- Data-driven: You learn from existing data and use sophisticated models to predict the future. The underlying data are usually a combination of CRM and increasingly data "mined" from Social Media. Learn more...
- Expert-driven: Here, you aim to embody the soft knowledge of your organization's experts into large simulations of your entire market. Those allow you to test future scenarios risk-free. Learn more...
- Crowd-driven: Finally, the "Wisdom of Crowds" - the knowledge of your entire organization - is tapped in smart ways to achieve accurate predictions even if many participants are no experts at all. Learn more...
Not all of them are equally useful for all kinds of predictive analytics problems or business issues. Their selection depends on what you want to achieve, what inputs are available to you and which timeframe you have to achieve the predictive quality that you desire.
In his 2009 SMR article, Tom Davenport made a similar
distinction but took a more fine-grained lens on the first bucket
that we tend to summarize in our client discussions:
http://sloanreview.mit.edu/the-magazine/2009-winter/50208/the-prediction-loverrs-handbook/
All of these approaches will be discussed in an interactive way and applied to participants' issues directly at our London workshop on November 28. Learn more...
The link to the annual Churn
& Segmentation conference itself:
http://www.segmentationandchurn.com/Event.aspx?id=583936
And the pdf brochure:
