Insight

Why the human ‘last mile’ of analytics can make or break the work of data scientists and their business sponsors

March 10, 2020
Read the full article

Not every result of an algorithm is welcome. I received some digital marketing recently wondering if Mr. Horny would be interested in some potential acquisitions. This, and a similar corruption of my first name, is so frequent that I now claim to have an alter ego called Ron Horny. Ron is not a big buyer of the offers sent to him.

But what if the algorithm had been accurate and told me something potentially valuable that I didn’t know? Who could have an issue with that? My recent experience tells me many people, and quite often.

For most data scientists, creating insights is an exciting and important task. We are increasingly sophisticated in our capabilities, moving from simple data description a decade ago towards prediction and even prescription (explaining how to action an insight, or even doing it automatically).

We have mostly assumed our insights are objective and true. At heart, many of us are engineers and scientists, who behave as if we inhabit a sane and rational universe.

 

Meet the author