10 Tips for Successful Analytics

Tip 1: Identifying the Right Problem; Tip 2: Tip 2: Getting Stakeholders onside

Date: Jun 2, 2010

10 Tips for Building Successful
Predictive Analytics Solutions

Richard Boire, Boire Filler Group

Data mining cycle Tip 1: Identifying the Right Problem

Listen, Listen, Listen

This means the ability to conduct a very sound information-gathering process

Conduct Interviews with Key stakeholders
Gather all information and reports that are pertinent to the issue

Example:
A telecommunication company has experienced 50% growth in their overall acquisition costs in the last 2 years.Their acquisition campaigns have been using direct mail and telemarketing as their key distribution vehicles for about 10 years. This distribution component now comprises 75% of their marketing budget. Attrition among the new customer group within the 1st three months has increased by 50% in the last two years.

Key Information:

Increasing acquisition costs
Direct marketing/telemarketing are their key marketing vehicles
Attrition is increasing

What is the problem?

Identify the "right" customers
We need to identify what 'right' means?
"Right" in this case means that we identify prospects who are both, most likely to respond to an acquisition campaign and remain a customer after 3 months.

The Business Problem/Challenge is: Develop an acquisition model which optimises the likelihood of a prospect becoming a new customer and remaining an active subscriber after 3 months.

Tip 2: Getting Stakeholders onside

Tip 3: Create Quick Wins

Here is the full presentation:

10 tips for Successful Analytics (PDF, 37 slides).