Seeing Major Serp Changes
| 6:16 am on Jul 20, 2005 (gmt 0)|
Is it just me or has there been a Yahoo update in the SERPs?
Seems my positions are fluctuating all over the place at the moment.
And, why can't we name Yahoo updates? Yahoo seems to be updating every couple of weeks now and it's spiders won't leave my sites! :>~
| 8:30 am on Jul 30, 2005 (gmt 0)|
Did anyone else just feel that?"
Suddenly the oddest results I have ever seen in my niche. Not necessarlly irrelevant, just odd.
Kind of interesting to see sites that I have NEVER come across before.
| 9:02 am on Jul 30, 2005 (gmt 0)|
Can anybody make any sense of this yet?
My site has been in Y search for about 5/6 years, never at the top, but always page 1 or 2.
Can't find it now for my major (competitive) keyword phrase.
If I search for company name, it's still there, but on page 2 or 3 or lower.
It looks like I've triggered something, but don't know what.
| 10:36 am on Jul 30, 2005 (gmt 0)|
Serp is changing...
It's interestint to note that browsing from Italy serpo is about the same; only few new entries.
Browsing using an anonymizer software ( so that I can see Overture PPC result too ) results are different!
Results seems to me fine. For the niche I monitor I can say:
- Big sites with a lot of links and a lot of satellite sites are scoring fine
- results are similar to msn.com
- it's better to use css
- New sites are scoring fine as well
- Fresh and continuos links are important
| 12:02 pm on Jul 30, 2005 (gmt 0)|
I'm now showing luck number 13 pages left in the index. The only pages left are in cache - that was not true before.
I'm still hoping for a rally! One way or the other, I'll be glad when this one is over.
| 3:03 pm on Jul 30, 2005 (gmt 0)|
There has definitely been another decent sized change today (or yesterday, wasn't online much) since it initially started updating
| 3:48 pm on Jul 30, 2005 (gmt 0)|
Doesn't look like on page counts at the moment.
Not sure about incoming links as factor, have site with about 5 incoming, PR0 and ranks in top 3.
| 5:01 pm on Jul 30, 2005 (gmt 0)|
"Whatever they're doing, with all good intent no doubt, is killing relevant subpages in favor of mass/bulk sites with little content matching 3+ word searches in way too many cases."
This is what I'm seeing as well. The result is poor relevancy on specific searches.
| 5:04 pm on Jul 30, 2005 (gmt 0)|
Where are you guys seeing this - Yahoo.com - where from UK?
I cant see anything different - is it consistently on there or jumping backwards and forwards. Might be certain sectors I guess.
| 5:06 pm on Jul 30, 2005 (gmt 0)|
My results suggest they have tried to apply some kind of over-optimization filter. I'm losing pretty much everything with a two-keyword filename and two-keyword anchor links. It's really crude and dumb.
| 5:25 pm on Jul 30, 2005 (gmt 0)|
One of my sites has been steadily climbing from 60+ to most recently #12, just found it at #101.
Don't really understand, nothings been changed...
| 6:08 pm on Jul 30, 2005 (gmt 0)|
>Where are you guys seeing this - Yahoo.com
I think that from Europe you should go to Yahoo.com using an anonymizer software.
For example from Italy if I search on Yahoo.com, normally I do not see the Overture/yahoo listing.
If I use an anonymizer software I can see the Overture listings and "unfortunately" the new Serp.
Unfortunately because i lost few positions on my preferred keywords.
I do not think the serp is ugly; I think is good.
| 6:09 pm on Jul 30, 2005 (gmt 0)|
Revisiting some observations I made in msg #138, plus one big new one (or at least I haven't seen it mentioned)...
|On page 2 - 3 of one of the searches, incidentally, Yahoo had previously returned 9 identical pages (judging by idiosyncratic filename) from different sites. Many were redirect pages. They've gotten the count down to 3 so far on this update, so that's obviously an area they're working on. |
These pages are now completey gone. Not to say that other low quality pages haven't risen up in their place, but these have vanished.
|Generally, but not always, the index page is a little bit up and the inside pages are a little bit down. |
- I'm now seeing some index pages down, and some inside pages up. Maybe I just didn't notice these before.
At this point, I'm finding it hard to notice general patterns... the results seem that chaotic, and obviously factors on the web are different as well, but here are some interesting observations...
- On some searches, the 800-pound gorillas (eg, Orbitz and Expedia in travel) are now up at the top for searches that they never previously ranked on.
- And, on some searches, the -asdf style text string, of the type that once worked on Google, seems to affect searches on Yahoo too, but not as much... ie, they won't get you back to where you used to be, but they will shift things...
...eg, on one index page search, where we've dropped from #1 to #15, applying the text string moves us up to #10. On the same search, though, an inside page moved up from #29 to #12. With "-asdf" or whatever added in a bunch of times, that page then drops to #18.
