Forum Moderators: open
<added>if you find this popular lake keep in mind that while the first wto have several external backlinks each it weighs out when the next three have NONE. I'll go check av and msn and atw.</added>
[edited by: heini at 10:09 am (utc) on Nov. 2, 2002]
[edit reason] no specific searches/urls as per TOS & charter please / thanks! [/edit]
I honestly think the travel sector is worse this time than last and that's saying something!
of course, what else would you expect when searching for 'popular vacations';):)
I honestly wonder what any engine could do to clean those areas up.
Affiliate sites are nothing illegal in any way in the first place. In most cases they will give the user what he wants.
Anyhow, whatever a search engine would do to cut down on the number of semi duplicated sites selling the same stuff as the next site - they wouldn't stand much of a chance. For each site getting dropped 10 new are built in no time.
I honestly wonder what any engine could do to clean those areas up.
I don't think server location would be good because just one variable would make it kind of a wild card but perhaps making use the backlinks' servers locations in relation to on page factors. Perhaps something as simple as number of pages within the subdomain or domain, since that seems to be something that is related to both local content and quality(more, therefore, more descriptive content). From there leave it up to the duplicate content filters.
I honestly wonder what any engine could do to clean those areas up
[webmasterworld.com...]
Google seems to be edging towards only using directory entries for some search terms. In other words if you try a search in Google for a hotel in a particular city (particularly one of the top 10 cities) , then you get a disproportionatly high number of directory sites (no doubt someone can demonstrate that I am entirely wrong with this assertion, but that is the joy of the board)
Certainly I believe that Google has made a number of improvements in its algo to reduce the problem.
It can feel like things are stacked against you, but largely it just comes down to the link map of the Web.
Google is designed to give prominence to pages that are linked from places that are well linked, so this can mean that the 'little guy' finds it hard to compete. The only answer is to make your prominence bigger than theirs.
I'm not suggesting that it's easy to be #1 for a popular term, life's just not like that. What you may be able to do, is to take some part of your wordspace and focus on it in a way that a more general network of sites can't. This can take a long time.
> I don't think server location would be good because just one variable would make it kind of a wild card...
I agree, server IP geolocation is not a good answer. You and I might quite quickly realise that we need to host in a particular place (I've made this decision before with UK related .com sites), but the average mom 'n pop won't know that they need to do it.
> ...but perhaps making use the backlinks' servers locations in relation to on page factors
That could work better, also inclusion in a regional category in Yahoo! or the ODP would be a much better indicator than server IP (I'm not saying perfect, just better).
> Perhaps something as simple as number of pages within the subdomain or domain, since that seems to be something that is related to both local content and quality(more, therefore, more descriptive content).
That would be far too easy for a spammer to work around. If you have a 100 page site of high quality content, someone can put up a 1,000 page site with low quality content with ease.
As heini points out, search quality is not an easy problem to solve. Google need to be better than the competition. For most phrases, most of the time, I think they are.
The effect has been the browsers spend hours looking for cheaper deals, hence the plethera of affiliate sites (usually hooked to a common data base)
One major hotel group has recently tried to get over it by offering a pledge that the cheapest offer will always be on their own hotels page
Every hotel reservation booked through a <snip> Hotels web site is guaranteed to have the lowest rate publicly available on the Internet. If you find a rate lower than the best available <snip> Hotels Internet rate that is viewable and bookable on another web site for the same hotel and accommodations, and for the same dates, and advise <snip> Hotels within 24 hours of booking your room, <snip> Hotels will honor that rate for the nights for which the lower rate was found, plus give you an additional 10% discount off that rate, upon its verification by <snip> Hotels.
If more hotels did this, then the affiliate sites would disappear, as they would have no raison d'etre.
As the example above is with one of the largest worldwide hotel groups, it is likely to be followed
1) As others have pointed out, a site in (and about) Timbuktu might be hosted in Pittsburgh for reasons of cost, convenience, or reliability.
2) The best travel sites may not be produced locally, just as the best guidebooks aren't. To use an offline example, would you automatically assume that a Michelin Green Guide, Lonely Planet, or Time Out guide to Rome was inferior to a locally produced guidebook from an unknown publisher?
