We did a lot of studies of this, and one of the more
interesting ones was that if you told people they were
seeing Google search results
This is super-interesting; where might one read more about these studies?Also, when were the studies done?
I'm curious about the "when" because I think Google's reputation has slipped quite a bit.
In general though, consumer confirmation bias is not that surprising, right? Particularly with technology brands, it's pretty well established that confidence in the brand leads one to be more likely to attribute failure to one's own "mistakes" and therefore invest in further exploration of the product and/or modifying one's interactions with it. It also diminishes one's belief that another product would produce better results, therefore diminishing interest in exploring alternatives; it also leads one to perceive a product's performance as "better" when it is exactly the same.
But in 2001-2002, I remember Google having a significantly better search experience. And it was massively better than all it's competitors from the 90s.
A large factor were the snippets, which gave you a decent indication if the result was relevant before even clicking. Before the other search engines copied that feature, they would just show the contents of the meta description tag.
It also seemed to index a much wider range of sites and ranked them better.
Yup, I would say it was better through most a good chunk of 2002, which garnered it [checks notes]... about 16% market share at that point. If you include the AOL traffic, that'd be 21%. To put that in perspective, MSN's search market share grew by more than that entire share just in 2003.
> A large factor were the snippets, which gave you a decent indication if the result was relevant before even clicking. Before the other search engines copied that feature, they would just show the contents of the meta description tag.
Yup, the impact of that UX was huge. It really improved conversion rates. Everyone else figured that out pretty quickly and adopted that.
> It also seemed to index a much wider range of sites and ranked them better.
Really, it was the ranking function. As their competitors rapidly discovered, having a broader set of sites actually made having a good ranking function so much more crucial.
No way Altavista or Inktomi offered better result than Google. I was a heavy user of both, having adopted AltaVista mainly for similar reasons that I adopted Google, because AltaVista had way fresher and more extensive content than competitors at the time, it offered query modifiers, multi-lingual queries, and media search (images, etc), it even had Babelfish translation. It was obviously, unequivocally better than everything else before it. So the value proposition was unambiguous.
Google was similarly an obvious value proposition when it launched. PageRank was objectively better at returning the most salient results. AltaVista did ranking purely based on how the query phrase matched the index. As the Web grew, the amount of junk it would return piled up, it's easy to spam or game such a ranking algorithm.
As for Inktomi, you can read Inktomi's own employee account: https://www.searchenginewatch.com/2012/05/10/why-google-beat... Inktomi employees were using Google Search as their primary search engine. If you're not using your own product, you've got problems, how can you expect your customers/users to use it if you won't?
Yes, Google's brand is strong, and marginal returns in search quality mean most consumers won't notice the difference between engines, and will judge based on brand or UI presentation now. But to claim Inktomi and AltaVista matched Google Search results I think is rewriting history, when Google launched, the difference in quality was quite obvious and striking.
These days search engines have other requirements given how big the web is, how rapidly it changes, and how many adversarial attacks are made against it (blackhat SEO/webspam). You've got to have the infrastructure to index the entire web frequently, you've got to have a robust defense against spam, and you've got to have an even better ranking strategy beyond PageRank because a lot of the long tail may be good content, but doesn't have many authoritative sites linking to it.
There's still the possibility of someone upstaging Google, by building a search engine that not only indexes the Web, but has the reading comprehension to understand it, and understand queries from the context of gained knowledge. The first company to do that will kill Google Search, as the results will be self-evidently better. There are queries today that GPT-3 can do that Google can't, they're amazing when they work, mindblowing. If you can scale that to the Star Trek level computer, you win.
But you can't displace Google with a 2% improvement in results.
Yes, that would be a foolish claim. Fortunately I made no such claim.
The differences at launch were indeed striking, much as they had been with HITS. That was in 1998. Four years later, at the time the article you linked to was posted, Google's market share was... 16%. While at that point they may have dominated mind share and brand value, they didn't dominate the market. That would come later.
Google's situation in 2002 was that they were no longer such a clear cut winner. Per the article: "Google much like Inktomi, must re-evalutate what itβs currently doing... ...search is going strong but the Penguin update has raised a lot of questionable results to the surface". Their search quality was already faltering. Of course, in the next twelve months, they'd literally gain as much market share as they'd garnered in the entire history of the company (and yes, that was a one time leap in market share).
I'll say it again: most of the market share gains came from the distribution channels. The cost of getting customers to switch was gigantic compared to the cost of just presenting them with your search engine first. The number of people using the Internet was growing by leaps and bounds then.
> There's still the possibility of someone upstaging Google, by building a search engine that not only indexes the Web, but has the reading comprehension to understand it, and understand queries from the context of gained knowledge.
That's been tried and failed. The technical difference didn't garner the market share needed to support it. I wouldn't say it is impossible for it to happen, but the odds are long, and ultimately the market largely does not appear to care.
> But you can't displace Google with a 2% improvement in results.
I guess it depends on how you are measuring 2% improvement, but by conventional metrics of the time, a 2% improvement now would be well neigh impossible to achieve. Again, that's part of the problem: the head of search engine queries have effectively been optimized to the point where there is little room for improvement beyond tiny fragments of the tail that amount to a percent or two of the average person's search history.
If you look at the queries in GPT-3 that are used to extract structured data, or to synthesize answers, the technical difference is huge, but it just has a high failure rate.
The various vertical search systems are examples of this, but the key to a next-gen system isn't to hand-code such verticals, but to have the AI be able to create an infinite number of them on the fly. Vertical based on Julia code? Yes. Vertical based on Pokemon cards? Yes. Today these things are hand-curated collections.
There's also the general switch from searching for information, to doing things. We've gone from 10 blue links, to answers, to assistive functionality. The current state of assistive functionality is pretty shitty and has a long way to go IMHO.
Yup. Though there are other systems that arguably had more technical success.
> There's also the general switch from searching for information, to doing things. We've gone from 10 blue links, to answers, to assistive functionality. The current state of assistive functionality is pretty shitty and has a long way to go IMHO.
Yes, I agree there's a great deal of opportunity for technical innovation. I expect to see a lot of it from Google themselves. Just don't kid yourself about the impact of those innovations on market share.
No question about the "feel", but it's amazing how perception fuels reality. We did a lot of studies of this, and one of the more interesting ones was that if you told people they were seeing Google search results (regardless if they were), they'd be more inclined to rewrite their queries and look farther down in the search results, rather than try using another search engine. Beyond the obvious revenue & marketshare benefits, this helped Google two-fold technically: it gave them better data on query rewrites and search results. The hoops you had to jump through to match our exceed their performance were substantial.