6 ms·

I think you have some great feedback here but for me it also highlights how subjective search results can be for individuals - for example, these false positive

I think you have some great feedback here but for me it also highlights how subjective search results can be for individuals - for example, these false positives that you mention (b2, b3) appear as the top result on Google for me for that query. It makes me think there must be some fairly large segment of the population that want that domain returned as a result for their query, no?

Archelaos 5y ago
I would not deny that a large part of subjectivity is involved. This is why I used several markers of subjectivity in my evaluation ("what I can see", "that leaves me", "they seem to me", "I would say", etc.). And related to that: I also agree with other responses that a search often needs to be refined. So my four examples where in no way an exhaustive evaluation, but an explorative experiment, where I just used two proper names, one for a city and one for a historical person, and two general disciplines as search words, in order to see what happens and what is noteworthy (to me). So much to the subjective side. But what can be said about ideal search results for these terms beyond subjectivity? I do not think that we can arrive at an objective search result, but are nevertheless allowed to criticise search results with respect to their (hidden or obvious) agenda. Let me give an example of the good old days: When I was searching for my surename on Google in the early 2000s the search results contained a lot of university papers or personal Web-sites (then called "homepages") from other people of that name. But suddenly, I can't remember when exactly this was, the search results contained almost exlusively companies that contained that surename in its company name. The shift was not gradually, as if it were representing a slow shift in the contents of the Internet itself, but abrupt. It was apparently due to an intentional modification of the ranking algorithm that put business far above anything else on the Internet. My explanation for this is the following: The objective metrics for Google search results is the stream of revenue they generate for Google. But not only for Google. The fundamental monetary incentive to Bing (and its derivative Ecosia) is more or less the same. And how different the impact of the somewhat different business model of Duck Duck Go is, is open for debate. If maximum revenue is the goal, the aim is to provide the best search results according to the business model (advertisment, market research and whatever else) without driving the users away. But the best search results according to the business model are not necessarily the optimal search results for the typical user. And as long as all relevant competitors are following the same economic pressure of maximizing revenue the basic situation an thus the qualitiy of the search results for the user will not improve above a certain level. If we want this situation to change, we need competitors with a different, non-commerical agenda. Either from the public sector (an analogon to the excellent information services about physical books provided by libraries) or from non-profit organizations (an analogon to Wikipedia or Open Street Map). To answer your question about b2 and b3: I checked with other search engines; besides Google they appear for me also on Bing (as #8, same product but on a different Web-site) and Duck Duck Go (as #10); Bing also has a reference to them in the right margin as a suggestion for a refined search (this time exactly b2 and b3). Although I do not think that the results from those search engines should be considered as a general benchmark for good search results for the reasons given above, we may speculate why they appear on the first page of search results. I would guess that it is a combination of gaming the search engines by using a generic term as a product and domain name to get free advertising, and search engine algorithms making this possible by generally ranking products and companies high in their search results.
- Terretta 5y ago
  You inspired me to try an even less specific search: thing Subjectively felt the gigablast results were a relative delight.
  - Archelaos 5y ago
    No bad idea. At the risk of being sidelined: "philosophy" was not so a bad term either. Start with an arbitrary Wikipedia link and click on the first keyword of the summary after the linguistic annotations (or other annotations in brackets) and repeat the process until you reach a loop. You will almost always end with "philosophy" -> "metaphysics" -> "philosophy" -> ... This works for "Berlin", "history" and "Caesar" as well as for "thing". For the latter very fast: "thing" -> "object" -> "philosophy".
    - gbmatt 5y ago
      that's tripped out. where did you hear about that?
      - Archelaos 5y ago
        I can't remember. Probably on Hacker News.
- dgivney 5y ago
  Oh of course you can criticise the result, I more found it interesting that a billion dollar, optimized search experience thought your false positive was actually a top result. A huge variance in the subjectivity between your experience and their invested reasoning. But while we're speculating on how the domain the appears at the top of the list, let me hazard a guess... Philosophy.com was registered in 1999 and according to waybackmachine, has been selling cosmetics on the site since 2000 (20+ years). The company sold in 2010 for ~$1B to a holding company with revenues of $10B+ today (Unfortunately I couldn't find how much it contributes to that revenue). According to Wikipedia, the Philosophy brand has been endorsed by celebrities, including "long-time endorser" Oprah Winfrey, possibly the biggest endorsement you could get for their industry/demographic. I think it is a long established business, with strong revenues and there's more people online searching for cosmetic brands than for philosophers. In the same way (admittedly in the extreme) when I'm researching deforestation and I query to see how things are going for the 'amazon', the top result is another successful company registered pre 2000, with strong revenues that most likely attracts more visitors..

