In this post I share a presentation on principles of business analytics.
My colleague Diana Limburg asked me, not along ago, to run a session on business analytics, as part of her module on eBusiness. This is a very popular course, attended by 2nd and 3rd year students. The session on analytics lasts about 45 minutes, and takes place towards the end of the module.
Whenever I teach an introductory session on the topic of metrics or analytics, I follow the approach presented in the book ‘How to Measure Anything’ by Douglas Hubbard. Indeed, if your work involves producing or reading metrics, I warmly recommend that you read this book. It is a really straightforward introduction to measurement methods for various scenarios, using very simple examples.
Hubbard warns that, to decide how to measure something, analysts first need to consider what it is, exactly, that they want to measure. Moreover, they can’t really tackle the ‘what’ unless they have considered the ‘why’. So, to measure anything, analysts need to start with the WHY and the WHAT, before attempting to tackle the HOW.
For instance, if I were asked to identify ‘online influencers’, I would first need to understand what exactly the client means by influence and why s/he needs to identify such influencers. As discussed in this post, influence is not only a function of the number of social connections in the network, but also of the type of network. Some networks make messages travel far, whereas others are very effective to change behaviours.
In this presentation, too, I followed the WHY > WHAT > HOW approach.
I started by discussing the rationale of using analytics (the WHY). Then, I highlighted the areas that eBusinesses should monitor (the WHAT) – namely, investment performance, the processes, and their customers. I also mentioned key analytics for each of these areas (the HOW).
In the end, I addressed a few caveats about the analytical process, noting:
– the limitations of quantitative insight (for more on this, see here)
– the dangers of relying too much on historical data
– the need to understand human behaviour (for more on this, see here).
As usual, I am looking forward to reading your thoughts on the presentation and the approach.