Google Changes Search Queries to Show More Ads
Megan Gray (via Jason Kottke):
When you enter a query, you might expect a search engine to incorporate synonyms into the algorithm as well as text phrase pairings in natural language processing. But this overhaul went further, actually altering queries to generate more commercial results.
There have long been suspicions that the search giant manipulates ad prices, and now it’s clear that Google treats consumers with the same disdain.
[…]
Google likely alters queries billions of times a day in trillions of different variations. Here’s how it works. Say you search for “children’s clothing.” Google converts it, without your knowledge, to a search for “NIKOLAI-brand kidswear,” making a behind-the-scenes substitution of your actual query with a different query that just happens to generate more money for the company, and will generate results you weren’t searching for at all. It’s not possible for you to opt out of the substitution. If you don’t get the results you want, and you try to refine your query, you are wasting your time. This is a twisted shopping mall you can’t escape.
[…]
This system reduces search engine quality for users and drives up advertiser expenses. Google can get away with it because these manipulations are imperceptible to the user and advertiser, and the company has effectively captured more than 90 percent market share.
Previously:
- Unlimited Kagi Searches
- Google at 25
- The Enshittification of All Things
- Google Search Is Dying
- Google Ad Numbers Don’t Add Up
- Introduction to the Ad Auction
Update (2023-10-09): Nick Heer:
This article was understandably quoted and linked to across the web as its allegations seem to support complaints about a decline in quality of so-called “organic” search results while bolstering rumours that these results are influenced by ad spend.
There are so far no exhibits posted by the U.S. Department of Justice which make this claim. The best match seems to be UPX0204, an internal presentation called “Ranking for Research” (PDF). The version made publicly available is “redacted […] and abridged”, which suggests parts of a longer presentation were shown in court but were deemed too sensitive to show online. Hey, at least Judge Mehta is relaxing the trial’s secrecy a little bit. In any case, the last available page of that exhibit has, as part of a flowchart showing how search works, a “query rewriter”, but there is no suggestion anywhere that it is doing anything more than “interpret[ing the] query”. If I search for “children’s clothing”, I would want to see results from websites selling “kids’ clothing” and “child-sized t-shirts”, for example — “semantic matching” of the type suggested by the quote above.
Zoë Schiffer (via Jason Kottke):
“Google does not delete queries and replace them with ones that monetize better as the opinion piece suggests, and the organic results you see in Search are not affected by our ads systems,” the company told us.
I asked Google PR if they could share the trial exhibit that @megangrA’s Wired piece referred to (which this tweet responds to ⬇️). Here’s what they shared[…]
Via granzymes:
The author appears to have gotten the slide exactly backwards. She said the slide showed a query of “children’s clothing” that Google rewrites to be a “Nikolia kidswear” query so that it can sell more ads. But in reality, the slide is describing a fuzzy keyword matching system that takes a query of “Nikolia kidswear” and allows it to match ads with “children’s clothing” keywords.
I’m surprised WIRED allowed such an obviously incorrect article to be published in the first place, particularly when it was by a known partisan (the article discloses that the author is a former Duck Duck Go executive with an obvious bias).
When you see the slide, two things are clear:
Based on the template, this slide is probably not from the “Ranking for Research” (PDF) presentation. However, it is possible — Google’s presentations often mix and match slide formatting, and do not even have a consistent display of the company’s own logo.
The title of the slide posted by Kovacevich is “Advertisers benefit via closing recall gaps”. This is a slide about how advertising in Google results can target synonyms and contextually related phrases. It does not appear to relate to “organic” search results at all and, as I mentioned before, is a publicly documented feature.
Wired (via Hacker News):
WIRED editorial leadership has determined that the story does not meet our editorial standards. It has been removed.
Though it faced immediate skepticism and Gray presented no proof, the claim was widely re-published; it feels true. Despite days of questioning, the article stayed online without updates or changes — until, it seems, the Atlantic’s Charlie Warzel asked about it.
[…]
Gray also said nothing publicly in response to questions about the article’s claims between when it was published on Monday morning to its retraction. In an interview with Warzel published after the article was pulled, Gray said “I stand by my larger point — the Google Search team and Google ad team worked together to secretly boost commercial queries” — but this, too, is not supported by available documentation and it is something Google also denies.
Update (2023-11-22): Davey Alba and Leah Nylen:
Google maintains a firewall between its ads and search teams so that its engineers can innovate on Google’s search engine, unsullied by the influence of the team whose goal is to maximize advertising revenue. […] As part of the emergency, which lasted for seven weeks, engineers from Google’s search and Chrome browser teams were reassigned to figure out why user queries had slowed, according to the documents.
Via Nick Heer:
This and the related internal presentation (PDF) is the closest this trial has so far come to suggesting the kind of juicing reported by Megan Gray earlier this month. The slide on page five of that April 2019 presentation, for example, notes that while there is a “strong separation between Ads and Search”, those two teams are now “working together to recapture this commercial intent”.
