Wired had an op-ed last week that apparently accused Google of artificially modifying search queries on the back end to add keywords that would increase the number of commercial results you see. They've since redacted the entire article:
Editor’s Note 10/6/2023: After careful review of the op-ed, "How Google Alters Search Queries to Get at Your Wallet," and relevant material provided to us following its publication, WIRED editorial leadership has determined that the story does not meet our editorial standards. It has been removed.
There's an archive of the post available. Google's Search Liason (an employee doing outreach / explanations of how search works) responded with a Tweet saying that the article was conflating ad-matching with matching for organic results, and that the systems were separate..
An ex-Googler asked Google PR to provide the referenced slide that is the source for the original article, which they did.
Slide Contents
Advertisers benefit from closing recall gaps
New matches for keyword*: +kids +clothing
| kids → children | kids clothing → kidswear | clothing → apparel/ outfit |
|---|---|---|
| clothing for young child | nikolai kidswear | creative apparel for kids |
| children's clothing in singapore | tj maxx kidswear | kids outfits |
| kids clothing canada | kids winter wear for girls | kids apparel in citywalk |
| best children's clothing brands | sean jean kids wear | |
| childrens beach clothes | kids wear online | |
| newborn children's clothing | kidswear outlet |
Note: Table is a sample of matches, not exhaustive.
* Includes both S&R, SNE, and SemPhrase & SemBMM matches (all are new).
Without the surrounding slides or accompanying presentation audio, it's hard to tell exactly what this slide is showing. The article seems to have interpreted it as Google effectively adding brand names like "sean jean" into a search that originally didn't have any, while the slide's intent seems to be more to show that matches that happen to have brand names in them already that would have been missed because they used words like "children" instead of "kids" would now be included. Whether it's one the other depends on some baseline understanding of what a "match" is in this system.
Regardless of the answer, Google's assertion is that this whole system is only used for selecting what sponsored ad to use, and isn't at all used for organic results in the first place. The slide isn't specific enough to tell one way or the other.
Tech discourse is rife with this kind of stuff. People assume that being published on Wired.com means the article has been researched and fact-checked, even if it's an op-ed by an external contributor. They see the headline "How Google Alters Search Queries to Get at Your Wallet" and not only do they accept the direct claim, they accept the framework under which the claim was made as well.
The headline implies something a little subtle: your search query is sacred and Google shouldn't be replacing anything in it without your consent. Your searches are a representation of your intent and it's wrong for Google to change that, especially to do so in order to profit. This is, of course, kind've ridiculous, because the "change" is happening within Google's own system for finding matches. Your queries aren't being posted anywhere or shared to some external audience that will misconstrue your intent. Google literally can't do a search in the first place without ingesting and splitting up your query a bunch of different ways.
But most people reading the headline won't be thinking directly about this and will just kind've accept in their head that "modifying my query == bad". Their mental model of searches and how they work (or should work) has shifted in a way that doesn't actually match reality.
This is ultimately the same as those myths and rituals artists believe about posting on Twitter and working with The Algorithm. Stuff like: you will get de-prioritized if you post direct links to your Patreon, so instead you should subtly reference "you know where to get the full art" and only link to Patreon on your bio.
Without a full, accurate understanding of every detail about how feeds are constructed on Twitter, the only way to discover whether things like this are true is through experimentation, and Twitter (and most websites) do not give individual users enough access to run rigorous experiments because there are too many confounding factors. A single test with two different posts on your account can vary wildly depending on the time of day, day of the week, what news was happening that day, what similar users were posting, what country you live in, what language you're posting in, and a billion other factors.
The only way to get reliable data on whether something as small as "post contains a patreon link" has an effect would be to run controlled tests across thousands of accounts with identical post content, and even then that test could be invalidated by Twitter identifying the test itself as a spam attack and de-prioritizing all posts with similar content.
This does not stop folks from writing tools for testing if your account has been marked as bad by The Algorithm, and other folks with large followings posting the results of those tools and treating them as unimpeachable. And soon it becomes "common knowledge" about how The Algorithm works that everyone knows and works around.
The point is: These systems are built such that external people almost never have the ability to reliably determine how they work, yet we are constantly making claims to know how they work as part of discussing them. We can't trust claims made by the companies building the systems themselves because they have an interest in obscuring or lying about how they work for their own profit. Instead the best we have is the stamp of authority from news outlets by Wired that whatever story is trying to explain how a system works can be trusted because they probably fact checked it, whatever that means.
I'm not saying we should uncritically trust claims from companies or employees that build these systems, because that incentive does totally exist. But I will say that we aren't critical enough of claims from people external to these systems, especially when they match our preexisting views like "Google is a dog chasing advertising bux".
Which, to be clear, is broadly true. We don't need a little factoid like this to see that when we have very visible and obvious things like their moves against ad-blocking in browser extensions, their placement of ads on search pages, their obfuscation of where ads come from and who they're being shown to, etc.
But when we amplify these claims we're chipping away at a proper understanding of how these systems work, which piles up over time until our collective ability to be critical of them is hurt by our misunderstanding of how they work in the first place.
