Hyper-personalisation: Has AI finally made retail’s ‘holy grail’ possible?
Hyper-personalisation has long been the holy grail of retail: to recreate the customer experience of a local, everyone-knows-your-name store on a global scale.
Now, with the combination of data tracking, AI and our ever-increasing tendency to consume online content, genuinely personally-tailored ads are just about within touching distance.
These ads, introduced in their earliest form in the 1990s with email personalization, are the “biggest opportunity that hasn’t really been fulfilled”, Asam Malik, head of digital & risk consulting at Forvis Mazars, says. “It’s not there yet… but it’s the direction of travel.”
The ads undeniably work: McKinsey research has found that people are significantly more likely to buy a product, recommend a product, and repeat-buy a product if their ads have been personalised.
What is hyper-personalisation?
Imagine a local store in a small village. The shopkeeper definitely knows your name, and they might know that you bought a bag of flour two weeks ago, and that you’re probably about to run out. They might even suggest you buy a new type of flour which is especially good for bread – which they know you make a lot of, because you talk to them every week.
This access and knowledge is the goal of hyper-personalisation, albeit driven by data tracking and large-language AI models rather than repeated contact.
“[Companies have] the data not from just your interactions with that particular retailer, but numerous data feeds – be it from social media or from other websites that you’ve interacted with, and can give you something that’s very tailored to what you want,” Malik says.
For example: “Based upon your shopping habits, we think you’d love this product. And by the way, we have it in your size, in your favourite colour, click to buy it,” he explains.
Eagle AI, for example – the AI company behind many of the UK’s biggest supermarket loyalty schemes – provides location-based offers, behaviour-triggered promotions and event-driven incentives.
Each of these represent real-time, personalised push offers to consumers, with the latter aiming to harness “real-time issuance” by sending notifications to customers during events to “create urgency by delivering immediate offers that prompt customers to act quickly”, according to its website.
“The more businesses can learn from consumers, the better… this is how they can give the most relevant offers, and this is how they can personalise product recommendations,” Zsuzsa Kecsmar, co-founder and chief strategy officer at Antavo, said.
Antavo research has found that personalised loyalty programmes deliver an average of 5.2 times advertisers’ return on investment.
Where are these personalised ads?
Historically, the main channel that retailers have sent personalised ads to customers has been emails. But, as anyone who has ever had an email account will agree, the saturation point for email ads is very easily reached. There are pop-up ads, too – personalised with cookies and also with a low customer tolerance point – phone push notifications, and ads on social media.
It is this last set of ads which retailers are finding the most success with – particularly on Youtube. With 46 million adults on Youtube and over half watching on a large TV, targeted ads are becoming a must-have.
With AI, retailers are now able to link the product someone is researching for with purchase intent and their prior internet behaviour, and automatically feed the data into the adverts shown to that person.
“Youtube is one of the most powerful ways to raise awareness,” Google’s head of retail Sophie Neary told CityAM last year.
Long-form and short-form content are both growing “exponentially,” Neary said. “It’s a great way to get brands in front of the British public.”
Retailers are investing an increasing amount of money in generating content as a vehicle for showing customers personalised ads (and building their brand), with “huge teams just building social media content,” Malik says.
Companies like Gopro, which has 11.2m subscribers on its Youtube, and Foot Asylum, with 2.9m subscribers, are some of the most popular. Ferrari, too, has 1.85m subscribers, and Lego has 19.5m.
These accounts don’t primarily publish ads, but content which is likely to appeal to their consumer base. One of Hermes’ most successful videos in a mini-documentary on silk marbling, while Gopro has 88m views for a helicopter skydive.
“The hyper personalisation happens in the adverts in the middle,” Malik explains, and is based on “what you’ve clicked on, which episodes you’ve watched, and what they know about you.”
What’s next?
There’s a “massive push” on AI, but the data retailers have collected is often “inaccurate or incomplete”, Malik says, meaning that at the moment, the ads don’t always work as the retailer would like them to. “If it’s not relevant, you’ll disengage straight away and say, ‘I would never be interested in something like that'”.
But retailers are becoming increasingly savvy. Many made the mistake of jumping into AI and data collection without a strategy, meaning they failed to make the most of the early opportunities.
In the next few years, both the technology and retailers’ strategy will have improved to such an extent that hyper-personalisation will “get there”, Malik says.
But there are stumbling blocks: the line between showing someone something they already want and manipulating them into buying something they don’t need is a thin one, and something people are watching out for.
Tesco boss Ken Murphy’s suggestion that the grocer would use AI to “nudge” customers of its loyalty scheme towards healthier options, for example noting if they had high sodium intake and pushing ads for low-sodium foods, did not go down well with much of the public.
Jake Hurfurt from Big Brother Watch told The Telegraph it was “astounding” that Murphy wanted to use the data, and said his comments should “alarm everyone”.
Data privacy is therefore a significant concern, as is the bias inherent in AI models, and the unbelievable amount of electricity AI data centres can get through.
There are government attempts to regulate these concerns, like the EU AI act, which opens retailers up to significant fines if they breach rules. Civil rights groups have pointed out loopholes, though, with the European Centre for Non-for-Profit law saying that the Act “fails to effectively protect the rule of law and civic space”.
Despite the very real danger around these issues, it’s unlikely the personalisation train will stop anytime soon. After all, tailored ads have always been a holy grail – it just so happens that now the technology is here to build them, and that we’re supplying our own data.