5 ms·

I remember a conversation I had with an engineer working on Bing around the time of the Microsoft/Yahoo deal... their core metric was the number of searches whe

by dsl 8y ago

I remember a conversation I had with an engineer working on Bing around the time of the Microsoft/Yahoo deal... their core metric was the number of searches where they matched 8 out of 10 or more of Google's results on the first page of the search. Qwant isn't even that close. The results are absolutely poor across a number of different queries.

pluto9 8y ago
> their core metric was the number of searches where they matched 8 out of 10 or more of Google's results on the first page of the search. That's a really defeatist way of entering a new market--"our competitor is the gold standard, so instead of trying to be better than them, we'll measure our success by how closely we can imitate them". It's like Pepsi and Coca-Cola, which I also always found strange. Why advertise Pepsi as tasting just like Coke? Why not advertise it as tasting _better?_ > Qwant isn't even that close. I don't think that, by itself, says anything about the quality of their search results.
- TylerE 8y ago
  > Why advertise Pepsi as tasting just like Coke? Why not advertise it as tasting _better?_ That's exactly what they did. Remember the Pepsi Challenge?
  - pluto9 8y ago
    This is weird--I remember Pepsi commercials with a blind taste test on TV when I was a kid. Oddly, I distinctly remember them framing it as people not being able to tell the difference, and that being seen as a good thing. But now I can't find any reference to those commercials or that marketing angle anywhere. Now you've got me wondering if I misunderstood the commercial as a child and have had a tainted perception of Pepsi's marketing my whole life because of it.
    - TylerE 8y ago
      Yes, the whole point was they were given two unlabeled samples, and they preferred the one that ended up being Pepsi. This is not surprising as Pepsi is much, much sweeter so in small doses it does tend to be preferred. However, when given a full glass, many go back to preferring Coke. https://www.youtube.com/watch?v=ghMYzo0rgrw https://www.youtube.com/watch?v=ghMYzo0rgrw (1983) https://www.youtube.com/watch?v=QRQxzQGwEcI https://www.youtube.com/watch?v=QRQxzQGwEcI (1977) https://www.youtube.com/watch?v=QK26LU5ZhmE https://www.youtube.com/watch?v=QK26LU5ZhmE (1982) https://www.youtube.com/watch?v=8YHFJB72vi8 https://www.youtube.com/watch?v=8YHFJB72vi8 (2003)
      - stephenhuey 8y ago
        The story of what went wrong with New Coke was that Coca Cola's research was based entirely on sip tests and people preferred the slightly sweeter taste. I read in Malcolm Gladwell's "Blink" and people ultimately didn't want to drink an entire can of the stuff. This went down as one of the most famous blunders in market research history. https://www.printrunner.com/blog/sip-testing-potential-employees-new-cokes-marketing-blunder/ https://www.printrunner.com/blog/sip-testing-potential-emplo... https://www.coca-colacompany.com/stories/what-new-coke-continues-to-teach-us-30-years-later https://www.coca-colacompany.com/stories/what-new-coke-conti...
        TylerE 8y ago
        One aside that might bear mentioning: Coca-Cola Classic wasn't. It was when they slipped in the change from cane sugar to corn syrup in the US market.
    - rectang 8y ago
      That's a fascinating (and probably harmless) misperception. :)
zeugmasyllepsis 8y ago
In some sense that's kind of the point right? If Qwant isn't tracking user information than we _expect_ the results to be different, since the engine doesn't have a chance to build up an information bubble about its users.
- vntok 8y ago
  Different is fine, however I certainly don't expect my search results to be atrocious.
- nine_k 8y ago
  There can be a number of queries where no context is required to give a relevant answer, because the query carries enough context. E.g. html5 working draft when new york city was founded aapl ticker us consulates in germany can i haz cheezeburger there is no spoon tyger tyger burning bright hurricane florence I did not try any of these, though.
- dsl 8y ago
  > If Qwant isn't tracking user information than we _expect_ the results to be different We expect the answers to be correct. Google can use everything it knows about me to turn an abstract term like "giants" into SF Giants or New York Giants, but when I explicitly enter "san francisco giants" as part of a query and still get poor results, that is a failure.
  - gleenn 8y ago
    If 2 pages have similar topics and coverage, and 1 is ranked higher on Google and the other higher on Qwant, who is "correct"? Search is a fuzzy thing, and while some results might be objectively, I challenge your notion of correctness.
- pbreit 8y ago
  If tracking is going to give me much better results, seems worth it. Would a better positioning be "safe" tracking (whatever that would exactly mean)? I'm not sure matching Google, but a little worse, is best objective.
eitland 8y ago
> Qwant isn't even that close. The results are absolutely poor across a number of different queries. Don't know what you tried, I just tried it on a number of searches like: Logging framework .net Edward Snowdon (misspelled) Space shuttle Not only was it fast but it delivered good results as well. Will try it next week with some of the stuff Google struggles with now: respecting when I use doublequotes etc. Edit: the speed difference might be related to the fact that I am in Europe.

by dsl 8y ago

pluto9 8y ago

> their core metric was the number of searches where they matched 8 out of 10 or more of Google's results on the first page of the search. That's a really defeatist way of entering a new market--"our competitor is the gold standard, so instead of trying to be better than them, we'll measure our success by how closely we can imitate them". It's like Pepsi and Coca-Cola, which I also always found strange. Why advertise Pepsi as tasting just like Coke? Why not advertise it as tasting _better?_ > Qwant isn't even that close. I don't think that, by itself, says anything about the quality of their search results.

TylerE 8y ago
> Why advertise Pepsi as tasting just like Coke? Why not advertise it as tasting _better?_ That's exactly what they did. Remember the Pepsi Challenge?
- pluto9 8y ago
  This is weird--I remember Pepsi commercials with a blind taste test on TV when I was a kid. Oddly, I distinctly remember them framing it as people not being able to tell the difference, and that being seen as a good thing. But now I can't find any reference to those commercials or that marketing angle anywhere. Now you've got me wondering if I misunderstood the commercial as a child and have had a tainted perception of Pepsi's marketing my whole life because of it.
  - TylerE 8y ago
    Yes, the whole point was they were given two unlabeled samples, and they preferred the one that ended up being Pepsi. This is not surprising as Pepsi is much, much sweeter so in small doses it does tend to be preferred. However, when given a full glass, many go back to preferring Coke. https://www.youtube.com/watch?v=ghMYzo0rgrw https://www.youtube.com/watch?v=ghMYzo0rgrw (1983) https://www.youtube.com/watch?v=QRQxzQGwEcI https://www.youtube.com/watch?v=QRQxzQGwEcI (1977) https://www.youtube.com/watch?v=QK26LU5ZhmE https://www.youtube.com/watch?v=QK26LU5ZhmE (1982) https://www.youtube.com/watch?v=8YHFJB72vi8 https://www.youtube.com/watch?v=8YHFJB72vi8 (2003)
    - stephenhuey 8y ago
      The story of what went wrong with New Coke was that Coca Cola's research was based entirely on sip tests and people preferred the slightly sweeter taste. I read in Malcolm Gladwell's "Blink" and people ultimately didn't want to drink an entire can of the stuff. This went down as one of the most famous blunders in market research history. https://www.printrunner.com/blog/sip-testing-potential-employees-new-cokes-marketing-blunder/ https://www.printrunner.com/blog/sip-testing-potential-emplo... https://www.coca-colacompany.com/stories/what-new-coke-continues-to-teach-us-30-years-later https://www.coca-colacompany.com/stories/what-new-coke-conti...
      - TylerE 8y ago
        One aside that might bear mentioning: Coca-Cola Classic wasn't. It was when they slipped in the change from cane sugar to corn syrup in the US market.
  - rectang 8y ago
    That's a fascinating (and probably harmless) misperception. :)

zeugmasyllepsis 8y ago

In some sense that's kind of the point right? If Qwant isn't tracking user information than we _expect_ the results to be different, since the engine doesn't have a chance to build up an information bubble about its users.

vntok 8y ago
Different is fine, however I certainly don't expect my search results to be atrocious.
nine_k 8y ago
There can be a number of queries where no context is required to give a relevant answer, because the query carries enough context. E.g. html5 working draft when new york city was founded aapl ticker us consulates in germany can i haz cheezeburger there is no spoon tyger tyger burning bright hurricane florence I did not try any of these, though.
dsl 8y ago
> If Qwant isn't tracking user information than we _expect_ the results to be different We expect the answers to be correct. Google can use everything it knows about me to turn an abstract term like "giants" into SF Giants or New York Giants, but when I explicitly enter "san francisco giants" as part of a query and still get poor results, that is a failure.
- gleenn 8y ago
  If 2 pages have similar topics and coverage, and 1 is ranked higher on Google and the other higher on Qwant, who is "correct"? Search is a fuzzy thing, and while some results might be objectively, I challenge your notion of correctness.
pbreit 8y ago
If tracking is going to give me much better results, seems worth it. Would a better positioning be "safe" tracking (whatever that would exactly mean)? I'm not sure matching Google, but a little worse, is best objective.

eitland 8y ago

> Qwant isn't even that close. The results are absolutely poor across a number of different queries. Don't know what you tried, I just tried it on a number of searches like: Logging framework .net Edward Snowdon (misspelled) Space shuttle Not only was it fast but it delivered good results as well. Will try it next week with some of the stuff Google struggles with now: respecting when I use doublequotes etc. Edit: the speed difference might be related to the fact that I am in Europe.