Google Updates and SERP Changes - March 2013

[gyppo]

< continued from [webmasterworld.com...]

I'm seeing daily switching between 2-3 datasets across around 600 keywords. Seeing roughly 10-20 keywords swapping in & out of positions.

[edited by: tedster at 2:47 pm (utc) on Mar 1, 2013]

[Panthro]

Interesting thought AlyssaS!


(and this is about a page late, but..)

I, for one, enjoy reading the back-and-forth even when it does get slightly off-topic. Somewhere in the midst of it all are always great arguments and information worth thinking about.

A thread full of "I see this. I see that" reporting would not be as helpful imo. So, thank you TOI and Wilburforce for making us think a little bit more about statistics while webmasterin.

[AlyssaS]

I'm not seeing anything unusual. Are you sure the Analytics code wasn't updated/edited on that date?

That's a possibility - that they changed the reporting (perhaps the length of the cookie for how they define a return visit), so it's not a real loss of traffic, just a reporting issue.

Then again if it was a reporting issue every site would show this, instead of just on the one site that appears to have lost traffic (the other sites seem to be fine).

[Bones]

that they changed the reporting

I meant the Analytics code snippet that appears on the website in question, which would perhaps explain why you're not seeing the same issue with other sites.

[TheOptimizationIdiot]

@TOI, do you remember the site i mentioned yesterday ( with the php-login and nothing on it )? Guess what, it climbed 2 positions from 10 to 8! But a page from me that was on page one got a -30 position hit!

I'm seeing the -10% indexed pages data set today, but the SERPs I've checked haven't budged.

@ - TOI & MIW - definitely seeing some very odd movement and even "Day Parting" (see, I'm not using the Zombie word anymore) or throttling periods where G must be diverting traffic through more paid channels. I believe if they had their way, they'd eliminate organic results overnite for a 100% paid system. The Wild West days are over my friends!

I'm seeing 24 hour switching on indexed pages (day 5 now and on the - dataset. I'm not worried though, they'll go back up tomorrow am just like they did yesterday lol).

Oddly quiet in the SERPs though. I'm seeing almost 0 movement in the ones I've checked and still not a single new page indexed (start of day 6 since they were added).

A thread full of "I see this. I see that" reporting would not be as helpful imo. So, thank you TOI and Wilburforce for making us think a little bit more about statistics while webmasterin.

Sure thing and I had fun with it too and it definitely made me think more than once.

Had fun with it except the little part where I was thinking since the decks were individual and each person was drawing from a unique deck the calculation was * 6 rather than ^ 6 like it is for 6 people picking the winning lottery numbers for the same drawing. Gotta remind myself to double check the maths before replying next time! lol

One of the biggest reasons I read here is it always challenges me to think about the concepts more since I can't just go looking at the sites people are talking about and I definitely believe it's a good thing that's helped me quite a bit over the years I've been a reader here.

I've seen scraper pages removed, yet Google still trusts them enough top place even more of their ill gotten content on page 1. What happened to the "Trust" metric?

The trust metric is still there. They obviously trust if they show the scraper instead of the original their visitors will likely never know or care and even if someone does notice they trust there's nothing anyone can do about it, because if I remember right when things like that happen "The algorithm did it!" not them.

I think we just misunderstand what "trust" and "quality" mean and how they are applied sometimes lol

[Lorel]

Here is another idea to consider. I checked my site in MajesticSeo yesterday and noticed that my site has lost a lot of back links in last few months and gained very few (because I've been too busy trying to FIX whatever MIGHT be wrong with my website to pursue that avenue or even participate in social stuff, except on WebmasterWorld).

With all the Panda/Penguin penalties the 2 years (and sites disappearing or their ranking dropping), it seems to me that could be the reason for the downgrade in traffic and ranking, i.e., the whole internet has been downgraded except for spammers floating to the top intermittently. Maybe we are looking at the new Normal.

[backdraft7]

I think we just misunderstand what "trust" and "quality" mean and how they are applied sometimes

It's not just us, I don't think Google understand it's own guidelines.

BTW - just another observation - I'm now doing better (conversion wise)from 12:00am to 6:00 am than I do all day. Looks like I have Zombies in the day time, Vampires at night. ;^D

Daytime traffic is almost non existent ATM. Where's it going when my serps still look the same?

[ScubaAddict]

No changes in the SERPs, but I have noticed quite a drop (15-25%) in traffic the last couple of days... I also wonder, where does the traffic go?

[frankleeceo]

How many people see that their SERP stays consistent while traffic falls off?

Is it personalization at play? It seems like when I observe SERP position with WMT, my KW positions tend to rise while total impression stays the same. You would think that rising SERP means increased impression and clicks. But no.

