[Spivak]

I don't think free/paid is the thing that will stop the decline because the downfall is adversarial input. If Kagi got huge people would start trying to SEO against them and they would end up like Google. Because I really don't think Google is bad because of the ads, it's bad because people desperately want their shitty site to be the first organic result rather than pay for the ad spot.

Having a huge behemoth that competitors like Kagi can fly under the radar with might be the only way for any search engine to maintain quality.

[lolinder]

It's mostly an academic exercise since paid search will always be small, but I think both ads and SEO are at fault here.

Google profits from ad impressions and ad clicks. That is what they're incentivized to maximize, which means that for Google the ideal mix of search results is a few high quality ads on top of a lot of pages that are loaded with Google advertising. Far from having an adversarial relationship with SEO content farms, Google is symbiotic with them: the more time people spend on the content farms, the more ad impressions Google gets to sell. If the content farm doesn't actually answer the user's query and they return to Google for another round of first party ads, all the better.

They only need to provide relevant enough searches to keep people from trying a different search engine, which as we've seen isn't difficult.

[freediver]

What would be an adverserial tactic against Kagi and its business model? What would be the motive?

[squeaky-clean]

Make a ton of test websites to try and reverse engineer how Kagi ranks websites, and then prioritize that over actual good content. The same way people manipulate Google results through SEO.

I've been at a job where we did this with Google. We had about 20 websites and 40 ad campaigns on adwords for fake companies where you could buy used dirty dishes. We wanted a topic that would have zero real-world results aside from us so we could easily compare our hypotheses and someone suggested pre-dirtied dishes. It was hilarious and had almost no search results so we went with that.

Strangely enough after a few weeks we noticed eBay and Amazon were buying the top ad slot for those queries. They probably had some automated way of tracking new trending Google search keywords matching "Buy ____" or "____ for sale".

Seeing "Amazon is the best place to buy Dirty Dishes" as a top ad was also hilarious. In fact if you Google "Buy Dirty Dishes" right now, Etsy is the top result. "Check out our dirty dishes selection for the very best in unique or custom, handmade pieces from our dining & serving shops"

We took our findings from that "research" and applied them to our real customers websites.

[gwd]

> I've been at a job where we did this with Google. We had about 20 websites and 40 ad campaigns on adwords for fake companies where you could buy used dirty dishes. We wanted a topic that would have zero real-world results aside from us so we could easily compare our hypotheses and someone suggested pre-dirtied dishes. It was hilarious and had almost no search results so we went with that.

But... is that manipulation or genuine optimization? I mean, if there were people who wanted dirty dishes, and you were selling them, then helping Google help dirty-dish searchers find you is making Google better (and overall making the world a better place).

SEO is a problem when it causes mediocre, crappy, or completely irrelevant sites to be promoted over actually good sites.

[squeaky-clean]

We managed websites for hotels/inns/ other travel related things. We obviously did not sell pre-dirtied dishes and there is no one buying those.

We figured out what made some experiments rank higher than others, and then used those techniques on tourism related websites.

The point is our customer's websites did not rank higher based on actual merit. They ranked higher because we learned how to abuse Google. And no, the sites were not high quality. We shuttered out higher quality websites who were not abusing SEO.

[freediver]

Kagi's ranking algorithm is already public. Kagi downranks sites with ads/trackers and promotes sites that do not have them.

The motive people manipulate Google is financial gain, most commonly monetizing traffic with ads/affiliate links.

That does not work for Kagi, first because there is no financial gain, second the amount of traffic you would get would be a fraction you would get from Google and that is not going to change any time soon.

Not saying it is immune to manipulation, I would love to know what a good attack vector and motivation behind it would be.

[CSMastermind]

Presumably the same as it is for Google?

SEO spam doesn't exist because Google sells ads. SEO spam exists because there's money in getting people to visit your site and Google is the largest search engine.

SEO spam is also something Google themselves are already generally incentivized to eliminate. It's not like their ad model does better if users lose faith in their results.

[alex7734]

> It's not like their ad model does better if users lose faith in their results.

Yes but not really: if the results are so bad that they're useless, yes, the users leave, but Google profits maximally when the results are only barely good enough that you don't leave but still filled with a ton of ad-bloated crap.

[freediver]

Where do you users leave?

[lolinder]

Google generally sells the ads on the SEO spam sites, so it actually does profit when it serves those results.