On another site, an inside page that had long been #1 has dropped to #8. Applying "-asdf" repeated puts us back at #1. I should mention that this site had known dupe content issues (ie, a distributor had quoted a bunch of our pages verbatim, put them all on one page), but I don't know whether the dupe issue entered into it.
From these examples (and others I've tried), it appears that there are several sets of factors at work, some of which can be affected by the non-sense text strings, which affect "filtering," and some of which are independent of "filtering."
- Re page movement and patterns, I'm seeing that some index page searches that were surviving in the top 5 have dropped to the mid to bottom of page 2. Other index page searches for similar, but not exactly the same searches have remained untouched. Some inside pages are also dropping down further. Personal biases aside, I'm agreeing that the results are currently terrible.
| 6:17 pm on Jul 30, 2005 (gmt 0)|
This is a very strange update that Y! has done. 4 of my sites went from #1 for their major keyword to 24, 45, 124 and nowhere to be found respectively.
I am now convinced that Y! is doing major updates every few weeks. I hope this stablizes over the next few months.
In the mean time, I will have to monitor the links, pages indexed and any changes in the results.
| 6:29 pm on Jul 30, 2005 (gmt 0)|
Well there's been some word that their recent algo is valuing backlinks more, and on page less.
I can buy the "onpage less" part from what I'm seeing, but the backlinks alone doesn't seem to fully explain it. The backlinks may explain the large number of mega site pages showing up now, instead of more targeted relevant subs from smaller and mid sized sites. But we have some sites that are doing really well now for some important terms (for the first time ever) where those sites have minimal backlinks (except the scrapper backlinks, which I suppose could be playing a role here).
Personally, I think that over optimization filters were dramatically tightened here.
Various threads have popped up around the Web wondering at the possibility that Y has also introduced 'sandbox like' elements. I've noted that I see some evidence of that, but more related to trying to cut out extremes (not so much age related, in other words).
| 6:33 pm on Jul 30, 2005 (gmt 0)|
I too have noticed a "sandbox like" feature being implemented in Y!
| 6:38 pm on Jul 30, 2005 (gmt 0)|
|Well there's been some word that their recent algo is valuing backlinks more, and on page less. |
When Google Florida happened, I felt initially that they'd simply reduced the boost given by internal anchor text way down. This has much the same feeling, particularly with the "-pfdrq" text strings bringing results back up... some variant of Hilltop (maybe we can call it Hillside, for patent reasons).
| 6:41 pm on Jul 30, 2005 (gmt 0)|
|simply turned the boost given by internal anchor text way down |
I don't think this is true, at least not universally. My largest, oldest site's homepage moved up from page 5 to page 1, and it has identical internal link anchor text on every page to the homepage.
I tend to go with the higher weight going to backlinks theory, and possibly a "regular updating" element. The pages I add content to regularly seem to have moved up. But who knows.
| 7:14 pm on Jul 30, 2005 (gmt 0)|
|My largest, oldest site's homepage moved up from page 5 to page 1, and it has identical internal link anchor text on every page to the homepage. |
Chances are, though, that your homepage also has the most external inbound anchor text of all your pages. In some cases, I'm talking about internal pages that have either no, or relatively few, inbound external links.
| 8:12 pm on Jul 30, 2005 (gmt 0)|
RC I tend to agree, meaning I'm more or less guessing same as you right now ... but I also see clear evidence that they're actively excluding some of the most relevant subpages from many sites, while still showing, for example, the site's homepage and possibly another related but less important subpage for a two or three word search - when it's the missing page that is the one most relevant to the search. So in other words, they're still showing the same SITE for the search, but not the site's most relevant page.
That can only mean one thing: OOP. Actually, I don't like that phrase, because it contains the word 'penalty.' It's not a penalty. It's OOF (over optimization filters).
Oddly, these SERP's seem to be a slight positive for us, because the lost subpages are offset by some homepages rising to levels that they have no business rising to. Oh well, go with the flow, I say.
| 8:27 pm on Jul 30, 2005 (gmt 0)|
Regardless of what changes may still be in the works in the two week window, it is discouraging to look at "evolving" serps and not see one single positive element (aside from perhaps a bit more pages indexed).
The current serps are a complete joke. For one big money phrase the first three results are "account suspended" pages that show thousands of blog comment links, that were nonsense text, that were keyword-keyword.subdomains, that were setup just to redirect... not one single element might make you think "well that is a positive direction at least".
Still hopeful about it being better next week, but in a way, the thing to do to improve this mess is to value the near exact opposite of what is being valued now.
| 8:36 pm on Jul 30, 2005 (gmt 0)|
"When Google Florida happened, I felt initially that they'd simply reduced the boost given by internal anchor text way down"
I think it's more than that. I think they may be penalizing for what they perceive to be internal anchor text link spamming.
