Why all profiling is not created equally

An influx of generic profiling systems in New Zealand is great for businesses who want to scratch the surface of business analytics, but can the data be relied upon using such broad categorisations and pre-determined segments? A tailored profiling system will always be more specific to your business, customers and campaigns, says New Zealand’s leading datamining organisation, Datamine.

“One size doesn’t usually fit all, despite what the labels say,” explains Paul O’Connor, Datamine’s founder. “The same applies to profiling in the data world – a generic system that uses predetermined segments won’t ever fit a business’ needs as well as one designed to meet the specific needs of the business. It’s a bit like grabbing any old suit off the rack and anticipating it will fit you well.”

Generic segmentation profiles are static and based on household characteristics, usually demographic information. The segments offer a generalised snapshot and targeting is based on the assumption that the people within those segments fit the general profile of the target group.

“Already established external data sources are often gathered for a different purpose and can have little specific relevance to many businesses and their customers,” explains O’Connor. “While restaurant habits and car make/model offer intriguing anecdotes, what must be top of mind is the relevance of such information fields to your specific product/service offering and of course the application of it.”

One such generic profiling tool has just 11 general segments for the entire NZ population of 4,000,000 people!

Datamine's profiling and acquisition modeling in contrast is derived from customers own data and external sources such as the 2006 Census. Datamine determines which variables are statistically significant for the specific population group in question, looking at your existing customers and their specific regions.

“While knowing how many bedrooms a customer has may seem an interesting variable, if it is not statistically significant to the population, it is about as much good as knowing what each customer's preferred brand of toilet tissue is,” says O’Connor.

Datamine is also able to lend behavioural insight to the picture. This is especially valuable for retailers; instead of just speculating what people might do, what they might spend, or where your specific customers might come from, actual behaviour gives real validity to the profiles.

Modelling and profiling common characteristics of customers is part art, part science. Datamine have a dozen or so gifted statistician types with uncanny abilities to identify rich information within data, using multiple methods that have statistical validity.

“Many of the generic profiling tools have spent a lot of time creating very general profiles for people that make for entertaining reading, and could make a really useful guide as to where to put leaflets if you wanted to mass mail in a shotgun approach and weren’t worried about the cost or return,” says Sonya Crosby, Datamine’s managing director.

Datamine’s approach is to include datamining as a long-term strategy. This changes the focus from being a one-off mail-out, to a programme of activity that optimises the use of targeting across the business and is based on generating positive returns to the business. The long-term return on investment from using datamining as a strategic tool is the key focus.

Datamine use business intelligence to identify and address business challenges. They help New Zealand companies wade through and interpret large volumes of data, taking the guesswork out of business decision-making.