I was struck by the comment “Instead of relying on the subjective opinions of salespeople, critics, or even well-meaning friends, customers could receive personalised, data-driven suggestions that align with their tastes.” There’s a big assumption here. It suggests that recommendation is a data-acquisition task, based on some - notionally - objective framework. I don’t buy that (pun intended). It’s entirely possible that sales assistant recommendations and family & friends are favoured simply because people like talking to people and value the exchange inherent in the experience. And even more that what people are looking for is not something that “aligns with their tastes” but that wine is often an expression of “taste” that can ONLY be accurately deduced through the cues of interpersonal exchange. There is a vast amount of psychology that would back up the idea that “flawed” forms of “subjective” wine recommendation are a feature not a bug.
I agree 100% The whole business of recommendation is highly complex and involves an equation in which both sides are essentially variable - in knowledge, aspiration, motivation etc. One might even try to argue that no two wine (or book, movie, car etc) recommendation situations are entirely alike.
I enjoyed the reference to Brave New World. I might argue Edward Bernays’ Propaganda was equally apt: “The conscious and intelligent manipulation of the organized habits and opinions of the masses is an important element in democratic society.” He’d have loved the power of LLM’s
I feel like I have missed something in your choice of a digital pure play like Spotify as your straw man for AI and choosing wine. Why not a riff on Vivino? I have my preferences, can get the recommendation, can find the wine and buy it. Not a big step to delivery. The big supermarkets have taken the lions share of wine ecommerce and have strong platforms. Here Dan Murphy’s can connect my physical and online sales and provide pick up, click and collect, and delivery but currently have terrible recommendation engines. They also connect hundreds of small producers to consumers via drop shipping. Why not a Vivinofication of big retail?
I have followed Vivino since it began, and have used it, both as a wine drinker and as co-owner of a brand. The percentage of wine drinkers who use it is very small - 68m versus 10x that number for Spotify. But the difference is even bigger, because the Vivino ‘users’ rarely ‘use’ the app, while the Spotify figure is for monthly users. I can see Vivino surviving as a platform for enthusiasts, but I cannot see it ever having the general appeal and ubiquity of a platform like Spotify - or what the tech giants are planning for us. My guess/bet is that we will rely on the giants rather than on smaller players. I’m sure a few people do straight to the online version of the Oxford companion when wanting to find out about a grape variety, but I’m also sure that 99.9% rely on Google. And my points about delivery stand. Supplying someone with a single bottle of wine, unless you are already a logistics business like Amazon, is not cost effective.
I was struck by the comment “Instead of relying on the subjective opinions of salespeople, critics, or even well-meaning friends, customers could receive personalised, data-driven suggestions that align with their tastes.” There’s a big assumption here. It suggests that recommendation is a data-acquisition task, based on some - notionally - objective framework. I don’t buy that (pun intended). It’s entirely possible that sales assistant recommendations and family & friends are favoured simply because people like talking to people and value the exchange inherent in the experience. And even more that what people are looking for is not something that “aligns with their tastes” but that wine is often an expression of “taste” that can ONLY be accurately deduced through the cues of interpersonal exchange. There is a vast amount of psychology that would back up the idea that “flawed” forms of “subjective” wine recommendation are a feature not a bug.
I agree 100% The whole business of recommendation is highly complex and involves an equation in which both sides are essentially variable - in knowledge, aspiration, motivation etc. One might even try to argue that no two wine (or book, movie, car etc) recommendation situations are entirely alike.
I enjoyed the reference to Brave New World. I might argue Edward Bernays’ Propaganda was equally apt: “The conscious and intelligent manipulation of the organized habits and opinions of the masses is an important element in democratic society.” He’d have loved the power of LLM’s
I feel like I have missed something in your choice of a digital pure play like Spotify as your straw man for AI and choosing wine. Why not a riff on Vivino? I have my preferences, can get the recommendation, can find the wine and buy it. Not a big step to delivery. The big supermarkets have taken the lions share of wine ecommerce and have strong platforms. Here Dan Murphy’s can connect my physical and online sales and provide pick up, click and collect, and delivery but currently have terrible recommendation engines. They also connect hundreds of small producers to consumers via drop shipping. Why not a Vivinofication of big retail?
I have followed Vivino since it began, and have used it, both as a wine drinker and as co-owner of a brand. The percentage of wine drinkers who use it is very small - 68m versus 10x that number for Spotify. But the difference is even bigger, because the Vivino ‘users’ rarely ‘use’ the app, while the Spotify figure is for monthly users. I can see Vivino surviving as a platform for enthusiasts, but I cannot see it ever having the general appeal and ubiquity of a platform like Spotify - or what the tech giants are planning for us. My guess/bet is that we will rely on the giants rather than on smaller players. I’m sure a few people do straight to the online version of the Oxford companion when wanting to find out about a grape variety, but I’m also sure that 99.9% rely on Google. And my points about delivery stand. Supplying someone with a single bottle of wine, unless you are already a logistics business like Amazon, is not cost effective.