Want to do analytics on large data volumes?
An then we ask: So... How is your data?
How is it organized? How is the quality, what is the level of integration, standardization and how is it related to your well described processes?
The point is obvious, in my opinion there is no use trying to get information out of unmanaged data. When you can’t tell the level of quality of your data it is impossible to say something about the quality of the analytics results.
As Thomas Redman put it in his book “Data Driven, Profiting from your most important business asset”: “We have not even begun to understand the potential for analytics and data mining. Yet it’s reputation may be sullied, in some companies anyway, by half-hearted efforts that don’t produce extraordinary results, just as it is generally considered unwise to put in only enough energy to leap halfway across a stream, so too with analytics and data mining.”
I think this means that when we wanted to follow the Big Data hype too fast and start running analysis software on large volumes of unmanaged data, the results will be disappointing and the hype will pass by for there are few people harder to convince than disappointed business managers.
Boring how it may seem to some, organization, standardisation, description, in short management of your data is the only way to go.
If management is hard to convince of the proposed priorities, would it not help to make the value of the data and the potential value of the information that can be made, visible, to make the case?
On the London Data Management Conference I will be glad to discuss this with you.