Visibility
Every kid loves to peek through a cardboard tube. The world
looks different and bizarre. There's no context — nothing
but a tiny, limited circle of reality.
One would think that software product designers would remember
this experience from childhood — namely, that an environment is very difficult
to navigate or comprehend without context to provide clues and
peripheral structure. The more context there is, the better the navigation.
|
There's a technical reason why most data analysis
products don't do what BIQ does. It's because what BIQ does
is very expensive.
|
|
So why do data analysis products promote the notion of
"drilling down" on a single dimension at a time?
Suppose you wanted to see IBM and HP spending by the top
three cost centers, for the months of May and June. In a
typical product, you would perform the following operations:
1) Make the Supplier dimension visible ("pivot" to Supplier)
2) Drill on "IBM" and "HP"
3) Make the Cost Center dimension visible ("pivot" to Cost Center)
4) Drill on the three top cost centers
5) Make the Time dimension visible ("pivot" to Time)
6) Drill on "May" and "June"
Now, what if we wanted to see how this spending split across the three cost centers? We would have to:
7) Make the Cost Center dimension visible again ("pivot" to Cost Center).
Wait a minute. What was the split between May and June again?
8) Make the Time dimension visible again ("pivot" to Time)
Wait a minute. How does this break down between IBM and HP?
9) Make the Supplier dimension visible again ("pivot" to Supplier)
One can almost visualize the cardboard tube moving back
and forth through the data — in this case, "pivoting"
through the data. It's ridiculous. What is the purpose of this obfuscation?
With BIQ, you simply:
1) Drill on "IBM" and "HP"
2) Drill on the top three cost centers
3) Drill on "May" and "June"
Every dimension is visible at all times, at every step.
Every dimension adjusts itself to every drill. You can
position the key dimensions wherever you want — in a grid,
on top of the grid, wherever. Visibility is key.
Of course, there's a technical reason why most data analysis
products don't do what BIQ does. It's because what BIQ does
is very expensive in terms of machine resources. Instead of a single OLAP query on a
single dimension, BIQ performs that OLAP query on every
dimension, every time you drill. If there are ten
dimensions, BIQ works ten times harder.
We think it's worth doing the extra work, so we built BIQ
to be as fast as possible. BIQ finishes ten drills in the
time that others finish one. We're so fast that you won't
notice — but you'll sure notice the visibility difference.
We've heard various black beret-clad, octagonal sunglasses-wearing
"user interface designers" argue that high visibility solutions are
"too difficult" for users to understand. "There's too much data
on the screen," they argue. These people don't work for BIQ, and
there's a reason — we think that point of view is nonsense. Smart
users of data want to see more, not less — that's why we
allow the BIQ user to "pop out"
data blocks so that even more information can be made visible. If you're
afraid of numbers, BIQ is not your cup of tea. But if you want
to figure out what's going on, we think it will be.
