How BA can avoid marketing campaign fights

Westpac uses business analysis to target sections of the customer data base

Westpac has used business analytics to improve the impact of its marketing campaigns, by targeting particular campaigns to the section of the customer base most likely to respond favourably.

In the process, says customer intelligence manager Bonnie Law, the bank can avoid turning off its customers by targeting them with too many campaigns or with campaigns not relevant to their needs. At the same time, the use of analytics minimises infighting among teams, each anxious that its campaign should be pursued first.

Law spoke to an audience of users of the SAS suite of analytical tools at the annual conference of local SAS users in Wellington recently.

Processes for targeting and prioritising marketing campaigns are blunt and unscientific and the campaign that wins out is not the one with the greatest likely value in increasing sales of the chosen product. Rather it is the one whose champion puts the case most forcefully with organisational status to back it up – “the loudest GM voice”, Law believes.

By approaching the process with rigorous statistical analysis, the customer intelligence team can choose non-overlapping sets of customers with a high likelihood of responding positively to a campaign, so more than one campaign can be run concurrently. “Each campaign gets enough relevant customers in its target pool without a fight,” says Law.

Showing the reasons behind the analysis can also break down the stereotypes in marketers’ minds, that, say, all customers over 40 will be interested in term deposits, because they will have funds to spare, having paid off their mortgage. In reality there are a lot of other factors and gradations segmenting the class of over 40-year-olds into those more and less likely to take up the opportunity. “It’s not all black and white,” she says.

Marketers are by now used to using predictive modelling in their work, but “‘optimisation’ and ‘linear programming’ are still big words for them,” she says. Time needs to be put into explaining the concepts.

At the end of the campaign, statistical software tools can be used to measure response rigorously and judge the success of the exercise, Law says.