Google Updates and SERP Changes - October 2010

[tedster]

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

There are many different kinds of ranking disruptions being reported right now - dropped domain roots, unexplainable traffic spikes and dips. These things often come right before some kind of algo shift. What do you think - is something brewing?

[edited by: tedster at 12:38 pm (utc) on Oct 1, 2010]

[Shaddows]

We've been able to get a slight increase in traffic since MayDay. It was done through lots of hard work [to get to] pretty much normal in terms of traffic

Ok, so you have nailed down the precise nature of the Mayday update, analysed the effects on your site, identified an escape vector and followed it dilligently, always increasing your traffic along the parameters you initially projected?

No?

Then an equally plausible statement is
"We got hurt initially before traffic slowly returned to normal as per many updates. In the meantime we did some frantic fiddling which may or may not have had an impact, whether positive or negative"

I'm not trying criticise you directly, but this thread has been sadly lacking in critical analysis for some time. Too much of it is anecdotal, or just simply an opportunity to moan.

FWIW, I see a BIG uplift in traffic since 1st SEPT (20% above trend) with a corresponding drop in conversion rate, so sales are broadly static (on trend). Referrals shifted at precisely the same time. No visable change in ranking.

A referral shift on 16th Sept, another 6th October- both traffic and conversion neutral (relative to the 1st Sept).

Then the biggy. 12th October, huge referral shift. Traffic-neutral, but conversions back at pre-Sept level. In other words, we are now 20% up on sales. The referrals are NOT the same (even particularly similar) to the the pre-Sept level. It looks alot like a 6-week algorithmic test-program for our site, and we passed.

By the way, we did NOTHING fundamental, just added new pages at the normal rate- although we did get a nice link boost being first-to-market with a number of product releases.

[Jez123]

Then the biggy. 12th October, huge referral shift. Traffic-neutral, but conversions back at pre-Sept level. In other words, we are now 20% up on sales. The referrals are NOT the same (even particularly similar) to the the pre-Sept level. It looks alot like a 6-week algorithmic test-program for our site, and we passed.

What do you think you passed? What was the test for? What did you prove to google? I don't understand what the point of such a test would be. Can you elaborate please Shaddows?

[scottsonline]

@ all
Way too many sites are being effected for it to be a site by site basis. I've given links to other members in a nice not my own. In the week since I've done that the results have changed a few times. Companies with one or two big back links from parent companies are rocking. Imagine if you sold oil cleanup supplies and had a link from mobil which owned you. The other sites could have 500 rock solid white hat links, but that one link would rule. It really favors the multinationals, just like "branding" is doing. It also seems to negate penalties. In one case I'm seeing a number one serp domination from one of these companies that takes every keyword possibility and jams that into a blatant hidden text field. Doesn't matter. A company at the top for 14 years went from 1 to 6. Another company that buys links all over is 2, the 3rd ranked company got moved to sixth. It's not one site that got reevaluated, with newer content under six - eight months old it all got reevaluated.

@shaddows the specifics of this site prohibit me from posting what I do know following a few different industries. What it does show is that duplicate content filters seem to be relaxed between sites, although i cannot be sure cheap paid links are helping I am absolutely sure they aren't hurting some. Google alone seems to be a victim of it's own spidering capabilities and it's been too easy for the imanalmosyexpireddomaim dot com to be fired up and interlinked with 100 others, each having 500 pages where ten paragraph form text links back to subcategory pages of sites. In one case I looked at 60% of the sites on one of these ranked first for the keyword term. It didn't appear to be hurting them.

When I search for a product and get a serp for the google store version of it with a YouTube logo on it instead of a store I know they turned up something. I didn't search for it with a YouTube logo nor will most anyone else. This factor combined with the ease at which spam IS being indexed and in mass counted towards the SERPS is likely what we see shifting. It will take weeks for any of us to be able to even see the new backlinks just created effecting the SERPS and with the speed of these sites appearing I have a hard time thinking google can react. It's enough with their super fast indexing to keep the results in a constant state of flux.

