Perspective on Predictive Analytics Products from a Data Scientist
The traditional higher education industry will be faced with many challenges over the next decade. These include generational declines in student populations, declining state and federal funding, and new education models competing with traditional higher-ed institutions.
Institutions are looking for new technological solutions for sustainable growth and revenue plans while operating with tightening margins. Inevitably this means leveraging data for informed decision making for nearly all university leadership teams.
Are you considering internal hires?
Many colleges and universities may consider building their own in-house data science teams. The benefits of having an in-house team includes reliability on meeting timelines set by the institution, availability for ‘ad-hoc’ analysis, and developing long term relationships with the professionals who are supporting your strategic initiatives. On the other hand, this presents an entire new set of challenges which we will discuss and weigh below.
The following is a quote regarding data science hiring, irrespective of industry, demonstrating initial barriers to entry.
According to Robert Half Salary Report for 2021, the data scientist median salary is $129,000. And for those managing a team of 10–15 members, the salary goes as high as $350,000. And a recent Dice Report shows that demand for data scientists increased by 50% in many sectors in 2020. Add to this the current shortage of data scientists and you can see it's simple supply and demand.
- Frank Howard Group
How do you find higher education talent with a data science background?
When we look at the higher education ecosystem, we are introduced to generational talent challenges. Higher-ed institutions must find and hire qualified higher education experts from an aging industry with many experts in the education space approaching retirement within the next decade.
According to Cupa HR the age distribution of higher education staff is skewed to the right, or towards older ages more so than the average US workforce. As of 2018 there was a small proportion of people under the age of 25 gaining employment in higher education.
Conversely, according to KD Nuggets, developers, both data scientists and non-data scientists have a higher proportion of younger talent. This indicates there is an exceedingly small overlap of higher education professionals that may also have a data science skillset.
What other options does higher-ed have for data science solutions?
The current data science marketplace for higher education includes institutional wide software services that are costly and are difficult to implement. Implementing software solutions require excessive time and resources to integrate into the university and overcome the steep learning curve of interpreting their own data and insights. Therefore, working with consulting firms or hiring an in-house data science team may appear more attractive.
Traditional consulting services have maintained a hands-on approach with universities but have not adopted modern machine learning (ML) and artificial intelligence (AI) solutions needed for precise forecasting and decision making. Consulting firms must adapt to a changing technological landscape and begin to incorporate more advanced predictive analytics capabilities into their current offerings which will drive their already expensive services up in cost.
Universities seeking to develop their own data analytics solutions will face other types of barriers such as time and resource constraints and hiring of personnel whose costs may even exceed that of working with external vendors. All solutions, especially in-house analytics development, will struggle identifying data science experts who can provide precise insights with uncertain estimates.
What kinds of data science solutions are a good fit for colleges and universities?
At SightLine we believe in a hybrid model between software-as-a-service and hands-on consulting. We keep our software in-house, allowing us to leverage the power of ML and AI, while delivering results to our partner institutions through presentations, reports, and training sessions. We find that this model makes predictive analytics adoption at least 3X faster for our partner institutions. They appreciate having direct contact with the data scientists who are working with their data daily. Sightline does the heavy lifting by following this formula:
- Listen to our partner institution’s goals, challenges, and vision for the future
- Train and test 100’s of machine learning algorithms to select the optimal predictive analytics model for each of our individual partner institutions
- Deliver intuitive visuals that demonstrate the institution’s positioning in the market, factors that impact student recruitment, enrollment, and success, and how to optimize these factors
- Summarize recommendations for pricing, marketing, branding, target student demographics, messaging, and individual student intervention for long-term success
SightLine never requires their partners to use complex data dashboards, that leave the institution to derive their own insights. We work together to make a data-based plan that works for our partners.