| 9:01 am on Aug 8, 2005 (gmt 0)|
It certainly is an incredibly easy way to fix glaring problems, or situations where "getting it right" would be particularly appreciated.
I haven't seen more hand results, but I'd sure love to (and yes, even more so after this not good update).
| 10:39 am on Aug 8, 2005 (gmt 0)|
Yeah, apparently Yahoo! is not confident in their algo based search at all. I did not think they would quit this early.
I wonder who selects the sites that manually rank? Seems like once an area becomes hard coded the serps NEVER EVER change.
Ah well, sad day in the search engine world...
| 1:59 pm on Aug 8, 2005 (gmt 0)|
Well as I've stated in several threads in the past I think it's a good thing to hand edit the more competitive serps for spam. There will always be black hat seo and hand editing is the only way to get rid of it. (although it will be back, and will need to be hand edited again and again) Humans are still better than any algo.
| 2:13 pm on Aug 8, 2005 (gmt 0)|
|Humans are still better than any algo. |
I assume you use mostly directories for search?
Google does a decent job of evaluating quality sites and ranking them accordingly. Why can't yahoo? 99.99% of searches are on terms other than the one's they edit. Why should we believe they can get that right if they are forced to hand edit some terms?
Yahoo is simply missing almost every single niche authority site. The hand edits are basically mega sites with a page or two about the term (in many cases).
It is not about fighting "spam", rather, it is about Yahoo! figuring out ways to identify authority sites without having one of their customer service reps scroll through the directory and add them to the once "natural" listings.
We all knew this was coming. Yahoo! and algo based search do not go well together. They first added their own pages to the top of the serps, now they are hand picking the others...it is simply sad how badly google is whipping them. They should be ashamed.
Oh, and about the quality. Sure some areas look good when they put the sites there, but many, many, better, more authoratative sites will never see the light of day as they simply do not update the hand fed serps.
| 2:31 pm on Aug 8, 2005 (gmt 0)|
I am in agreement with mfishy here. Could this spell the end for Yahoo search?
Oh well! One less algo route to negotiate ;oP
| 3:41 pm on Aug 8, 2005 (gmt 0)|
I can't imagine that Y came all this way on the search front to rely permanently on this sort of thing. Lot of work went into buying, sorting out and combining the other branded search properties that led them to this point. Plus until this update, it seemed to me that they were hot on G's tail, albeit not there yet.
Maybe this is a short term stop gap, maybe not. But the majority of the SERP's we watch still don't have much or any hand editing. Besides, this update is so off the mark that they could not rely on hand editing to fix it all anyway. For quality, reliable, manageable SERP's, the vast majority of the results they show must be algo driven. They have no choice but to get it right.
And with the growing realization that search (not desktop) may be the way to control the Internet, they have a strong incentive to stay the course. They're still in the second best spot to have a shot.
I also think that an ongoing aspect of Y's SERP's is hurting them right how. Y has always been notoriously spotty in terms of how many pages they serve from a given site. They feature a widget site, and show the page on red widgets, but the equally important page on blue widgets is nowhere to be found. Tim mentioned in N.O. (I think) that more deep linking to a site encourages more thorough/deeper spidering (Tim feel free to correct that!). There is of course some logic to this as it follows the basic concept of PR or page importance on the WWW. But sitewide indicators of quality matter too, and in this update, those seem to have been too heavily discounted or bent in some odd way. Y! ought to be far more concerned about grabbing all the d*mned pages from any site worth featuring, and show them. That alone would fix about 25% of the current problem. It's a bigger issue than most talk about, I think.
For now I'm certainly prepared to view this as a sidetrack on their march to better, more interpretive SERP's. This update seems to be a fundamental shift in the way their algo operates, just as the Florida update and sandboxing were for G. It may take Y! a while to get it right. It took G a couple months after each major update since Florida update to get the SERP's looking decent.
That said, this hand editing thing is worrisome, since it does suggest the possibility that they may not see their way to a short term fix with this algo.
| 4:05 pm on Aug 8, 2005 (gmt 0)|
If hand coding serps makes them more relevent then more power to them.....
