Picture this: a customer lands on the homepage of your online fashion shop. At her choosing you know her age, gender, relationship status, location, and income. You also know that she’s an extrovert, a social butterfly, and a risk-taker.
From the moment you start engaging with her, through to the final stages in the customer journey, every image, word, and call-to-action you present has been carefully selected through algorithms that take into account her personality and interests.
The product catalogues you highlight on the homepage are ripe with bold, statement items. Imagery shows those items being worn in group, social scenes, while a call-to-action on the basket page dares her to ‘bite the bullet’ or ‘do it’ rather than ‘proceed to checkout’.
This paints the picture of a new breed of customer experience that’s a lot closer than we realise. That’s according to Vesselin Popov, business development director at the University of Cambridge Psychometrics Centre which is making key developments in the area of personalisation.
Customer experiences can be shaped using psychological profiling – sample creative from a study run by the Psychometrics Centre.
Research shows that people spend more money on things that fit their psychological profile – and this makes them happier, explains Vesselin. Brands understand this – in fact 94% of CMOs say it’s important to understand the psychological attributes of their customers.
To this end, The University of Cambridge Psychometrics Centre – a multi-disciplinary team of psychologists, statisticians, mathematicians, computer scientists, linguists, engineers, and entrepreneurs – has been developing a personalisation engine capable of predicting psychological traits and behaviour from digital footprints.
A new class of personalisation
The Psychometrics Centre’s application program interface (API) Apply Magic Sauce builds upon a 30-year legacy of leadership in computational behavioural science. It can accurately identify an individual’s personality, intelligence, life satisfaction, political views, religious views. And it does so by building psychological profiles based on the digital footprints we create through our online activity – from posting on social media networks, to streaming music, buying products, and consuming content.
Vesselin Popov and team are developing a smart personalisation engine
Unlike most data-driven marketing tools, the Apply Magic Sauce API focuses on the psychological traits and emotions that drive a given behaviour, not just the fact that someone has clicked or ‘liked’ something. Crucially, it also empowers the consumer to learn about themselves and demand greater personalisation, rebalancing the usual power dynamic in predictive marketing.
Based on the myPersonality database – which contains psychological ground truth and social media information from more than 6 million volunteers worldwide – the API benefits from data and methods that have been proven in more than 45 peer-reviewed scientific articles in the last five years.
Mining data for new customer experiences
Equipped with such technology, the opportunities are endless for brands looking to provide new and personalised experiences. Imagine being able to adjust the presentation, delivery, and content of your message to suit the psychological makeup of the person viewing it – or being able to offer real-time feedback and recommendations that set your brand apart.
You don’t have to be a scientist to understand the power of using data in this way, according to Vesselin.
You don’t have to be a psychologist to understand how you can deliver your messages and content differently, based on an individual’s profile. It’s a very human way of looking at customer experiences.
‘You don’t have to be a psychologist to understand how you can deliver your messages and content differently, based on an individual’s profile. It’s a very human way of looking at customer experiences and it brings new perspective to your content and creative’.
Early market research reveals how effective these new customer experiences could be. A proof-of-concept ad campaign conducted by The University of Cambridge Psychometrics Centre in partnership with an online beauty retailer found that personality-optimised ads were twice as profitable as those that were not optimised.
Further ongoing work with Hilton Hotels demonstrated the impact of these techniques on click-through – which rose 1,750% in a real campaign application – and social engagement, where optimised ads were shared three times as often.
An example of images and copy tailored for high extraversion
Getting closer to the individual
The results give serious food for thought to online retailers who are striving to differentiate through the experiences they deliver. ‘Online retailers are doing a reasonable job of consolidating the data they have to try and get an end-to-end view of consumers’, says Vesselin. ‘They’re making search better and more relevant, and they’re building in some data science and recommendations as part of the experience’.
But as those recommendation algorithms get more complex, and with a growing amount of content to serve, retailers risk getting further away from understanding the individual. ‘We’re still putting people in buckets for the most part, and that’s not good for personalisation’, says Vesselin.
Online retailers need to move beyond customer segments and understand the individual
What ecommerce personalisation needs, he argues, is to move beyond segments and demographics as the norm. ‘Yes it’s efficient to have 10 segments for which you deliver a different user journey, set of products, or content. But it can feel jarring from the consumer’s point of view when they’re lumped into one of those buckets’.
For this reason the Apply Magic Sauce project is focused on collecting its information reliably and transparently, incentivising consumers to part with their data in exchange for being able to individualise their own experiences.
Creating a dialogue around data
As consumers become increasingly reluctant to share personal data with brands unless they can see clear benefits, this transparent dialogue about data between brand and consumer will become hugely important. In Vesselin's view, algorithms will work better if people are motivated to interact with you and tell the truth.
Those brands that create effective dialogue around data will stand out, gain trust, and be able to deliver a more relevant service to the end user.
Those brands that create effective dialogue around data will stand out, gain trust, and be able to deliver a more relevant service to the end user’.
But today’s recommendations are not only fairly inaccurate, they’re also being made without brands explaining to the consumer what they’re doing and why they’re doing it, according to Vesselin:
‘If you describe what you’re doing and the data points you’re using – the fact that you’re serving products you think are relevant for someone aged 18-45 in London, for example – the consumer won’t be creeped out, and they’ll also realise you don’t have very accurate information about them. In being transparent, that consumer might be motivated to provide their precise age, or access to a digital footprint, so that you can recommend something more relevant’.
Consumers should always be able to see the results of what they’ve opted in for, Vesselin argues, and they should never have a prediction made about them without their knowledge. Moving towards this more transparent model will go a long way in terms of earning trust, whilst enabling algorithms to explain themselves and learn from their mistakes, just like humans.
Transformation starts with experimentation
So how can brands begin to make these changes to the experiences they deliver? Having an infrastructure that enables you to experiment is vital, says Vesselin.
You need to be able to quickly swap things in and out and measure those initiatives against your business goals.
As organisations evolve how they interact with consumers online, the winners will be those retailers that don’t try to go to market with a fully-developed solution, but instead take a ‘build, measure, learn’ approach, using small experiments and explorations to discover what works – and ultimately, to get to market before the competition.
This article first appeared on Figaro Digital.