CDS will be exhibiting at next February’s Marketing Show North.
The personalisation of digital content has grown into one of the hottest digital topics in the last few years, and the technology is now becoming widely available.
At CDS, we work with a number of platforms in this space, notably Episerver’s Content Management System (CMS) and associated tools. In 2016, Enterprise CMS platform Episerver announced its acquisition of Peerius, a leading provider of intelligent omnichannel personalisation in the cloud, accelerating the third wave of smart personalisation technology called ‘autonomous personalisation’.
Since then, the Peerius platform has been tightly integrated with Episerver’s commerce and CMS platforms to become Episerver Personalization.
These are rules-based tools that bring machine learning capability to content management enabling digital managers to personalise search results, product descriptions, calls to action, website content, and navigation.
Google understands that relevance is extremely important to consumers. They have been delivering personalised organic search results since 2004, alongside their paid-for Adwords service. Their business model depends on them getting the right message in front of the right person, at the right time and in the right place, and the Google search algorithms are continually evolving to refine this targeting.
This type of personalisation combined with marketing automation technologies can be a powerful tool for marketers. However, its uptake amongst even the most technologically equipped marketing teams has been somewhat slower than we might expect. The reason for this slow uptake is simple – marketers simply don’t have the time to create and test multiple versions of the same content.
However, the day is almost here when AI can deliver what humans simply don’t have the time to create.
Indix have a product that can create bespoke marketing copy to create product descriptions – entirely written by machines.
And in 2016, a world first for AI writing occurred at the Rio Olympics. The Washington Post reported on 300 events using Heliograf, their in-house AI software. These robot-generated articles were published alongside ones written by humans, and crucially, no-one spotted the difference. Since the Olympics, The Washington Post’s robot reporter has published 850 articles in the past year to cover congressional races on Election Day and many local sports events that would have gone uncovered.
It is only a matter of time before these complex algorithms will be capable of producing long-form marketing copy that is indistinguishable from that created by human copywriters. And marketers; those of us who have invested in the tools to profile, segment and target our customers – will reap the benefits.
In 2015 Rishiraj Saha Roy published research into the automated personalisation of targeted marketing messages, mining user-generated text on social media to create unique copy designed to evoke positive sentiment in specific audience segments. Crowdsourced experiments verified that these personalised messages were almost indistinguishable from similar human compositions.
The moment is approaching where automated content, written by a machine, and AI’s personality profiling tools will converge – meaning that not only will the content we see be produced by an AI, but will, potentially, be uniquely constructed to appeal specifically to you.
Our online behaviour reveals a huge amount about our personalities. Researchers have created an algorithm which can accurately predict personalities simply based on Facebook interactions. And, surprisingly, it knows your character better than your close friends. The team found that their software was able to predict a study participant’s personality more accurately than a work colleague by analysing just 10 ‘Likes’. Inputting 70 ‘Likes’ allowed it to obtain a truer picture of someone’s character than a close friend, while 150 ‘Likes’ outperformed a parent or sibling. At 300 ‘Likes’ the programme was able to judge character better than a spouse.
A personalised experience is nothing new. Google has been personalising search results for many years, and the potential impact of these artificially constructed ‘echo chambers’ were identified as long ago as 2011 – a phenomenon described by Eli Pariser in his 2011 book and TED talk as the ‘Filter Bubble’.
However, 2018 has been the year that the world began to take the power of data seriously, and we begin to understand the power that personalised messages can have on a reader’s beliefs and behaviour. In the wrong hands, these tools can have dangerous consequences, and can even shape world events and swing elections.
Cambridge Analytica whistleblower Christopher Wylie stated this in the clearest terms:
“We exploited Facebook to harvest millions of people’s profiles. And built models to exploit what we knew about them and target their inner demons.”
Just because your personal data (hopefully) wasn’t included in the Facebook data mined by Cambridge Analytica, this doesn’t mean that you are immune to its effects. Wylie stated that the data was used to model interactions and behaviour across entire populations, and whilst we all like to think of ourselves as individuals, we are all susceptible to suggestion and bias. In 2017, Facebook supplied to Congress over 3,000 examples of Russian Facebook banners designed to sway American political opinion, all aimed at different slices of American society with the targeting made possible by Facebook’s advertising algorithms. Unpicking who saw what (and when) is a mammoth task for lawmakers.
As marketers, we have a responsibility to behave ethically. Recent GDPR legislation goes some way to protecting customers from unwanted and inappropriate marketing, but personalisation based on user behaviour and anonymised profiling often falls outside the scope of this legislation. There is still a lot of room for sharp practice in our field.
We have a human duty to use these tools and powers with great care, and take responsibility for the messages our robot servants create on our behalf.