3: Google Updates and SERP Changes - May 2010

[tyler756]

< continued from [webmasterworld.com...] >

I have been attempting to follow this thread to the best of my ability but must admit there is information in here that is a little out of my league.

Has anyone been able to do anything to help increase rankings? I lost rankings for 3 my keywords (I am not ranking for a ton of them yet) and these 3 drive a lot of my traffic. I dropped down from the 1 spot to position 5 and 6 for two of the keywords, and from the 3rd spot to the third page for the other major keyword. I am not sure what to do and we are about to enter the "slow months" for my industry and my traffic is dropping every day.

Just thought I would post to see if anyone is having any luck

[edited by: tedster at 7:41 pm (utc) on May 22, 2010]

[imbckagn]

If you want to figure out whats going on everyone needs to be on the same page and solve one problem at a time. I think the best place to start is to hear from those who have built links from May 4th till now. I'm choosing May 4th because the Google interface/algorithm changed for most people on May 4th.

So I think the focus should be:

1. What day did your rankings drop?
2. Did you loose main keywords, longtail or both?
3. Have you acquired backlinks after May 4th?
4. What type of links did you acquire?
5. Have you seen an improvement in your rankings?


At least this way we can figure out if Google is discounting links. I'll start and if anyone feels this is a decent idea they can follow.

1. May 4th
2. Main and some longtail
3. Yes
4. 65 Blog post contextual links
5. Haven't budged

[tessmac]

1. May 5th
2. Everything
3. Yes
4. Blog posts, social network links
5. A few keyowrds have returned to a much lower position, but seem stuck there. Lots of poor sites ranking above, results in serps "appear" outdated.

## I think its worth adding here that we only target about 12 main keyword terms. These are fabulously competitive. However this does make it easier to spot trends and changes in the serps as we check positions manually through search. Sites who are targetting 100's or 1000's of keywords may struggle to get a picture of what has changed in the serps for a particular keyword ( by that I dont mean the position of the keyword, but what is going on around it )

[bowdeni]

In regards to the devaluing of sidebar & footer links, I follow a very competitive keyword SERPs. I know of two competitors who have link profiles heavily consisting of sidebar & footer links and they have dropped significantly in rankings, but this drop commenced end of March and not Mayday.

I don't know if there is any value to this, but thought best to put it out there since people seem to only like putting thesis supporting evidence out there.

[anand84]

I have very few inbound links from footer/sidebar of other websites. Most of the links I have are from within the article content on other websites. And I have still lost traffic.

Additionally, another website I know has several thousand IBLs from sidebar blog-rolls and the website does not seem to have been affected at all. So I am really not sure if the drop in SERPS is because of this.

but this drop commenced end of March and not Mayday.

Website owners have been seeing the 'Mayday' drop since Mid-February. So, it is possible that the drop in traffic is because of the factor you have mentioned.

[bowdeni]

Ok I'll throw a couple of more things in.

The website I'm working on used to have a very similar link profile to the competitors, i.e. links on same pages (sidebar/footer... previous seo agency). In the past our rankings fluctuate with a reasonable level of correlation as one would expect. Since we've worked on the website, we've been removing such links and creating out own link profile through content creation and distribution.

I mentioned that our ranking fluctuations were similar, well the drop our competitors had since late April marks the first shift away from the correlation. We haven't had the same (big) drop they had. (Maybe as a result of shift from sidebar/footer links to content, but many other variables involved.

We have dropped a little too. . We have our own problems too, so it's hard to put these things together and draw any correlations, but perhaps someone else will analyse something here.

Going back a month or so when many in the UK including myself were being served different indexes from DCs, I had an indicator I could use to tell apart the two indexes. In Feb I had a trailing slash on a popular on URL, and we changed the preferred version to non trailing slash. Google quickly indexed the non trailing slash. With the caffeine results however, it drew on the older slashed version but only for a week or so.

In the past two weeks, I've only seen the trailing slash version in the SERPs, indicating to me that the results are a little old, but hey, only hunches!

[seoN00B]

1. It started to dive on April 28.
2. Much on the long tail probably.
3. Yes
4. themed links
5. yes but not stable, it will up and drop.

[tessmac]

If "Old Results" are being served, then logically new links wont move you in the old results as they are being driven by old links, amongst other things.

Sorry to bang on about this but I think this is what I am seeing.

Another theory is that I'm losing my mind...

[edited by: tessmac at 9:45 am (utc) on May 26, 2010]

[Hissingsid]

Results here in UK, in my niche, have now gone back to something like pre MayDay. Sites that I previously saw get a -10 or -6 penalty have now gone back to -10 and/or -6.

google.com and .co.uk serps are so close it looks like they still are not applying the UK filter.

Once they switch back UK filter I'll be pretty satisfied with these results.

