Instead of fundamentally solving the problem of discovery, Amazon has laid a solution on it: advertising.

Platform, platform, platform

Bezos did not carefully assemble Amazon into the high-growth business aggregation it is today; he "just" designed Amazon's algorithm. His first act of genius was to make it unfettered; his second was to devise a solution to bureaucratic complexity that would otherwise implode. Amazon's disorderly expansion is no longer a bureaucratic burden, but a huge area of customer contact from which Amazon can derive dozens of sources of revenue.

With an established model for solving the problems and bureaucratic problems caused by unlimited shelf space, Amazon began to systematically eliminate growth bottlenecks. It found that sellers in the market were not particularly good at shipping directly to Amazon customers, which led to a bad experience for customers and a frustrating experience for sellers themselves. At the rate at which new sellers sign up, Amazon cannot convert all of these sellers into its supplier plans, nor does it have enough capital to carry all its inventory on its balance sheet; instead, FBA allows sellers to ship their inventory to Amazon's fulfillment center, giving Amazon complete control of the customer experience. It has the huge added benefit of honing Amazon's own rapidly expanding fulfillment networks, which are certainly at risk of being bloated and loose; like AWS, Amazon has become one of many "customers" in its own fulfillment center.

The platform has become Amazon's answer to every growth obstacle it has encountered. The platform becomes part of the algorithm. Sellers are limited by access to funds? Launch Amazon loan. Customers can only buy things in front of a computer or mobile phone? Establish an Echo. UPS and FedEx can only deliver within 24 hours? Launch faster Amazon Flex and Amazon Logistics.

Amazon has assembled a huge machine and deployed its algorithms over and over again, and its momentum is unstoppable. Every obstacle on its way was solved by one platform until one of those platforms led Amazon into a catastrophic mistake.

Advertising

Unlimited shelf space brings a question: how do customers find new products? If you think of Amazon's website as a physical retail store, in other words, how does Amazon decide which products are on the top shelf and which are on the bottom shelf in the middle of the aisle? For specific categories such as Aquarius, Amazon may have thousands of pages of search results.

Amazon will never be able to effectively plan such a large product category. It is not particularly good at selling, and even if it is good at it, it is impossible to build a team of salesmen large enough to plan such large-scale sales. Instead, Amazon relies on a ranking algorithm that largely weighs product reviews and sales speed. The more reviews and sales of a product, the higher its ranking. Of course, this creates a positive feedback loop: the more a product is exposed to customers, the more it sells; the more it sells, the more reviews it gets, the higher it ranks, and starts the cycle all over again. Yes, this is a serious oversimplification of Amazon's extraordinarily complex ranking algorithm.

This poses a big problem for Amazon's customers, who want the latest and best products, and sellers, who want to develop and sell exciting new products. If these requirements are not met, Amazon's dominant position in e-commerce will be threatened.

Amazon answers this question in a typical way: with a platform. Amazon ads allow sellers to launch "sponsored products" (Sponsored Products) (paid ads that appear at the top of search results). Sponsored products solve three problems at the same time: discovering new products for customers, introducing new products to sellers, and providing Amazon with net gross profit of $8 billion a year as an additional reward.

The problem with sponsored products is that sponsorship lists are actually not good for customers, they are good for sellers; more specifically, they are good for sellers who are good at advertising, but not for others. Paid digital advertising is a very specific skill; the chances that brands with the best products also happen to hire the best digital marketers or agencies are very low. In addition, buying a top position in search results is good for the product with the highest gross margin (the highest bid), rather than the product that best satisfies the customer.

The problem is further complicated by the inability of ordinary customers to distinguish between "organic" search results and sponsored products. Now, the first four results of Amazon's search are dominated by sponsored products, which means that Amazon's average customer is very likely to buy sponsored products. And because sponsorship lists tend to be highly profitable products driven by shrewd digital marketers, Amazon's customers are highly unlikely to buy the best products available in the market.

To be sure, very bad products will be poorly rated and quickly eliminated, but overall, sponsorship lists drag the average quality of products sold to mediocrity and further away from good products. This is bad.

Another way to frame the question: what are the benefits of sponsorship lists for you as an Amazon customer? The only answer I can think of is "find new products," but there is a better way to solve this problem; part of the solution is the life cycle of product reviews.

For high-sales categories, product reviews should not be permanent. After a period of time, product reviews should be deleted, and the faster the project is sold, the shorter the time. This has two purposes: 1. Ensure that product reviews are applicable to the latest status of the product (solving the problem that manufacturing quality increases or decreases over time); 2. Provide a level playing field for new entrants. For a fast-growing project, the deadline for comments can be set to, for example, a rolling six-month period; instead of competing with five-year cumulative reviews, a new entrant will only need to compete with the latest comments from the past six months.

Instead of fundamentally solving the problem of discovery, Amazon has laid a solution on it: advertising. This is usually a reversible decision, but the large amount of advertising revenue it generates may prove that it is impossible for a well-capitalized company like Amazon to become addicted. One way of thinking is that the $8 billion generated by Amazon advertising is about 1/3 of Amazon's entire R & D budget.

