“Our predictive dashboard helps us evaluate our work, and better understand what we’re doing well and where we need improvement.”
Each year, College Forward reaches more than 12,000 students from low-income homes around Austin and Houston, and its 85 percent college graduation rate is far above the national rate for students from more affluent backgrounds.
Salesforce has been College Forward’s platform since 2010, when it received a special grant for a systems upgrade. “Selecting Salesforce was a no-brainer,” Buchan said. “It was customizable in the cloud, and we have 130 young people working for us who are just out of college themselves. The usability had to be there.”
College Forward entered the pilot program for Einstein Prediction Builder in the summer of 2017.
After a simple four-step process — using clicks, not code — Einstein sifted through the non-profit’s rich six-year trove of student data and generated a valuable new datapoint for College Forward counselors: a risk score assessing each student’s likelihood of dropping out of college.
“For the first time ever,” said Buchan, “our counselors could rapidly identify on a day-to-day basis which students were in most need of their support.”
Einstein derived the dropout-risk score by detecting patterns in College Forward’s structured metadata — student record field entries for missed days, days late, changes in GPA and other factors. However, those student records also contain written counselor notes about each student, which often refer to common non-academic problem areas such as an unstable housing situation. Einstein’s natural language processing extracts such useful information from the counselor’s notes, factoring the information into the dropout risk score.
“Now we’ve built a workflow with Salesforce process builder that flags a student’s record for these kinds of severe risk factors,” said Buchan. “A case might be automatically opened for a meeting with the student’s guardians, or perhaps a housing case might be forwarded to a specialist who can help with that situation. Our predictive dashboard helps us evaluate our work, and better understand what we’re doing well and where we need improvement.”
Since 2013, College Forward has used AppExchange to share a version of its customized platform with other nonprofits that have a similar mission.
“We realize, as we’ve grown to 12,000 students, that’s still just a fraction of the need that’s out there,” said Buchan. “So we took all these incredible customizations we’ve made on the platform and packaged them into this incredible tool.” Called CoPilot, the app is now being used by forty organizations serving a total of 240,000 students in 20 states and the African nation of Rwanda.