Hacker News Reader

Comment by whakim | Hacker News Reader

It seems weird to me that Google would take "search quality" seriously for 15+ years (including nearly a decade as one of the biggest companies in the world) and then suddenly stop. Are you able to share any of those objective measures of quality? Because it seems to me that most of the discussions around the declining quality of Google search amount to anecdata backed up by reasoning that doesn't really make a lot of sense to me (e.g., "Google only cares about short-term revenue!")

View on HN

> It seems weird to me that Google would take "search quality" seriously for 15+ years (including nearly a decade as one of the biggest companies in the world) and then suddenly stop.

Is anybody claiming that it’s sudden? Everybody I’ve seen complain about Google search results has seemed to think that it’s been getting slowly worse over a long period of time.

It seems Google started decaying when they gave up their social platform G+

Nowadays all those tweaks to the algorithms, specially search and gmail spam filters, are like a desperate attempt to make the ship move faster by removing parts of the hull while keeping the water out. It will just sink at some point.

I don’t think google search has gotten worse, I think the SEO abusers have just started winning. So much content on the web is automatically generated and made to look just like real hand crafted content. So much of it has titles and headings which have little relation to the search.

I was recently searching for a piece of ceramic cookware, specifically looking to avoid non stick coatings. And google search showed me lots of listings with my exact search terms, but when I click the page it shows their generic product range which has nothing to do with ceramic despite the title saying so.

>I don’t think google search has gotten worse, I think the SEO abusers have just started winning. So much content on the web is automatically generated and made to look just like real hand crafted content. So much of it has titles and headings which have little relation to the search.

right, I mean we have nearly daily posts here about almost human level quality texts you can generate with various machine learning solutions.

Add to that sites who serve users needs must focus on doing that, the sites that just game Google search rankings can focus on that.

I mean maybe the problem is that the SEO abusers are only 35% of the way towards being constructive, a la XKCD https://xkcd.com/810/ - so the thing that wasn't considered is as your spammers get closer and closer to being constructive the more stress is placed on the systems that have to deal with fighting non-constructive spam and the worse it is for the users.

But maybe at some point there will be a breakthrough with SEO abusers being able to be as constructive as the actually useful sites, at which point there will be a short time when the SEO abusers are the only sites returning results, causing all other sites to go out of business, causing the downfall of the SEO sites which rely on actually useful sites for a great deal of their stolen content / content seeding corpora.

I know another possible explanation, way simpler.

Vast majority of these SEO abusers are monetized through google ads. For every advertisement dollar spent through google, google takes 30-50%. And 80% of Google's income is from internet ads.

I would be surprised if google's search quality would not deteriorate - they have very strong incentive for that.

I still get strange results from flaky websites that have strange TLDs and the content is literal garbage. Like just words mangled together. Ads on there are the worst and reach into malware territory. Results like these are on the first page. I have no explanation for it, but 100% of my searches for personal use are DDG or Bing now.

Not sudden for me. It has been progressively worse over time, and unfortunately the competition is not much better.

Searching today is not as bad as it was during the Altavista/Netscape/56k modem era, but it is getting dangerously close, from my personal and anecdotal experience.

i has felt quite sudden for me ~ within the past year

> Is anybody claiming that it’s sudden?

I mean, there are two ex-Googlers in this thread claiming firsthand that search quality was considered sacrosanct as of a few years ago, so it seems like it'd have to have been sudden if indeed it has happened at all.

Four and a half years is quite enough time for search results to get worse that I wouldn't call it "sudden".

I didn't notice a rapid decline until early 2019, when the search results stopped containing the keywords I used. Before then there was a blend of synonyms and actual keyword matches.

This was roughly when Google's new AI began "interpreting" the "meaning" of the specific jargon and product serial numbers I was inputting, and then decided what I actually needed was song lyrics.

Your post exactly.

> decided what I actually needed was song lyrics

Or random gossip about media celebs. I'm searching for SCSI cable stuff and getting Kim Kardashian (whoever the #@$%@ that is).

True, but I didn't assert "it was great right up to the moment I left."

The process might have started before, or indeed after I left. And I wasn't in Search, anyway.

Well, it's possible they attempted to keep search quality high, and they were unsuccessful but they thought they were successful.

After all, if me, my team and my boss are rewarded when an imperfect measure of search quality goes up and it's going up, why would I fight to switch to a different measure that wasn't rising?

Could just be that sundar pichai wants a bigger paycheck and tunes the knob on more ads money but less good search results

To me, it seems crazy that a company would actively try to self-sabotage their bread-and-butter product in search of a bit of extra revenue when they're already making record profits. Not to mention that, as the author of the article points out, Google published a well-known quantitative paper on why that's a bad tradeoff.

It happens all the time though. Look at amazon, they flood their marketplace with cheap knockoff products that hurt the credibility of their ecommerce offerings, but the money generated is too much to shut off. So in the short run, amazon makes more money for a worse reputation, and in the long run ??? its unclear

I don't think that's analogous: you'd have to argue that Amazon spent a ton of time and effort on keeping cheap knockoff products off their marketplace and then did a sudden u-turn and decided not to do that anymore. Even then, I don't think the analogy holds, because the quality of Amazon's marketplace listings isn't anywhere near as synonymous with Amazon as the quality of Google's search results is with Google.

>Even then, I don't think the analogy holds, because the quality of Amazon's marketplace listings isn't anywhere near as synonymous with Amazon as the quality of Google's search results is with Google.

What? Why not, and anyway Amazon I'm giving money. I'd think that quality there would be more deeply linked than anywhere I'm just giving some time.

My dad frequently says that sooner or later the finance guys get in charge of the company and then they cut the fat and keep cutting into the product. Sooner or later they go too far. Seems fairly inevitable, seems to happen a lot. I assume that what happens is that the product matures, so the obvious thing to do is make things more financially efficient. Problem is, the company sells a product and the finance CEO's core competency is reducing costs, and reducing costs isn't something that customers actually buy.

The CEO and upper management might simply not care about the long-term viability of the company. Cashing out in the short-term sounds more like what these executives generally do.

> Cashing out in the short-term sounds more like what these executives generally do.

It only sounds like what they do because you only ever read stuff written by people who believe that without any evidence.

Sundar Pichai has been with Google since 2004.

Sundar Pichai is probably the weakest of the FAANG executives - if not one of the weakest Fortune 100 executives. Google's search result quality, reduction in overall prestige and lack of new major products are just a few indicators of that.

But speaking in general, there is limited correlation between executive compensation and company performance. You can DuckDuckGo it if you want to find the evidence for this statement.

For whatever it's worth: I had a talk with a WSJ reporter, who said he's met all the FAANG chief executives, and Sundar was the least impressive of them all.

Having everyone you've worked with love you and find you brilliant is not at all the same thing as being brilliant.

It's rational and commonplace behavior for a monopoly to prioritize profit maximization over product quality, there must be thousands of case studies over the years. If a firm has no viable competitors, why do they care whether their product is good or not?

Google Search has 86% market share in the US.

The cases you mention involve maximizing profits by saving on costs. The theory above is that lower quality leads to an increase in revenue, and it is tenuous at best.

That happens when companies have monopolies. They can sacrifice quality for profits. Happens frequently. Obviously you can't sacrifice quality to zero, but you can a bit. It would make sense that they would keep pushing it down little by little (returning higher and higher profits) to find the turning point. I don't think they're yet in danger though because there isn't a real uptick in the competition.

> you can't sacrifice quality to zero,

GM tried in the 80s

Key word: "would". We don't know and I highly doubt this is the case.

Reply on news.ycombinator.com