We talk about personalisation as if it’s about the person. It isn’t. It’s about prediction.”
That line sits at the centre of this conversation with James Taylor, CEO & Founder of A Particular Audience and once you hear it properly, it’s difficult to go back to how most marketing teams currently think about relevance. Because for years, personalisation has been framed as something close to one-to-one messaging. The idea that if we just had enough data, we could tailor every experience to the individual. It sounds right. It feels intuitive. And yet, in practice, it has largely disappointed.
What James lays out here is a different way of understanding the problem.
Not who the customer is — but what they are doing.
Not static segments — but real-time signals.
Not demographics — but behaviour.
Drawing on his experience building AI-driven recommendation systems used by global retailers, he explains how the most effective ecommerce experiences are not built around people, but around patterns. Around the relationships between products, actions, and intent. Around what millions of other customers have done before you, and what that makes likely next. This fundamentally changes how you think about websites, search, media, and even creativity.
Along the way, the conversation explores why so many early personalisation efforts failed, how Amazon and Netflix approached the problem differently, and why most retailers are still playing catch-up despite having access to the same underlying data.
There’s also a more grounded thread running through it — the reality of AI in practice. Not the version you see in product demos or LinkedIn posts, but the version that still requires constraints, rules, and human oversight. The version that gets things wrong. The version that can be incredibly powerful, but only when properly understood.
For marketers, there’s a useful tension here, on one side, the promise of hyper-relevance and automation, on the other, the discipline required to make it actually work.
This episode sits right in that space.
⏱️ Key Moments:
00:00 – Why “you are not your demographic” changes everything
02:15 – From investment banking to building AI products
08:00 – The real meaning of personalisation (and why it’s been misunderstood)
12:30 – Behaviour vs demographics: what actually drives relevance
18:00 – Building a recommendation engine from scratch
26:00 – Why most retailers still lag behind Amazon
30:00 – How AI is changing marketing teams
34:00 – The limits of AI (and why rules still matter)
36:30 – “Personalisation is just good prediction”
What you’ll take from this episode:
Why most personalisation strategies fail to deliver
How recommendation systems actually work in ecommerce
The difference between explicit and implicit customer signals
Why demographics are often a poor proxy for behaviour
How AI should (and shouldn’t) be used in marketing today
What marketers need to rethink about relevance and experience design
Brought to you by Tracksuit
Tracksuit is the always-on brand tracking platform helping marketers understand brand health, measure impact, and make better decisions over time.
👉 https://gotracksuit.com
Listen / Follow That’s What I Call Marketing:
🎧 Spotify: https://open.spotify.com/show/7MXhujDpTzbSRRbyQFgdWp
📩 Email:
[email protected]Subscribe for weekly conversations with leading marketers.
Hosted on Acast. See acast.com/privacy for more information.
Ireland