top of page

It’s Time to Redefine Early Intervention

Over the past several weeks our Co-Founder and CEO, Carolina Recchi, has outlined the urgency around why we must tackle the college completion crisis. She’s detailed some of the biggest challenges facing students today (backed by nearly 4 decades of research).

Throughout it all it has become abundantly clear that current systems and resources alone are not enough.

Prior to joining the team at EdSights I spent nearly my entire career working in enrollment marketing; focused on helping institutions find, engage and recruit prospective students. Very little time or effort was placed on ensuring student success once they enrolled. Over the past year I’ve spent a lot of time learning about existing processes and approaches to supporting student persistence. Today I will discuss some of the gaps between what I refer to as “legacy” approaches and the opportunities modern technology unlocks.

Surveys vs. Real-Time Engagement

For the longest time institutions have relied on surveys to understand their audience. Unfortunately, when it comes to effectively supporting retention and persistence, student satisfaction satisfaction surveys fall short.

While their practice is well intentioned, consider this scenario:

Mary, a sophomore, receives a survey from her college asking her if she feels the institution’s financial aid support is adequate for her needs. Mary is doing OK because in addition to her Work Study job on campus she has a part-time job waiting tables on weekends. So, feeling in that moment that things are tight but fine, she responds that financial support is adequate.

The next day Mary gets bad news. The restaurant she’s working at has cut her hours leaving her with no source of additional income. The statement for the credit card she put her textbooks on is coming due and now she has to figure out a solution.

In this situation the institution is led to believe that Mary is fine financially and takes no action to reinforce resources available to her. Ultimately she finishes the semester but now is not in a position to be able to return in the spring because she fears bringing on too much more debt.

What if there were a way to check-in with Mary on a more consistent basis and get updates on how she is feeling and the challenges she is facing on a regular basis?

​Hear how Franklin Pierce University is tackling student persistence leading to a 14% boost in retention.


Predictive Model Scoring vs. Proactive Updates

Another common practice is the use of model scoring to identify students who are potentially at risk for not persisting based on historical data. Once again, while this practice may be well-intentioned it is ripe with bias and also reliant on data that is likely no longer an accurate predictor in light of the COVID-19 pandemic’s impact on these students’ educational experiences.

Like the example above, the legacy approach of model scoring alone falls short.

Consider this scenario:

Sharon, a first-year biology major, was identified as a student potentially at-risk of not persisting because of the combination of her home ZIP code, SAT score and math curriculum on her high school transcript.

Mark, another first-year biology major, was not identified as at-risk because he had a strong SAT score, has a ZIP code from a more affluent neighborhood and took all AP courses in his curriculum.

Sharon is excited about the opportunity to be on her own. She applies herself in class, studies more frequently than she ever did in high school, makes friends and finally feels confident in her abilities. Because she was flagged by the institution’s predictive model she is highlighted as a success story even though no interventions were ever actually taken to support her success.

Mark has a hard time making friends, negatively impacting his sense of belonging at the institution. While he knows he has strong skills in math and science he starts missing classes as he is not motivated to do well. Because he was not flagged by the institution’s predictive model he does not receive any type of proactive outreach and it isn’t until he fails his midterm that anyone flags him in the system for needing support. Mark ultimately drops out.

What if the institution could proactively reach out to both students on a regular basis and tailor follow up and interventions based on the responses they receive directly from them rather than relying on bad data to guide decisions about their support?

Dr. Will Miller shares his perspectives on the shortfalls of legacy predictive models for supporting student outcomes.


Incorporating the Student Voice

As we developed our Adaptive AI framework we looked at numerous resource studies on the key drivers impacting student persistence. This enables us to ensure that the SMS text messages each institution’s branded chatbot sends is effectively addressing the challenges students are facing.

These factors include:

  • Financial Distress

  • Academic Engagement

  • Wellness

  • Engagement

Additionally, the Collected Intelligence shared across all of EdSights’ partners ensures that the AI bot supporting your students learns faster and becomes more precise with each student interaction.

In order to effectively support student persistence we need to be asking students about the challenges they face earlier and more often. This cannot accomplished with legacy approaches alone.

​Hear from Dr. Vincent Tinto about how the COVID-19 pandemic has shaped student perspectives about including the student voice into every conversation about their success.


Align Your Team’s Efforts With Student Experiences

I used to say when I was working in enrollment marketing that it seems odd that as consumers ourselves we leave our personal perspectives about being marketed to at the door when we walk into an admissions office. This is why so many recruitment practices like large scale direct mail campaigns and impersonal email campaigns are still so prevalent even in an age when student demand personalized and real-time connections. These types of efforts also lead to burned out teams who are pulled in multiple directions and are using bad data to make decisions.

The same can be said about the legacy approaches to student retention, persistence and early intervention.

  • How often do you fill out long surveys?

  • How annoyed are you when you get a phone call or an email you aren’t expecting because it doesn’t match your needs?

  • In the age of Quiet Quitting, what are you doing to ensure your team is not burned out?

It’s time to redefine our approach to early intervention by listening to the student voice first, faster and more frequently.

Hear how Western Illinois University is boosting team efficiency using Adaptive AI.



By Gil Rogers

Head of Marketing at EdSights

Gil is a recognized leader in higher education enrollment management having served in leadership roles at the University of New Haven (CT) as well as Director of Enrollment Marketing at Chegg, Director of Marketing at NRCCUA (now Encoura) and most recently as Executive Vice President at PlatformQ Education.

Gil has conducted numerous research studies on student engagement published by Inside Higher Ed, The Chronicle of Higher Education and presented at numerous regional and national conferences focused on student recruitment and engagement including the National Association of College Admission Counseling (NACAC), Ruffalo Noel Levitz National Conference, ACT Enrollment Planners Conference, TargetX Summit and more.



bottom of page