The Disappearance of Search: How Shopping Will Work in the AI Era?

This article explores how AI agents are transforming commerce and their impact on Google's business model, affiliate marketing, and online shopping. This article is derived from an article written by a16z and was compiled, compiled and contributed by PANews. (Synopsis: Bloomberg shouts "AI is already a long-term demand": power equipment company Caterpillar soars to a new high for eight consecutive days) (Background supplement: Musk predicts that xAI will launch the first "AI-generated game" at the end of next year, and the game production model will be subverted? a16z General Partner Alex Rampell and Partner Justine Moore discuss how AI agencies are transforming commerce and their impact on Google's business model, affiliate marketing, online shopping, and more. Why think about the role of AI in business? Erik Torenberg: You've been thinking about the role of AI in business for a long time. Can you start by talking about what inspired this article and how did it come about? Alex Rampell: A long time ago, I started a company called TrialPay. In fact, I've been selling things on the web for a very long time, even before the web came along. I was thinking, what will happen to Google in the first place when AI comes along? Because many people are thinking about this question, is the search volume going up or down? Personally, my search volume is declining, but not commercial searches, but all non-commercial searches. This is one thing, and TrialPay is one of the largest affiliate marketers in the world. The essence of affiliate marketing is that you refer customers to others, which is the oldest business model on the web. It's actually a little bit earlier than AdWords and AdSense, and you direct someone somewhere, and you get a share of the proceeds, or you get a commission. But will the affiliate marketing model really be the driving force for the new business landscape? On the other hand, like these very expensive items. You're digging into them with AI. But there is no alliance model. So how do you conduct business and transactions? So, the first point is that the ontology of business is very interesting. The second point is, the whole alliance model, will it still make sense? Because this seems to be an area that ChatGPT and other companies are stepping into. Thirdly, this is just my personal behavior, and it's interesting that I probably use ChatGPT three orders of magnitude more now than I use Google. Justine Moore: There are several very large consumer markets right now, the biggest of which is probably online shopping. But I think so far there have been relatively few startups trying to get into this market with AI, although there are more opportunities now because with these very intelligent language models and agents, they can help you make better decisions than you can and even buy goods for you. You might think that this will create an opportunity for more people to package this content into products and then make it available to consumers, but we haven't seen many people doing that yet. So I think this part is about digging into why this system is so complex. In which different types of procurement AI can play a role. And what do you expect to see when thinking about the broader market? To what extent will AI enable dynamic pricing? Erik: Alex, can you imagine a world of extreme dynamic custom pricing? It's like seeing the same thing on Amazon, but it will charge you a higher fee, probably because you have more money. Alex: Well, I mean, people have tried it many times. From a beginner's point of view in economics, this is certainly a very intelligent world. For example, how do you get consumer surplus? Consumer surplus is good for consumers and bad for producers. Apparently, Delta has tried this as well, or they have tried to do so. It's like a poor man's version, for example, if you're using an iPhone, you should pay more than someone who uses an Android phone, because iPhones are more expensive. This actually shows that your elasticity of demand is different from those with less money. I think this could lead to regulatory challenges as a result. Of course, this will certainly be met with a backlash from the customer base, but someone has tried this practice. But usually, this approach is rarely successful. Why is e-commerce so small? Erik: Before we get into that, let's review previous platform changes. E-commerce currently accounts for only 16% of total retail sales. If you had discussed and predicted the share of e-commerce in business 20 years ago, you might have thought that the percentage was much higher. But why is this not the case? Alex: It turns out that the demand curves for immediacy and non-immediacy are different. I once served on the board of a company called Wise, and I observed that the remittance market that arrived immediately and the remittance market that arrived two days later were completely different. Because of what arrives two days later, the demand may not be so great. Surprisingly, you say only 16%, which seems very low. Justine: I think that's a high number. I'm not questioning your research on the data. I think so because I think there's a lot of these situations where people do their research online and then go to a brick-and-mortar store to buy it. Alex: Actually, the hardest part, and related to this topic, is attribution. It's a pain point for everyone, like, how do I assign attribution to Justine's MacBook sales? And I think the most pervasive and corrosive business model like this on the web is this "last click attribution." You give 100% credit to the last click. A lot of people simply can't tell the difference between correlation and causality and fall into this trap, and this business model is the one I hate the most, like Honey. It's the kind that will appear when you're already on the web page ready to buy. It will ask you, do you want a coupon code? Of course you want it, why not? Nine discounts. Click here. Just click here. It will redirect you to an affiliate page. It places a cookie on your device. It will then redirect you back to the page you were on just now, actually stealing the attribution of that visit. Why are e-commerce brands struggling? Erik: Let's look back at the industry. It seems that the biggest winners are at the aggregation platform level, like Shopify, or Amazon, and big independent brands like Allbirds or Casper. They seemed to quickly reap huge revenues, but they didn't become as durable as they had been. As they scaled, their performance didn't get any better. Can you talk about the industry in general and why this is the case? Alex: Maybe it's because you don't really make the product, it's usually someone else who makes it, and it's not a good business model. E-commerce brands are just resellers of goods, and I think that's the problem. As happened in the Internet 1.0 era, the long-tail effect of merchandise retailers is essentially starting to die. Justine: I think there are shoes or cosmetics like Allbirds that are very dependent on trends, especially in the age of the Internet. Nothing lasts that long. For example, one year Allbirds was a big hit...