The more the world of Big Data/novel analytic techniques/machine learning is internalized, the greater the likelihood assumptions move to presumptions and technical terms unwittingly become marketing terms. The Gartner Hype Cycle is a great illustration.
In the context of “predictive analytics,” it’s worth knowing what people actually mean. First, predictions–obviously the term is about the future, something that has not yet happened. If you are able to “predict” GDP growth this quarter, does that mean you are saying what it will be prior to the close of the period or are you reducing the lag between close of period and your claim? The former is predicting, the latter is reporting/measuring. A picture is worth 1,000 words:
When inserting something about analytics into a business-oriented discussion, my inner statistical background grows a little closer to death. Much of what follows in these conversations are graphs and charts that illustrate totals, maybe an average and segment. This info can certainly yield valuable insight, but don’t conflate it with anything analytical. Statistical tests, starting with descriptive tests, can be invaluable, but unfortunately the true meaning seems to have been co-opted.
Do you wildly concur with me? Wholeheartedly disagree? Do you bemoan the state of analytic or embrace its democratization?