Google Updates and SERP Changes - February 2013

[petehall]

< continued from [webmasterworld.com...]

Not just location screwing things up... all web browsers are now using your history to distort results, even with Web History switched off. Only way to check the real results now is Chrome Incognito.

That and the location issues are really screwing things up. If I type in weather, G gives me weather for Windmill Hill, UK... I don't even know where that is. I looked and its 200 miles away at the other end of the UK!

[edited by: tedster at 2:08 pm (utc) on Feb 1, 2013]

[superclown2]

I didn't actually check - and anyway, I should have thought anyone wanting a haircut would search for a barber or hairdresser - but now that I look, I couldn't get a haircut from any of the top ten results

To be fair, one out of the page 1 serps showed me where I could get one, provided that I was a student (well past that stage I'm afraid). I learned a bit about high finance, music, olympic heroes and ladies' modern hair styles too.

With Mr Schmidt's knowledge of marketing at this doleful level it's a wonder that Google ever outgrew the garage it started in.

[Pudders]

What also amazes me is that we got a PR increase on both our home page & some sub pages and also on our Blogger Blog site around the 8th Feb.... and then a drop in SERPS a week later... really?

[Jez123]

The current penalties seem to affect a lot of people that previously enjoyed favour (not just posters here: as I trawl through to my current page 46 position, I notice that at least half of the sites that used to be vying with me for the top space on page 1 are on e.g page 38 now).

Yes, it's the same in my SERP. In fact only one site is left from the lot I used to comoete against.

I would love to hear some theories why this has happened. Why are the sites now in favour that google always seemed to dislike? At least in my SERP the sites that are now top 10 were all kept out as google didn't like their tactics though now it does OR just page 2 transferred to page 1 as page 1 has dropped off the map. Anyway, I digest - but I would love to hear some theories.

[Rasputin]

I can also confirm that both us and several of our previous page 1 competitors (and between us the best sites for the subject, even if G don't like my sites they must like one of us) are all huddled toogether around page 60.

Some kind of consequence of favouring brand names, allowing brand name sites to have tens or hundreds of listings, and implementing very broad synonyms rather poorly taken together seems to be at the heart of the problem.

Someone came close perhaps in a recent comment in a different thread when they said that G now want to 'upgrade' good sites rather than 'downgrade' poor sites - so many of us with 'average' sites (ie not major brand names) are now pushed down.

[Wilburforce]

I would love to hear some theories

Something that was once either assigned a positive value or ignored is now assigned a high negative value.

What could it possibly be?

1. Content (including e.g. keyword density)
2. Backlinks
3. Internal Links
4. Outward links
5. Bounce rate
6. Content renewal rate (age of content)
7. Any suggestions?

It is certainly related to short-phrase key-terms. For example, on my site Key Term has gone from the top of page 1 to page 46, but Key Term Prices is at #1, and Specialised Key Term is at #3. This makes 1, 2, 3 and 5 better candidates than 4 and 6.

Key Term is one of my highest density on-site terms, and one of my most common backlink terms. I don't see that as unnatural or manipulative: it is what my business does. However, I have done several checks on the competition in the competition in the last year, and my site has higher key-term density than any currently in the top 20. Some in the current top 20 don't include the term at all. It looks to me like Google has developed an aversion to key terms, and is favouring sites with very low density.

The Haircut results show that this approach really doesn't work on one or two words, but I think they are determined to find alternatives to key terms on their own (but have no objection to them in context) in defining "relevance". Bing haven't followed suit (yet?).

FWIW, I think Google are still much better than Bing for longer phrases (Bing returns anything with any of the words in it, whereas Google's results look much more like an intelligent rendering of the entire phrase).

For my money, concentrate on longer phrases (i.e. market more specifically).

[tedster]

What could it possibly be?

A machine learning algo ranging over several factors. Something like Panda but with a different target and seed set - and many fewer factors involved.

I think "almost" everything Google does these days follows that pattern and it's very unlikely for most algo changes that we'll ever pin it down to some one definitive factor.

[Wilburforce]

A machine learning algo ranging over several factors.

Using Henry VIII as a role model?

