Predictive Marketing: Is your daughter pregnant?

This article is a review of session 16105 at the Adobe Digital Marketing EMEA 2012 Summit, titled ‘Predictive marketing: No PhD in statistics required!’.

Speakers:
Bill Gassman, Research Director, Gartner
Katrine Kieldsen, Consultant, Adobe
John Bates, Product Manager – Predictive Analytics, Adobe

Is your daughter pregnant?

Target, the US retailer, probably has a better idea than you do.

They set out to increase revenues by going after a segment of people buying a category of products for the first time and willing to spend lots to get them. They identified pregnant women as the perfect group.

Target needed build a model of the shopping habits in this group of people before the birth – before they started receiving catalogues through the mail from other baby goods providers who had got the details from the birth register. In order to do this they used “predictive analysis” to answer the question: Are you pregnant?

By looking at historical purchasing data of customers known to be pregnant (because they bought baby related products) they identified a pregnancy prediction score based on new shopping behaviours (e.g. buying calcium tables for the first time) and/or a change in shopping behaviours (e.g. switching from regular to additive free body lotion).

Through this analysis Target was the ability to predict with a high degree of accuracy if a women was pregnant, just two weeks after they became pregnant. The amazing thing was that sometimes the women wouldn’t even know they were pregnant and yet Target would!

The result – an uplift in sales and customer acquisition by covertly targeting these people – for example placing nappies next to additive free body lotion on the store shelves.

Be pro-active

Predictive marketing is about being pro-active. Rather than simply reporting on the outcome of your campaigns, it allows you to see into the future, and act before the event happens.

For example, rather than waiting for a customer to cancel their contract with you, why not use predictive marketing to identify those people most likely to cancel and attempt to re-engage them?

Focus on retention – it’s far cheaper than acquisition.

Another example given in the session was a client who wanted to improve their website sales on Black Friday and Cyber Monday. The past two years of trended data forecasted a decline on those days. Rather than simply accepting this, the company used the analysis to re-distribute their marketing channel budgets ahead of time, resulting in a record breaking year.

Key takeaways

  • Look at historical data. It gives you you a clue about what is going to happen in the future. Then take appropriate action to change the future to your liking.
  • Have a little faith. Whilst you can’t trust the conclusions of predictive analysis 100%, a little work in getting close is most often better than doing nothing at all.
  • Copy the traditional retailers. Offline businesses have been using predictive modeling for years. It’s time the online world caught up.

1 comment en "Predictive Marketing: Is your daughter pregnant?"

Leave a comment

Name
E-mail (not displayed)
Web page
Comments