How to create a more agile analytics program
The rate of change in the business world is increasing. This is not news.
A corollary is that it won't be enough to "do big data". You're going to have to do it quickly, and agility needs to be a defining characteristic of your overall analytics program.
On Datanami, Alex Woodie lays out five steps toward creating that kind of agile program.
The step with the catchiest name is number two: "Abandon Ye Olde Data Warehousing Notions". Woodie explains that old-timey collection and normalization processes have to go by the wayside in the prevailing culture of immediacy. He does note that some of the technologies associated with warehousing will still have a role to play.
But the subsequent steps speak more directly to the article's headline. Woodie prescribes flexibility in data schema and workflows. In particular he advocates for a "schema-on-read" approach, as not all the potentially useful schema for a given data set may be fully understood at the time the data is loaded.
He also points out that data from outside sources may be "faster" than information you can gather yourself. An agile program focuses on data from whichever sensors or sources can provide it most quickly.
- read Datanami's article
More on business intelligence and big data: