3 ms·

The real mystery is why nobody has ever been able to do better search than Google, or at least just as good. Of course search is really hard, but how is it pos

by bambax 6y ago

The real mystery is why nobody has ever been able to do better search than Google, or at least just as good. Of course search is really hard, but how is it possible that huge companies with nearly infinite money and talent (MS / Apple / Amazon) were never able to do it? MS actually tried; it's unclear if the others even gave it a thought. The rest of Google is different. I use Gmail but could easily switch. I've not used Chrome for over a year now and really don't miss it (either on desktop or mobile). Google maps are the best but others would be good enough if need be. But search? Every time I try to use Duck Duck Go I'm extremely disappointed. Google search is something I don't think I could live without.

mjburgess 6y ago
I suspect people optimising for google over the last, c. 20 years has made the difference. ie., people have spent years writing their sites explicitly with google SEO in-mind... add-in all the data google has via adwords...
- why_only_15 6y ago
  Why can't other search engines just hook into those same signals though?
  - dmitriid 6y ago
    Because they don't have the traffic/market share to do that. And it's not just search. Every single Google property ultimately feeds into Google's search engine.
    - why_only_15 6y ago
      The traffic and market share shouldn't matter if these are SEO signals inside the application itself, like in the HTML. Those are independent from traffic.
lazyjeff 6y ago
I did my Ph.D. in information retrieval and had studied the workings of the 3 major search engines from the inside during various points about a decade ago. Without giving away any trade secrets, part of it is the power of search behavior data (for example, what you click on after a search, how long before you go back to the search page, etc.). The more popular your search engine, the more search behavior data you have, which makes the search results better especially for tail queries (the less common searches). So then you attract more searchers, and it's a cycle that is hard for a newcomer to break into. It's why academia has given up on competing with commercial search engines in the web space, and why alternative search engines are just begging at the table right now.
dmitriid 6y ago
> The real mystery is why nobody has ever been able to do better search than Google, or at least just as good. Google's search is the way it is now because they've now had 20 years to work on it. To create search of similar quality you'll need to spend a comparable amount of time (during which Google will not rest on laurels, but will continue to evolve). It would've been great if someone took the pains to implement, maintain and evolve a search similar to Google's at the time when Google became available (or a few years after that). Then it's possible we would have two high quality search engines now. But even that is a simplistic view of things. Everything that Google ultimately has feeds into its search. Google Translate? Helps Google understand pages in multiple languages. Google Maps? Helps Google learn people behaviour, places of interest, businesses etc. Google Patents, Groups, Books? Provide vast amounts of natural text to train models on. And so on and so on. A "just as good" search needs all (or nearly all that) that to become just as good.
alexyz12 6y ago
according to the innovators dilemma you would have to find a market that google isn't serving with its search, use relatively lame search technology to serve it, then scale upwards into google's market from there. Maybe because google's market is everyone, disruption can no longer exist?
freediver 6y ago
> Of course search is really hard, but how is it possible that huge companies with nearly infinite money and talent (MS / Apple / Amazon) were never able to do it? MS actually tried; it's unclear if the others even gave it a thought. What if selling phones and cloud compute is a superior business to search and they are smart to realize it? In reality both Apple and Amazon are bigger companies than Google, and Google is actually the one trying to chase selling phones and cloud compute?

by bambax 6y ago

mjburgess 6y ago

I suspect people optimising for google over the last, c. 20 years has made the difference. ie., people have spent years writing their sites explicitly with google SEO in-mind... add-in all the data google has via adwords...

why_only_15 6y ago
Why can't other search engines just hook into those same signals though?
- dmitriid 6y ago
  Because they don't have the traffic/market share to do that. And it's not just search. Every single Google property ultimately feeds into Google's search engine.
  - why_only_15 6y ago
    The traffic and market share shouldn't matter if these are SEO signals inside the application itself, like in the HTML. Those are independent from traffic.

lazyjeff 6y ago

I did my Ph.D. in information retrieval and had studied the workings of the 3 major search engines from the inside during various points about a decade ago. Without giving away any trade secrets, part of it is the power of search behavior data (for example, what you click on after a search, how long before you go back to the search page, etc.). The more popular your search engine, the more search behavior data you have, which makes the search results better especially for tail queries (the less common searches). So then you attract more searchers, and it's a cycle that is hard for a newcomer to break into. It's why academia has given up on competing with commercial search engines in the web space, and why alternative search engines are just begging at the table right now.

dmitriid 6y ago

> The real mystery is why nobody has ever been able to do better search than Google, or at least just as good. Google's search is the way it is now because they've now had 20 years to work on it. To create search of similar quality you'll need to spend a comparable amount of time (during which Google will not rest on laurels, but will continue to evolve). It would've been great if someone took the pains to implement, maintain and evolve a search similar to Google's at the time when Google became available (or a few years after that). Then it's possible we would have two high quality search engines now. But even that is a simplistic view of things. Everything that Google ultimately has feeds into its search. Google Translate? Helps Google understand pages in multiple languages. Google Maps? Helps Google learn people behaviour, places of interest, businesses etc. Google Patents, Groups, Books? Provide vast amounts of natural text to train models on. And so on and so on. A "just as good" search needs all (or nearly all that) that to become just as good.

alexyz12 6y ago

according to the innovators dilemma you would have to find a market that google isn't serving with its search, use relatively lame search technology to serve it, then scale upwards into google's market from there. Maybe because google's market is everyone, disruption can no longer exist?

freediver 6y ago

> Of course search is really hard, but how is it possible that huge companies with nearly infinite money and talent (MS / Apple / Amazon) were never able to do it? MS actually tried; it's unclear if the others even gave it a thought. What if selling phones and cloud compute is a superior business to search and they are smart to realize it? In reality both Apple and Amazon are bigger companies than Google, and Google is actually the one trying to chase selling phones and cloud compute?