That is only possible through active and intentional algorithmatic SERP index shuffling isn't it?

[TheOptimizationIdiot]

It seems like when I observe SERP position with WMT, my KW positions tend to rise while total impression stays the same.

Some of that is possible depending on how much the rise is. If you go from 10 to 5 your impressions would stay the same, because it's all on the same page, but if you're going from 12 to 4 it would seem like the impressions should increase quite a bit unless something like the "page one full of results from one site bs" is driving a much higher % of people to page two than used to visit it and then they could remain relatively static even for a bigger jump in positions.

That is only possible through active and intentional algorithmatic SERP index shuffling isn't it?

It seems that way but I can't decide if it's personalization or just "split testing" and the average position of when the result set you're in stays relatively constant but that result set is only shown 1/2 or 1/3 the time, which would make it look like you're in the same position all day every day when you're really only showing to people in the position you're seeing listed a fraction of the time you would if there was only one result set. (I think I'm saying what I'm thinking right, but please try to go with the point if it's not coming out in a way that's quite making sense.)

[TheOptimizationIdiot]

Just saw one move from not on the first three pages to position four, but not until after I switched browsers.

So when I searched for it 20 minutes ago in browser 1 the result wasn't on the first 3 pages. When I searched again in browser 2 it showed on page 1 position 4. Then I searched with browser 1 again and it was page 1 position 4 there too.

Uh that seems a bit odd to me.

Update: Also seeing a difference in results when searching for an area depending on whether I use the word in or not. So if I search for "widget area" I'm getting different results than when I search for "widget in area". I have noticed this before but didn't think to mention it until now.

[mirrornl]

hoping this does not stick

[backdraft7]

@TOI - Google is trying to read your mind, but keeps guessing instead. ;^)
Welcome to the new millisecond based Everflux.

I've noticed the semantic problems for quite some time...probably since 2010. Plurals also have issues.
I used to be able to search on keyword 1+keyword 2=result 1 and keyword2+keyword1 = result 1, but now the second case gives me result 2, perhaps in the name of "diversity".

G has price tags on all the known key phrases in the universe. I've noticed my organic position for the high priced phrases dwindled when Larry stepped in. But, they were kind enough to leave me the low hanging fruit...for now.

[TheOptimizationIdiot]

@TOI - Google is trying to read your mind, but keeps guessing instead. ;^)
Welcome to the new millisecond based Everflux.

LOL!

I'm guessing the algo doesn't get "half crazy and 50% nuts yet", but maybe someday.

And I really wish it could figure out I want 10 results on each page with a diversity of sites instead of 8 to 10 results with 5 to 8 from one site, especially when it's 5 to 8 from Amazon. If I wanted that many results from Amazon it's pretty easy to go to one page and click their links. I don't need all their pages in the SERPs for any reason.

If I was going to be really nit-picky I really don't want any bleepin youtube videos or pictures when I search for a business, I'd just like to see a group of websites as choices I can decide from.

[TheOptimizationIdiot]

How many people who are seeing huge fluctuations have made significant changes to their pages lately?

I'm wondering if the "pop" I saw earlier from not in the top 30 to number 4 was the "rank modification for spam detection" expiring, because there were some fairly significant changes to the page not too long ago and it's the only significant change I'm seeing in the results I've been watching.

[Saffron]

I really can't understand what they are playing at. My best performing article which has always ranked no. 1 or 2 is now down to 10. It is an extensively written article of over 1,000 words. A huge amount of work went into it. It is now beaten by a publishing company with the same name and a furniture company.

Maybe I am a little bias, but how is that improving user experience by actually providing irrelevant links over the genuine topic?

All other search engines are showing the article as no. 1. Google have a screw loose. I've switched to Bing. They aren't putting user experience first, that's for sure.

[Wilburforce]

@Saffron

How many words are in the search phrase you are looking at?

[Saffron]

Just one. It's actually a one word topic. Guess some people mis-spell it and use two words. It's gone from 23,000 hits a month to 12,000 and is continuing to drop.

I am incredibly disillusioned with Google. Not just the fact they've done this to me, but because it has now demonstrated that their algorithm is screwey and they can and do prioritise irrelevant pages now. Other search engines are still ranking the article in first place. But Google have decided a furniture company is more relevant.

[Saffron]

I just looked at the top 100 keywords, and the article in question is only searched for with one word.

I should also add the article has had 1,200 Facbook likes. I don't understand why Google have decided it's trash.

[netmeg]

Umm, if it's still on the first page, Google does not consider it trash.

