Anyone following my "tweets" on Twitter during the Salesforce.com Dreamforce '07 conference probably saw my references to lead scoring. The presentations and discussions inspired me to revisit this topic and explore how to implement lead scoring in Salesforce using an existing i-Dialogue web event data mart.

My past experiences with lead scoring were based on assigning "A", "B", "C", or "D" scores to leads derived from their online interactions. This prioritization gave sales reps some insight into which leads were supposedly more qualified than others.

A much simpler approach to lead scoring is to simply assign a score to each interaction and let the cumulative score provide a relative indicator. For example, in the illustration below, 2 leads enter a web site and initially appear to be equally qualified. But over time, and with enough data collection, we see that one lead was actually more interested in seeking employment with the organization rather than becoming a customer.

Defining lead score rules in Salesforce is possible through the use of a "Web Event Scoring Rule" custom object (see below), which is a prototypical definition of an existing Web Event object (already included with i-Dialogue).

A periodic evaluation by the i-Dialogue-to-Salesforce integration process evaluates each interaction and checks to see if it matches a lead score rule prototype. A match results in a Lead's score being increased or decreased by a defined amount. Rule prototypes may have any number of evaluation parameters, allowing for simple or complex rule definitions. If it can be captured in i-Dialogue, it can be evaluated.

A custom field on the Lead record named "Lead Score" is updated by the scoring process. Scores in Salesforce are updated in near real-time and are accurate to within 30 minutes.

This score then becomes the basis for Salesforce Reports and Dashboards. Once the technology is in place, an iterative/incremental relationship marketing process is implemented to routinely review and adjust scores and online event collectors. Lead scores will never be perfect, but the balance of online interactions and evaluation rules can achieve an 80% "good enough" state to be widely trusted by Sales Representatives and Marketing Managers.