Should an enrolment forecast be subject to your data governance policies?
A few weeks ago, I ran a poll on whether your enrolment forecast should be covered by your data governance policies. While this poll has a small sample, there were some great comments from Stephen Dove, Associate Director of Resource Planning & Budgeting at the University of New Brunswick (Canada). He suggested that there’s a bit of a split – that the forecast itself is created from data. The data part can be included in data governance, but the information part is actually just standard records management – no different than posting a budget on a website. I really appreciate Stephen sharing his insight in this space.
Similar to Stephen, I’d argue that the data of the enrolment forecast should be subject to data governance policies and practices. The question then becomes what kinds of policies would need to apply, which depends on the kind of forecast you’re doing:
- If it’s a behaviour model, where you’re trying to predict individual student behaviour, such as with an early alert system (i.e., is this student at risk of leaving), then many institutions guard the results with the same kinds of policies that they use for student personal information.
- However, if your model is more of an aggregate prediction (i.e., how many students will we have) then the outputs are likely to be okay for a wider audience.
Even with the aggregate prediction though, there may be some components that have to be guarded more closely. Often, we see institutions that have a pipeline of data feeding their model that resembles this:
P.S. I am aware that I am vastly oversimplifying here. These four blocks also describe two years of my job at a large university :)
So why do I suggest data governance from stem to stern in this process? Because a huge part of data governance is creating a culture around creating, using, and sharing data. While some might argue that a forecast is just data about students who aren’t even real (yet), I’d argue that the forecast is a great place to tie in data governance.
Not sure where to begin? In our research, every data governance leader recommended starting small, with a pilot. An enrolment forecast is a great pilot initiative. Over the years I’ve learned a lot from George Firican of LightsOnData. One of his suggestions is to start with a business glossary. Let’s get started together. Here’s a few terms I see over and over again in enrolment forecasting, but only rarely do I see definitions:
- New intake: student new to your institution.
- New to program: student who is new to a program but not to your institution; usually an internal transfer.
- Continue: a student who is continuing their studies at the same level relative to the prior time period in your forecast.
- Progress: a student who is continuing their studies at the next level up relative to the prior time period in your forecast.
- Exit: a student who leaves your forecast model at a specific time period. Note that an exit may well mean someone who changes to a new program, graduates from your institution, or leaves. Depending on your context you may need to disaggregate this term further.
There are of course other terms related to forecasting, and your institution may have its own twist on the above, but this gives you a start. Stay tuned for my next post that focuses on enrolment forecast data lineage, which helps answer the question of “where the heck did you get the data for this forecast?”