We Cannot Rely On Gut Instinct Any Longer
The days of relying on gut instinct alone are over.
Don’t get me wrong – I’m not saying there’s no role for a non-qualitative assessment of your work – but it’s just not reliable enough to be central to your decision-making process throughout the cycle.
We’re not the only ones for whom this is true. As I discussed last week – admissions leadership has gone through a similar transformation that professional poker has – the best of the best are relying on assessing, calculating, and recalculating probabilities of student enrollment decisions.
This goes far beyond guessing the final yield rate at the end or even attempting to calculate how a financial aid adjustment might impact students in a particular discounting matrix cell.
For a long time, the most advanced data analysis in college admissions offices looked at three or four variables. FAFSA, visit scholarship amount, etc. We didn’t know it, but we were off on orders of magnitude regarding how many variables we needed to analyze.
The current best practice is to assess hundreds of potential variables and put together the right mix of student behaviors by which we can calculate probability. Then, each day, we need to refresh the calculation and develop the newest number for every student. Then, we need to be prepared to do it again tomorrow.
Of course, there are human limits to accomplishing this, which is why our profession needs to embrace data science and machine learning to guide us more fully.
Some of us will be slower to embrace this new model; others will outright oppose it. But just like poker, the increased adoption of this technology will slowly rack up more and more “wins” for the colleges that lean into it.
This reality will lead to transitions in leadership positions where ultimately, the most significant data nerds will be the most successful among us.
You’ll know the transition is complete when, a generation from now, expect a NACAC conference where probability analysis is part of the “first-time attendees” track.