Bing seems less effected. In many cases with bouncing SERPS a quick check shows they haven't indexed the pages of spam google has. Coincidence?

IMO one of the huge issues is out of our control. The adaptive algorithms google has...in some instances I'm finding 7-10 new domain backlinks appearing each day most from expiring/expired/hacked domains with new barely detectable links. On one of these last night the site operator returned 1129 pages. I checked the first 40, nome were in bing. They were all setup perfectly with the page name being general to all the links on the page. Many were so new they weren't cached. They spanned every industry. Probably the $9 buy a 1000 link deals. With all these links appearing I think it's distorting SERPS directly, and indirectly by throwing off the calculations they use for the rest of us too when our sites are viewed in pools.

The battle that will rage now will be between the programmers who are developing these link tools which are shockingly complex, and googles team. Some of these sites are so well done I almost wonder if part of this is being done to monkey with google.

At the same time the resurgence of Asian export sites is "interesting"

Edited: if you really want to see the flaw in branding, take an XML file of a hotly contested 100 products in your niche. Put those on sears buy amazon and other sites that allow you to do that. Within a few days you (well them) will occupy the top spots on Many of those products.

@Robert so you're seeing new content issues around as well? I am seeing sites that link to others appearing above the destination site. Let's say time magazine wrote an article linking to our site by company name...they'd take the 1 spot for our name. I'm seeing that all over by products, articles etc

[Shaddows]

I used "and we passed" a bit flippantly. I do think there was a test program.

Here's my working theory.

Using a multivariate dataset, across a range of different keyphrases, user intents and user types, Google exposed our site in marginal but significant ways (putting us up one place, dropping Universal search, above or below shopping results, etc). They did this with (at least) four seperate sets.

They started with one categorisation of us, tested, retested and finally were finally happy with their results, and re-categorised us.

I imagine we have lost a lot of rankings on some phrases, for some users, some of the time. However, we're now getting non-intuitively improved ranking for a different group of people, in particular contexts.

I think this speaks of much more extensive profiling of users and sites, and more intensive inferring of user-intent. There will be losers, no doubt.

Funnily enough, bounce rate has been more or less static thoughout, apart from a BIG spike of 5% on 16th October.

[Jez123]

I suppose I have seen something similar, more traffic (or better targetted) for worse SERPS and now I am seeing worse traffic (less likely to buy) for better SERPS so I suppose it makes some sense.

When did the traffic (or quality of) return, Shaddows?

[Shaddows]

When did the... quality... return

Went 1st Sept, came back in spades 12th October. Raw traffic up 20% the whole time.

@scottsonline
You can "Buzz" a domain really easily atm. Just hit blogs, forums and SM- NF doesn't matter, and it doesn't necessarily need to be linked.

I'm seeing alot of domain-burning. Buzz a domain, wait for penalty, buzz the next. Rinse and repeat. If you're in an SM-friendly niche, you need to get ahead of the curve.

[Jez123]

Went 1st Sept, came back in spades 12th October. Raw traffic up 20% the whole time.

How do you know you haven't just entered the next phase of testing? I'd call it the FSOS phase.

[scottsonline]

@Shaddows, can you elaborate on buzzing?

I'd believe that this wasn't web wide if it weren't for the fact that I can search formabout any product from any category right now and get the same results. Two weeks ago when I was searching for an led type t v I typed in a specific product and got results. Just now I got a review site, auction site, a 100 top xyz site and then a bunch of pages with links out.

There are a lot of articles showing not from the original content point but instead as within content as a link on other sites. That's just weird.

[Shaddows]

Create the perception of interest. You've heard of "social media buzz". Google gives disproportionate value (albeit for a very short period) to coverage from social media. If you can create a buzz, you get a boost. If you FAKE a buzz, you still get the boost, followed by a filter which looks like a penalty, because there is no underlying value.