I also disagree with the suggestion that hotels could make affiliate sites irrelevant by publishing the best deals on their Web sites. In theory, that might be true, but there's one practical reason why it wouldn't work:
Many travelers are searching on "Springfield hotels," not "Hotel Springfield Plaza" or "Springfield Inn." They'd much rather land on a well-organized directory of hotels by location and/or price than have to wade through a random list of Google search results. This makes it important for hotels to gain exposure in directories (whether those directories are sponsored commercial directories, guidebook-style editorial pages, or well-designed affiliate sites).
Many travelers are searching on "Springfield hotels," not "Hotel Springfield Plaza" or "Springfield Inn." They'd much rather land on a well-organized directory of hotels by location and/or price than have to wade through a random list of Google search results. This makes it important for hotels to gain exposure in directories (whether those directories are sponsored commercial directories, guidebook-style editorial pages, or well-designed affiliate sites ).
I agree with that, I was not arguing against that ...except for your last few words "well-designed affiliate sites"
By the nature of the hotel business, few hotels will release blocks of rooms to large numbers of data bases to enable online booking to take place. Therefore the number of independent data bases on the web is small, and few affiliate sites do more than try to funnel their browsers through to the same place...hence user frustration at wading through sites that are supplying essentially the same information.
Browsers are looking for the best deal in the hotel that they finally choose. If that is most likely to be in the hotels own web site then they will go to " sponsored commercial directories, guidebook-style editorial pages" but not "well-designed affiliate sites" which will take them directly to the hotel's own booking system, and not to one of the few affiliate data bases that affiliates funnel them to.
Google serps with the actual word 'travel' in them went somewhat weird starting last month. I find that related searches with similar words give much better results.
I don't think that many people are going to be searching under "travel" (happy to be proved wrong on that), though it continues to surprise me how many use the one word "hotels". One assumes that they then qualify the search to narrow it to hotels in a particular location.
Having said that I tried a few searches using "travel" followed by a location. The results seemed fairly innocuous. What was weird about the sort of the serps you were getting?
>I honestly wonder what any engine could do to clean those areas up.
One question heini: Is any search engine actually succeeding in doing this? Google may well have decided that it isn't worth the bother. With 99.9+% of all searches done by Google users having nothing to do with travel, should Google worry inordinately about how well they do for these rare searches, or worry more about the things people typically search for? Also, if the other search engines also are failing, Google has no relative disadvantage to these other search engines. No reason to switch to Altavista if things are just as bad at Altavista.
My point was exactly this: all engines have to deal with this problem. There is no easy way to fix this problem for spidering engines.
With 99.9+% of all searches done by Google users having nothing to do with travel
Err...where did you get that statistic? Travel is the second biggest business on the web. I only search for a few things on a regular basis on the web:
1) News
2) Finance
3) My web sites' progress on SEs
4) Webmasterworld
5) Airline fares and other travel related information
6) Telephone numbers and addresses of companies
That's about all I can think of. The rest is just a bit of this and a bit of that. Just an indicator of how big travel on the web is: our company gets at least 95% of bookings through the web, despite the fact we are on every travel agent's screen through their non-internet booking system.
Whilst everyone does not have the same search profile as me, I guess I am pretty much john doe.
Browsers are looking for the best deal in the hotel that they finally choose. If that is most likely to be in the hotels own web site then they will go to " sponsored commercial directories, guidebook-style editorial pages" but not "well-designed affiliate sites" which will take them directly to the hotel's own booking system, and not to one of the few affiliate data bases that affiliates funnel them to.
When searching for hotels, readers may be looking for any or all of the following:
1) The right hotel (based on location, appeal, price, recommendations, etc.)
2) A good price.
3) A means of booking.
Different readers will approach hotel selection in different ways. For example:
A) The people who read Karen Brown's CHARMING INNS & ITINERIES guidebooks look for the perfect hotel first and worry about price afterwards. They want cozy country inns, hotels in converted monasteries, rooms overlooking canals in Venice, etc. Saving $50 on the room isn't as important as having a memorable experience. If these people search for hotels on the Web, they're probably looking at editorially driven "content sites"--and if they see an affiliate link, there's a good chance that they'll use it to book a room.
