Case Study - AirTank Uses Machine Learning to Increase Applications for Higher Education Institude

Artificial intelligence and machine learning are critical technologies to help businesses accelerate their decision-making and improve overall results. One of the ways machine learning can help us to determine user flows is to improve the customer experience. Here is just one example of some work that AirTank did with a higher education institution to help improve application rates. 

Situation

This institution had 2 flows for potential students; 1) Request for Information - This was something that was used earlier in the funnel to attract students and get them into the lead flow journey. Allowing them to send information about potential programs and success stories as the potential student went along their journey. 2) Submit an application - This was when the student was ready to apply for the program. Typically once the student had done enough research and was ready to make the final decision (apply). 

This created a choice for potential applicants. Which path should they choose? How do they know which path to choose? 

Solution

The hypothesis was fairly simple, can inbound data be used to help make this choice for the applicant?  This specific institution was looking for ways to improve conversion rates and bake Machine Learning into daily business strategy. 

The proposal was to use available 1st party cookies and browser data to choose a path for the applicant. 

For example, if the applicant had been to the site previously and was in a higher converting application state we could make an assumption that their ideal path would be into the application.

If the applicant was new to the site and was in a poorly converting state the better path might be to provide them with more information and start them on their way down the path of an application in the future.

Together with the higher education institution, we took 20 different data points (weather, time of day, location, device type previous site visitation, etc…) and combined it with application rates. This data set was used to determine which visitor's path was most likely to achieve the desired outcome. The machine learning model would then adapt from the training set of data against live actual data and improve the model over time.

Results

This model resulted in an increase in application starts by 7% and total application conversion rates by 25%. The model was proven to not only be effective but to drive continuous improvement over time. Reinforcing the value of testing in the organization, but also the value of machine learning and artificial intelligence. 

Curious as to how you can leverage AI to exceed goals for your brand? Let the AirTank team help you navigate this new and exciting channel by reaching out to us here!