I am finding Yahoo much more relevent than google these days.. today is a good example.. I was looking for a hotel in a famous us gambling destination by searching for
Widget Hotel, City, State
I have read on a post a few months back about a reporter commenting how Google is is way off track here but until today, I didn't realize how bad it was.. I couldn't find the hotel in first 3 pages...
I went to yahoo and it was NUMBER 1 where it deserves to be for this search....
The hotel in question does not have a domain name that exactly matches the name so I had to search for it...
BTW, in google the hotel in question was on page 9. (90-100 results)....
I hear complaints that the serps are not perfect.. and now I'm hearing complaints that people are trying to improve them manually... Which one do you guys want? You can't have both perfect serps and non-human-edited serps...
| 4:24 pm on Aug 8, 2005 (gmt 0)|
I think you are missing the plot here.
This is more of a "big picture" discussion. Whether or not a few serps look better or worse is not that important. The fact that Y! may not be able to run an algo based SE is what is significant.
| 5:18 pm on Aug 8, 2005 (gmt 0)|
mfishy - I agree with you on the issue of authoritive sites, but still think a combination of algo and hand editing will ultimately produce the best quality serps.
Yahoo is prone to this more than Google, as hand editing was how they did it once upon a time.
Right now though the serps in Yahoo are a mess. Across many categories too, not just one or two.
I hope they fine tune things alot from what we seeing today.
| 5:31 pm on Aug 8, 2005 (gmt 0)|
Great question pmac! It's loaded with an assumption, just like saying, "Are aliens still abducting humans?"
|Is Yahoo Hand Coding More Serps? |
If anyone has PROOF that any SERPs are being hard coded/hand coded, I'd love to see it.
YPN is contexually driven and Y is moving to contextual search. There will be a period of tuning and refining as there was post-Florida.
[edited by: martinibuster at 9:31 pm (utc) on Aug. 10, 2005]
[edit reason] clarify the wording [/edit]
| 5:39 pm on Aug 8, 2005 (gmt 0)|
|If anyone has PROOF that any SERPs are being hard coded/hand coded, I'd love to see it. |
Are you kidding? Yahoo themselves do not even deny it. If we could post searches here, some are so glaringly obvious that my dog could see it...
| 5:44 pm on Aug 8, 2005 (gmt 0)|
I took a look at the serps in question, and according to pmac, the changes to a specific industry were done yesterday. It looks like a large amount of these serps are top heavy in the top 5 with corporate and .gov sites.
One of these .gov sites keeps popping up for these serps if it contains the word on the page which is fine because it's relevant, but what is funny is that, to me, it looks like this .gov site has been given so much mojo credit, that it is popping up in completely unrelated searches.
Trusted sites in the top five?
To me, it looks less like a hand coded serp, and more like they gave a group of sites more bonus points if they contained part of the query string. It's like having a large group of trusted sites that they can rely on for answers to queries normally spattered with spam.
If Yahoo gave extra credit to certain sites, then that might explain some of what you are seeing. Might also explain the weird search I spotted the other day with the .gov ranking way higher than it has any reason to, despite the the query terms in the backlinks.
Just my hunch.
The bottom five of the top ten don't seem to be affected.
| 6:36 pm on Aug 8, 2005 (gmt 0)|
|To me, it looks less like a hand coded serp, and more like they gave a group of sites more bonus points if they contained part of the query string |
Nah, this has been going on for well over a year. Some serps do not move...I mean at all....ever...through every update. Many are not .gov sites or anything close to that - simply old sites from the Y! directory.
For a while you could actually see the h=1 in the url but they hid that...This is not exactly a secret within the SEO community btw (or among the employees of all the SE's) :)
| 6:55 pm on Aug 8, 2005 (gmt 0)|
|For a while you could actually see the h=1 in the url but they hid that... |
Specific to the H1 in the url string, that's an old wive's tale. A number of my spammy ass sites had that in there and I'm positive Yahoo hadn't hand coded that in there. lol
I'm well aware that Ink has always been out about hand tweaking serps. But I think it's a gross oversimplification to say that Yahoo hand tweaked the top five for hundreds of searches within this niche. I am not saying it's impossible. It's entirely possible.
