In the late 1950s, as American Airlines was looking for a new, non-manual way to take ticket reservations, it came across a product IBM was developing for the U.S. Air Force called SAGE, which helped the military better coordinate attacks. The airline realized that a similar version of this system could be used for ticketing, and it engaged IBM to build it. SABRE launched in the early 1960s, transforming American Airlines’ ticketing operations from manual to increasingly automated.

In 1972, American extended the system from internal use only to travel agents, allowing them to lookup flights from any airline and book the reservation for clients through SABRE. Though the company could have offered only American flights through the system, extending the booking platform to all airlines meant that it could own both a crucial piece of supply (flights) and the distribution channel that people used to buy them (SABRE), becoming the primary tool for many travel agents. By 1982, over 2,000 agents were using the system to book flights for clients.

But while American wanted most flights offered within its system, it didn’t want all of them to be displayed with equal weight. Soon enough, it started prioritizing American flights above others, often listing its own flights in the first line of any query, even if another airline had a faster or cheaper option It also banished competing flights from the system that served highly contested routes. American also paid travel agents higher commissions to those who booked the most aggregated business with the airline, a meaningful amount of money that further tipped the scales in American’s favor. While travel agents and other airlines picked up on this questionable practice, it went mostly uninterrupted until 1984 when the Department of Transportation banned airlines from inserting bias into the system. But bias persisted.

This story is not unique to the airlines, and aligns closely with how Amazon has approached its own growth, especially when it comes to prioritizing its own gains—and products—over the rest. Amazon’s brands and exclusive products often outrank competitors on its site and are advertised on the product pages of competing items—practices that echo American’s moves decades earlier.

Amazon, however, is taking on novel approaches to indoctrinating companies into its ecosystem, while promoting its own products and services above all else—behavior which is contributing to mounting criticism. Last week Amazon announced Moments, a service for companies that sell content, games and other digital services to offer specific digital or physical rewards to users who complete certain actions. Say someone watches the first three episodes of a show on Amazon Prime, the content creator could decide to send this person the soundtrack from the show as a way to encourage future consumption. Moments, in effect, is a tool that allows companies to build their loyalty program on Amazon’s back, which is exactly what it wants them to do. However, the risk for these brands is clear: Farming out a loyalty program to another platform further reduces them to yet another player in a commodity market, whether in content or clothing.

Amazon is in a challenging, but entirely self-inflicted, place where every new “innovation” it brings to market is seen as degrading healthy competition in online commerce. Not every platform has this problem—Shopify and Stripe, for example, are known to prioritize new products and services that, yes, make money, but also provide value to customers. Not everything they do fits the bill, but by and large these platforms don’t have the perception as a predatory player like Amazon does. What do you do when almost every move you make its viewed as an attack?

Everyone knows that Amazon calls itself the most customer-centric company on earth, and it’s clear that its merchants fall outside of this definition. While marketplaces usually have a bend to them—the open secret of Airbnb, for example, is that it’s built for hosts, not guests. But as Amazon continues cutting down merchants across its endless categories, and promoting its private labels in the void, this credo of customer centricity is starting to wear thin. Despite the inherent bias of marketplaces, they can’t be entirely lopsided because they still need supply and demand to function. Amazon is edging closer to forcing the dynamic out of balance, and potentially beyond reproach.

Amazon has escaped most regulatory scrutiny because it has focused on lowering prices and doing things that are “pro-consumer,” but as with the American case, that sentiment cannot last forever. Recent price increases at Whole Foods, the failed HQ2 bid in New York, further prioritization of Amazon’s own private labels and other developments are pointing to the slow collapse of this balancing act.