Get in touch

Using learning analytics to improve student retention

24 February, 2021

Underpinned by an OES Customised Learning Analytics solution, Swinburne Online observed improved student retention and revealed valuable insights to safeguard revenue. Here’s how the partnership worked:

Our client

OES delivers courses on behalf of Swinburne University of Technology – an internationally recognised university, ranked in the top 1.5 per cent of universities worldwide according to The Times Higher Education University World Rankings 2021 and QS World University Rankings 2021.

Student retention is an important business objective for Swinburne Online, which delivers high calibre online programs to more than 10,000 students each year and receives extremely high student satisfaction scores (QILT SES) from its online student cohort.

As Swinburne University’s trusted partner in online education, OES is always pursuing innovation, in this case to help improve the retention rate among its online undergraduate students.

The challenge

The majority of Swinburne Online’s undergraduate students are mature-age and studying part time. Juggling a degree with other commitments – such as work and family responsibilities – means the experiences and needs of this cohort are unique.

“Unlike on-campus undergraduates, many of Swinburne Online’s students are trying to fit university into an already busy schedule and it’s been many years since they last completed study,” said Sacha Nouwens, Executive Director of Student Experience and Insights at OES.

“By effectively integrating data from across the Swinburne Online ecosystem, our aim was to better understand and meet the needs of this non-traditional undergraduate cohort, to help them stay engaged and realise their degree and study goals.”

Our solution and approach

The OES analytics team captured and combined approximately 5,000 data points per student from across Swinburne Online’s many platforms and channels.

“By collating, interrogating and analysing this single view of each student, we were able to establish which metrics were the most likely indicators of student engagement,” explained Sacha. “Our sophisticated analysis of large and diverse data sets was instrumental to the success of the actions we subsequently took to improve student retention.”

“Once we had the right data, our next step was to build a machine-learning propensity model to pinpoint the critical metrics impacting engagement and identify those students at risk of dropping out. This propensity model brought the right data points together to establish a view of each student’s level of engagement, to accurately determine whether any intervention was necessary and to inform what that intervention should be.”

“Our next step was to empower teaching and support staff to easily access the analytics and monitor student engagement week-to-week, using custom built automated trackers.”

Finally, OES developed best practice intervention guides to enable both teaching and central support staff to take the right type of action, at the right time. Different approaches were used for these two staff groups:

  • Teaching staff used OES’s bespoke engagement tool, which was embedded within Swinburne Online’s LMS, to decide which students would benefit from direct contact and what the best way to reach them would be, for example a phone call, an email, an SMS or direct message within the online classroom.
  • Support staff, through our central outbound team, were able to contact and encourage students at key points along their study journey. For example, contacts were triggered for new students who needed assistance with orientation, students completing their first major assessment, students returning from a study break and students whose behaviour patterns had changed compared to previous teaching periods.

Monitor, measure and refine

Continual improvement is a key pillar of the OES approach to learning analytics and involves ongoing measurement of impact and outcomes.

“If we find that a particular action or intervention isn’t having the desired impact on student engagement or retention, we either modify it (for example by altering the timing, the message or the channel) or stop doing it.”

“This constant evaluation also helps us identify new interventions or support services we can provide to assist students and achieve our client’s student retention goals,” Sacha said.


This Customised Learning Analytics solution featured all the hallmarks of a successful OES/client partnership: expertise in advanced analytics, constant collaboration and our proven, objectives-focused process:

  1. Track and understand to generate meaningful insights
  2. Take tailored, targeted action
  3. Measure to achieve continuous improvement.

The outcomes for Swinburne Online were impressive, including a 5% improvement in retention of commencing students into their second year.

This comment from a Swinburne Online student who was contacted in the right way, at the right time, reflects the way that effective application of learning analytics can impact individual outcomes:

My [Learning Advisor] reached out to check why I did not submit my last assignment and tried her best to encourage and help me, which took me by surprise. If it weren’t for that, I would not have returned in [the next teaching period]. Her actions gave me hope that someone does care about my progress even though it’s entirely my responsibility.

Swinburne Online undergraduate student

Find out more about how OES tailors Customised Education Solutions to meet the diverse needs of our clients.