"I don't think this is true, at least not universally. My largest, oldest site's homepage moved up from page 5 to page 1, and it has identical internal link anchor text on every page to the homepage."
How many pages do you have, though? LOTS of page with identical internal anchor-link-text may
be making a huge difference.
| 8:40 pm on Jul 30, 2005 (gmt 0)|
|How many pages do you have, though? LOTS of page with identical internal anchor-link-text may |
be making a huge difference.
That site has about 275 pages. It gets updated regularly, too, and I'm thinking more and more that has something to do with this update. It rose 40 positions to page 1 and has 16,000,000 competitors in a highly competitive industry.
| 9:38 pm on Jul 30, 2005 (gmt 0)|
|is discouraging to look at "evolving" serps and not see one single positive element |
Yes, it is. Seemed like until now they were on a slow but steady climb to better SERP's. Poof. Gone. The results are so off the mark, in so many ways, that it's hard to even make constructive suggestions.
|may be penalizing for what they perceive to be internal anchor text link spamming. |
Anything's possible but I would seriously doubt this. Reason: This approach takes tons of innocents down too. Another approach is simply to discount the value of these links, which often feels like a penalty, but in fact it's just removing something that worked previously.
That said, the same kw, occuring too many times, too many ways, in too many places, may be enough to get a page filtered out of the SERP's (which is not the same as a penalty).
But this is pretty much going to be all guesswork for a while now anyway, plus we don't know that it's over either. Presumably not.
| 10:01 pm on Jul 30, 2005 (gmt 0)|
>>>but I don't know whether the dupe issue entered into it
Based on some "test" sites I run, I've seen a change in the dup filter come and go during an update, but it usually goes back when it settles down.
>>>where those sites have minimal backlinks
But maybe they have the right recipe for anchor text...
>>> that Y has also introduced 'sandbox like' elements
I can't buy that one. I see new sites making it in - what I'm seeing more of is it taking quite a while for backlinks gained to give the site any type of power - whether that is "aging" or "dumping" who knows. On page changes, until today, have been enough to boost a site quite a bit in the serps - so it seems this "issue" of it taking a while to see effects (to a new or old site) is more related to backlinks, IMHO.
>>> Another approach is simply to discount the value of these links, which often feels like a penalty, but in fact it's just removing something that worked previously.
That's my theory caveman.
>>>That said, the same kw, occuring too many times, too many ways, in too many places, may be enough to get a page filtered out of the SERP's (which is not the same as a penalty).
Devalued until they hit oblivion.
| 10:14 pm on Jul 30, 2005 (gmt 0)|
Maybe I'm way off base here, but it seems that some of the update is using a very old index, with a mixture of new. Am I the only one that thinks this is possible?
| 11:31 pm on Jul 30, 2005 (gmt 0)|
Take that to mean basically 5-10 backlinks plus whatever scrapers there are, FWIW. Really doesn't seem like it should be showing up. I'd have understood it if onpage were still important, but that doesn't seem to be so much now (or, it's being trumped by some negative factors). If it ain't onpage, and it ain't backlinks (unless it's scrapers), I'm not sure what it is. Scrapers seems a stretch, but maybe.
Rae, I think aging is only one aspect of sandboxig. Have to keep reminding myself many/most don't see it that way. It is the other aspects of sandboxing that I wonder about here. Might as well not make that reference any more though. When I say it everyone thinks age mainly.
|discounting value of internal links |
Possibly; maybe even probably. But if so, there's more to it than that, because I see other subpages from the same site showing where the missing ones from that site ought to be showing. The missing ones are being actively filtered out.
Let's say you've just searched on "blue widgets." What I'm seeing is that the widgets.com/blue page that used to rank #1 is gone. Instead, I see two pages: widgets.com at #3, and widgets.com/blue/tiny at #9. Where is widgets.com/blue? Nowhere. (Except that I can get it to come up for other searches. ;-))
IF they're cutting out widgets.com/blue for 'blue widgets' searches for dup reasons (i.e., if they see widgets.com/blue and widgets.com/blue/tiny as too similar, then their dup filters are badly constructed, because they're still serving up the blue/tiny page when that's the one that should be filtered out. But I don't really think that that is what's happening. Still think it's over optimization related. Speaking of which...
|occuring too many times, too many ways, in too many places ... oblivion. |
How exactly it's being done I dunno yet, but that seems to explain much of what I'm seeing WRT lost pages that were good, relevant results. FWIW, I see the same thing in both commercial and non commercial cat's.
Like I said though, a lot of this is still just guesswork.
Frankly it's all such a mess right now, I have trouble believing it'll stay. All I can say is that if it does, I'm happy we kept those other sites alive and well, if neglected. Some of 'em are rockin' for the first time in, umm, well ... forever. ;-) Crazy, man.