@Jez123
Could be, but the referrals aren't the probing, cylcing, seemingly-random slow churn of the previous 6 weeks. If we're still on the cycle, a new phase should be in place by 1st November.

Personally, I'm expecting the pre-holiday updates to roll out over the weekend, so I know my data is going to look a mess.

Out of curiousity, is anyone else in ecommerce putting test pages up now, to see how they move during the inevitable holiday updates (whenever they arrive)?

[tshirtdeal]

Hi,

My site has been killed the last 2 days... homepage which has ranked in the top three positions for 6 years for a not so competitive keyword knocked back to page 7.

All my pages which were doing well for a long time all knocked back 6-7 or 8 pages out of the blue... Pages that were on the first page all gone... Even the most non-competitive pages which ranked in the top 5 for years...

Ugly, ugly, ugly...

any ideas?

btw, this is a white hat site pr4-5 which I only write my own unique content for,

[Jez123]

I am obviosly having "that" day of the week when almighty G decides to throw me some scraps! :-) Last wek it was Monday, the week before it was Tuesday.

For this afternoon all has been back to normal order wise. But I bet the lights go off again tomorrow

[scottsonline]

@tshirt I've been following a friends site perhaps in your niche. Are you in the business of your name?

One thing for all of us in the US, the economy is not what it was at this time last year. Instead of growing it's steady or shrinking.

@Shaddows I'm not sure about the update. You could be right that what I'm seeing is the update still cycling.

Everyone assumes because John mu said there was no update there weren't global
Changes. Doesn't mean they didn't change filters. It's awfully suspicious that at the moment this all happened blogs around the world got knocked out.

[ohno]

Jez-interesting, all of a sudden at about 4.30 we have people all over the site, some sales, some add to carts.

[tshirtdeal]

Hi Scott,

I actually do not to t-shirts anymore, did though... Now my site is on entertainment, reviews of concerts, shows and so forth... I have had success for 6 years now and built it slowly with content, all white hat... The economy as well as new competition (classic supply and demand) has cut me 50% over the last year or more...

This has just killed everything... I spent 2 years building my ranking with fresh unique content, adding news to the pages and so on... Got many pages in the 1st and 2nd page for the keyword... All work has been just knocked away... This is pretty much a total disaster for me...

Everything was fine on Monday, loss in rankings occurred yesterday...

[danimalSK]

Hey tshirtdeal

If it is any consolation I think things are pretty f*cked at Google at the moment.

MySpace has had 100 ish pages indexed in the last 24 hours (out of about 100 million):

[google.com...]

Timeout - around 80
Eventful - around 50
IMDB - around 120

These sites normally have thousands of pages added to the index every day. It is not just the site: operator bugging out either - if you paginate through the results you'll see how few pages have been added.

[danimalSK]

Google is also reporting only ~150 pages indexed on Facebook in the last 24 hours....

Am I the only one who is seeing this?

Is Caffeine / Google broken?

[tshirtdeal]

Thanks Danimal,

I have handled all algo changes and tweaks well over the years, only one time did I have something occur like this where my SERPS just dropped off the map, but the next day things were back....

Still none of my competitors have been effected... Think things will return or am I stuck back here now to try and figure a way to fight back?

[danimalSK]

As far as I can tell there have been 2 clear events:

1) indexing is #*$!ed :)
2) an algo change / filter has rolled out that seems to have predominantly affected European sites. i suspect this is why there is less noise than normal, whatever has happened didn't hit the US. some of JohnMu's comments on the GWT forum are pretty cryptic, but he seems to be alluding to an update.

[scottsonline]

@danimal I don't think he was saying an update occurred but it's kind of clear something did happen or about to as indexing is mush right now.

Ohno and others, how was traffic the rest of the day?

[imbckagn]

I think this thread would be much more productive if people stopped posting every time their website does or doesn't make a sale/have traffic. Maybe if you have some analytics to go with it but all those posts just clutter up the thread.