What I am saying is that it makes more sense to automate the process by creating a batch of trusted sites, which would explain why I see one of these sites popping out in unrelated searches. That's an anomalie that happens within an automated process.
| 8:43 pm on Aug 8, 2005 (gmt 0)|
Hand ranked serps are from the Directory, or drawn from a megasite for obvious reasons, with "microsoft" the example used here before.
| 8:56 pm on Aug 8, 2005 (gmt 0)|
|But I think it's a gross oversimplification to say that Yahoo hand tweaked the top five for hundreds of searches within this niche. |
What niche? I see it across dozens of areas. Anyhow, this would have been a better thread if we could maybe argue elswhere about whether or not they are doing it. I was more interested in talking about the impact since they obviously are hand editing serps...still. I hope those that do not know this are not charging clients for SEO...
|Hand sorting is the future. It has to be. |
The future of what? There are 100's of milllions of queries daily. How are the Se's going to hand sort this again?
| 9:06 pm on Aug 8, 2005 (gmt 0)|
|Hand sorting is the future. It has to be. |
Given the speed at which web is growing in terms of the content that is going online every day and what is already there waiting to be indexed, I don't think hand sorting is the answer or the future. As mfishy mentioned, already there are SERPs in which top-5 sites have not moved in last over 1 year. This is what hand sorting can do to the SERPs.
Ok, mfishy also updated that before me :)
| 9:49 pm on Aug 8, 2005 (gmt 0)|
Yahoo has been hard coding for years, this is nothing new.
What is new, is the fact that hundreds of the most competitive keywords have been hard coded in the last 72 hours. To invest hundreds of millions of dollars into search technology, and then just throw it aside and let human editors pick "the best" sites, seems like a waste to me.
They might as well just use DMOZ listings for their search results.
| 11:44 pm on Aug 8, 2005 (gmt 0)|
I think I've seen these called authority sites somewhere... Could it be that Yahoo is part way into a switch to contextual based search and they're setting up some lexical reference points?
|To me, it looks less like a hand coded serp, and more like they gave a group of sites more bonus points if they contained part of the query string. |
Nah! Has to be hard coding. ;-)
| 12:42 am on Aug 9, 2005 (gmt 0)|
|Could it be that Yahoo is part way into a switch to contextual based search and they're setting up some lexical reference points? |
| 1:18 am on Aug 9, 2005 (gmt 0)|
"How are the Se's going to hand sort this again?"
Um like they are. No need to be blind about this. They have millions of searches but that is irrelevant to the discussion. Hand ranking is plainly easy to do if you are an engine if you restrict it to virtually any criteria. For example: what are the fifty most common one word searches? It's absurd to assert this wouldn't take minimal humanpower to manually check and sort, even with a few levels of backup (three levels of hand sorting or whatever; supervisors checking regular sorters).
Then, once you do this, you even have a seed pool of results. Links to and from the hand sorted results would get a higher authority value than normal ones within the niche, or even within search terms involving the one word.
It's an excellent practice that becomes more inevitable everytime one of the algos craps all over itself.
Also, just because the hand sorted results don't get updated doesn't mean they can't. That's just silly. It would be a trivial amount of work to keep the 50 or 100 most popular search terms hand sorted. These are companies making hundredsof millions of dollars. They can hire a handful of people and not die.
| 3:34 pm on Aug 9, 2005 (gmt 0)|
I see some evidence of what looks like hand edited serps in the niche I follow. For the major 2 keyword phrase in my industry the serps in the top 10 cleaned up a bit a couple days ago. When I use quotes with the keyword phrase it reverts back to the really bad serps that existed without the quotes prior to the cleanup. I really do not think these hard coded serps will stay, but rather these are just a temporary fix for major keyword searches until they fix whatever problem they are having at the present with the algo. Using humans to edit serps creates and unfair bias whereas an algo can not be bias one way or the other.