B) A traveler who's visiting New York and wants to stay within walking distance of the theatre district may be open to any of a dozen or more hotels within the desired price range. For such a traveler, a well-designed directory is highly useful, and the traveler probably doesn't know or care whether the site is supported by paid advertising or affiliate sales.
C) A turnpike traveler who's stopping for the night in Montpelier, Ohio may simply be looking for the cheapest room. For such a traveler, typing "Montpelier Ohio motels" in Google and getting half a dozen motel pages may be ideal. The traveler can then call a motel's 800 number and say "Do you have a room for the 29th with a queen-size bed and a rollaway?" This traveler has no need for a travel writer's description or a hotel directory's interface. It's easier and quicker to just look up motels in Google.
For traveler A, affilate *sites* aren't useful, but affiliate links are.
For traveler B, affiliate sites may be useful if they're well-designed (i.e., if they have a convenient interface and a good selection of hotels and competitive rates).
For traveler C, affiliate sites and links are irrelevant.
As far as Google is concerned, I think it's fair to say that readers (depending on their needs) are looking for:
- Editorial content about hotels (which they may be finding in a destination site).
- Well-designed hotel directories (which may or may not use affiliate links).
- Individual hotel or motel Web sites.
It's obviously hard for Google to know what type of content a reader who searches on "Paris hotels" is looking for, but a good place to start is by filtering out the directory sites that use boilerplate text and photos.
It's obviously hard for Google to know what type of content a reader who searches on "Paris hotels" is looking for, but a good place to start is by filtering out the directory sites that use boilerplate text and photos.
I think we could all agree on that (unless any readers of the forum write directory sites that use boilerplate text and photos...)
You then have two problems:-
1. Your definition of such sites. Don't think I have ever met the owner of such an affiliate site who would not argue (vehmently) that their site offered unique content of some sort
2. How you suggest Google, in particular, removes them.
As I said earlier, only using DMOZ directory entries is about the only way it can be done. DMOZ entries are not altogether foolproof in weeding out such sites, but are the best we have!
How you suggest Google, in particular, removes them.As I said earlier, only using DMOZ directory entries is about the only way it can be done. DMOZ entries are not altogether foolproof in weeding out such sites, but are the best we have!
DMOZ's travel categories are pretty weak, IMHO. Many categories don't have editors, and some have links that are outdated by a year or more. Plus, I'd hate to see Google outsource something as important as quality control to a third party--especially a third party owned by AOL Time Warner. :-)
Besides, categories in DMOZ or Yahoo don't necessarily coincide with what people are looking for in search, because Google users are often looking for "deep links" rather than home pages. A person searching on "Springfield hotels" isn't necessarily looking for Springfieldhotels.com; he may be looking for hotel descriptions on editorial sites (such as Timeout.com or Karenbrown.com) rather than e-commerce sites.
No; if Google is going to filter out (or de-emphasize) boilerplate affiliate sites, it will simply have to do it the old fashioned-way: by comparing page content and using logic to determine that 500 words of boilerplate copy about the Hotel Springfield + a TravelNow link + links to 500 other pages with boilerplate listings and TravelNow links = a boilerplate affiliate site.
And, what makes you think you necessarily are typical? More typical is people are searching for things like chocolate chip cookie recipes, Britney Spears lyrics, football statistics, and other such non-commercial topics. Many people never travel at all or very rarely, and most who do travel only do so rarely, maybe once every year or 2. For the rest of the year these people are searching on non-travel topics. And, many who travel will not bother with the web at all, and go to a local travel agent, etc. I'd be surprised if the frequency of searches where the object appeared to be travel was more than 1 in 1000 searches. Unfortunately, to do so would need a random sample of say 10,000 Google searches and try and figure out exactly what the person was looking for. I doubt this has been done, and the analysis would tend to be subjective. How can you know if someone enters searches about Detroit, MI if the person wants to travel there, or if they already live there and just want to know more about things in town?
Should you come up with a solution, please do publish it. I doubt the spidering engines have either the resources or the inclination to make this a priority. SEs will always have problems in any area where there is a propensity for people to try and spam the SERPs. The spammers will find a way to get around attempts by the SEs to stop them. And, if they whack one of their sites next month they'll be back with 2 more. Take a look at the e-mail spam problem. Lots of it, and at least in theory you can complain about the source(s) as being abuse OF the Net. SE spamming isn't abuse of the Net.
