Analytics in the Real World Delivery Cycle

Much agree with the general premise below.  But I take it a step forward structurally.   I have given several talks in the last thirty days which have emphasized how analytics in general needs to be connected dynamically with the real world.

I represent that with the circle diagram shown here.   You can start your journey at any point: Data, Decisions ....   You then travel around the circle to deliver, improve, test and further adapt.  Even dynamic delivery becomes important. You know the data will change, so this is a continuous experiment.

Why not include this same kind of structure in the test benches and 'sand boxes' being constructed today?

"Analysis: Unlocking the secrets of data requires real-world experimentation 
With the rise of enterprise Big Data, more businesses are turning to data science to give them new insights into the behavior of markets and customers, but experts say analytics is not just about asking how things happen but why they do. According to commentator James Kobielus, this means parlaying findings into real-world experiments that unlock secrets hidden behind the numbers ....  "