<added>Don't want to forget: Those weak subpages from YP sites etc are way too visible in the current SERP's. That part at least seems to reinforce the notion that bigness/backlinks in some fashion are important currently, beyond what was previously so. I'm very curious to hear what the blogspammers think of this update right now. Not that we have any blogspammers in here.</added>
| 12:35 am on Jul 31, 2005 (gmt 0)|
|...I also see clear evidence that they're actively excluding some of the most relevant subpages from many sites, while still showing, for example, the site's homepage and possibly another related but less important subpage for a two or three word search - when it's the missing page that is the one most relevant to the search. So in other words, they're still showing the same SITE for the search, but not the site's most relevant page. |
I'm seeing the same thing about subpages, but...
|That can only mean one thing: OOP. Actually, I don't like that phrase, because it contains the word 'penalty.' It's not a penalty. It's OOF (over optimization filters). |
...I'm not sure about this, particularly on the home page searches. I've been thinking it's curious, considering what I'm seeing on the subpages, that some the home pages are only surviving for searches that are absolutely right on... all the words in inbound links, titles, headings, onpage content.
Previously, some home pages were doing well if they were right on for, say, two words of of a three word search, and had some onpage and linking boost for the rest. Yahoo has never been as amenable to this as Google, eg, but by the end of June they appeared to be digging a little deeper than they had, say, at the beginning of the year.
Earlier in the year (and I haven't tracked dates), Yahoo was in fact ridiculously susceptible to onpage factors, at least on index pages. They seemed to have fixed this recently and were in fact looking very Google-like. Now, what I'm seeing is 800-lb gorillas that don't appear to be very relevant for the query dominating page one, and on page two I'm seeing results that look a lot like what I was seeing on page one at the beginning of the year, again with some ridiculously right-on index pages outranking other sites that deserve to be there.
So, in the small sample I'm looking at, it's not as simple as just OOF. Hope this makes sense. It's hard to describe all this, because it apparently is chaotic.
Again, all this reminds me of discussion after Florida, where OOP was discussed, but there were too many counter-examples for many to believe it.
Keep in mind that there is some filtering (of the kind that can be circumvented by overloading the computations with strings of excluded random text), and it's apparently working differently with different sites... It's affecting different searches to different degrees. This is a two week update, so there have got to be many layers.
Just as, after Florida, it was clear that Google couldn't keep the results they were displaying, I think in this Yahoo update, it's clear at this point that we can eventually hope to see something else.
| 12:55 am on Jul 31, 2005 (gmt 0)|
|Maybe I'm way off base here, but it seems that some of the update is using a very old index, with a mixture of new. Am I the only one that thinks this is possible? |
No, actually, I see the same thing. Earlier this year, Y! was shifting back and forth from an old index and an updated version. We even saw an old cache in the results of some of our sites.
This update certainly does not appear to have corrected Y!'s shortfalls.
| 1:16 am on Jul 31, 2005 (gmt 0)|
using -asdf actually pulls in results from a different data center. I used the same function with the Overture and got identical results. Use the -asdf on any search term or phrase for both
Yahoo's parallel network search engines have not been switched to these horrific results yet and are still using serps from before 7/29.
This may be strong evidence that what we are seeing is not complete or just a test since yahoo serps are on isolated data centers.
| 1:34 am on Jul 31, 2005 (gmt 0)|
|Maybe I'm way off base here, but it seems that some of the update is using a very old index, with a mixture of new. Am I the only one that thinks this is possible? |
I have a pure content site; new page appears five days a week. 950 pages and it will be growing again on Monday.
After the April 1st update, I went from 85 pages to over 700. In June I dropped to 450, but stopped ranking for everything, even the site name. Today I've got only 13 pages left. On the positive side, best Saturday ever (thanks MSN and Google).
As far as OOP goes... Each page has a meta description that matches the page name that matches a single H1 tag. If you call that over optimization, then I'm guilty - and so is W3 - cuz that's that that site does too.
Perhaps my downfall in Yahoo might have been attributed to my popularity elsewhere. Scrapers love my site for some reason - they outrank me for my own site name.
My site's white hat, and I write quality articles (much better than About, which covers some of the same topics) if the SERPs stick, so be it. As the great Governor of California once said - "I'll be back."
By the way, why does Dmoz love About so much? If you examine the site closely, they put up pages full of nothing and even include intentional misspellings. I use them as a benchmark - if you can't top them, then you might as well not write...
| 2:49 am on Jul 31, 2005 (gmt 0)|
"If you examine the site closely, they put up pages full of nothing and even include intentional misspellings"
Absolutely. About pages really suck.
About this Y update, I would characterize it as a major *-up.