This may seem like a small footnote in Amazon's grand picture, but it is definitely a devastating mistake for Amazon's retail business. This is not "just" search results; search results are the entire driver of Amazon's retail engine. Keep in mind that in a world with unlimited shelf space, the ranking algorithm is actually the entire commodity sales strategy. The organic, customer-centric product ranking (the strategy to bring Amazon's retail revenue to $250 billion) has been permanently distorted. And everyone is praising Amazon for it.

Platform problem

More broadly, Amazon's mistakes show that something strange happens in the end when an unconstrained search encounters such a huge problem space. Amazon has opened itself to sellers'"games" as it builds markets and removes restrictions on its entry as a supplier. Another theory is that once the rules of a system are understood, it will play games according to the rules that have been created.

Amazon has unlimited shelves, constantly expanding and filling, unrestricted, and it is impossible for Amazon to control the growing universe it has created. Another way of thinking is that while Amazon's SKU catalog is growing, the best locations for popular categories are not growing; no matter how many water bottles are added to Amazon's catalog, the number of search results on the first page is fixed. However, the growing army of competition has put great pressure on the whole system; the company is always trying to bring down the current king from the mountain. Bad guys tactics inevitably emerge, and Amazon is in a constant war to keep its platform in line with its customer-centric mission. This is a war that Amazon can never "win"; all it can hope for is the strategy of trying to keep up with the changing bad actors, which seems to be a high requirement given the growing total number of sellers (call it the "IRS problem").

The hidden cost of SKU proliferation, then, is that the overall quality of the SKU category begins to be affected. This is not a particularly meaningful issue when it comes to categories such as phone cases, because the cost of buying suboptimal goods is small, but when customers are buying, even slight differences can lead to great frustration. If customers buy shoddy accessories for their sports cars, motorcycles, rifles, boats or other items related to passion or entertainment, customers who are a little annoyed with the shoddy phone casings will be even more annoyed. not to mention more sensitive categories such as food or health products.

With its marketing platform, Amazon has created a wild west for sellers, with huge profits and traps that follow. It continues to accumulate market sellers at an alarming rate. Amazon's next decade (in retail) will focus on taking advantage of this opportunity while mitigating all the extraordinary challenges that come from trying to maintain order in a non-human market.

It's easy to see Amazon's explosive market numbers and insurmountable success, but we're just beginning to see the price it pays by building a world with unlimited shelf space and opening its doors to anyone who wants to open a store. The daunting challenge for Amazon is to maintain its crown as the "most customer-centric company on the planet" given the market dynamics it has created.

In other words, Amazon hasn't figured out how to extend its internal incentive structure, which has been very successful in keeping the company focused on customers, to its external platform participants: sellers.

Antitrust problem

To put it bluntly, there are no competitors on the horizon, and without a particularly strong competitor, Amazon's mistakes are unlikely to cause any problems affecting its survival. The point is, two years ago, it was even hard to think of a theoretical approach in which Amazon might be caught; today, there is a gap, a real gap, a meaningful gap. At present, Wal-Mart is the only company with the resources to take advantage of it, but as a company, it has not yet found a way to accumulate the unfair advantage of the scale of the Internet.

I'm not saying this against Amazon. I'm a person who has been unabashedly bullish on Amazon, but for the first time in more than a decade, at least in my estimation, Amazon's expected value has fallen.

The question on my mind is whether Amazon has grown so big that hidden under the golden goose that produces 8 billion dollars of eggs and tens of thousands of new market sellers each year, it may ignore the mistakes of its retail business (loss of focus on customers). The most interesting unknown here is the threat of antitrust action across the industry.

We have seen Amazon's strategy when it sees an imminent regulatory threat; it voluntarily began collecting a sales tax in many states before states could force it to do so, and before provoking outrage from the populist movement, adopted a minimum wage of $15. With multiple headquarters in multiple cities and external access to all its key products and services, Amazon seems more likely to disassemble itself along its preferred route than if regulators forcibly dismantle it.

Review Bezos's decree of 2002: 1. All teams will provide their data and functions through interfaces in the future. Teams must communicate through these interfaces, 3. All interfaces, without exception, must be designed from scratch to be made available to outside developers. Those who do not do so will be fired.

If every meaningful Amazon product or service is provided to external developers through an interface, does it matter whether Amazon is a single company or many companies? Taking antitrust action to break it up has no effect on itself; this is undoubtedly designed.

To put it another way. I'm just as happy to own half a dozen Amazon shares as I am to own one, and maybe the former makes me happier because each of the resulting subsidiaries is smaller and more focused.

So, what is Amazon? It started as an unconstrained Wal-Mart, an unconstrained algorithm for finding global optimality in the world of physical products. It becomes a platform for adapting the algorithm to any customer-centric value creation opportunities it encounters. If it can figure out how to maintain the integrity of its incentive structure when it exposes itself through expanding external interfaces, Amazon (or its various spin-off subsidiaries) will dominate the economy for a generation. If not, it will become a seemingly unstoppable company until it is pulled down from the altar.

I still want to know what happens next.

Translator: Ti Kewei