As I talk to people about utilising Endellion’s services to leverage the text and numerical data within their businesses, I often hear stories of their previous experiences with data analytics.
A lot of the time I get told first or second hand tales of requests being made of consultants to examine data only to be shown things, based on their years of experience, they can already interpret from the data themselves.
After sitting down and thinking about this I have come to the conclusion that this is a communication issue between the analyst and the client. If you outsource your analytics (and lets face it many companies are starting to), there’s an obligation for the consultant and the client to discuss expectations. At this point the client can tell the consultant the things they already know. Regular meetings can also make sure that the consultant is heading in the right direction and not making conclusions that are just not correct.
With analytics, taking advantage of someone else’s experience is extremely important to achieving a practical and useful result. Unless an analyst is a specialist in a particular field then how are they to understand the nuances behind the data, or what is or is not appropriate?
So, if someone already knows the answer, why hire an analyst? I think intuitively we know we are leaving insights on the table. We generate huge volumes of data on a day-to-day basis, and have an uneasy feeling we are not leveraging it properly.
We can easily establish the first order relationships between data, but what about the underlying relationships we don’t yet understand?
The problem we have with data is we only have the things we measure to work with. There’s only so many x, y (with the occasional z axis) plots we can generate and relationships we can establish with the information we measure. But what about the information we don’t measure?
I was recently told by a friend about an experience they had in the mining industry in the early 1980’s. He worked on an optimisation project where they were asked to determine why a piece of equipment in one mine in South America was outperforming a similar piece of equipment in Australia. They were measuring all of the same things but the data could not help them determine the difference between them.
It turns out the operators at the mine in South America had tuned their hearing to detect a change in operating performance and would adjust the operating parameters of the equipment to achieve the optimum performance based on what they heard.
So the answer wasn’t discovered by analysing conventional measured data. It took someone to ask a question or read shift reports to establish the reason.
My point is, unless we use different sources of data we won’t see anything except for the first order relationships we already know.
At Endellion we collide text and numerical data to discover underlying relationships. We can leverage your existing data through these Data Collisions so you don’t leave anything on the table.