I saw this little post over on Angoss’ blog: The Score – Utilizing predictive analytics for marketing in 2009. In it they identify three areas in marketing where you could put predictive analytics to work. Each one struck me as deserving of a comment on the value of decision management in that context:
Marketing more optimally means you can market less
> Segment prospects and effectively market to each
Segmentation can help you do this but only if your marketing supports different decisions for different segments. If you assume one marketing decision for everyone (sending the same newsletter, for instance), then knowing how to segment your customers won’t help. Focus on the decisions that impact your prospects to use the segmentation.
Sifting through prospects and finding high-conversion ones means you will spend less
>Uncover sales patterns in historical data
Again, only if the decisions you make about who to try and convert, where to spend time/money converting customers can be tailored based on the patterns. If decisions about resources for conversion cannot be tailored based on the sales pattern analytics then knowing your sales patterns won’t help.
Retaining customers means less marketing to existing customers
>Gain insight into customer purchasing trends and maintain optimal contact with customers
When customers call to cancel or when you become concerned that they are at risk you need to be able to make a custom decision or this won’t help either.
Decisions matter if you are to apply analytics. I am speaking and teaching at next week’s Predictive Analytics World so, if this kind of thing interests you, come along to the lovely Hotel Nikko in San Francisco and hear all about it.
Comments on this entry are closed.
James,
I think I disagree with part of what you said. You said “If you assume one marketing decision for everyone (sending the same newsletter, for instance), then knowing how to segment your customers won’t help.”
I argue that analytics can help you know to whom to send and not send the offer. If a big credit card company has the data they can create an analytics profile for the kind of prospect that is more likely to respond to their offer. Then they filter a list of names using the score so that only the more likely people are sent the solicitation.
Doing so means they can send fewer mail pieces (lower costs) and achieve a higher response rate (if the analytics are accurate). I’d call this a kind of “marketing ROI” or perhaps decision-yield improvement. Perhaps this is what you intended with your thoughts on segmentation, but I took it to mean different offers by segment.
Cheers,
mee
Mark
I guess that using analytics to decide yes/no on a specific offer would add value but it seems a small return on the analytics investment. Unless you at least have the control to send one of several offers based on the analytics I would argue your ROI will be small
JT