Hacker News Reader

Comment by dsl | Hacker News Reader

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.

View on HN

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.

> 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.

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.

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.

Different is fine, however I certainly don't expect my search results to be atrocious.

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.

> 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.

> 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.

> 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?

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.

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 (1983)

https://www.youtube.com/watch?v=QRQxzQGwEcI (1977)

https://www.youtube.com/watch?v=QK26LU5ZhmE (1982)

https://www.youtube.com/watch?v=8YHFJB72vi8 (2003)

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-emplo...

https://www.coca-colacompany.com/stories/what-new-coke-conti...

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.

That's a fascinating (and probably harmless) misperception. :)

Reply on news.ycombinator.com