Reuters’ flawed attack on Amazon is weak

By Mark Jamison of AEI.

"These days, mainstream journalists are taking aim at Big Tech. The Wall Street Journal conducted investigations critical of Facebook, Alphabet Inc., and Amazon. Politico obtained leaked internal Federal Trade Commission (FTC) memos and published criticisms of the FTC’s decision to not prosecute Google. The House Judiciary Committee’s majority staff report arguing for regulations and breakups of Big Tech relied heavily on journalistic input. I have written about some of these investigations, explaining that the actual information found did not support the outlets’ conclusions.

Reuters is the latest outlet to take shots at Big Tech, namely at Amazon’s operations in India. The headline of Reuters’ latest “special report” reads, “Amazon copied products and rigged search results to promote its own brands, documents show.” Copying others’ ideas and rigging results sound evil. However, as with other journalistic Big Tech investigations, the headline and summary statements overly dramatize what the investigation actually finds.

Let’s begin with the copying claim. The journalists say they found that Amazon employees conducted detailed studies using “proprietary data” on products sold on Amazon.in to identify which were best for customers and how these products could be improved. The authors leave readers with the impression that the data belongs to the third-party sellers, but the article doesn’t directly say that. It then states that Amazon employees used their knowledge to create Amazon products that “equaled or exceeded the quality of the competing brands but were 10% to 15% cheaper,” which I am sure consumers appreciate.

Assuming the article is accurate, Amazon’s conduct is wrong only if Amazon violated non-compete clauses or intellectual property rights, if the data was used in violation of agreements with third-party sellers, or both. The article never states that either occurred; it only directly says that Amazon knew more about customer behavior than third-party manufacturers did and that at least some of the third parties were unhappy competing with Amazon. This is normal in retail business.

It appears from the article’s factual statements that all Amazon did is learn and then innovate, which is both legal and good for consumers. It is almost always true that retail outlets know more about customer conduct in their stores than do the product makers, even in brick-and-mortar stores. For example, The Guardian reported in 2013 that grocery stores gather extensive shopper data and use it to sell product research to suppliers. Indeed, Reuters all but admits that Amazon’s tactic is a standard, profitable retail practice when it said in a paragraph about other retailers: “Private-brand products often have higher profit margins than normal retail brands because production and marketing costs can be lower.”

Absent Amazon being deceitful or breaking commitments, it is hard to see how learning and giving customers bargains is a bad thing.

Now let’s turn to the Amazon search issue. Reuters accuses Amazon of “rigging” and “manipulating” search results. This begs the question: Compared to what? They don’t offer a clear or objective standard, and for good reason. Search is complex and hard to do well, which is why Bing, Yahoo!, and others struggle to outcompete Google search. Also, it is dynamic: The pursuit of better quality is an essential feature of creative destruction.

According to the article, Amazon says its search results are unaffected by “whether such products have private brands offered by sellers or not.” That is probably true because that form of favoritism could be unprofitable.

Consider this: Suppose you spent millions of dollars building search algorithms for an e-commerce platform that you own. You would want the search results to improve profits. There are at least two ways of doing that. One is to guide each customer to products he or she is likely to buy, and of those products, promote the ones that have the highest profit margins. Reuters admits that private-label products often have the highest margins, so it would be no surprise if Amazon promoted these. But this would be based on profit margin, not private-label bias.

The other search algorithm consideration is the customer’s buying experience. Amazon’s search results influence whether customers search on Amazon. A 1996 Harvard Business Review article explained that a retail outlet has to carefully balance its mix of private-label and brand-name products to remain attractive to customers. It may be tempting for Amazon to simply favor its own products, assuming they provide the highest margins, but this could be a losing strategy if it attracted fewer customers.

Based on the factual information Reuters presents, its findings appear overdramatized. That is understandable: As historian C. John Sommerville notes in his book “How the News Makes Us Dumb,” news outlets use drama to attract readers. Fair enough — but perhaps they could find stories with actual drama rather than dramatize and demonize practices that actually benefit customers."