This article is meant as a comment to the following McKinsey Article:
Data Lake: or the challenge of pulling data together in one place.
In my data analytics project most of the effort was putting together all relevant data sources. I was actually forced to build and I am forced to maintain a small subsystem. The effort that goes into this exercise is far from minimal. We then had to make the data talk to each other: accessing data does not equate to having information. You accomplish that task once you understand what you are accessing. Finally we’ve had to cleanse it .. and man that was challenging. Making sure segmentation was accurate. At one time we’ve had as a team to go through almost 40.000 client record to check whether their actual activity was the one we had on record for them.
Now I can safely state that we have one of the cleanest business customers dBs in Italy. We should be actually thinking about selling this information to others who might find it useful.
Translators: the key skill in the game.
You need someone in the organisation that speaks both languages, business and technical. If the same guy also has the required strategic vision as to know what to do and why, it means you really got lucky.