Data shows that machine learning can boost international admissions conversions
- “Artificial intelligence” (AI) or “machine learning” is a type of technology that can evaluate large pools of data and learn from it, especially with regard to identifying patterns or predicting outcomes
- New data from both the US and UK shows that AI-enabled admissions systems see significant improvements in conversion rates
"Technology is not the be-all-end-all," says Ashish Fernando, the CEO of iSchoolConnect. "Technology is not the answer to all questions, and artificial intelligence is not something that can work autonomously."
That may seem like a strange thing for a technologist to say, not to mention the CEO of a technology company. But it reflects a growing understanding of the role of technology in society, and the contribution that it can make to international enrolment management in particular.
Speaking at ICEF Toronto last month, Mr Fernando asked his audience to imagine a school where all prospective and current students get undivided attention at all times, and where there is an artificial intelligence (AI) companion to help staff make key candidate decisions. (Please note that we are using the terms "artificial intelligence" and "machine learning" interchangeably to describe a technology that can evaluate large pools of data and learn from it, especially with regard to identifying patterns or predicting outcomes.)
To illustrate the point, he noted that a 3% conversion rate throughout the admissions process is "a good market standard." That is, out of every 100 prospective students who, for example, visit an institutional website, three actually end up enrolled. Based on actual enrolment data, however, Mr Fernando observes that institutions that employ AI systems typically see that conversion rate move from 3% up to 7-8%.
Taking that one step further, the iSchoolConnect data also reflects a higher graduation rate for AI-supported institutions (from 68% for "non-AI" institutions to 78% for those using AI tools).
"Typically, institutions are looking at students only from an enrolment perspective," he adds. "But what you want is for the student to go through the entire two-or-four-year graduation – to complete the entire course. And that is where we are heading. We want to focus on not just enrolling any student; we want to focus on enrolling the right student."
What Mr Fernando is describing in part is an AI-supported admissions process that crunches vast amounts of data to help admissions staff identify and focus on the candidates that are best fitted to the institution or school in question. And "best fit" in this sense can include sweeping considerations around not only which students are most likely to be successfully admitted, but also those that are most likely to secure a study visa, graduate, and even have a successful career after their studies.
"At no point when you deploy technology, including AI, will it take the place of a human," he cautions. "Technology is not a replacement; it is an enabler. It can make a process faster, cheaper, and more convenient."
A recent white paper on the subject from QS echoes the point by drawing on aggregate data from QS client-institutions in the UK. Within that sample, AI systems were found to have a noticeable impact on conversions for the September 2021 intake.
Across a sample of nearly 50,000 offer holders, machine learning led to a 3% increase in conversion.
As QS explains, "Without machine learning, the conversion rate from offer to enrolment was 8.01%. With machine learning, the conversion rate was 11.3%. If a typical institution with 2,500 international students per intake increases conversion by 3% then this equates to more than £1.5 million in additional first-year tuition fee income."
Faster, cheaper, and more convenient
There at least two important levers that allow AI systems to boost conversion rates. First, they can be trained to evaluate large amounts of data across the institution, and to use that information to evaluate prospects so that admissions staff can then better focus on the candidates that have the greatest chance of following through to enrolment and then to be successful during and even after their studies.
Second, through the use of chatbot services and other automated functions, technology can assist international offices in better managing the increasing volumes of enquiry and admissions traffic that most institutions and schools are experiencing these days.
The QS paper makes the point clearly in comparing 2021 and 2020 admissions traffic among client-institutions in the UK, "We saw new enquiries spike by 82% compared to 2020 and new application numbers increase by 179%. Not only are there more prospective students asking questions, but they are also asking multiple questions and expecting quicker and quicker responses. What we are hearing from several institutions in the UK is that they have not been able to keep up with demand and the result is that thousands of enquirers are not receiving replies. Similarly, admissions teams are experiencing significant growth in applications but are not able to process the higher volume of applications they are now receiving in a timely manner, if at all."
In that context, any automated communications to students can make the difference between losing and engaging a good prospect, especially given that many enquiries or other student communications will arrive outside of normal campus hours from students around the world.
As QS also points out, students are only becoming more demanding in terms of the timeliness and relevance of responses from institutions and schools. As such, any improvement in this aspect of student service can be a substantial competitive advantage.
"In the most recent QS International Student Survey, 84% of prospective students expect to receive a complete and personal response to an enquiry within a week, with 18% anticipating this level of response in just 24 hours. After submitting an application, 45% of prospective students want to be informed of the result of their application within a week, with a further 39% expecting the outcome within a month. It’s clear that already-stretched admissions teams often cannot deliver to these challenging timescales, leaving applicants feeling ignored and frustrated."
For additional background, please see:
