gianfranco cuzziol quote web

If you put 200-plus of Europe’s email and CRM cognoscenti into one Chamonix hotel, you can be certain that, as sure as après follows ski, you won’t get more than two sentences into a conversation without the words ‘big’ and ‘data’ coming up in close conjunction.

We’re all perfectly aware of, and more than happy to talk about, the transformational potential big data has in guiding how we might better talk to our audiences. However, one of the tales that struck me most, a tale told by a fellow CRM marketer, was about a new AI algorithm that would crunch the data from a pizza company client and tell them exactly what content they should concentrate on. With fanfare, the AI-infused insight revealed… the subscribers to the pizza-based newsletter were most interested in… pizza.

And that’s why smart brands are still making high-profile punts on human intuition.

Look at Channel 4’s new Walter Presents service, showing international TV content curated by Walter Iuzzolino. The former film producer and agency bigwig has, it is claimed, spent 3,800 hours “locked in a dark room, handpicking the most compelling, explosive, funniest and sexiest shows”. While it must have cost a fortune to keep topping up poor Walter with pizza and beer during a box-set binge that long, C4 is betting that the selection algorithm operating in Walter’s brain will be better attuned to its viewer demographic than Netflix’s is to mine. Apple, meanwhile, bet heavily on the future of human curation with music experts recommending songs when it launched Apple Radio. Indeed, it was a key point of differentiation against the algorithm-based Spotify.

Try looking at it another way: it would be easy for the BBC to provide us with a Premier League team of the week based purely on Opta stats, but it’s more interesting reading about why Garth Crooks this week picked De Gea, Mahrez and Firmino.

Personalisation is beginning to be realised by the likes of Homebase, who can tell if you are about to undertake a big DIY project because you bought a packet of birdseed (I kid you not!) and EasyJet, whose excellent 20th anniversary email reminded me of the miles I’ve travelled with them, my favourite destination and whether I prefer a window or aisle seat. But it isn’t a panacea - and we should stop considering it as such. It’s ironic, given we’re talking about data being processed and analysed by hugely sophisticated programmes, but it’s still a problem of complexity.

Yes, I trust a machine to suggest to me I might want to buy some long-bristle paintbrushes to go with the Farrow & Ball satin I just added to my shopping basket. And yes, I also trust it to suggest the shade of gloss I might want to use on the room’s skirting board, based on both my previous shopping behaviour and those of others who have followed similar purchasing paths. However, I still don’t trust it to tell me what film I should watch when the decorating’s done… or what to have on my pizza while I’m watching it.

It’s about using machines to recommend recipes to customers based on previous things they’ve tried, but using the human touch to make suggestions that take us outside of the obvious. After all, we’re in the business of speaking to human beings, and when it comes to knowing what complex human beings really like - there’s nothing like complex human beings.

Gianfranco Cuzziol is head of CRM at Zone