Algorithmic decision making was a hot topic during the summer. This was partly thanks to the way in which it was used in estimating the final grades for students whose education had been interrupted by the COVID-19 pandemic. Rightly or wrongly, the conclusions the algorithm had been programmed to deliver were deemed unfair to the students who had been unable to demonstrate their abilities.
There was disbelief that such an important decision might be left to a piece of software at all. After all, the teachers knew better: surely the people with the closest personal and professional relationship with the students was best placed to determine what they might have achieved had they managed to sit the exams in person?
Like it or not, algorithmic decision making plays a huge role in our day to day lives – its just that most of the time it’s so unobtrusive and accurate, we barely notice it.
I’m writing this post from a train to Manchester while listening to my Daily Mix on Spotify. The track list is determined by Spotify’s personalisation algorithm. It analyses my listening patterns and picks a bunch of tunes that it knows I like (because I listen to them) and then adds in a few tracks that are liked by people who also like what I like. The music’s probably a bit on the beige side for most people, but it’s better than alright, and it’s introduced me to a fair amount of (admittedly very vanilla) songs that I missed first time around.
When they have a lot of data, recommendation algorithms like Spotify (or Netflix, or YouTube, or Kindle) serve up content that appeals to us. The more we share with them, the better they get at giving us things we like – whether that’s letting us discover a new TV show, meeting an old friend, or learning that “Lost Weekend” by Lloyd Cole & the Commotions was always your favourite tune.
We’ve grown accustomed to checking the “yeah, whatever” box when we’re asked about our privacy preferences because it’s generally more convenient to be followed around by ads for things we might actually buy rather than things we could never need.
And that’s where we get into the marketing part.
Facebook, Google, LinkedIn and every other major digital ad platform collect vast amounts of information about users so that they can package those users up into convenient and receptive audiences for advertisers to promote goods and services to. It means that advertisers can build highly targeted campaigns for clients that help them reach new customers efficiently.
The audience definitions that we as advertisers have access to in 2020 are the wet dream of our marketing ancestors. They’d have killed for access to even a sliver of what we have today, and yet, we often squander the embarrassment of riches we possess.
Without knowing as much about our customers as the platforms do, we can never truly understand their real needs and the triggers that will encourage them to make a decision we like. Properly analysing customer intent, investigating who they are, what they respond to and the nexus between what we have and what they want allows for something much more profound than a well targeted advert, it allows for a genuine relationship where we don’t get things nearly right, we get them really right.
If you want to reach customers, you need to know your customers. Our Customerology approach provides powerful insights, meaning we are able to create content that engages and drives behavioural change across paid, owned and earned channels. Find out more about Customerology here.