Many results from one site - Host Crowding vs Brand Authority

[Brett_Tabke]

Over the last year, Google has switched from the "host crowding" method of displaying multiple results to the same site, to a system of displaying all those results on SERPs.

eg: Host Crowding was the old method of:

Site Result
--- indented 2nd site result

The new system displays all the results inline. In many cases, up to ten results can be for the same site.

The result has been (for many of us), a frustrating - if not - fatal loss of Google search relevance. The only way many of us have gotten around it, is to become very good friends with the minus operator "-" (which has also quite working at times) to remove the site from the results.

Google has not talked about this change much at all. Despite numerous people talking about the loss of relevance with the new system, Google has continued to turn up this dial.

For me, the new system has all but been a nail in the coffin of my Google usage. I can't recall a SINGLE time other than a navigation query, that I have EVER clicked on a result from a site showing 4 or more results from the same site. I had been using Google for about 20-40% of my queries (always starting at Bing) - which was actually a pretty good track record for Google if you consider that means Bing was failing for me 20-40% of the time.

The only real new thing here is that for the first time, Matt Cutts talks about the issue in a video, but never does in fact answer the question entirely. I have yet to hear any major voice say this is a plus.

[youtube.com...]

Matt gives quite a few comments about whey host crowding is a plus. In fact - listen close - seems almost as if he prefers the old method like all of us do...

[santapaws]

the funny thing is, search for a product and you get 70% amazon extreme domain crowding, search just for amazon and you get a variety of sites.

[petehall]

I'm seeing signs of this being corrected on a couple of badly affected searches (UK).

[johnhh]

In our niche getting worse today ( in UK )

[santapaws]

yes getting worse for me too.

[diberry]

Re: the mac and cheese search... does Google Recipes View just pull from the normal SERPs somehow, or is it running on a different algo altogether?

[zehrila]

After going through this thread, i observed a weird sort of instance where exact same url is listed twice for one search query. I took a screenshot, checkout first and last listing in the image.

[s16.postimage.org...]

Weird much?

[MikeNoLastName]

Came in to see why our traffic was up 35% across 4 domains today. Especially since G has been moving our pages lower and lower on the SERPS every week. Saw this thread. Now I know why. Everyone has switched to searching on Bing and Y where our pages are still listed on page one!
Searched one fairly competitive 3-word term and 16 results on the first 3 pages were all from the same site. At least it finally seems to have cut down on Wiki results a lot. Matt, if people didn't find the answer in the first 2-3 "BEST" results from a site, chances are they aren't going to find it in the next dozen!
One large part of my internet use requires finding answers to obscure (non-shopping) questions quickly, often and preferably just from the snippet. G was previously great at this. Over the last few weeks I've literally HAD to switch to Bing to get the work done. The latest results are ridiculously bad, and I'm saying that from a USER point of view not an SEO view. The results they are showing tend to have cruddy, useless descriptions and I've gotten to the point of almost automatically skipping to page 3 or 4 to get useful results. Guess it's good for their ad serving. Remember what G does with sites with no useful information above the fold... Something about not a good user experience.

[Zivush]

Now I get it.
Been working lately on a website design small project and made ~ hundreds of searches for many CSS, HTML and PHP terms and queries during these 3 last days.
I confirm everything that was discussed here.
Instead of receiving diverse results (& solutions) from many sites and forums, I’ve seen the same sites all over again.
When it comes to a heavy Google user, it is frustrating.

[arikgub]

Everyone has switched to searching on Bing and Y

I wish ... doubt that ..

Been working lately on a website design small project and made ~ hundreds of searches for many CSS, HTML and PHP terms and queries during these 3 last days.

The problem is that Bing is even worse for these terms. I do a lot of web dev related searches, and what i have been doing lately is going directly to StackOverflow.com and searching there - I always find an answer.

In some area Google has become an unnecessary link in the chain, waste of time

Books - directly on Amazon
Hotels - TripAdvisor
and coding questions - StackOverflow

...

[bloard]

This host crowding thing has gotten so annoying over the past week or so. Today I was searching for an electronic component using a somewhat vague search query. I know there are hundreds of retailers selling items like my query described. After the top few amazon results there were literally 20 alibaba results in a row from different importers.

I don't want to import 1,000 of the thing. I just want to see some options on where I can buy it. This happened multiple times as I tried to refine my search with the same results. I went to Bing and the results were way more reasonable.

This can't be something that they intentionally want to be this way is it? It truly makes search annoying using google. It's happening over and over across most of my searches across the spectrum.

So I checked on a random keyword I'm targeting. My site would be 7th if every domain had only one listing. Instead it's 28th. So on average every domain for that search term had 4 listings. Crazy.

I search with 100 results returned because I know it's going to take a fair amount of scrolling to find what I'm waning anymore. It seems the 100 setting magnifies the problem even more.

[Robert Charlton]

Re: the mac and cheese search... does Google Recipes View just pull from the normal SERPs somehow, or is it running on a different algo altogether?

As soon as you either click the Recipe "tab" or click one of the recipe ingredients which highlights "Recipe" in red, the serps change considerably... my assumption is only returning recipe pages identified with rich snippets, but, for whatever reason, no longer "brand-crowded". Each further ingredient selected whittles the number of results way down, depending on number of results satisfying the selections.

What's curious to me is that I don't remember the check-boxes appearing initially... I noticed them about 10 hours after first checking the search... but I'm not absolutely sure they weren't there initially. Does anyone else have a recollection, or have familiarity enough with "recipe" searches to know if the ingredient check-boxes had appeared previously, without the "Recipe" tab being selected?

It's almost as though the results aren't granular enough to be useful without those check boxes... as if the "brand-crowded" results need another level of search refinement to make them really useful.

