3 ms·

I wonder how many "legitimate" links a site like RapGenius gets. I've never seen anybody link to them, except maybe in social networks. Similar for the stackexc

I wonder how many "legitimate" links a site like RapGenius gets. I've never seen anybody link to them, except maybe in social networks. Similar for the stackexchange sites. Sometimes people link to good questions from blogs, but a large portion of links are probably from other stackexchange sites, twitter, and Jeff Atwood's blog (which is a marvel of SEO and affilate marketing in itself). For many of these high quality content silos (for lack of a better word), the PageRank paradigm seems pretty broken. People don't feel compelled to link to well-known sites precisely because they are well-ranked. And when they do, it is in form of viral posts in social networks, which is not really a great input to determine site quality. I wouldn't be surprized if Google's real secret algorithm nowadays consists of millions of special cases maintained by thousands of poorly paid interns.

judk 13y ago
Why poorly paid?
- captainmuon 13y ago
  It's the cliche. But you're right, the most menial worker at google earns probably a lot more then I do (in academia). I don't know...
jlees 13y ago
I was thinking something along the same lines. PageRank was great back in the day, but in today's world, the link is not the only way of determining whether a site is good or not. I wonder if a rethink of the HITS model or a complete overhaul of search ranking based on social and reputation cues is in order? StackOverflow does well despite few people linking to them because the search terms are so niche, if I search for a fairly generic programming term I maybe get one StackOverflow post in the first 10 but if I search for an error message, they'll all be SO. Other StackExchange sites maybe don't benefit quite so well from that long tail, though. Sites like Quora presumably have the same problem. It's interesting how sites like these, and in turn mobile apps, are turning information into vertical silos (which want you to stay within their bounds) rather than the horizontal web of equality we're used to. (Also, the interns are pretty well paid, but yes. They start out as autocompleters: http://www.youtube.com/watch?v=blB_X38YSxQ http://www.youtube.com/watch?v=blB_X38YSxQ)
- jrochkind1 13y ago
  The genius of the original PageRank is that links serve as machine actionable artifacts of 'social and reputation cues'. But of course Google's algorithm is way more complicated now, and takes into account all sorts of things. They of course have all sorts of smart experienced people working on improving it, and apparently still determine that links should count for a lot. I'd guess there are actually lots of links to stackoverflow.
jonahx 13y ago
Are you saying SO questions high google rank is mainly a result of links from within the SE network?
- jaredmck 13y ago
  Yes, internal links can be extremely valuable especially in narrowly relevant niches like SO.
- captainmuon 13y ago
  No, but I guess e.g. physics.stackexchange.com or judaism.stackexchange.com benefits a lot from the SE internal links, and I guess for those smaller sites it is the main source of page rank. That is, if page rank still works like Google claims. I'm pretty convinced that there is a lot of manual curation going on. At least for popular terms and for the first page of results.
porker 13y ago
> and Jeff Atwood's blog (which is a marvel of SEO and affilate marketing in itself). Really? What makes you claim that?

judk 13y ago

Why poorly paid?

captainmuon 13y ago
It's the cliche. But you're right, the most menial worker at google earns probably a lot more then I do (in academia). I don't know...

jlees 13y ago

I was thinking something along the same lines. PageRank was great back in the day, but in today's world, the link is not the only way of determining whether a site is good or not. I wonder if a rethink of the HITS model or a complete overhaul of search ranking based on social and reputation cues is in order? StackOverflow does well despite few people linking to them because the search terms are so niche, if I search for a fairly generic programming term I maybe get one StackOverflow post in the first 10 but if I search for an error message, they'll all be SO. Other StackExchange sites maybe don't benefit quite so well from that long tail, though. Sites like Quora presumably have the same problem. It's interesting how sites like these, and in turn mobile apps, are turning information into vertical silos (which want you to stay within their bounds) rather than the horizontal web of equality we're used to. (Also, the interns are pretty well paid, but yes. They start out as autocompleters: http://www.youtube.com/watch?v=blB_X38YSxQ http://www.youtube.com/watch?v=blB_X38YSxQ)

jrochkind1 13y ago
The genius of the original PageRank is that links serve as machine actionable artifacts of 'social and reputation cues'. But of course Google's algorithm is way more complicated now, and takes into account all sorts of things. They of course have all sorts of smart experienced people working on improving it, and apparently still determine that links should count for a lot. I'd guess there are actually lots of links to stackoverflow.

jonahx 13y ago

Are you saying SO questions high google rank is mainly a result of links from within the SE network?

jaredmck 13y ago
Yes, internal links can be extremely valuable especially in narrowly relevant niches like SO.
captainmuon 13y ago
No, but I guess e.g. physics.stackexchange.com or judaism.stackexchange.com benefits a lot from the SE internal links, and I guess for those smaller sites it is the main source of page rank. That is, if page rank still works like Google claims. I'm pretty convinced that there is a lot of manual curation going on. At least for popular terms and for the first page of results.

porker 13y ago

> and Jeff Atwood's blog (which is a marvel of SEO and affilate marketing in itself). Really? What makes you claim that?