{"id":32809,"date":"2021-06-11T20:58:40","date_gmt":"2021-06-12T00:58:40","guid":{"rendered":"https:\/\/mjtsai.com\/blog\/?p=32809"},"modified":"2021-06-13T18:19:27","modified_gmt":"2021-06-13T22:19:27","slug":"removing-the-manual-boost","status":"publish","type":"post","link":"https:\/\/mjtsai.com\/blog\/2021\/06\/11\/removing-the-manual-boost\/","title":{"rendered":"Removing the Manual Boost"},"content":{"rendered":"

Keith Collins<\/a>, in 2019:<\/p>\n

But as Apple has become one of the largest competitors on a platform that it controls, suspicions that the company has been tipping the scales in its own favor are at the heart of antitrust complaints in the United States, Europe and Russia.<\/p>\n

Apple’s apps have ranked first recently for at least 700 search terms in the store, according to a New York Times analysis of six years of search results compiled by Sensor Tower, an app analytics firm. Some searches produced as many as 14 Apple apps before showing results from rivals, the analysis showed. (Though competitors could pay Apple to place ads above the Apple results.)<\/p>\n

[…]<\/p>\n

Spotify complained to European regulators in March that Apple was abusing its role as the gatekeeper of the App Store. By April, all but two of Apple’s apps disappeared from the top results for “music.”<\/p>\n

An Apple spokeswoman said the company could not verify the data because it did not keep a record of historical search results.<\/p><\/blockquote>\n

Just like it has never tried to calculate whether the App Store is profitable<\/a>. Phil Schiller and Eddy Cue said that Apple wasn’t<\/a> putting its thumb on the scale. Rather, “the algorithm had been working properly. They simply decided to handicap themselves to help other developers.”<\/p>\n\n

However, the Epic trial has revealed an e-mail from Tim Sweeney<\/a> reporting that Apple’s then-new Files app<\/a> was ranking ahead of Dropbox when searching for “Dropbox.” In the replies from Apple employees, we learn from Debankur Naskar, the lead for “SEO experiences in the App Store”:<\/p>\n

I think the Files app was manually boosted on the top for the query “Dropbox” during last WWDC. We are removing the manual boost and the search results should be more relevant<\/a> now[…]<\/p><\/blockquote>\n\n

It would be nice if we could chalk this up to one rogue employee, but the e-mails show that Naskar’s boss, VP Matt Fischer, didn’t admonish him for gaming<\/em> the search results, but rather for doing so without first getting his approval<\/em> to do so. It sure doesn’t sound like Apple has ruled out such manipulation.<\/p>\n\n

Somewhere there’s probably a similar e-mail explaining Bing’s recent accidental<\/a> human error<\/a>.<\/p>\n\n

Sean Hollister<\/a> (Slashdot<\/a>):<\/p>\n

Apple tells The Verge<\/em> that what we think we’re seeing in these emails isn’t quite accurate. While Apple didn’t challenge the idea that Files was unfairly ranked over Dropbox, the company says the reality was a simple mistake: the Files app had a Dropbox integration, so Apple put “Dropbox” into the app’s metadata, and it was automatically ranked higher for “Dropbox” searches as a result.<\/p>\n

I’m slightly skeptical of that explanation — partially because it doesn’t line up with what Naskar suggests in the email, partially because Apple also<\/em> told me it immediately fixed the error (despite it apparently continuing to exist for 11 months, hardly immediate), and partially because the company repeatedly ignored my questions about whether this has ever happened with other apps before.<\/p>\n

[…]<\/p>\n

Besides, the distinction between a “manual” boost and any other kind of boost may be purely academic. Algorithms are written by people, after all. If Apple can build a 42-factor algorithm that gives its own apps favorable results, why would it need to override that algorithm and risk its emails getting caught up in a lawsuit years from now?<\/p>\n

It could just tweak that algorithm at will — which is exactly what it did to resolve the WSJ<\/em> and NYT<\/em>’s scrutiny two years ago.<\/p><\/blockquote>\n\n

Previously:<\/p>\n