It's incredibly difficult for a non big-brand to show up on the home page for a one word keyword in a competitive niche nowadays.

[Saffron]

It's always swapped no. 1 spot with Wiki. Sometimes mine, sometimes theirs. Never moved past that. Just comparing the past 12 months to the 12 months prior. The rot really started with the last panda release. 800 visits on 20th Jan, 250 yesterday. Far out. Like I said, other search engines still ranking it no. 1.

I don't see how a furniture company is more relevant to such a detailed article. They've lost the plot, but it's come at my expense.

I get really cranky at people who say that you only get hit by Panda if you write crappy content.

[TheOptimizationIdiot]

You didn't get hit by Panda if you're in the top 10.
How many total results are there for the query?

[Saffron]

If it's not Panda what else would cause a site wide drop?

From Feb 2012 - March 2012, I had 7,280 queries, Feb 2013 - March 2013 I had 1,195. Down 83%. Looking at the other searches, they're all down around the 50% mark.

Actually, looking at the past two years, dip end April 2012, dip end Sept 2012, big dip 21/22 Feb 2013.

[TheOptimizationIdiot]

Any of the 500+ updates a year could cause the type of drop you're seeing without it being Panda.

If you have a page in the top 10 for a competitive one word query, the likelihood of it being Panda is slim in my experience. The sites I've seen get hit by Panda tank.

The difference in "overall score" given to your page in the top 10 dropping from 1 to 10 relative to the other pages in the index (they stop indexing at 1000 pages) is 0.09%.

That's hardly a huge glaring out-of-control drop by any stretch of the imagination.

The rest of your traffic could easily be the same type of drop, like going from an average position 4 to an average position of 11 or even an average of 8 with more results in position 11 and 12 that used to be 9 or 10. Even a tiny percentage of "score" awarded to your pages relative to the rest of the pages in the index can cause the type of drop your seeing.

There's probably some "fine adjusting" somewhere that could get you back where you were, in my opinion.

It's even possible the drop is due to some link network you're unrelated to being totally discounted and the "ripple effect" of PageRank passed to your site from other sites has decreased from it so you don't quite have the ranking power via links you once did. (That's the type of "small" I'm talking about I think it could easily be.)

The main point I'm trying to make is you're not seeing a huge drop going from 1 to 10 in any way.

There's something "relative small" that changed somewhere and to get back where you were in my opinion you're going to have to find a way to "make some small score improvement" on your site, but I wouldn't say you've been hit by Panda, because in my experience sites that get hit by Panda Drop Like a Rock to something like page 5 or 50 or out of the index, but not from position 1 to position 10 or position 4 to position 11 or anything small like that.

[Leosghost]

Saffron..
I just had a KW1KW2 less than 10 letters overall EMD single page site that had been at #1 for years, bounce around page 1 for the last week or so..it only has an inbound from me ..and some crappy whois inbounds..
it has about 200 words only..very concise very targeted..KW1 and KW2 are in the domain , the title, and a couple of times in the content..it is a dot net..

It settled at #3..of around 2 billion..( you can pay games with adding a hyphen or not, in between the KWs to change that figure )..

It is "tech" sector..money term..

#1 An article from wired..with a gazzillion on subject backlinks..like netmeg said..no-way am I going to out perform a brand like wired with all those backlinks for a term in "tech"..

#2 is another KW1KW2 EMD identical to mine.. except theirs is a non English TLD..they have less words than I do on that page..their title is only KW1KW2..( these are English words )..and they use KW1KW2 a couple of times in their otherwise totally non English content..
But they are targeting the same words as the KW1 and KW2 in that "combo" are what almost anyone , no matter what their first language was would type in..they'd put "them in English"..

( I missed the dotcom when it dropped a couple of years back..now it is parked with premium adsense feed running on it )

The new #2 came in from nowhere, it is an aberration, but foreign language EMDs with the KWs in English apparently seem even more concise to G..

Emds with matching titles and only the KW on page are quite easy to reach #1 with..even in competitive areas..makes sense..otherwise "Ford" wouldn"t rank for "Ford" etc....this effect can be "played" ;)

Wired at #1 isn't strange at all to me..

And I don't think that the foreign EMD will last at #2 longer than 6 months..( most of the traffic responding to the English EMD and KWs won't be able to read the rest of the page..so the user engagement will look bad )..and as English is not their first language they won't want to alienate their home market to chase people that they'll have trouble communicating with, plus, if they switched to all English text..I'd outrank them on "concisivity[sic?] anyway ..

They also may not even realise how and why, and even if, they are ranking in English language serps on English language Google sites..