[helenp]

It is certainly related to short-phrase key-terms. For example, on my site Key Term has gone from the top of page 1 to page 46, but Key Term Prices is at #1, and Specialised Key Term is at #3. This makes 1, 2, 3 and 5 better candidates than 4 and 6.

I agree with that,
I have same problem.
Highest density words in order
mycity
villas
apartments
Doing bad in mycity villas (in some countries only)
doing ok in mycity villas other words
but doing well in mycity apartments as apartments does not have that density.

And there is no keyword stuffing, those are the key words with most density naturally.

[Martin Ice Web]

I can only tell from my daily searches and the statements about panda/penguin before they have been released.
Nearly all of my searches does not give me -> the <- pages back where i do find the needed information. The returned pages have little information, near by information or are not relvant. But the serps do not give me back the pages with the "total picture" for my question, only pieces.
The anoying thing is, that this algo seems to think i am to stupid to ask a good query. It ignores one or more words or replaces it with a ( in most cases ) not relevant synonym. Even to put a + before a word or set it in quotaion marks will not bring the algo to search for it.

The obvious thing in serps is, that the returned sites are most brands or big websites or live in a near nightbar hood to them. The not relevant pages are the diversity that MC did mention but they are all from "trusted" big brands.

tedster - why should a machine learning algo degrade all the former realy good sites to page 50+ even if they are white hat and do have a very good UE, metric ...?
Google would have already pointed out that they know changed their system to machine learning algo!
And even a machine learning algo needs some human direction input?!

[helenp]

In my search tearms in spanish, it is a real desaster, searching for in spanish mytown rent apartments, the first 4 pages I checked 90% are searchpages, looks like searchqueries from big nacional or international sites. But the worst, there are at least 10 results from same company, just diferent kinds of search queries.
No need to do the searches anymore, google does it for you. If a result from a searchquery is content......
What they do is to do a link with a search query that says view all apartments in xxxx and then the result from that search comes up a a page, they had hard work doing that page.
Is that quality pages?
I can also do many pages like that.

[Rasputin]

It won't help to do pages like that yourself unless you are a 'big name' - more likely you will be down-rated in the next penguin update for producing poor quality pages.

Perhaps the more interesting question is, do people think G themselves are happy with the results as they are (multiple results from one domain, very high rankings for low quality content if it comes from well-known brands, 'curious' use of synonyms), and these results are here to stay, or is it an 'error' that is likely to be fixed (or at least 'turned down') in due course?

The answer of course affects the best approach we should be taking to our own sites, if 'high quality content' is no longer the main key to success.

[helenp]

It won't help to do pages like that yourself unless you are a 'big name' - more likely you will be down-rated in the next penguin update for producing poor quality pages.

I have clicked on contacto google on the bottom of page and it did a screenshot, but I know its for nothing, I have reported so many spamsites and google ever done anything about it.

[Wilburforce]

is it an 'error' that is likely to be fixed

I don't think it can be viewed in those terms. Similar effects will occur repeatedly, I think, until either the public goes away or they get it right.

I still think short search phrases are an issue. The mean number of words in my top 25 Google referrals yesterday was 4.9.

[tedster]

tedster - why should a machine learning algo degrade all the former realy good sites to page 50+ even if they are white hat and do have a very good UE, metric ...?

Because it is a "learn-ing" algo and not a "learn-ed" algo, I'd say. I will be very much different in another two years or so.

Google would have already pointed out that they know changed their system to machine learning algo!

If you read their comments in a certain way, Amit Singhal and Matt Cutts have been dropping big clues about this for well over a year. They certainly don't owe us a clear look at their algorithm. I'm just happy they say anything at all - there could be no communication whatsoever!

And even a machine learning algo needs some human direction input?!

Certainly. For one thing, humans decide what it needs to learn. And then humans create a seed set. And finally, humans measure the level of success or failure. A lot of that goes on even before a new module of the algorithm goes live.

--------------

All of this means we are no longer able to isolate specific factors for the algorithm changes very well. However, through this thread we can still tell WHEN an update has happened. It's just that we're better off looking for the target ideas for the see sets - what kind of things Google is trying to reward in the sites that rank better after that update. Of course, as long as the machine learning is in a primitive state, that can be very difficult.

