Will predictive behaviour be a part of your Facebook game?
You’ll be forgiven for missing this bit of news (most people did), but back in 2016, Facebook introduced an advertising product that could be used to predict users’ behaviour and give brands the opportunity to send highly targeted ads in an attempt to alter their behaviour.
This ‘Loyalty Prediction’ product was part of a suite of features included in a machine-learning ‘AI’ tool called FBLearner Flow. But, whilst the technology has been quietly in the public domain now for a couple of years, the advertising techniques this enables have only just been revealed.
So what is ‘Loyalty Prediction’?
To paraphrase, FBLearner Flow’s machine-learning platform can use the platform’s 2 billion customers and their data points, to identify individual users who are at risk of moving to a competitor brand. Thus giving advertisers the opportunity to influence their behaviour through targeted advertising in a way that would be very cost efficient.
This isn’t Facebook giving a brand the opportunity to advertise brand A because they’ve been reading about brand B all week. This is Facebook using the facts of your life, revealed by the personal data you share on the platform, to predict that you’ll get sick of brand A and give this knowledge to brand A-Z.
This technology is a reminder of the power of Facebook – predicting how users will behave offline having interpreted their online behaviour.
How could this be applied to student marketing?
Facebook has an incredible reach amongst the student audience, so any new ad development is worth your attention. There could be huge opportunities (and risks – more on that later) for early adopters.
We know how important Facebook is as a channel. The National Clearing Survey found that 60% of students going through Clearing use Facebook on a daily basis, and 58% use Facebook Messenger daily to communicate with their friends. We also know about the challenges universities face retaining prospects. We’ve recently blogged about the challenges universities face keeping leads warm over a long time (check out the webinar), the Summer Melt phenomena, challenging retention rates nationally and the growing consumerisation of student behaviour and the Clearing ‘marketplace’. Understanding and influencing brand loyalty directly would be a big weapon in any institutions arsenal.
Imagine being able to advertise your course during Clearing to an identified ‘Mind Changer’, before any university had even identified that particular student as vulnerable to changing their minds. Or targeting students who would be willing to switch institutions during Christmas term or after their 1st year. Perhaps your institution banks on a certain amount of HND students taking your top-up, but imagine the possibilities this product could provide if you are a competitor, marketing a third year BA (Hons). You could promote your provision in a highly targeted way to those who are most likely to apply to a different institution.
It could help with retention as well. You could send targeted advertising to your cohort who are vulnerable to another institution’s overtures, with broad messaging, reminding them about the experience of your university – lifestyle, clubs and societies, accessibility, support, identity – these are the broad themes students care about and could keep them engaged.
OK, sounds great, but what’s the catch?
There’s a big one. In a post-Cambridge Analytica era, the use of this kind of technology and audience understanding doesn’t feel like a good look. And particularly for HEI’s, whose core values often include principles of trust and transparency, this may be AI and audience understanding gone too far. As marketers, we’re always balancing cost versus outcome versus risk, and in this equation would the use of personal data taken directly from the platform and applied to a machine learning tool really sit well with your brand and how it is perceived?
Utilising ‘Loyalty Prediction’ might not be a million miles away from what Cambridge Analytica did. Taking mundane details based on demographics, likes and dislikes, to accurately model election behaviours. Whilst Cambridge Analytica had to jump through the hoop of being a third party, Facebook would have unfettered access to 2 billion consumers. And those consumers may react just as angrily towards any brand that adopts this new tool in their strategy to influence their behaviour, no matter the cause, product or brand.
But there’s more…
Let’s not forget that Facebook is still the biggest social media network in the world and incredibly important to any university or college for student recruitment, engagement and reach. There are lots of brilliant tactics and strategies that we recommend, using the very best that the platform has to offer. We’ve written a lot about how universities can protect both their brand, their audience and prospective students’ by treating data ethically and with dignity.
We’ll be talking about the latest innovations that we recommend for Facebook and other platforms in our live webinar on the 10th May, check it out here. Facebook will once again be one of our headline speakers at our upcoming event for education marketers, ASDIE, on the 19th July. You’ll hear all about the latest in how you can utilise the platform for your brand, direct from the platform.
If you would like to find out more about Facebook, machine-learning or how to use cutting-edge innovations in your marketing, get in touch.