analytics methodsGiven the interim results of our recent survey, I thought it might be helpful to share some content; on less common analytics methods & applications.

It was striking, when looking at those early results, that certain uses of analytics methods were rarer than expected. So, using that as my initial challenge, I’ve been trawling the wealth of analytics blogs (to see who is sharing relevant content that might help).

Here are some of the under utilised methods & applications of analytics, that appear to deserve more of a hearing.

In the same order as our survey, the first topic is consumer segmentation. Interim results showed less use of attitudinal & geo-demographic segmentations. This short post by Uzra Edery on Merkle’s blog reminds us of the benefits these can offer instead.

Segmentation and Customer Strategy | Blog | Merkle

Marketers predict that 2016 will be about “overcoming data and technology challenges to connect with real people.” We live in the Age of the Customer, where connecting with consumers in a meaningful way means doing so on their terms, in their language, wherever they are (online, offline, cross-device), in real time.


It’s become popular, in recent years, to consider that traditional approaches to segmentation should be replaced by a focus on ‘jobs the consumer needs to get done‘. This idea does have some merit and was popularised by articles published in Harvard Business Review. If you too have started down that route, of focussing on the job to be done not any concept of a stable customer segment, here is a great critique as to the pitfalls of such an approach; from leading researcher Kevin Clancy:

Jobs-based market segmentation: not a cure for marketing malpractice

Jobs-Based Market Segmentation Is Not a Remedy to Marketing Malpractice We share the same frustrations as innovation guru Clayton M. Christensen, Intuit’s Cofounder and Chairman Scott Cook, and the Advertising Research Foundation’s Chief Strategy Officer Taddy Hall, the authors of an article in the Harvard Business Review, “Marketing Malpractice: The Cause and the Cure,” with conventional approaches to market segmentation.

Useful cautions, I’m sure you’ll agree.

So, our next question (and results) related to descriptive analytics. Here, the under-utilised methods included univariate analysis, which was a surprise. Here is a useful refresher, on outlier detection, using two different techniques. Handling outliers effectively can add real value to both data quality & your analysis hypothesis:

Outlier Detection with Parametric and Non-Parametric methods

Dealing with Outliers is like searching a needle in a haystack An Outlier is an observation or point that is distant from other observations/points. But, how would you quantify the distance of an observation from other observations to qualify it as an outlier. Outliers are also referred to as observations whose probability to occur is low.

Hopefully that was helpfully specific & something you or your team can action right now.

The next scores, from your interim results, covered behavioural analytics. Focussing more on the application of analytics, it was surprising to see no votes for applying behavioural analysis to product purchase order, channel switching or relationship networks. To help you make sense of the latter, here is a fascinating article from Peter Perera on the benefit of a Master Graph or ‘connectome‘ to explore your customer networks:

How to Expand Your 360-Degree view of customers and their data | SmartData Collective

Today’s CRM and Master Data Management (MDM) technologies don’t enable complete customer knowledge. In fact, they unwittingly turn customer focus into customer tunnel vision. We need an epistemic graph database – a context-aware master graph – to make possible richer, fuller customer stories and expanded 360-degree views for total awareness.

That was well worth a read & inspires me to try new approaches with clients in future.

Last, in my search of relevant blog content, is that related to our last survey question. That was on predictive analytics. The surprise, in votes for this question (so far), is the lack of using machine learning & other forecasting techniques. Not as surprising, but still a missed opportunity, is no usage of survival analysis. In case you’ve not come across this econometric method previously, here is an introductory video from Venko Rao explaining the basics:


No Description

I hope those resources are helpful. Perhaps they will encourage you to expand your toolbox & try new analytics methods. What could you or your team do differently this week? Which of the above articles resonates with a business challenge you are facing right now?

Best wishes for an analytically productive week.