Mastering Mass Personalisation with Artificial Intelligence8 min read
Becoming the marketing boss you’ve always wanted to be starts today!
We know that personalisation is a powerful way of influencing consumer behaviour.
But a few years ago, the generation of personalised content used to require much more effort on the companies’ side.
Now, the increased availability of artificial intelligence for business combined with marketing automation made sophisticated segmentation less costly and faster to implement.
So let’s dive in and see how AI can help you to achieve mass personalisation.
Personalisation at Scale through Digital Footprints
First of all, computers are now able to perform profiling and classification of consumers, based on the data they actively or passively provide, the so-called digital footprints.
In the last few years, this practice became known as personalisation at scale.
McKinsey summarised what personalisation means to customers according to this simple equation:
Personalisation value = ((relevance + timeliness)/Loss of privacy)* Trust
So let’s analyse this formula.
Mass personalisation is directly related to the relevance of the content.
For example, consumers have been reported to prefer suitable recommendations that they wouldn’t have thought of by themselves.
The travel industry actually makes use of text mining (a form of natural language processing that derives high-quality information from text) to create and test recommendation systems based on the similarity of destinations.
For this, they use the reviews that previous travellers have made and the most common co-occurring words they used to describe a destination.
With that, they can set retargeting campaigns suggesting similar destinations to previous travellers, something that users would see as their “unknown unknowns”.
Timeliness is also quite important. Users report that they to be approached when they are in a “shopping mode”.
Despite people constantly checking their emails using their smartphones, sending automated messages late at night isn’t timely, as consumers are unlikely to be thinking about shopping.
Privacy and Trust in Mass Personalisation
So how do we design these things in such a way that we do not interfere in their privacy and deteriorate their trust?
After all, what are the boundaries of digital mass persuasion?
Recent research has shown that very frequently, people do not behave logically when it comes to privacy.
For example, we humans often share intimate information with strangers while we keep secrets from our closed ones, the so-called Privacy Paradox.
“People don’t always behave logically when it comes to privacy. For example, we often share intimate details with total strangers while we keep secrets from loved ones. “ – Harvard Business Review
This helps to explain why, on average, just 65 liked Facebook Pages allows behaviour analysts to understand someone’s personality traits better than their friends do, 120 to understand them better than their family members, and 250 to understand them better than a partner or spouse.
Nevertheless, behavioural science has identified some factors that predict whether people would be ok with the use of their personal information.
And I’ll illustrate this based on some experimental examples.
Don’t talk about people behind their backs.
When we know that a friend has revealed something personal about us to others, we naturally get upset.
Those norms also apply in the digital channels.
The researchers used a dimensionality reduction method to find groups of practices that consumers tend to dislike.
They did that based on a list of common ways in which Google and Facebook use consumers’ personal data to generate ads.
The results suggest that:
- Obtaining information from third-party platforms, and
- Deducing information about someone from analytics.
Are more frequently disliked.
Copywriting for Mass Personalisation
Previous research has also tested whether varying the copywriting would affect buying behaviour, for example running the same ad but with different disclosure designs.
So in one design, a group of participants saw an ad that had the following copywriting:
“You are seeing this ad based on information that you provided about yourself.”
A second group of subjects saw:
“You are seeing this ad based on information that we inferred about you.”
And a control group saw the “business as usual” no disclosure ad.
Participants who viewed the ad framed as inferred behaviour analytics showed much less interest in purchasing than the other groups did.
Also, Booking.com tested variations of the copywriting to check which was more effective at increasing conversions in retargeting email campaigns.
In their case, a less intrusive variation was less effective than the one obtained by
A LDA model, a type of natural language processing, applied to the users’ reviews.
This variation said that: based on your past trips, their team of travel scientists thought that you probably have a passion for romantic landscapes, local food, or shopping.
I find this example interesting because it shows the importance of testing how humans react to copywriting variations.
Artificial intelligence is helpful here because
“Machine learning improves our ability to predict what person will respond to what persuasive technique, through which channel, and at which time.” – Behaviouralscience.org
This combination of behaviour analytics and automation is called digital nudging.
For instance, digital nudging can help some companies to reframe their services within a “personal advice” approach. This can make it easier for them to acquire customer data.
Also, it helps to comply with the GDPR regulations concerning privacy issues and information sharing.
Also, in a recent podcast hosted by the channel Behavioural Grooves, Rebecca Blank, the Chief Behavioural Officer at Maritz, discussed how customers might react differently according to the way companies communicate with them through personal content.
Mastering Mass Personalisation
In conclusion, we see that personalisation is an efficient way of influencing consumers, especially when it is powered by artificial intelligence.
But we have to mitigate backlashes by testing different layouts, images and text that provide information on how your personalised content was obtained.
The key takeaways to mastering mass personalisation are:
- Any disclosure is less creepy and will convert better than no disclosure
- Deduced information on the customer will convert less than an open explanation of why they are seeing an ad
- Trust is a key factor in dealing with the new world of shared information and data gathering
Looking to get started with A.I. for Business & Marketing? Our A.I. for business starter pack is full of resources to help you with the first steps!
Design a chatbot that your customers will actually want to chat with.read more
If you’re not thinking “mobile-first”, you’re not doing it right.read more