Understanding the Data
A thorough collection and exploration of the data should be performed to allow those building the application to get familiar with the information at hand, so they can identify any quality problems, glean initial insights, or detect relevant subsets that can be used to form hypotheses suggested by the experts for hidden information. This also ensures that the available data will address the business objective.
Preparing the Data
To get data ready, IT organizations must select tables, records, and attributes from various sales and marketing systems; transform, merge, aggregate, derive, sample, and weight, when required; and then cleanse and enhance the data to ensure optimum results precision. These steps may often need to be performed multiple times to make it truly ready for the modeling tool.
Modeling
Once the information has been prepared, various modeling techniques should be selected and applied, with parameters calibrated to optimal values. For example, application developers can choose regression, decision trees, or other modeling methods. Choosing a modeling technique should be done according to the underlying characteristics of the data, or the desired form of the model for scoring. In other words, some techniques may explain the underlying patterns in data better than others; therefore, the outcomes of various modeling methods must be compared. In addition, a decision tree would be used if it were deemed important to have a set of rules as the scoring model, which are very easy to interpret. Several techniques can be applied to the same scenario to produce results from multiple perspectives.
Evaluation
Thorough assessments should be conducted from two unique perspectives. A technical, databased approach should be performed by statisticians, while a business approach gathers feedback from business issue owners and end users. These will often lead to changes in the model. But while the technical/data evaluation is important, it should not be so stringent that it significantly delays implementation and use of the model. The business value of the model should be the primary test.
Turning Results Into Action
Most importantly, companies must be able to transform the results of modeling efforts into actionable insight that sales and marketing professionals can apply to enhance planning and decision-making. The insight provided by predictive analysis initiatives must be shared with key stakeholders across the entire sales and marketing landscape – including any third-parties that may be involved, such as consultants and advertising agencies – to create an analyticsdriven culture.
Delivering Predictions to Business Users
By incorporating models into dashboard and reporting environments, organizations can ensure that the results are readily accessible to not just marketing analysts, but all sales and marketing professionals, whenever they need them. To maximize usability and value, these results must be presented in an intuitive way, such as a list of target customers who are most likely to participate in certain cross-sell offers, or a list of high-value clients with a high likelihood of churn who require specialized attention to ensure loyalty.
Performing “What If” Analysis
Sales and marketing strategies must shift frequently to keep pace with rapidly changing customer and market demands. To deliver the most advantage to sales and marketing departments, predictive analytics applications must enable the analysis of “what if” scenarios. This will allow professionals to understand the potential impacts of price adjustments or discounts, alterations to product positioning or messaging, and other changes – before they are implemented.
Feeding Downstream Systems
Even further value can be derived from the results of predictive models when they are dynamically shared with other applications and systems. For example, the results of proposed content changes for an e-mail can be automatically fed to a campaign management system, while the estimated impact of discounts can be sent directly to pricing systems to enable automated adjustments.
