The Facebook of ecommerce⁠↗

Amazon, very obviously,  is Google for products. It’s good at giving you the best-seller you’ve heard of or the water filter for your fridge (the long tail). It’s not so good at the things in the middle. Amazon is great at selling you what’s on the table in the front of the bookshop, and at selling one copy a year of a million or so obscure titles, but it’s not very good at showing you what’s on the shelves at the back of the bookshop. It’s not so good at selling the mid-list - things that you didn’t know existed, or didn’t know you wanted, before you saw them. It does have a recommendation product, but it’s not clear how well it works, and indeed an interesting question for Amazon is how far it can grow before running into categories for which its commodity merchandising model doesn’t work so well. (Even in print books, Amazon’s market share only reached a quarter of the market after 20 years of ruthless execution).

we’ve now reached the point that it’s not clear that there is anything that cannot be bought online. But actually, the real barrier now is often not touching it, but knowing about it. These are things where a lot of the role of a physical shop is curation, recommendation and demand generation rather than logistics: the shop shapes your choice but it also tells you about things you didn’t plan to buy, and you can’t search for things you hadn’t thought about.

However, there is clearly also an opportunity to do this at massive scale, for many people across many products - that’s kind of the point of the internet. From that perspective this is not a particularly solved problem. Google doesn’t do demand generation. Etsy, ProductHunt and Pinterest (the latter two both A16z portfolio companies) are building solutions from different directions, and a great many people trying to replicate some form of the boutique model online as well, as I suggested here - to create some kind of specialist curation. As machine learning ripples out across the tech industry, entrepreneurs will probably also find entirely new and better ways to answer ‘what would I like’. That might include Google, indeed.  That kind of scalable automation, though, could also go in completely the opposite direction for some things - away from any kind of decision at all.