Ryan Moulton, a Google Search engineer who works on Google's search rankings algorithms for over 18 years now, went to X to call out some of what the courts are saying as not true. He went on to say, "This is now the second time I've seen my work cited prominently in a major lawsuit against Google, and I have to say, it has not given me a lot of faith in the court process around issues like this."
The degradation of search quality aspect is super interesting to dive into as well.
As a reminder, the U.S. court system ruled Google a monopoly multiple times.
But I found Ryan Moulton's rant interesting because we don't often see technical engineers go off publicly about this. Here is what he said on X:
(1) "This isn't true. When search results are worse, people attempt fewer tasks. When they're better they attempt more."
This isn't true. When search results are worse, people attempt fewer tasks. When they're better they attempt more. https://t.co/XFOjN9HtjU
— Ryan Moulton (@moultano) May 11, 2025
(2) "Apparently the root cause of this article is that the judge in one of the antitrust lawsuits thought these numbers were small instead of big. (They're big.)"
(3) "This is now the second time I've seen my work cited prominently in a major lawsuit against Google, and I have to say, it has not given me a lot of faith in the court process around issues like this."
(4) "It's particularly Kafka-esque to use the experiments that the company runs because they're obsessed with making the product better and want to make sure they do it well to argue that they don't have an incentive to make the product better."
(5) "Web Ranking: Look how important we are! It's like we added a couple Wikipedias to the web!
Web Ranking (separately): Look how important we are! We cause significant user growth!
DOJ: What If we take the ratio of those two to say you aren't important?"
(6) "Also particularly perverse because the metric/growth ratio is comparing two extremely noisy heterogenous things, so the correlation substantially understates the underlying effects. It's not "how much does improvement/degradation matter" it's "how well are we able to measure it.""
(7) "Also worth noting that the firewall between search and ads is strict enough that I never learned about the effect on ad revenue (despite being the person who set up the experiments) since I work in search. I only learned about other metrics unrelated to money."
(8) "This is a really useful paper on how to measure this stuff. Worth paying attention to in particular that short term and long term measures of usage can be anti-correlated. "
He also responded to some questions after he posted:
Roughly. The translation to units of "removing Wikipedia" is supposed to give an intuitive feel for what the metrics based on human ratings measure. Nobody ran an experiment removing Wikipedia to my knowledge. But yes, 1% is basically zero does seem to be what he concluded.
— Ryan Moulton (@moultano) May 11, 2025
I don't know exactly what they documented in court, but it definitely was not anything equivalent to the statement in that paragraph. I would have to see a citation to figure out what they are talking about. I have personally measured this.
— Ryan Moulton (@moultano) May 11, 2025
I can't find any source for that text other than that article itself. pic.twitter.com/4G0tGX8IB3
— Ryan Moulton (@moultano) May 11, 2025
Oh hey! Here's the slide I worked on that showed enormous user growth from ranking improvements. pic.twitter.com/3cXUKEfvNj
— Ryan Moulton (@moultano) May 11, 2025
I think the first sentence in the paragraph is just made up. I don't know what they think they are summarizing but it's impossible to tell without a link.
— Ryan Moulton (@moultano) May 11, 2025
It's complicated. I think it's both higher expectations and a declining internet. People expect a lot more from their search results than they used to, while the market for actually writing content has basically disappeared. https://t.co/efLjvjNK8J
— Ryan Moulton (@moultano) May 11, 2025
FWIW people have been saying that Google has gotten worse non stop for 20 years. I think that's mostly because their expectations keep getting higher, even if there is substantial truth to individual complaints.
— Ryan Moulton (@moultano) May 11, 2025
Yeah, that I know less about and can't really speak to. I've only been tangentially aware of how it has developed over time.
— Ryan Moulton (@moultano) May 11, 2025
Search has a longstanding and well publicized strict firewall between search and ads such that people in search don't even see revenue numbers. They understood that changing ranking to influence revenue would be bad for the product and for the company in the long term.
— Ryan Moulton (@moultano) May 12, 2025
Yes, that's what that model would predict. I was just responding to you initially describing it as a miniscule change, and thought that you thought it was a 1% change from just doing it for a fraction of users, and maybe made things more confusing.
— Ryan Moulton (@moultano) May 12, 2025
It's complicated, and I can't share any details, but your intuition is correct that spending more searches and more time to accomplish the same thing is a cost, not a benefit, and we try to count it that way.
— Ryan Moulton (@moultano) May 12, 2025
That's exactly the sort of thing we work hard to prevent. Next time it happens if you feel comfortable, send me a query I can debug.
— Ryan Moulton (@moultano) May 12, 2025
The people who set up Google's basic principles understood that changing ranking to influence revenue would be bad for the product and for the company in the long term.
— Ryan Moulton (@moultano) May 12, 2025
Here is a screenshot of what he posted this past Sunday, just in case it doesn't load for you:
Forum discussion at X.
