On Valentine’s Day this year, Deliveroo launched what it thought was a charming campaign. The company sent handwritten love letters to select customers, posing as a secret admirer, along with a 15% off flowers voucher. Meant to be cheeky and playful, the gesture left many creeped out. Customers started to wonder how Deliveroo knew so much about them.
The campaign has since become a cautionary tale, illustrating how easily personalisation, when misused, can backfire.
The problem with over-personalisation
According to Andy Kaiser, VP of product at Contentful, the root of many personalisation fails is simple — brands try too hard.
With access to everything from recent location history to past purchases, some brands overstep, using information in ways that feel invasive rather than helpful.
“The most obvious mistake is over-personalisation. Brands straddle the line between being helpful and, in the consumer’s eyes, a touch creepy,” said Kaiser.
“Consumers appreciate relevant recommendations. But they also expect a level of discretion,” he added.
On the opposite end of the spectrum are brands that collect volumes of data but don’t use it in a meaningful way. This leads to impersonal, irrelevant experiences, like recommending products a customer has already bought or showing promotions that no longer apply.
“Tastes and buying patterns evolve. If personalisation strategies aren’t updated regularly, they become not just useless, but frustrating,” explained Kaiser.
Instead, marketers need to keep recommendations fresh, timely and rooted in actual customer needs.
Focus on first-party data, not creepy tracking
To build effective personalised campaigns, brands need to know what data they need to collect and why, and they need to empower their customers to opt in and adjust their preferences.
Personalisation should feel like a service, not surveillance. Allowing people to reset their data, change preferences or even opt out entirely builds long-term trust.
With growing concerns about privacy, companies to prioritise first-party and zero-party data — information willingly shared by the customer—over invasive third-party tracking.
“The days of relying on third-party cookies are ending,” said Kaiser.
“And that’s a good thing. It’s forcing brands to get more honest and more creative,” he added.
Kaiser stresses that transparency must go hand-in-hand with value.
“If customers understand how their data improves their experience, they’re more likely to share it,” he added.
Don’t let automation kill the human touch
Even with the best data practices, personalisation can still feel robotic. “Artificial intelligence can process massive datasets and predict preferences, but it lacks emotional intelligence,” said Kaiser.
The answer lies in a hybrid approach: AI for efficiency, humans for empathy. “Let automation handle repetitive tasks. But when things get complex or emotional, bring in real people,” he commented.
Without human oversight, personalisation can feel flat, or worse, wrong. “AI can make incorrect assumptions,” warned Kaiser. “If there’s no human in the loop, customers are left with irrelevant or irritating experiences.”
A simple example that got it right
When personalisation works, it feels effortless and natural. Kaiser highlighted a homeware brand that tailored content specifically to pet owners. Dog owners saw images, reviews, and messaging about dog-friendly rugs. The cat owners saw the same, but for cats.
By aligning the ad campaign, landing page, and product suggestions, the company saw a 25% increase in conversions and a 7-fold boost in click-throughs, while cutting content creation time from two days to 30 minutes.
“The takeaway is that effective personalisation isn’t about knowing everything. It’s about making people feel understood,” he concluded.