Cheers

Sid

[Andylew]

Just an update on the route I am taking ref earlier post, convinced there is nothing wrong with IBLs and sitemaps still reporting a substantial number more indexed pages than site: (930k v 80k) indicating there is an index being serverd which is different to the one wmt is reporting we took a hardware route.

The reasoning is if google is reindexing or taking more account of speed or perhaps sites per server = authority we wanted to make sure that they could reindex our entire site in the fastest time possible.

So a new server later and promissing news, within hours of switching we have more than trebbled the gbot activity on the site. Although the old server was not struggling google was clearly throttling the speed that it crawled considerably, perhaps an equation between crawl speed and no pages resulting in a drop - perhaps they are thinking a site with massive amount of pages that cant handle intense gbot traffic is clearly not very good. We used to see 4-5 distinct ips crawling at any time in the day, we are now seeing between 12 and 20 distinct concurrent gbot ips.

[WebPixie]

@CainIV

Certainly some interesting changes. A website we rarely worked on just leap-frogged several other authorities in a genre I watch, with little to no work done on it (now position 3, up from 8-10)

I noticed that, congrats. This is a very competitive keyword phrase and two of the sites his site leap-frogged have been top 3 for years. Those two authority sites use similar SEO methods, which is to mainly buy SEO/traffic links from related sites so that might be a clue as to part of what has changed.

[ohno]

Well this week is tracking the terrible week of the 10th for us. Sorry but in no time since we started (2002) have we had sales up & down like this. Good week bad week good week bad week. Will this continue? Will it settle down & give us bad weeks until we tweak something? Will it settle down & give us the good weeks we always used to have? Is there ANYTHING coming from Google as to what is going on? All i know is on a bad week & may as well close & go out :(

[cien]

I also went Andy's route. Been playing with the idea for awhile and yesterday I bought a VPS to see if things get better. Although traffic was starting to "slowly" pick up and things were slowly getting better, I have the suspicion that something related to the server I am on has been the problem. I cannot get more than one Google spider stuck on the same page even though there is a lot content to index. I have tried almost everything to get their attention. Hope this new server move helps.

[edited by: cien at 1:56 pm (utc) on May 26, 2010]

[walkman]

So we got two theories:
- G has devalued ROS /Blogroll type links
- Google is in the process of refreshing its db and some data on links is still missing.

[Andylew]

Definately on to something with a server upgrade (allthough time will tell if it is related to mayday) now looking at 36 distinct googlebot ips on the site concurrently! In comparison the 'massive googlebot activity' reported just after mayday was relating to around 5 ips up from 1-2 concurrant in april.

I wonder whether mayday was in fact a speed update, ie sites were lost from the top ranks if google couldnt crawl them quickly enough to give them credability. Ie a large site that could only be crawled at the rate of a small site would be of less worth as new content could take months to find and therefor be outdated. Also a large site that could only be crawled at a rate of a small site would have less credibility - harware infrastructure missing to support such a number of pages so unlikely to be seen as authorititive.

[Reno]

I'm still wondering about this post from chrism:

1. Approx 40 days, with a brief return exactly a month after it dropped off.
2. have returned

Does this mean that after 40 days the traffic returned to near normal? I'm looking for straws to grasp, so am hoping this is the interpretation...

..........................

[chrism]

reno - yes within normal bounds from before all this starting. still there for now too - note my confidence!

strangely its less traffic per keyword over more keywords now too.

[imbckagn]

FYI: I moved to a very fast dedicated box on May 10th and it didn't do a thing for me. I now score 93 with firebug and very high with other speed tools. I too thought it may have something to do with speed but it hasn't helped me.

I also validated my entire website with W3C and CSS at the same time as the sever upgrade.

[edited by: imbckagn at 4:18 pm (utc) on May 26, 2010]

[Reno]

Thanks chris for the prompt answer -- I can now go into this holiday weekend with a sliver of hope that the worst may be over sometime in early June!

.......................

[suratmedia]

Anyone noticed inside Google Webmaster Tools, Sidewiki thing is gone from Labs section.

[tedster]

A couple of comments:

1. If you're concerned about the site speed factor causing a massive traffic loss - I'd say relax about that one. First, it's a minor factor. And second, server speed is only the small part of the Speed Speed metric. The bigger part is all about the page's rendering speed in a browser, and those fixes are within your ability to change without moving servers.

2. As for the date, yes there have certainly been ranking changes all along. But in this case we are calling Mayday, we're looking at a roll-out right over the end of April and beginning of May. At the Google I/O conference, Matt Cutts confirmed that it is an algorithm update - something he has very rarely confirmed in the past. < see this article from Vanessa Fox [searchengineland.com] >

So I'd say it's more than a bit of dial twisting around the factors we already know. Although that may also be involved, Matt never called that kind of event an "algorithm update" in the past. Instead it may be the introduction of a new ranking factor[s], or it may even be a whole new methodology.

[edited by: tedster at 5:45 pm (utc) on May 28, 2010]

[ohno]

Yep sidewiki gone, my sites are showing indexed pages have gone DOWN in the last day or so too.

