Analytics – pluralist, but not a plural

(Some nit-picking on terminology.)

The term analytics means different things to different people. The term goes back to Aristotle’s Prior and Posterior Analytics, which is out of copyright and on Project Gutenberg, who knew. Francis Bacon refers to it too. Wikipedia says that in the current sense it means something like “the discovery, interpretation, and communication of meaningful patterns in data”, which is fair enough, and it’s good to explicitly mention maths-n-stats and computing too, even if in principle they’re only implementation details.

Some people take a narrow view of analytics, that it’s basically statistics applied to business, but I and the UCD MSc in Business Analytics take a broad, pluralist, catholic view, including pretty much all of traditional operations research, such as exact and heuristic optimisation, and simulation methods; broader than just quantitative methods, e.g. including decision science including multi-criteria decision-making, behavioural approaches, and so on as part of analytics too.

For short, I like to say that analytics is about making decisions – better. That dramatic em-dash is my way of distinguishing myself from Davenport, who rightly reduced it to this essence a long time ago in “Make better decisions”.

Regardless of your favourite definition, there’s one bone I have to pick with Davenport, and that is his use of analytics as a plural. For example, in Harris and Davenport, Competing on Analytics 2nd edition, writing about HMH, they say that before HMH’s analytics transformation, there were “few analytics” to inform and motivate employees and management. (It’s not a once-off or a typo.)

Now gives: “-ics: a suffix of nouns that denote a body of facts, knowledge, principles, etc., usually corresponding to adjectives ending in -ic, or -ical: ethics; physics; politics; tactics.”

Further examples include mathematics and economics. Analytics fits in this pattern. It’s a singular noun, just like all the rest. We don’t say economics are so difficult, or I used a mathematic earlier today in my everyday life (not just because is that even possible), and nor should we say analytics help managers to make good decisions or which is your favourite analytic?. I think people who say these things are slightly conflating analytics with metrics. Analytics is not just metrics. (Metrics is a genuine plural, and I won’t go into the distinction between metrics as used in business (more like some kind of generalisation of norms, if anything!) and metrics in the mathematical sense.)

How does analytics differ from analysis? (Whatever you do, don’t confuse business analysis with business analytics!) Well first, what is analysis? Analysis means understanding something by taking it apart, in some ways the opposite of synthesis. From Oxford dictionary: Analysis (noun): detailed examination of the elements or structure of something … late 16th century: via medieval Latin from Greek analusis, from analuein ‘unloose’, from ana- ‘up’ + luein ‘loosen’. One definition, which is at least internally coherent and coherent with terms like economics and mathematics, is that analytics is the study of methods of analysis.