When you have more than 100 different products on offer, how do you decide which ones to focus your market efforts around? You can promote those that have the highest margins, but doing so might come at the expense of higher sales volumes from lower cost products.
This is the problem that faced Rakesh Vemu, head of digital acquisition for Open Colleges, an Australian online provider of more than 100 TAFE and other nationally-accredited courses.
Vemu’s role is to acquire active leads that are then passed over to a sales team to be converted into student enrolments. In recent years, however, that task has been made more difficult by a desire within the organisation to reign in its overall marketing spending.
“We spend up to 90 per cent of the budget on search, and it has proven itself over the years as a great channel for driving leads,” Vemu told CMO. “When I joined the business about a year ago, I wanted to make sure we are doing everything we can to make this and other channels as profitable as possible.
“The easiest way would be to put more budget to the most expensive products, but then there may not be as much demand for the most expensive products.”
A solution was needed that would enable Open Colleges to determine the ideal allocation of spending across its numerous courses. So Vemu briefed Open Colleges’ primary search advertising partner, Google, to help build a simple AI model that would enable him to weigh up the various internal data points, and enrich those with some external market data provided by Google.
“It was not humanly possible for myself and my team to cover all those data points, so we needed a layer of AI,” Vemu said.
An algorithm was created that could examine all data points and provide an output on how to spread the marketing budget to maximise profitability. This was ready in time for Open Colleges’ peak conversion period at the beginning of the following year, and was first deployed in January this year with immediate uplift to profitability on a year-on-year basis.
“For January we had a 90 per cent improvement in profitability,” Vemu said. “And then for February it was around the 60 per cent mark, because we had more competition in the market. It led to higher revenue and higher profitability from a lower budget, essentially helping me derive more from less.
“And that is where data democratisation was critical – in that, I was able to allow the right parties access to confidential business data along with Google providing the external market data that was used in the algorithm to spit out an ideal allocation of budget.”
By the end of March, however, improvement had fallen to 4 per cent. While this was still a good result from a decreased budget, Vemu said it provided an important learning. In both January and February, Open Colleges had been regularly feeding internal and external market data into the model during the month, but it only did so once in March.
“In a way it was good, because it proved the point you have got to be more real time with this,” Vemu said. “The real-time aspect of this whole exercise was manual, and my team had to put in that data and run the model to get the most real time recommendations out of it. What we are doing now is looking at building a more automated way of looking at the signals, and I am working with Google to automate that part.
“You’ve got to trust your partners. And if you empower your team and your partner with the right amount of data you can achieve some good results.”
Read more: Open Colleges weighs in on Salesforce Google Analytics partnership
Follow CMO on Twitter: @CMOAustralia, take part in the CMO conversation on LinkedIn: CMO ANZ, join us on Facebook: https://www.facebook.com/CMOAustralia, or check us out on Google+: google.com/+CmoAu