I don't want to move the discussion away from web spam, but I disagree with your statement. I have email for several domains hosted on Google (as well as a regular Gmail address) and they all suffer from legitimate emails regularly going into the spam folder. A quick search has shown that I'm not the only one suffering from this problem.
(i use gmail primarily, but have an old Yahoo! address for bacn)
For example, I get a few very similar Canadian pharmacy junk messages every single day. Even though I usually mark them as spam, they continue to appear in my inbox. It never learns. Additionally about 1 in 4 order confirmation emails ends up in the Junk folder.
My understanding of Bayesian spam filtering is that it's supposed to get better over time, not worse.
It took me a while to figure it out and wasn't sure if my order was properly registered, if I mistyped my email or if it was a problem on their side. Lesson learned, now I systematically scan and empty the spams.
The original stackoverflow article on the other hand is on page 3 or 5 or greater.
I'm not sure what Stackoverflow is doing, but in this case efreedom is providing me a service by making this content easily found in a search engine. Sure its one more click, but in my mind the alternative is never being able to find the SO article at all.
My workflow involves looking at multiple articles, so I can't simply put 'site: stackoverflow.com' into my search.
Ah, here's one from the search history with EF above SO:
retaincount instruments
There are several eFreedom hits from the first page, and SO has one hit on page 2 and one hit on page 3.
I have screenshots of the first three pages.
Still, I think it's becoming apparent that SEO has broken PageRank, and we really are in need of something else. Whether that something else is algorithmic, social, or, most likely, a combination of the two, remains to be seen. Still, I don't think we'll ever have a system that can't be gamed.
Based on my search history I've clicked efreedom links several times, but currently the stackoverflow link is the top one for those searches now. I would be surprised if I had clicked efreedom if stackoverflow was at top.
http://pinoytech.org/question/1699545/unable-to-get-value-fr...
I've personally not seen efreedom ranking above stackoverflow. I've seen it 2 places down or one down. So this is a very valuable comment. IMHO. YMMV.
I received an eFreedom result on the second page of results, and no corresponding Stack Overflow article. The title of the eFreedom page is "DEBIAN - pbuilder storing dependencies - efreedom"
I then searched for: ppa in pbuilder site:stackoverflow.com
This gives one response - not the same article that was scraped by eFreedom. So that seems a little off - as a user I'd expect the Stack Overflow article that was scraped to meet the query if eFreedom does.
Finally, based on the title of the Stack Overflow article that was scraped in my first query, I searched for: pbuilder storing dependencies
Here the original Stack Overflow article is listed first, eFreedom second, so that's good. Unfortunately when you scan a little further down the page, the article "UBUNTU - pbuilder create fail - efreedom" is above "pbuilder create fail - Stack Overflow".
As a side point, my search is not made more valuable by the scraped versions appearing in the results. Making Stack Overflow appear before eFreedom is one thing, but why should I see eFreedom at all? I'd rather the other results be different content, that way I don't have to mentally parse everything to work out if it's the same content with a different skin.
The content was interesting, thank you. Good point about niche sites like dpreview.
I wonder if the search quality issue is not also linked to a vast increase of the size of internet - searchs are the same throughout the years (couple of keywords), for more and more websites. Maybe vertical search helps there just because they reduce the amount of data to look for ?
As noticed by some other people, I have used google all my internet life (since 2000 or so), and google has never ever be useful to look for things like hotel, off the shelve software information, etc... I did notice a decrease of quality in technical searchs like API, etc... (where google has always been effective up to now), I wonder if something changed the last few months.
OT: did you type your commment on an ipad?
I can't seem to remember Amazon coming up on product searches I've done. I have to wonder:
- Is this because I went straight to Amazon?
- Am I misremembering?
- Is there something wrong with Amazon's SEO?
- Is Amazon restricting the their content being indexed (I can't imagine this is true)?
- Is Google actually failing here?
OT: Yeah my typo rate on my iPad is pretty terrible. I'm too used to touch typing on a keyboard. Still, I find myself reading the Wen on my iPad instead of a laptop/desktop these days so I need to put up with the typos (and get better with any luck), not comment or move to a computer to comment (not going to happen).
http://www.google.com/search?sourceid=chrome&ie=UTF-8...
http://www.google.com/search?hl=en&safe=off&q=+Asus+...
http://www.google.com/search?sourceid=chrome&ie=UTF-8...
Etc etc. I see them all the time in Google, despite also searching Amazon directly quite a bit.
As someone who pays for advertising this is a distinction I pay attention to, especially if it generalizes to more people than just me: people going to Amazon are ready to buy and not just doing research.
If Amazon can make that stick I think they pose a significant threat to Google. Why pay as much for research clicks when you can get clicks that are closer to a sale?
Yes indeed, on the money.
Further, the attitude of these various whiners is pathetic.
The war on spam-in-a-general-sense is not never going to be won or lost. The war for attention simply, inherently, has too many dimensions to be winnable or losable. As long as there's benefit for merely getting people attention, there will be people who successfully put out low-content, attention-getting information in one form or another. Human beings themselves are often often satisfied with less-than-stellar-content, how then could a given algorithm be expected to filter for stellar content?
Google and other filtering-entity's ability to provide us with stuff we're satisfied with will fluctuate. They won't ever be done. The "good" providers won't ever be fully satisfied and the "bad" provider won't ever completely go away. Any "new Google" will be "the same as the old Google".
Let's move on.
Which is fine if you already have narrowed down your camera options to a top-5 or so, but the issue is getting to that stage.
This post basically complains about two things: the finer points of SEO and content farms.
Content farms is an easy one. They're the Web equivalent of spam and I'm talking about the likes of Associated Ontent and Demand Media. They re a relatively new (last few years) phenomenon.
My personal view is that no one is better placed to deal with this new threat than Google. Email spam is basically a solved problem on Gmail. Thats not ss there aren't false positives and negatives but it's oohing like it used to be or could be. It'll take time but I believe that content farms are a transitory and doomed business model.
As for product searches, this encompasses many things. Anecdotally I recently searched for "<camera make and model> review" and found what I wanted no problem. Prices I found on pricegrabber (they have an iPad app).
SEO is a trickier beast. For one it's a constantly moving target. A combination of suboptimal source SEO and content farm SEO gaming allows the scrapers to survive. I can't say that keyword position matters all that much. Anecdotally Jeff claims it does but many factors are at pay so it's always best to be careful about making absolute claims.
Jeff claims not to want to be acquired. I'm reminded of a story I heard. Basically: if you wanted money (from angels) ask for advice. If you wanted advice, ask for money (IIRC this story came from either Mark Suster or Jason Calacanis, can't remember).
So, if you want to be acquired, say you don't?
Lastly, I'll reiterate my own opinion that social search isn't the answer in the general case (ie it will have specific use cases).
Content curation is a mixed bag. I believe there will (for at least a very long time) be a place for niche verticals. For example, dpreview is a vertical for cameras. General purpose models like Mahalo I think are doomed for much the same reason that Jeff and Joel have contended that general Q&A sites are doomed.