Earlier quoted context omitted.
You're certainly welcome to disagree, but Google would be the first to tell you how relevant the number of pages indexed would be. When you actually measured the result relevance number, it turned out that all the major engines were pretty similar, and Google often was not the top one. More importantly though, the observed phenomenon was that no matter how bad the results you presented were, most people didn't switch…
I am no longer a Google fan. Haven't been for years so there's no bias. > More importantly though, the observed phenomenon was that no matter how bad the results you presented were, most people didn't switch to another search engine. That may be the case today because the results are similar now, but it wasn't the same case years ago. The ancient search engines prior to Google weren't much better than human curated d…
I'm sorry, I guess I inferred the wrong conclusion from your statements. I would argue that your perception of Google at the time would be a more significant factor influencing your perception of their performance, but that's actually largely beside the point.
> That may be the case today because the results are similar now, but it wasn't the same case years ago. The ancient search engines prior to Google weren't much better than human curated directories like Yahoo. Otherwise, why would anyone switch from them to Google?
I think your question at the end there underlines the central point of confusion here. While some people do indeed switch search engines (depending on how you measure, somewhere between 10-30% of users), the vast majority do not. Even when presented with what one would expect to be the worst search engine in the world (one that presented only paid search results), most people would not switch to another engine.
So the answer to your question is: most people don't switch from any particular search engine to any other search engine.
I would be the first to agree that there was definitely a window of time where Google was observably producing better search results than their competitors, and I'm sure you were amongst many others (including myself) who switched to them at that time... but again, that was a small subset of the population.
The search engine space was (and is) very competitive, and it did not take long for competitors to react and close the gap. When we measured performance in 2004, Google did not come out on top. Subsequent measurements produced similar outcomes, although occasionally Google would come out on top. Having more relevant results definitely lead to increased adoption, but only of a small segment of the much larger market.
Unless you count AOL searches as part of Google, Google didn't lead in market share until ~2004 (and it really wasn't until 2005 that they were ahead by more than a paper-thin margin). At that point, they had slightly less than 30% market share as compared to present market share which is closer to 90%. Even by 2004, the predominant way that search engines acquired new users was not through "switchers" changing search engines because of better results, but rather because of distribution deals (like the deal Google had with AOL) where a distributor would make a particular engine the first one users were presented with.
Think of it this way: how many times have you checked to see if other search engines were performing better than Google? After you switched to using Google, how regularly did you check to see if other alternatives were doing a better job? Most people I've spoken with never even heard of Powerset, let alone tried it to see if it served their needs better than Google. If consumers never check, it's hard to believe product performance makes a difference, right?
That's the reality of the search engine business. It certainly helps to be better than everyone else at search, but it is no where near the primary driver of market share.