Steve Jurvetson· @FutureJurvetson · X·· 3 小时前
事实摘要
26年前我写《亚马逊是时代错位》预测其会败给“个人购物智能体”,大错特错。(见下文) 所以今天我觉得挺有趣:“AI个人购物助手可能通过直接查找并购买商品,改变亚马逊700亿美元的广告业务”
正文 · 原文
I was very wrong 26 years ago when I wrote "Amazon is an Anachronism" predicting their downfall to “personal-shopping-agents.” (below)
So I was amused today: “An AI personal shopper could disrupt Amazon’s $70 billion advertising business by finding and buying products directly"
Reminds me of a CNET article I wrote in the 90's which provoked an inquiry by Bezos. It was titled Amazon is an Anachronism, and explored the idea of agents poaching the promotional dollars Amazon depended on at the time. I was a wee bit early in that. :) Here it is: "The whole concept of a “retail brand” is an anachronism on the Internet. The vertically integrated e-retailer is a holdover from the physical world that is entirely out of place online. The Internet empowers consumers over retailers for the first time. E-retailers will no longer be the force—and the bottleneck—that they are in the physical world. This implies a dramatic shift of who has power in the e-commerce domain. The first approaches in any new medium mimic what is already well understood in prior media. Marshall McLuhan first espoused this idea in the ’60s, and it has played out to this day with remarkable accuracy. The early movies were “filmed theater,” with a single camera angle capturing the view from the audience seat. Early TV shows were televised radio programs. People start with what they understand, and the unique expressions of the new medium take time to emerge. The Internet has grown much more quickly than any of the preceding media did, but it has not yet matured. Many millions of people are using the Web while it is still in the “filmed-theater” phase. Untargeted banner ads, like subway billboards, are simplistic and increasingly ineffective. We have yet to see the widespread adoption of business models that are uniquely derived from the Internet media and that would not have been feasible in the physical world. Over the next few years, merchandising and product selection will be decoupled from product fulfillment and support. Today they are smashed together out of habit, because that is what we are used to in the physical world. We browse the products on the shelves, ask a sales clerk for advice, pick a product, pay for it, and take delivery all from the same store. In the physical world, it is very cumbersome to segregate these activities to different companies, as the customer would have to shuttle across town for each step in the buying chain. The retail store is orchestrated around maximizing efficiencies of the physical supply chain—the logistical relationships that move the goods from production to the consumer. The consumer is only engaged in the final step with the binary option to buy or not to buy from a fixed set of options at fixed prices. Power rests with the store. Geographic constraints defined this optimal retail chain, one that has undergone little change for hundreds of years. What if your alternative was a “shopping agent” which has no economic incentive to sway you one way or another, will give you honest consumer feedback, and will get to know you and all of your purchase habits across the entire Web? For both objectivity and pan-Web personalization, there will be separate consumer companies that focus exclusively on product selection and reviews. In this new horizontal model, retail brands won’t exist on the web. No Wal-Mart or Toys ‘R’ Us. These brands have been built on the irrelevant physical world attributes of location and shelf-space. These brands are not strong in the areas of rapid fulfillment or customer support. They will continue to be strong in the physical world, but they have no meaning online. They will be hollowed out, and if they persist in a vertically-integrated mindset, they will find that their customers are being poached at the checkout stand by the front-end agents. Now, there will still be plenty of powerful brands—the old product brands, like Tide detergent, and the new personal-shopping-agent brands. These new agent brands will be customer-centric and built on trust. They are uniquely enabled by electronic networks and will engender the strongest of loyalties. It is possible that some forward-looking retail brands will migrate to becoming agent or fulfillment brands, but not if they think of themselves as a “retailer.” In a horizontal industry, value accrues from relationships. Positions of power can be most easily maintained at the boundaries of the supply network—at the consumer shopping agent interface and the interface to the product manufacturers. The valuable assets are customer knowledge and exclusive access to product brands and private label products. Although http://Amazon.com is an anachronism as a simple e-retailer, it is by no means doomed. Shopping habits change at a glacial pace. In the meantime, it might gravitate toward domination of a horizontal layer, like back-end product fulfillment from its enormous distribution centers. It might spin off its product-catalog front end to compete more effectively with the independent horizontal innovators. I don’t expect it to stick stubbornly to a vertical strategy when the horizontal pioneers start poaching customers. http://Amazon.com will likely recognize the transition sooner than IBM did in the mainframe days."在 X 查看被引用的帖子
来源:Steve Jurvetson · x.com