What really frustrates me is seeing the other countries web results look so much better:
And the other Yahoo supported portals:
Not perfect by any means, but at least these serps look like some logical algo ranking criteria is being used unlike the U.S. serps.
| 5:05 pm on Aug 9, 2005 (gmt 0)|
Not to side track convesation, but no SE has solved Spam. Google may be further ahead but really the sandbox is the moral equiv to hand editing SERPs. Both show that the underlying algo can't handle the strength of mfishy's army of tools ;-)
| 5:12 pm on Aug 9, 2005 (gmt 0)|
I am not necessarily seeing the hand coding... sorry.
What I do see is "Hey, intern, see that nob over there that says 'Directory Link Weight' under it? - Good, turn it all the way to the right, and that other one with 'DMOZ' under it? Yeah, that one too, and while you're at it go ahead and trun the .gov one a little. Thanks."
OH, BTW while I was checking results with a friend who is two states away, multiple searches in our niche resulted in these sets of results:
Two word general search:
1,3,7,9 DMOZ sites
2,4,5,8,10 Directory sites
Three word general search:
1,2,5,7,8 Directory sites
3,4,9 DMOZ sites
6 Yellow pages.
| 9:41 pm on Aug 9, 2005 (gmt 0)|
"Using humans to edit serps creates and unfair bias whereas an algo can not be bias one way or the other."
Of course it can, and always does.
| 1:29 am on Aug 10, 2005 (gmt 0)|
|Of course it can, and always does. |
Well in a way you are correct. However, the bias is universal without exceptions. I.E. programming an algo to favor 32" waist jeans and apply more weight to ranking for instance. When was the last time you picked up 5 pairs of jeans each from different manufactures that all had the same inseam and waist identified, then tried them on and liked how all 5 of them looked and felt without a preference?
Subtle differences in the way a webpage is laid out or the way it appears can effect the physiological aspect of how it influences an editor.
Point is even with guidelines and rules for humans you are introducing unpredictable outcomes when you allow the psychological human make decisions such as this.
| 5:01 am on Aug 10, 2005 (gmt 0)|
|What niche? I see it across dozens of areas. |
I'm looking at a specific niche that had their entire top ten changed five hours prior to this thread being made.
What this thread is about is if these recent changes to the serps (if you see recent changes) are a response to the algo not cutting it.
|...this looks like its creeping into more and more areas. With all the outcry over the latest update, has Y! decided that the algo just isn't cutting it and they now have to hand tweak in order to have clean results? |
mfishy, I think the frustration you may be feeling is that you may be having a different perception of what this discussion is about. Nobody is arguing with you whether or not Yahoo hand tweaks in general actually occur. It's well known that hand tweaking has occurred in the past.
What this thread is about is if anyone else has noticed an increase of serps that look suspiciously like a hand tweak, and if the recent changes are a response to the algo not cutting it.
Just a friendly heads up to bring this back on topic. Thanks.
| 12:03 pm on Aug 10, 2005 (gmt 0)|
|Nobody is arguing with you whether or not Yahoo hand tweaks in general actually occur |
OK, it looked like some people had their head in the sand on this one - thinking yahoo didn't often hand pick serps. Seemed odd as yahoo was when google was feeding them the serps - certain terms were just Y directory selected by people (100% different than the serps google showed). Those serps remain the same 1.5 years later, even with the addition of their new algo. Then, they started feeding msn serps after they bought INK, and the hand picked serps were not displayed on msn.
The hand picking of termsdefinitely increased, if that is what this discussion is about. I have received messages from people in many industries stating the same.
If Y! is going to go this route on more terms, I only hope they find a way to update the categories and pick a wider variety in terms of "types" of sites, as they seem to largely ignore the niche stuff thatr eally focuses on a singular topic in favor of mega sites with a page or two with a definiton or something of the like.
Actually one has to question whether the editors at Y! EVER look at the directory serps they hard code again, after they have been selected as no updating has occured in a long, long time.
| 2:05 pm on Aug 10, 2005 (gmt 0)|
I did. I just haven't seen it in results I follow, so I've paid little attention to recent discussions.
|OK, it looked like some people had their head in the sand on this one |
Must be we hang out in different neighborhoods! :)
| This 68 message thread spans 3 pages: 68 (  2 3 ) > > |