1 - Create a compelling business case
One of the major advantages of predictive analytics is that it has a tangible, measurable impact on your business - already in a short timeframe. By predicting customer behavior and leveraging this knowledge in each interaction with customers, you can achieve improvements quickly. Typical benefits that can be achieved are reduced costs, increased profits, and enhanced sales results.

Best practice : calculate existing costs and returns on a campaign, and define the contribution margin and profitability per customer. Based on similar predictive analytics implementations, the potential improvements on the current results can be calculated and demonstrated.

2 - Use test groups to validate the impact
The best way to demonstrate the business impact is to show what the results would have been without using predictive analytics.

Best practice : apply analytics to a small part of your customers and work with a test group to demonstrate how much better your business performs in that group.

3 - Get senior management buy-in
Deploying predictive analytics may significantly affect the way you do business. For example, if you are able to predict individual customer needs, this allows you to determine the best product or service to offer to a customer at any time - truly customer-centric marketing. Being customer-centric has many advantages: reduced churn, improved customer loyalty, and significantly increased sales. However, most organizations are product-centric: driven by product sales targets and basically selecting the best customers for a product campaign rather than selecting the best product offer for each customer. Making such a change requires senior management support.

Best practice : involve senior management in your predictive analytics project, illustrate how it fits in their business vision (e.g. moving towards a customer-centric company) and stress the quick benefits. Agree to report back to them in a few months to show the short term benefits.

4 - Get end user buy-in
However good your predictions will be, their success also depends on the willingness of your customer-facing employees to use these recommendations and act upon them in real time. For example, predictive analytics can be used to create targeted cross-sell offers in a service call center. Achieving optimal results depends on the skills and ability of these agents to bring these recommendations to the customer.

Best practice : start with a small team, train and support them well, and then extend to a larger part of the organization. Again, show the good commercial results that are achieved by them.

5 - Step by step implementation
Although predictive analytics can be applied to many of your business processes, it does allow for an incremental implementation approach. It is better to follow a phased approach, implementing and expanding analytics per business goal or need (e.g. reduce waste on marketing campaigns first, then increase cross-selling in the call center), rather than following a ‘big bang approach’ in which you will try to do it all at once.

Best practice : go for the low hanging fruit. Select a first, smaller-scale project, which will have a limited impact on resources but a large impact on the business. Then roll it out further across the enterprise.