Predictive Modeling: Public College

Digital marketing, analytics, data science, campaign management

Mission Summary

At Avoq, I led the digital strategy components of an admissions marketing campaign for on of New Jersey’s top ranked public colleges.

The mission brief was to create an engagement funnel that targeted the right kinds of content to prospective students at the right time, based on what their behavioral and demographic data told us about where they stood in the college choice process.

DIGITAL CAMPAIGN

We incorporated supplementary tactics, such as a lead gen and content marketing campaign targeted to parents and guidance counselors, to drive dinner table conversations.

We also built a full-featured web analytics environment to track interim conversions, like campus tour sign-ups and important button clicks, as KPIs.

PREDICTIVE ANALYTICS

Besides defining the targeting strategy and managing the digital acquisition campaign, I undertook a predictive analytics model using Google’s Lightweight Media Mix Modeling package (since product-ized into Google Meridien) that ingested data across all of our advertising channels - including offline ones like billboards - and used machine learning algorithms to map their impact on applications.

Once I had built, trained, and tested a predictively powerful model, I was able to use it to prescribe an optimized budget for the next phase of the campaign.

CAPI and 1st Party Data Strategy

In refining the campaign for yield better results in the future, I helped implement a conversions API (CAPI) within the client’s CRM environment and developed a 1st party data strategy to increase personalization.

Results

The campaign helped drive a 10% increase in application year over year. When the optimized media mix is launch, we’re expecting to see a ~17% decrease in cost per application.

Tools & Tech