For now I'm busy with other stuff so it doesn't matter to me ..I'm still getting traffic..wired is outranking me ..but they don't provide what they are talking about..so I'll get the English speaking traffic..:)

I'm guessing that the removals site has less text around the kW and so appears to be more focused upon it to G..

That you can deal with..

When writing content..less is more..

Look at their word count..look at yours..try editing yours to approach or be less than theirs..be concise..but do it in baby steps..you should find that there will come a point where it will "tip" and G will decide you are more focused and concise on the KW..and re-rank you both..

[edited by: Leosghost at 3:24 am (utc) on Mar 9, 2013]

[Saffron]

Okay, thanks. My friend said when her site was hit it was a 45% drop overnight. Mine hasn't tanked like that, but it's a 45% drop in 12 months.

I still rank one with some of my search terms but the bulk of the articles have dropped. I was in the top 3 for almost all my articles, now I'm more like in the top 10. But some crappy sites are outranking me.

What I've done so far is re-write any duplicate content (only about 10 articles), set up the authorship tag, made the Facebook page more active, removed the ad from the beginning of the content, added lots of new content, added new images where possible, and re-added social sharing buttons (they vanished a while ago & I didn't bother to put them back on).

But none of those may have an effect. My main article slips further down the page very day. It had the no. 1 spot for years. It's frustrating when you don't know what the cause is.

[Saffron]

I thought more was better with content? I aim for a minimum of 500 words because Google are saying the articles should be extensive.

With the main one, it's 1,100 words long, and very thorough. I can honestly say that I am proud of it. I have never written for Google or done keyword stuffing. The aim was always to actually provide the reader with what they wanted to know.

I live in hope that it will pick back up, it's a good site. But it doesn't solve the problem of our main source of income halving and it won't pay the mortgage unfortunately.

Part of me just wonders why I'm spending 16 hours a day trying to fix something that's still dropping. I could be hanging out with my kids before the poo hits the fan with us in a few weeks.

[TheOptimizationIdiot]

As far as content goes, it depends and changes from time to time, but to give you an idea of how small the difference between top 3 and top 10 like you're seeing is, it could be a word (yes one word), a phrase of two words, a couple of plurals instead of singulars or singulars instead of plurals.

It could be the order of the words in a paragraph or two. The difference in the positions you're talking about is small, tiny, almost unnoticeable.

I remember talking to someone here (or reading it here) who added one word to a page and it went from position 43 to the top 10 results. That's how small the differences in rankings are even a bit lower than the differences you're talking about seeing, so you're not off by much anywhere and it's definitely not Panda if what you're saying is the difference you're seeing in positions.

My advice is to look at the little things and see how you can improve those to be a bit more targeted for your terms (removing some words as Leosghost suggested might be exactly what you need to be more focused and concise about your topics).

Also don't expect anything to happen overnight or get frustrated when it doesn't. Things don't usually change that fast (in the up direction anyway) too often these days, so part of it's a change and wait thing.

[System]

The following 2 messages were cut out to new thread by robert_charlton. New thread at: [webmasterworld.com...]
.

[edited by: Robert_Charlton at 4:35 am (utc) on Mar 9, 2013]

[micklearn]

This is off topic, somewhat, yet, it was one of the most important comments that I've taken into consideration as Google moves forward.

[news.ycombinator.com...]

Interesting comment by ChuckMcM, former Google employee who is now the VP of Operations at Blekko Inc. (this is one of his comments on the page above):

"You may (or may not) be surprised but when I was there a lot of energy was spent by the rank and file on not being evil.

The challenge is of course in 'winning' and decades of research on this shows that people who really want to win will be a bit evil if they can get away with it and win at the same time.

So what Google has is a competitive culture (for a long time you couldn't get a raise without getting a promotion) which is peer balanced. They have a huge toolbox of technologies that can generate cash when tweaked. And they have certain groups whose success criteria is 'can you make more cash this quarter than last quarter.'

Unless actively worked against, like mold in a bathroom, evil sort of emerges out of that environment. Kind of scary to watch actually."

More info about him:

source: [blekko.com...]

Chuck McManis, VP Operations - Chuck McManis is a respected technical leader who brings two decades of experience to the blekko team. His experience includes four years at Google where he worked on a variety of initiatives in the operations group, five years at NetApp developing storage architectures that scaled into clusters, six years on the team that built FreeGate from startup to successful acquisition, and ten years at Sun Microsystems where he was part of the original Java group, the ONC/RPC architect, and key contributor to many of Sun's networking innovations. Chuck has an Electrical Engineering degree from USC with a minor in computer science.

[epmaniac]

I noticed a common element: all the sites that have suffered are wordpress based.