Every collection holds by nature more information and insight not included in any of its members. Finding out the value of such information is, simply put, the very essence of data mining. If you were to prove or disprove, for example, the statement in the title of this post, that the letter “Q” isn’t to be found in any of the US state names, you’d realize a few things: (a) a data set (AKA database) of all the US state names might come in handy; (b) A simple search for the letter “Q” in that database could immediately disprove the statement by showing a single instance of a state name containing the letter “Q”, and if that same search turned up no results – you would prove the statement to be true; (3) figuring out what others letters enjoy the same status as the letter “Q” can be time consuming.
That’s where question velocity comes in.
In business, the term “Question Velocity” is often defined as the amount of time lapsed from the moment you come up with a “need to know” until you have an answer. Question velocity can sometimes take into account not only time passed, but other resources necessary to get an answer such as money and computing power.
If you wanted to figure out the answer to section (3) mentioned earlier – you could, if faced with the worst case scenario, have to run/ask as many as 25 different queries on your US state name database (26 if you include the letter “Q”, but I already gave up that one).
The ability to have an extremely high question velocity when it comes to business data exploration is an attribute that is becoming more and more sought after these days for several reasons. The first reason is that datasets are becoming more and more complex and with that the variety of business value carrying questions grows accordingly. In addition, modern business need to re-adjust, re-focus, re-align and verify their conduct much more often than they used to, due to an ever faster shifting and competitive business landscape. And finally, higher question velocity means each question is cheaper to perform and the ROI for asking a question (answer value divided by cost) increases with it.
The time-to-answer for any given query (TTA) is pretty much divided into two sections – defining the question and performing the query. Ergo, if you wanted to cut down on your TTA you could either get a faster computer (or algorithm), or find a way to cut down on the time it takes to articulate and pose a question and submit it to your DB. Personally, I specialize professionally in the latter and I use natural language analytics to do so.
Finding out true business insight is very like looking for that elusive “letter Q” fact. It takes wits, intuition, experience, and if you’re a methodical type business executive – a high question velocity data exploration tool.
- Documentation Is a Map: Learn to Stay on Course September 27, 2016
- You call yourself a Database Expert? Know your history, dude. August 30, 2016
- Did you know? August 1, 2016