A Big Tech crackdown could just make the Internet worse

Innovation would suffer if Google and Amazon had to justify every service they add to their platforms.

Adobe Stock

Last month, a bipartisan group led by Senators Amy Klobuchar of Minnesota and Chuck Grassley of Iowa announced a bill that would, in many cases, ban one of the most common practices of large tech platforms like Google, Amazon, Apple, Microsoft, and Facebook (which now wishes to be known as Meta). The bill is being touted as a way to protect consumers from the power of Big Tech, but it would actually make things worse for users. It would undermine the trial-and-error process that has built the Internet we know today.

A lot of the value we get from digital platforms like the ones run by Amazon, Apple, and Google is from the combinations of services they offer. If you Google “dim sum near me,” you’ll get both a list of search results and a Google Maps box showing you the closest dumpling restaurants. When you take a photo on your iPhone, it automatically saves it in iCloud. You get the map and the cloud backup without needing to think about it.

While users might like these bundled services — known broadly as “self-preferencing” — they’re unpopular with competitors whose products get sidelined, such as Yelp and Dropbox.

Lawmakers in the House originally proposed banning self-preferencing outright. That probably would mean, for instance, that Google couldn’t link to Google Docs from Gmail and Amazon couldn’t sell its own brands on its website.

To its credit, the new Senate version offers an out clause. Bundling of services would be allowed as long as a company establishes that doing so is necessary to “enhance the core functionality” of a platform. That means it’s possible that Google could still include a Maps box at the top of a relevant search results page and that Amazon could keep selling Amazon Basics batteries.

This might sound like a reasonable compromise, but it misunderstands the trial-and-error process of innovation. Companies have ideas about what consumers might like, and they try those ideas out. The ones that consumers like are left standing while the others go bust. It’s often difficult for a business to know exactly why its product is successful. Requiring that a business must prove that a given feature improves its service would foreclose much of this experimentation, leaving both consumers and competition worse off.

After all, some of the most successful business ideas have come from divine inspiration or just dumb luck. When 3M tried to develop a super-strong glue, it ended up with one that barely managed to stick at all — and thus the Post-it Note was born. Steve Jobs eschewed market research in favor of insisting on what he believed customers would want. Turns out, he was right.

Customers themselves often can’t explain why they prefer Popeyes to KFC or red cars to green ones. But they don’t have to — the act of buying a product or using a service demonstrates that they prefer it. Markets function because businesses don’t need to know why customers want their products. Prices and profits tell them to keep doing what they’re doing. Anyone who has pitched a business idea (or watched “Shark Tank”) understands that even if you know why your idea is good, persuading another person of that fact is a whole other matter — let alone convincing a court or a hostile federal agency.

That’s what makes the new Senate bill so harmful. Under its approach, Google would have to explain why some practices were necessary to improve its products. If the company failed to do so, it would face fines of up to 15 percent of its total US revenue. This huge risk would, in many cases, make it not worth experimenting in the first place.

The bill’s authors might respond that MapQuest and Yelp have a chance to compete only if Google is forced to list their services alongside or in place of Google’s own Maps and reviews boxes. Under the bill’s logic, the more choices consumers have at every moment, the better it is for competition.

That isn’t how most consumers see it. Choice is valuable, but it comes at the price of complexity. We don’t want to deal with deciding among a million ice cream flavors every time we go to the grocery store; thus, through trial and error, stores narrow the selection to a reasonable number of options. If Walmart had to explain why it was necessary to stock coconut ice cream but not pineapple, or one brand over another, could it do so? If it stood to lose 15 percent of its total revenues for failing to convince a court that one flavor was necessary but another wasn’t, how much would it experiment?

Given the abundance of choice and information on the Internet, online platforms filter choices and information to reduce the complexity and noise that consumers would otherwise have to navigate. YouTube and TikTok, for example, display videos that are relevant to specific users, and most of us tend to like them.

None of this implies that companies should get away with behavior simply because they are ignorant of potential harms. The law’s current approach correctly makes some activities illegal on their face, such as price fixing by cartels, because that behavior is virtually always harmful to consumers. But otherwise, it is up to an antitrust plaintiff to show that the activity harms competition. The risk of losing beneficial behavior is otherwise just too great.

The principle that businesses, like citizens, are to be considered innocent until proven guilty comes from the fact that regulators, prosecutors, and courts often get things wrong and so we should be cautious about giving any of them the power to dictate in advance how private businesses operate. If we throw out that principle just to punish Big Tech, it’s users who will suffer.

Brian Albrecht is assistant professor of economics at Kennesaw State University, in Georgia. He writes a weekly newsletter on basic economics at pricetheory.substack.com. Follow him on Twitter @briancalbrecht. Sam Bowman is director of competition policy at the International Center for Law & Economics, in Portland, Ore.