Up the Odds when Gambling with Leads Modeling and Scoring Leads for Marketing in Education Presented at the DETC 2009 Workshop, Naples, FL October 20, 2009 by Ron Gregory, President and Senior Consultant, Gregory & Partners
© 2009 Outline of Presentation About Gregory & Partners What’s the problem? What’s the answer? What is lead scoring? Why score leads? How does modeling work? What’s a typical model look like? How does lead scoring work? How would you use lead scoring? Other uses of modeling and scoring How would you start lead scoring?
© 2009 About Gregory & PartnersGregory & Partners Offering game-changing solutions for distance education, specializing in: Marketing Strategy & Planning Lead Modeling & Scoring Talent Assessment, Recruiting & Training Clients are institutions ready to move to the next level Identify gaps in infrastructure, personnel, expertise, processes, and analysis Help define strategy and goals Construct a plan to close the gaps Help execute the plan, as needed Train or recruit personnel, as needed
© 2009 What’s the problem? You receive both good and bad leads from your lead generation efforts. Represents a bit of a gamble every time. You could make your conversion efforts more efficient by knowing which are which. But how do you tell the difference? Knowing some of the factors helps, but how to precisely weigh each factor to achieve a reliable prediction of results?
© 2009 What’s the answer? Lead Modeling and Scoring is an analytical tool for your marketing program. This tool allows you to fairly accurately predict which leads are good and which are poor. Generally used to predict the odds of converting into an enrollment. Using this tool, you generate more profit from the same level of marketing expenditure.
© 2009 What is lead scoring? Similar to credit scoring, companies and schools score their leads, to sort and rank them from good to poor. In a conversion model, good leads score high, and are more likely to convert into an enrollment than average or poor leads. Poor leads score low and are less likely to convert than good or average leads.
© 2009 Why score leads? With a way to predict conversion rates, you can choose those leads to which you want to devote a high level of conversion efforts, and those to which you want to devote few efforts. Example: Type of Effort Score Level High228 Middle 126 Poor014 Conversion Mailing Number of Efforts
© 2009 Why score leads? Saves money on trying to convert leads with low chance of enrolling. Identifies good leads so more marketing dollars can be devoted to convert them. This procedure brings greater efficiency to conversion-marketing programs Optimizing the program can generate more enrolls from the same group of leads. The vast majority of enrolls will convert from the top and middle-scoring leads. Squeezing the last few enrolls out of the poorer leads will cost so much, you’ll lose money in the long run.
© 2009 Few dollars spent to convert best leads. Many dollars gained from enrollments received. Conversion marketing (direct mail, , etc.) Many dollars spent to convert poor leads. Few dollars gained from enrollments received. Conversion marketing (direct mail, , etc.) Best-Scoring Leads Poor-Scoring Leads
© 2009 How does modeling work? Demographic and other data is overlaid onto the leads to serve as “grist for the modeling mill.” The model sorts through all these overlaid factors to pick out those that are most predictive and to determine what weighting to apply to them. Predicting conversion rates without a model or computer would be nearly impossible.
© 2009 How does modeling work? Leads are scored using a formula that combines the contribution of each model- selected factor. Formula is created by a model using regression analysis, neural network, decision tree, or other analytical technique. Objective of the model: Closely predict the element you want to optimize or maximize, such as conversion rate or revenue, for each lead.
© 2009 How does modeling work? Scoring and ranking of leads can be based on enrollments, starts, or revenue Modeling on enrollments (conversion rate) gives you a faster look at the accuracy and effectiveness of the model. For schools accepting monthly payments, revenue takes longer to see a difference between the scored and un-scored leads and their resulting enrolls. First Test New Students
© 2009 How does modeling work? WeightingContribution Factor in ModelValueAssignedto Score Age in Years Household Income $ 53, Years of Education Age of Vehicle Children in HH TOTAL SCORE (Predicted Revenue) Example of Simple Formula
© 2009 Typical Results of Modeling 2% = overall average conversion rate Over one third of leads in lowest 10% Top 20% converted at five times the avg. Bottom 20% received one third the avg. CR Grouping leads into five tiers (A-F) makes it easier to understand
© 2009 Typical Gains Chart
© 2009 How does lead scoring work? By controlling your expenses in conversion marketing, you can make more profit, even after counting the cost of leads that you don’t try hard to convert. Using this sorting and selective conversion technique optimizes for profit, even when receiving a reduced number of enrolls. All New Students Optimum New Students
© 2009 How would you use lead scoring? Substantial savings from controlling expenses and optimizing enrollments from the high- and mid- scoring leads. Detect good and bad lead sources quickly. Gain leverage in negotiating for good leads. Send the best leads to the outbound agents fast. Depending on conversion costs and other factors, marketing contribution (or profit margin) can be 10 percentage points higher. “Your results may vary,” but it would be worth finding out just what lead scoring can do for your marketing program.
© 2009 Other Uses of Modeling and Scoring Retention – Modeling can be used to score students to find those who will continue studying (and paying). With retention modeling, you can identify which students need early intervention or additional support. Subsequent marketing events, such as annual discount periods or special offers, can be targeted to those leads most likely to respond.
© 2009 How would you start lead scoring? Choose a reputable vendor in the area of lead modeling – one with plenty of experience in the technique and in your field of business. Possible vendors: Gregory & Partners Wheaton Group Acxiom Experian TargusInfo
© 2009 How would you start lead scoring? Ask questions and find the vendor that fits your school and marketing program best. Sample questions: What type of are data bases are you using (census block/ household/individual)? Is the information from a single database or multiple? Are you building custom models based on our enrollments? Do you use best-of-breed modeling techniques to build our models? What type of validation and back-testing do you perform? How often do you re-check the model? Can you score leads real time? What are your prices? Do you charge for the upfront modeling?
© 2009 Thank you! Any questions?