[TheOptimizationIdiot]

If you read their comments in a certain way, Amit Singhal and Matt Cutts have been dropping big clues about this for well over a year.

All those two ever seem to do is drop clues.

I know I just joined but I'm a long time reader here so I feel like I've been a member for years and one day I found a link to some youtube.com videos people were talkin about, so I watched tons of em that day tryin to figure out what I need to do to get this site I've been workin on to rank.

I remember in one Matt Cutts said "Wouldn't it be neat if we could find a way to tell someone really put a bunch of work into making a useful website for their visitors and reward them for it." or somethin like that, I probably don't remember exactly, because I was thinkin "No. What would be really neat is if you just told people what to do to rank #1 so we could all do it instead of puttin clues for people to try and find and figure out how to follow all over the place."

It's like they've hidden all these clues in what they say and expect us to track em down and figure out what the clues mean and how to use em instead of just tellin us exactly what to do.

I guess you're right though.
Some clues we can find and figure out is way better than not givin us anything at all to use.

[Wilburforce]

A lot of that goes on even before a new module of the algorithm goes live

Not nearly enough, in my view. The results are shockingly bad immediately after whatever-it-is-they-keep-doing (and often persist in that state for weeks). Why we couldn't be given the option of "search using Google Beta" escapes me.

I still don't see, either, why previously favoured sites are consistently dropping hundreds of places. It is as if the human input is saying "put a high negative score on whatever we did before".

[SEchecker]

I duno how this month so far is for you guys, but here by us, conversations are down 50%! Overall traffic was at the beginning of the month terrible but started to increase around 12.02.

Till now a slightly plus in impressions and clicks from G but as already said conversations are terrible low.

I noticed that G is assigning new keywords to our main site every 2 weeks or so and this is going on since around 3 month. C

Conversations last moth were superb, but this month the NEW keywords G assigned to us performing well looking to impressions and clicks but conversations are 50% down...

Duno what to think about this! Instead keeping the well performing keywords and add new sets G replaces the well performing ones with new ones... it seems it some kind of testing and G is willing to risk as the losses and risk are by us!

[Martin Ice Web]

tedster,

you did not explain why former ver good sites have been pushed down, only the algo is learning and not a learned algo is IMO to less. Even a learning algo should see that most of the serps are not matching to the query - assuming the algo is able to make a quality check but that would implify that the algo knows what the user is searching for but that is based upon metrics, search behaviour...

I am with you when it comes to rate a document. This part seems the learning algo. But the serps itself are a fixed algo that is feeded by the learning algo. And this two do not correlate together.

It like Wilburforce says

I still don't see, either, why previously favoured sites are consistently dropping hundreds of places. It is as if the human input is saying "put a high negative score on whatever we did before".

[HuskyPup]

because I was thinkin "No. What would be really neat is if you just told people what to do to rank #1 so we could all do it instead of puttin clues for people to try and find and figure out how to follow all over the place."

Personally no, I don't want Google to tell anyone how to rank #1, the SERPs are a big enough mess now, just imagine if there were hundreds/thousands of sites in every niche ALL implementing precisely the same/similar strategies?

Then you would see even more complaining with people saying "I've done precisely what you said and so has my competitor however why is he ranking above me?"

I have no problem with the algorithm per se, I know what is required to construct authority pages and for them to normally rank well and Google's given enough information in the past on how to do this, what I have a problem with, and this is Google's major challenge, is that they built the world's biggest spamming engine and they let it get out of control and now they're trying to reign it back in by attempting to get rid of spammers/scrapers/duplicates/garbage meanwhile trying to keep up-to-date with all the new news and sites that are being launched every day.

Sure, there is bound to be innocent collateral damge, friendly fire, call it what you like, and I've had sites on the receiving end of it so don't think I've been immune to it, nonetheless it is going to take time. Panda and Penguin must be extremely complicated otherwise we would see them run more often and as to whether it will ever defeat the majority of scrapers I very much doubt it since some of them are so well-ensconced in the SERPs even Google itself seems to believe that they are the original!

[seo company chennai]

I too noticed that i got the same results for both Google.co.in and Google.com.

Is it normal if we get the same ranking.