IMO

[scottsonline]

Are we allowed to post analytics?

Here's a metric: 2835 newly acquired links, an average increase in traffic of 40% per compete pro, alexa backed up by a sneak peak by an account rep at the competitors growth.

Or another with roughly a 6k increase in backlinks targeting a single two word phrase since 8/1 put this company from #19 to #3 for the last two weeks.

I think the rough data on the whole is good.

Something else that may help. An associate wrote a response on an on topic site on an article 48 hours ago on a hunch. Their position in the serp was taken by the prominent site that had a link the the article page. I think we see this around updates where the source is deemed more relevant that the link target.

[RP_Joe]

I have been reading these threads carefully. My sites have not lost but gained. They are tiny sites compared to the people who post here. But I see something in all of this. I only have about 3 years SEO experience. Prior to that I ran a small custom programing business for 20 years. So here is my opinion.

Some parts of the Google index are broke. Its not an algorithm change nor an update.
Lets step back and look what we are dealing with. This is a software program that is used across the planet. This is a service used by BILLIONS of people. The actual code is spread over (who really knows) tens of thousands of servers. Its data may be on another tens of thousands of servers. This is a scale we have never seen before. The largest application in the history of the world.
As others have said, its not one program, how could it be? Its a group of programs managed by a core program I assume.
So why do I think that parts of it are broke? Lets look at the evidence.

1)Some pages that should rank, don't. By now we have a pretty clear idea how to get a page to the top of Google, at least for non competitive term. People are reporting & I have seen in some cases this does not happen. Other search engines zoom them to the top, but not G.

2)Some pages that used to rank by top professionals are no longer in the running. As chronicled here, they have dropped dramatically.

This is where I believe the problem is. Some pages are not judged properly. (just my opinion)

3) Some pages CLEARLY should not be ranking. As noted here, parked domains, dead sites, link spam and others.
Why? I believe they have turn off various parts of the program to aid in troubleshooting. Reduce the number of variables in play to get a better understanding of whats broke.

4) Some traffic dies at certain times. Again because of live testing by the troubleshooters.

5) Some cases (only sometimes) of excessive foreign traffic. In the absence of local traffic, foreign fills the void. Turning off local (or turning down the vol) to troubleshoot, causes an adverse reaction.

6) Some areas seem to be worse than others? It seems from reading here that UK is hit worse than others. This indicates to me that they are trying to localize the broken code.

7) As a general observation, made by many, that the overall quality of the of the SERRPS is not what it used to be.

Now I will not criticize Google. I believe they have the smartest people in the world. They probably have the best version control system, backup system and testing system in the world.
However, when you have a program as large as this, with as many changes as they do, eventually you will have a problem. Talk to anyone who has managed a large software project. Look at Microsoft. Everyone has problems from time to time. No matter how good they are.

So the real question in my mind is, when will it be fixed?
How long did it take MS to fix Vista? I don't know the answer, but I suspect if its not fixed soon this will become famous. Maybe it will be known as the "October Crash".

It might be helpful if Google were to open up with a dialog about the problem. But this may be difficult to control. There may be politics and careers at stake.

Lets hope for the best.

[Robert Charlton]

They are tiny sites compared to the people who post here.

I don't know whether this applies to your sites... but small focused sites that go into depth in a particular area could well have an advantage over very large sites that don't go into depth.

This is not the same as saying build a lot of small sites. This is saying that Google likes really useful material... and that large sites that try to cover a lot of ground superficially may be having problems. Large sites that cover a lot of ground well, and make themselves visible to Google, will probably continue to do well.

[tshirtdeal]

Very good post RP_Joe,

I hope you are correct because I am getting paranoid my site got a penalty...

If it does become a major mainstream crash watch how quick the government jumps into waste tax payer dollars, create new legislation and on and on... They probably have been drooling at the bit for such an opportunity to get better control...

[scottsonline]

Rp Joe I think you're right on. They test things by removing filters. Thats why sites with th high quality but paid backlinks shot to the top. They gradually reinstate filters as they go along.

