Hacker News

points by 1970-01-01 13 hours ago | 0 comments

A few days ago a Google found me something I was searching for, but via ads and not search. In the middle of searching raw part numbers, the ad became 100% relevant. It was a weird experience clicking on the ad; something I never ever do. So maybe this is a bit of good news? Yardinflatables.com if you're curious.

[–] chamomeal 13 hours ago

That’s wild. It’s possible that Google shows me relevant ads. But if it’s formatted anything like a sponsored result, I think my eyes just skip past it. Like after so many years I just have ad block built in to my retinas.

I wonder if one was actually useful, maybe I’d notice it?

[–] delegate 11 hours ago

We easily adapt and block out stuff we don't want to see. In big cities we do this automatically - we don't even notice the dirt, parked cars, litter, people.. unless we consciously focus on them. I guess it is the same with ads, they lose their power as we adapt to them.

The solution is to force ads upon you and hope that you accidentally click on one due to sensory overload. Or more ads.

[–] gumby271 12 hours ago

Yeah Banner Blindness is definitely a thing, it's something to keep in mind with UI design where you can end up making something so in your face that the user just ignores it even harder. https://en.wikipedia.org/wiki/Banner_blindness

[–] immibis2 13 hours ago

If it was actually the best result for you, they wouldn't need it to be an ad.

[–] 29083011397778 12 hours ago

Competitors on Apple's App store can pay to show up when consumers search for the app they're after. To counter-act this, some apps will pay to have their app show up, as an ad, after you search explicitly for them

Meaning that, due to and capitalist incentives and double-dipping from tech companies, what you're searching for may actually be what's advertised

[–] solarkraft 12 hours ago

Oh yeah. I recently started with engaging with a vendor that has good SEO and I would just instinctively skip past their search result because it was one of those expanded ones.

[–] ModernTimes 13 hours ago

Google are removing vast amounts of pages from their search indexes, leaving only "primary sources" that AI chat can reference.

So if you have a long-tail page ("obscure part number") then the only way to show it is buying an ad for it.

[–] Sopel 13 hours ago

Is this related to the fact that I'm regularly unable to find many github repos via google?

[–] scottyah 13 hours ago

I think github hides repos from indexing by default.

[–] sznio 12 hours ago

that never happened before. and kagi still shows them

[–] creato 12 hours ago

Kagi might be (intentionally or accidentally) working around GitHub's attempt to prevent indexing of repos.

[–] bigwheels 13 hours ago

A questionable move on Google's part, if true. Can you share any evidence?

[–] shevy-java 11 hours ago

Evil is what Evil does.

Does any of us have any insight in Google manipulating the world?

[–] DeluluDon 9 hours ago

Other than it's leading to their losses, no.

[–] Xenoamorphous 13 hours ago

I’ve always wondered what % of ad clicks are just misclicks.

[–] netcoyote 13 hours ago

When we created battle.net at Blizzard for Diablo, there was a bug where the game would launch in full-screen mode and hang, and the only way to escape back to the Win95 desktop was to click the ad displayed by battle.net.

We weren’t aware of this.

The ad company we used notified us of this issue and gave us three days to fix it. Afterwards our click-through rate went from something like 3% to .25%.

[–] Gareth321 13 hours ago

That's hilarious. I miss the old Battle.net.

[–] PaulHoule 13 hours ago

"All clicks are click fraud" is related to the dead internet theory.

[–] saghm 13 hours ago

I get what you're saying, but fraud is different from accidents by otherwise real people.

[–] PaulHoule 12 hours ago

Well... I'd say that

Many of those "accidents" are not really accidents: I've seen two kinds of ways sites engineer those "accidents"

(1) many sites, like phys.org, have crowded and confusing layouts that try to trick you on clicking on an "x" or something to close an ad but really clicking on the ad, there's a lot of deniability on this on mobile where the touch accuracy is poor

(2) a few years back there were a lot of sites which seemed to always change the layout when you were trying to click on something in the sidebar, anandtech was one of the worst, i could swear they were watching the mouse events and predicting the click events. this stopped all of a sudden, could be Google put the smack-down on this behavior

[–] deltoidmaximus 12 hours ago

I've long suspected Amazon's search is so crappy because it makes their ads necessary to actually surface a product. I guess Google got the idea as well.

The worst part is I've seen search other video sites that used to surface exact titles stop working around the same time. I'm not sure how to reliably find things I know are there anymore.

[–] QuantumGood 11 hours ago

I've had so much greater success using the built-in AI ("Alexa for search") that I no longer use the regular search except briefly. I imagine it will be destroyed eventually as regular search was, but for now it's dramatically better.

[–] burningChrome 12 hours ago

>> I've long suspected Amazon's search is so crappy because it makes their ads necessary to actually surface a product.

I thought it was just because they were forcing third tier Chinese knockoffs for everything I searched for. After a few months, I gave up completely. I now look for product reviews on YT, THEN go back to Amazon and search for exactly what I'm looking for. Even then? The search results are still jammed with cheap knockoffs.

[–] cobbzilla 13 hours ago

Fascinating; they must be bidding on the part numbers as search keywords.

Now I wonder what the median CPC is for part numbers (I’d guess very cheap), and which one is the most expensive (no idea, super curious).

[–] Aboutplants 13 hours ago

I’ve had this exact experience a couple of times in the past few weeks. Their “sponsored” ads of a kind of product that I was looking for (based mainly on description for the use I was intending) gave better results than any AI model I tried to use to find a product that fit my specific use case.

Found it interesting that I’ve flipped my search tool usage in such a short time. I use models for most “searching” but then utilize Google Search as the last resort.

[–] schainks 12 hours ago

This is exactly what the ad network is meant to do, and their ability to monetize this ability means competitor's ads (like meta) become less valuable.

[–] lelanthran 12 hours ago

> So maybe this is a bit of good news?

How is it good news?

Consumer searches for $FOO. $FOO does not show up in the search results.

At this point, the search is useless to users.

[–] ares623 11 hours ago

When StackOverflow tells question askers that their questions are off topic, everyone loses their minds and cheer that StackOverflow is now dead.

But when an algorithm does it, they go "wow such future"

[–] jedbrooke 13 hours ago

Is this how we fix google search? Make advertisers have to provide ads that are actually helpful search results, sort of like https://xkcd.com/810/ but for search ads

[–] barbazoo 13 hours ago

I'd still try to get to the result a different way, even if it involves doing another search. Same with Amazon.

[–] blks 12 hours ago

For me it didn’t find a file on GitHub with exact file name, but found it via llm summary.

[–] francispauli 13 hours ago

the bigger question is why the heck did you use Google, was it only after you exhausted all other LLM options

[–] shevy-java 12 hours ago

Google successfully trained this human to click on ads.

Thankfully I never see any ads - ublock origin kills these off. That includes the horrible UI choices Google made on its search interface. A custom ublock origin rule filters away the pointless "others also searched for pink cats" while I was searching for epic skyscrapers. Evidently Google tries to link-lose me here, wasting my time.