[TheOptimizationIdiot]

Personally no, I don't want Google to tell anyone how to rank #1, the SERPs are a big enough mess now, just imagine if there were hundreds/thousands of sites in every niche ALL implementing precisely the same/similar strategies?

Then you would see even more complaining with people saying "I've done precisely what you said and so has my competitor however why is he ranking above me?"

So you mean they really can't just tell us and they kinda have to do it the way they do cause there's no way all of us could possibly rank #1 all the time? Fascinating

Makes me wonder if that's what Matt Cutts was talkin about in that video.

Like maybe if people dig through things and figure things out more it'll show in the way they optimize and the more little clues they've figured out how to work into their site the more work Google will know they did and then they'll be rewarded with better rankings?

[edited by: TheOptimizationIdiot at 11:37 am (utc) on Feb 18, 2013]

[HuskyPup]

There can only ever be one #1 in each niche and letting everyone know how to be would be open season for the blackhatters and spammers:-(

[claaarky]

There's something pretty f'ing major going on in my niche in the UK at the moment......masses and masses of new sites, mostly overseas (generally .au or .nz).

All sites that were ranking on page 1 prior to Friday for almost every term I monitor have been scattered to the four winds. Top 3 sites seem to be fairly unaffected but everyone else has been pushed down by anything from a few places to several pages.

Can't believe nobody else has mentioned it here yet. Looks like a bomb has gone off.

[theskunk]

well well well ... here we go again with the google circus. how tiresome this whole thing has become... google please tell me why my site is so far off the radar.. it's a good site - it serves very well its customers.. but now you would rather show some USA spam wordpress - amazon affiliate sites (multiple forrm the same operator).. google why do you want to do this.. ? I have worked very hard on my business - pay my taxes - I know you don't owe me anything - but I thought we could (after all we have been through) now be friends... I guess not... it's El'crappo.

[Martin Ice Web]

claaarky, it started at 13 a clock.
Upsite down, left to right, x as an y.....

germany, ecom

[petehall]

@claaarky that's classic signs of a Panda update. Of course they won't announce it for a while yet, but when it's crunching I often see foreign sites seep into our UK SERPs. They most likely will be gone again after the announcement.

I had a break from G this weekend as I've found it quite frustrating lately. Returned this morning to a nice surprise - some pages from sites I work with had simply vanished from SERPs and these are now back. I think a little higher than they were previously as well, which can't be bad.

Looks like things are getting sorted.. I for one am very suspicious about the vanishing data in WMT not affecting SERPs.

[aboshakeeb]

but now you would rather show some USA spam wordpress

Interesting , anyone figured out what's special about Wordpress ! , it eats my SERPs .

[aristotle]

Jez123 said
Yes, it's the same in my SERP. In fact only one site is left from the lot I used to comoete against.
I would love to hear some theories why this has happened.

My theory is that Google added a section to the algorithm that can identify pages that are "too good". In other words, their information is "too relevant" and their quality is "too high". Therefore, says Google, there must be something wrong with them, since nothing could really be that good, and so we'll drop them 600 places in the rankings. This is my "Too Good" theory.

[Rasputin]

I'm not quite sure if you are serious Aristotle - but that is entirely consistent with my own experience, and i've often thought the same. I'll try and explain our circumstances...

Although I have travel sites, for one [destination] we have a local specialist travel writer who is also a professor of history and literature for [destination], and with each [place travel] article he also sends me a [placename history and etymology] article, typically 3000-5000 words, tracking the history of the place back 1000's years to its origins, explaining its ancient etymology in great detail, usually with 20-30 references etc.

Really, very professional articles and without any doubt the best (English language) articles about their subject, on line or off line. Although I have no expectation to make income from them (they are of course very low search volumes) they should definitely be shown very high for the appropriate [placename history and etymology] type searches. But they usually aren't.

Hence I had sometimes wondered if they are hit by a "it's not possible to write this much about [subject] so these must be poor quality" filter - even wikipedia will only have one or two sentences on the subject, if anything at all.

And if this filter does exist, I wonder what else does it affect, and to what extent?

[theskunk]

So - do we know if there was an update 1st Feb 2013 - since i lost 66% of my traffic overnight (14 year old UK ecom) - and its not bounced back. No messages in WMT... blah blah blah. Any clarification is very welcome.