Our branded traffic searches are down about 20% since last week. Ironically even though the site that replaced us also sells those products our manufacturer reports a 7% drop in sales this week versus last through mid afternoon. They are getting queries from the public that they refer out but their system was never intended to be frontline csr for consumers.

For us the quality relates to the brand filters. It gets released for 60-90 minutes at a time and the conversions come quickly. Then it's back on again.

The sites may harm your computer results have gone away. Replaced by amazon buy sears and other affiliate sites. That usually lasts a few days before it all settles out in our niche.

It is interesting that the largest supplier in a 50 billion dollar industry has had a sales drop since this push to branding in its niche.

I am not seeing the 3/6 results per domain tonight.

[brinked]

well I have something interesting to report.

I am getting different SERP's when I do a search through the google.com firefox search field than I am if I go to www.google.com and then type in my search there.

I noticed this because I was getting different results in IE than FF. But in IE I was going to google.com because it had bing as default. I added the google add on and now I am getting the same thing there as I am in FF.

[scottsonline]

This is probably because you have cookies on the firefox side or other stored data which personalized your search.

The results continue to clean themselves a little bit tonight. Strange influx of foreign traffic again.

I did talk to another friend tonight that heads a large - huge - firm. Sales slowed for them the last ten days so not all of this is on Google. This timemlast year the economy was really growing and we all thought the good times were coming back.

Even if nothing changed traffic might be down for some just due to a much different pattern in the business cycle. Don't underestimate the effects of a hotly contested election cycle on traffic.

[brinked]

no, cookies, web history has nothing to do with it.

I have been doing this for a very long time and I know how to track the SERPs.

I am not the only site affected so there is definitely something brewing.

BTW, when I do a test, if you look at my history on these forums, I do a very detailed analysis and I dont just post something when something changes on my site.

I am usually active on here when there are updates, as such when caffeine was rolling out.

I dont know what google is doing but right now there is a lot of changing in the serps on a daily basis.

I monitor over 30 different industries, so this is not effecting just one niche.

[Spencer]

We're seeing some of our regular checks showing some changes also. I think it's the pre-cursor to the usual pre-Christmas alterations. If my notes are correct they began on a Thursday last year and finally settled in by the following Sunday.

[theskunk]

Hi,

I'm not a regular poster here, but I can say that my 10 year old .com focused on a UK retail niche has had a similar 10 days.

Previously we had been running at around 85% of last years sales. In the last 10 days we've dipped to about 25%. We ave maintaind traffic but lost conversions. We always enjoyed top 1 - 3 for our desired keywords (globally).

In examining my usual search terms I see that:

01) My 'own brand' wesite is for the first time outranking my retail site. So now I shall retail enable it.

02) Visitors are arriving by off focus keywords, ie eg Cricket, or Cricket Bats, rather than cricket balls. Actual they are visit me for surge protectors (which is so far off my focus, but happens to be on my site.

The questions I ask myelf is, should I be trying to sell surge protectors, or should I try and get google back to understanding what I am all about.

I stongly agree that the results are favouring brand owners and manufacturers.

I suspect whereby a site might have additional content that is off focus from the actual niche, then this may be taking up some of the traffic being allocated by google.

Yes I think google is capping our traffic (throttling) based on authority, rank, revisits, bounce, other. Therefore i think 75% of my potential converting/purchasing traffic is being eaten up by off topic visitors. When I hit my limit of organic referrals I drop out of the SERPS.

This is my conclusion as the owner operator of a 10 year old niche .com domain, based in the UK, that has been trading online since day one.

This is the most focused thread on the topic I have seen. Well done to Scottsonline for the keeping it moving.

Google kind'ov sucks right now, not just for the against the wall retailers, but also for the junk results with scraped, spammy infected, popup BS websites that take me back to 1995. And yes, thankyou bookmarks.

ttyl

TheSkunk