Archelaos 5y ago

I would not deny that a large part of subjectivity is involved. This is why I used several markers of subjectivity in my evaluation ("what I can see", "that leaves me", "they seem to me", "I would say", etc.). And related to that: I also agree with other responses that a search often needs to be refined. So my four examples where in no way an exhaustive evaluation, but an explorative experiment, where I just used two proper names, one for a city and one for a historical person, and two general disciplines as search words, in order to see what happens and what is noteworthy (to me). So much to the subjective side. But what can be said about ideal search results for these terms beyond subjectivity? I do not think that we can arrive at an objective search result, but are nevertheless allowed to criticise search results with respect to their (hidden or obvious) agenda. Let me give an example of the good old days: When I was searching for my surename on Google in the early 2000s the search results contained a lot of university papers or personal Web-sites (then called "homepages") from other people of that name. But suddenly, I can't remember when exactly this was, the search results contained almost exlusively companies that contained that surename in its company name. The shift was not gradually, as if it were representing a slow shift in the contents of the Internet itself, but abrupt. It was apparently due to an intentional modification of the ranking algorithm that put business far above anything else on the Internet. My explanation for this is the following: The objective metrics for Google search results is the stream of revenue they generate for Google. But not only for Google. The fundamental monetary incentive to Bing (and its derivative Ecosia) is more or less the same. And how different the impact of the somewhat different business model of Duck Duck Go is, is open for debate. If maximum revenue is the goal, the aim is to provide the best search results according to the business model (advertisment, market research and whatever else) without driving the users away. But the best search results according to the business model are not necessarily the optimal search results for the typical user. And as long as all relevant competitors are following the same economic pressure of maximizing revenue the basic situation an thus the qualitiy of the search results for the user will not improve above a certain level. If we want this situation to change, we need competitors with a different, non-commerical agenda. Either from the public sector (an analogon to the excellent information services about physical books provided by libraries) or from non-profit organizations (an analogon to Wikipedia or Open Street Map). To answer your question about b2 and b3: I checked with other search engines; besides Google they appear for me also on Bing (as #8, same product but on a different Web-site) and Duck Duck Go (as #10); Bing also has a reference to them in the right margin as a suggestion for a refined search (this time exactly b2 and b3). Although I do not think that the results from those search engines should be considered as a general benchmark for good search results for the reasons given above, we may speculate why they appear on the first page of search results. I would guess that it is a combination of gaming the search engines by using a generic term as a product and domain name to get free advertising, and search engine algorithms making this possible by generally ranking products and companies high in their search results.

Terretta 5y ago
You inspired me to try an even less specific search: thing Subjectively felt the gigablast results were a relative delight.
- Archelaos 5y ago
  No bad idea. At the risk of being sidelined: "philosophy" was not so a bad term either. Start with an arbitrary Wikipedia link and click on the first keyword of the summary after the linguistic annotations (or other annotations in brackets) and repeat the process until you reach a loop. You will almost always end with "philosophy" -> "metaphysics" -> "philosophy" -> ... This works for "Berlin", "history" and "Caesar" as well as for "thing". For the latter very fast: "thing" -> "object" -> "philosophy".
  - gbmatt 5y ago
    that's tripped out. where did you hear about that?
    - Archelaos 5y ago
      I can't remember. Probably on Hacker News.
dgivney 5y ago
Oh of course you can criticise the result, I more found it interesting that a billion dollar, optimized search experience thought your false positive was actually a top result. A huge variance in the subjectivity between your experience and their invested reasoning. But while we're speculating on how the domain the appears at the top of the list, let me hazard a guess... Philosophy.com was registered in 1999 and according to waybackmachine, has been selling cosmetics on the site since 2000 (20+ years). The company sold in 2010 for ~$1B to a holding company with revenues of $10B+ today (Unfortunately I couldn't find how much it contributes to that revenue). According to Wikipedia, the Philosophy brand has been endorsed by celebrities, including "long-time endorser" Oprah Winfrey, possibly the biggest endorsement you could get for their industry/demographic. I think it is a long established business, with strong revenues and there's more people online searching for cosmetic brands than for philosophers. In the same way (admittedly in the extreme) when I'm researching deforestation and I query to see how things are going for the 'amazon', the top result is another successful company registered pre 2000, with strong revenues that most likely attracts more visitors..