[edited by: ohno at 6:37 pm (utc) on May 26, 2010]

[Bewenched]

1. First week of may +- a few days
2. appears we lost both
3. Yes
4. Social and forums
5. Not yet. Traffic still down about 22% from what we had in april and down about 50% of what we did in March.

Andylew- We recently upgraded our server back in March to a much much faster server and things were great .. google was loving us, but now 2 months later we are seeing large losses in traffic and crawling. Not trying to be a doomsayer .. but just when we thought all was good, it's almost as if the bottom fell out

[Andylew]

If you're about the speed factor causing a massive traffic loss - I'd say relax about that one. First, it's a minor factor. And second, server speed is only the small part of the Speed Speed metric. The bigger part is all about the page's rendering speed in a browser, and those fixes are within your ability to change without moving servers.

I agree partly, however basic bandwidth throttling on shared server would restrict googles ability to spider an entire site within a time period.

A very simple way of comparing with similar content and I would say improving their algo would be to add a ratio; site pages:max crawl pages pcm.

The logic is quite straight forward,

site 1: 5 million pages of which google can only spider 1mil pcm based on the server limitations not googles but all pages are updated once per month. That means some pages may only be crawled once every 5 months so google index could have pages listed that are 4-5 months out of date.

site 2: Same site but because of server resource google can crawl all 5mil per month. That means google would have pages that were never more than a month out of date.

Site 3: Same site but enough server resource for google to crawl 5mil pages per day. Resulting in near real time indexing of new pages.

If they were the only 3 sites in the world IMO they would rank, site 3, site 2, site 1.

Now on a large scale the same applies if their latest infrastructure and algo means they can now in effect real time list results then the time it takes to spider an entire site to find new content is going to be paramount or they will never be able to find the new content. Previously googles own infrastructure limitations would have probably meant that they couldnt spider a 5mil page site in a month (or pre defined period) so couldnt use this to penalise such a site. Now they have more resource they have removed their own limitations so can now use webmasters hardware limitations as a direct ranking factor?

[ohno]

May be just me but i checked some other things in webmaster tools & one of my sites was set to display URL's with the www and the other was not. When i tried to change this it said the site was not verified(despite the HTML file check completing and the history saying validation was completed days ago!), i had to upload a new meta tag to the home pages and setup up the non www as a site in webmaster tools. I had a 301 redirect setup anyway to avoid duplicate content, is this a good thing?

[Reno]

i had to upload a new meta tag

I've had to do this a number of times with various sites since starting with GWT, though it's not been necessary for quite some time, so I think we can say with complete certainty that it has zero effect on the up & down stats you are seeing. That's purely Caffeine, not verification related.

...............................

[SEOPTI]

Today again something shifted with long tail, I see less traffic on all my long tail sites starting today.

[walkman]

There's no way that speed can account for that much of drop, unless your page takes 2 minutes to load.

[Kristos]

I don't know if this was mentioned, please excuse if it was, this is a long thread and I have to get back to business.
One of our clients has a dotPRO domain in the computer field, 1 year old, 49k product and adding another 100k this month.
Rich user experience, relevant videos on every page, properly optimized.
we have him up to consistently over 100k unique visitors per month and this month things dropped dramatically as far as traffic, biz is still good.

so we made some changes that appear to have proven their worth and rankings are climbing
we had alot of long urls (trying to get each product on the page in the file name)so we shortend urls like
domain.com/red-widget-golden-widengets-rare-widegets-cheap-widgets.htm
down to domain.com/brand-widgets.htm
we performed this sitewide and we are seeing positive results

[Plan_D]

Matt Cutts confirmed that it is an algorithm update

Is there any possibility that MC could be being economical with the verite''. To me this feals like very serious teathing problems with implementing caffeine.

I don't think, whatever changes you were making to your 'algorithm' you would want deliberately to create the kind of mess that we are seeing. Especially out of the blue with no warning.

Last year google were bigging themselves up on how consultative they were being with the implementation of caffeine. Even letting us play with a sandbox version etc. But somethings gone wrong somewhere.

MC is googles ambassador to the web world.

What is it Sir Henry Wooton said about ambassadors: "An ambassador is a man of virtue sent to lie abroad for his country"

btw I had to look on bing to find that quote. google would not come up with anything even though I had typed in: "ambassador is someone who lies for their country" to try to find the quote.

Something not right me thinks...

[anand84]

we had alot of long urls (trying to get each product on the page in the file name)so we shortend urls like
domain.com/red-widget-golden-widengets-rare-widegets-cheap-widgets.htm
down to domain.com/brand-widgets.htm
we performed this sitewide and we are seeing positive results

That's an interesting observation. My URLs all contain these long phrases in them. Does anybody see a correlation between URL length and traffic?

But having said that, if I look back at the sites that have replaced my pages, they are no different. They too have similar long phrases in the URL.