Bob Regnerus was hired to generate leads for Dartmouth College’s Tuck School of Business executive education program, after it had seen the costs of advertising rise while its number of qualified leads shrank. The only advertising that was generating a significant amount of viable leads were full-page ads in the Wall Street Journal, costing roughly $100,000 per day. While the ad was effective at generating leads, it wasn’t cost-effective.
For a program that generated $25,000 to $35,000 in revenue for each student, that type of advertising was cost prohibitive. It wasn’t easy to create an effective lead generation website for the program. The version of the lead generation website we ended up using differed significantly from the one we started with. We did at least 20 A/B tests on the site, all with the goal of increasing not just leads but qualified leads.
The client was running a global campaign because it wanted to attract students from overseas. However, our tracking and analysis efforts determined that leads generated from certain countries were unqualified. Since advertising to those countries consumed a significant part of the budget, we revamped the campaign to stop advertising to audiences there.
Meanwhile, we discovered that the leads coming in through traditional paid search sites—Google and Yahoo—were not of the quality that Dartmouth was seeking. We tried placing ads on about 25 secondary sites—ultimately identifying six “hidden gem” websites out of the 25. On those sites, we A/B tested 50 different advertisements to find the best-performing ad. We also pared down an initial list of about 25,000 keywords to just 2,000 highly profitable and productive keywords. It took about six months, but the results were extraordinary.
Leads generated from this effort have increased dramatically. Cost per lead has been significantly more cost-effective than the Wall Street Journal ads. Feedback from the executive education sales staff shows that the prospects generated are exactly the kind of serious prospect they were hoping for.
What you’ll notice about this last case study is the sheer volume of testing involved—25 traffic sources tested (only six cost-effective sources discovered), 50 different advertisements tested (only one “most effective” ad discovered), 25,000 keywords tested (only 2,000 “most productive” keywords discovered). In a purely statistical sense, nine out of 10 things didn’t work (or didn’t work as well as the something we discovered later).
Thankfully, the process of analysis and optimization allows you to get rid of the things that don’t work. It’s like taking a 10-question test in school, realizing you got only one answer right, and then legitimately erasing the nine questions you got wrong. You end up with a perfect score every time. This is exactly how the analysis and optimization process works.