Britney spear lyric 14395
Vacation 261,538
Many people never travel at all or very rarely, and most who do travel only do so rarely, maybe once every year or 2. For the rest of the year these people are searching on non-travel topics. And, many who travel will not bother with the web at all, and go to a local travel agent, etc.
Travel agents are dropping like flies. Airlines won't pay them a percentage on most flights anymore cause they can sell tix easily on the web.
Many travelers are searching on "Springfield hotels," not "Hotel Springfield Plaza" or "Springfield Inn." They'd much rather land on a well-organized directory of hotels by location and/or price than have to wade through a random list of Google search results. This makes it important for hotels to gain exposure in directories (whether those directories are sponsored commercial directories, guidebook-style editorial pages, or well-designed affiliate sites ).
Yes but if an individual site has more relevant content than the directory or if the directory is the same content as 10 other directories should all the directories move to the front of the line simply because they amass huge numbers of keyword rich internal links.
If the site has more relevant content than the others (and original content, at that) then I'd hope that it wouldn't get shoved to the back of the line.
If it just duplicates all the other directories, then I'd hope that it--and the other directories that use boilerplate text--would slide way down in the search results.
What I'd really like to see is for the good, distinctive, or unusually useful hotel directories to rank high in the search results. In theory, PageRank should allow that to happen, because other Webmasters would link to the good, distinctive, or unusually useful hotel directories but not to the dime-a-dozen boilerplate affiliate sites. In practice, a handful of successful affiliate sites may be putting so much time and effort into short-term SEO returns that the worthwhile sites are falling behind. And Webmasters with editorial content sites may be so fed up with boilerplate lodging sites that they'll no longer link to hotel directories at all.
---
But this is just one example. Imagine if all searches on every music artist were added together? And, this is just one example of non-commercial searches.
What is the URL to do such searches on Overture? BTW, you botched that example search. The artist's last name is *Spears*, not *Spear*. I'd also think that that more likely people would use the plural "lyrics" in a search.
What is the URL to do such searches on Overture? BTW, you botched that example search. The artist's last name is *Spears*, not *Spear*. I'd also think that that more likely people would use the plural "lyrics" in a search.
The url is [inventory.overture.com...]
When you get ther search for "Britney Spears lyrics" you will see that it returns "Britney spear lyric" since it removes all pluralisms.
Not when they choke the top 20 results with SPAM they shouldn't. SPAM is spam and should be treated as such.
The answer is not to copy the techniques used to decieve the user & the robot, but to steadily continue to build quality.
As far as I know google is a level playing field that seeks to provide relevant results to match the searchers enquiry.
(The results do not always look as if this is the case, but we do know that google is constantly working at improving their algo/filters etc.)
In the long term anybody can rank well (no matter the size of their wallet) by building quality websites that provide content about their services/product and following the few simple guidelines that google makes public.
If the large international guides / affiliates / "fluff" sites continue to dominate by spam tactics then searchers and marketers will vote with their feet (or was that finger tips) and new solutions to bring quality sites to the top will be found elsewhere.. (IMO)
OK. Now I tried this:
Searches done in September 2002
Count Search Term
1821375 lyric
655788 lyric of song
334384 music lyric
These all beat:
Vacation 261,538
I would have expected that searches for song lyrics would be WAY more popular than ones for vacation, and if Overture can be believed that is what they indicate.
Additionally--in the www2 & www3 a few days ago, there were a bunch-a-bunch of spammy but high ranking affiliate sites, all the same network,for agents evidently selling, not renting, accomodations in these resort areas and the sites were L-o-a-d-e-d with hidden text. They had doorways for every southern resort area I ck'd. What the viewer saw, all looked the same, just a boring generic form to fill in with area of interest. Buy the time www2 & www3 rolled into www, the whole network was gone from the serps. It was big, high ranking briefly, irrelevant and broke a bunch of the rules.
--if it was removed by Google--i say--good work :)
Lets just say that travel is a sufficiently important category that when the serps are dominated by a particular type of site it shows.