I'm not seeing this in any other verticals, though... unless you think of some of the image result refinements as doing something similar.

[MustardDan]

I always run a full SERP check every Friday, results seem to be the same, a few ups and downs but still multiple results from same domains in searches. Nothing seems to have changed. One search query I found that one domain had 17 placements in the top 32 results.

[santapaws]

just searched for a product:
page 1: 2 amazon, 3 facebook (for a product!), 4 shopzila
page 2: 6 amazon, 1 shopzilla, 1 facebook
page 3: 7 amazon, 1 shopzilla, 1 facebook

lets face it, this is no mistake. They are doing this for a reason.

[Martin Ice Web]

@satapaws,

no that is just leveing the results, so you can better compete against big companies ( like amazon ) with quality content.

Oh no,no, that was just Matt Cutts in my mind...

[santapaws]

just wondering what part of the algo deals with on-page quality?
and i mean the actual page, not signals pointing at it. I know theres plenty of code dealing with demotion for on-page factors but quite honestly, if any part of the algo is coded to promote pages based on there on-page content then it seems to be switched off.

[lfgoal]

The ghost of Altavista calls to thee

[DigiSEO]

just searched for a product:
page 1: 2 amazon, 3 facebook (for a product!), 4 shopzila
page 2: 6 amazon, 1 shopzilla, 1 facebook
page 3: 7 amazon, 1 shopzilla, 1 facebook

lets face it, this is no mistake. They are doing this for a reason.

I see Facebook pages (for products) showing up everywhere for my industry. It seems people are baiting Google serps with Facebook pages on a wider scale than ever before. I can't imagine those Facebook pages convert very well though.

[peego]

Is anyone seeing a buttload of youtube results in top 30 as well? I did a few searches and they all had 10 to 15 spots out of the top 30... I had one result where page 2 was 10 youtube videos lol. /sigh

[santapaws]

yes lots of youtube along with lots news results.

[anteck]

Seems to me google are Doing everything in their power to filter out anything but what they consider a sure fire result.

Panda filtered out affiliate sites.

Penguin filtered out sites that had artificially high link profiles

Host crowding shows the same 'high profile' sites over, and over, and over.

Sad state of affairs, nothing gets better with this company. Gone are the days when a independent expert or leader in the field could make it big with a nice helpful site. Now it's all about brand brand brand and even that is t guaranted.

Google used to be about a wide variety of information, mom and pop stores, and finding what u needed in the first and second pages.

Now it's a capitalistic, profiteering, self promoting, locked down search engine.

They can do that. People will decide if they like it, or just do things differently. I think the days of ranking in the top 10 don't mean half as much as they used to. People Are shopping smarter. People are using google as secondary thought. google knows this, and thus is quite willing to clean up the results to the point of homogenization.

[diberry]

On what used to be my most popular keyphrase until Penguin demoted me to oblivion, I'm seeing an interesting switch today: several non-big-brand sites, including mine, got back up to page 1 after being on page 6 or lower. The repetition of big domains is more reasonable - three companies have 2-4 results in the top two pages, but they were having 4 on just page 1 a few days ago.

This looked exciting, until it hit me this looked a lot like a total Penguin roll-back which, let's face it, is never going to happen. And somehow, despite being back on P1, I got no searches for this term at all today. That didn't sound right.

So I searched my term through Google via some proxies, and it was beyond bizarre.

One US proxy brought up nothing but UK sites - loads of BBC, Guardian, etc. The brand repetition was mind-numbing.

The next proxy, again US, was even worse. Host crowding was so bad that a couple of pages had only three domains apiece on them, with lots of indented results below each. This clump was mainly US sites, but the host-crowding was just unbelievable.

I'm beginning to wonder if they're serving up lots of versions of the SERPs right now, maybe for testing purposes. Or maybe just because something's gone horribly wrong.

[outland88]

Host crowding shows the same 'high profile' sites over, and over, and over.

And over.

If you think about it there must be something related to restraint of trade by giving undue advantage to certain sites in the results. If it were one, two, or three mentions in the top fifty it would be one thing but ten-thirty results from the same domain gives unfair advantage to the competition. I wouldn’t be surprised if Google turned around and said those were paid results but we chose not to inform people till a later time.

[Rasputin]

A related (and perhaps even worse because it's more extreme) issue relates to the same 'brand' site but with lots of different tlds occupying all the results.

I am looking at a travel search (so you can probably guess which brand site I am talking about) for our local town that has all 10 first page results from the same group of sites - 4 from .com, 3 from .co.uk, 2 from .in and 1 from .ca

Many of the pages are pretty much identical except the tld - i.e. they are not offering different regional information for each tld

The same 'brand group' has 8/10 results on page 2 and 10/10 on page 3 and 6/10 on page 4. I didn't bother looking any further...

[santapaws]

yes but that site was getting results for every tld even before host crowding.

[johnhh]

@santapaws that is true - but now they are what I call "clustered"

[santapaws]

or what i call "more annoying"

[Robert Charlton]

Some quick observations... on a new thread that I spliced into the June Updates and SERP Changes thread [webmasterworld.com...] ...webindia123 posted that he was just noting the brand authority effect in India. There's also discussion on the Updates thread about lots of movement in the US.

I thought, therefore, that I should check out brinked's [mac and cheese recipes] search again, as well as some other searches I've been monitoring... and (from the SF Bay Area), I'm noticing a fair amount of shifting tonight.

brinked originally noted...

1 site has the top 3 results for this search, and then that same site has 5 consecutive results on the 2nd page and 7 consecutive results on the 3rd page.

This is what I've seen tonight for the same search over a fairly short period...

a) top 3 on first page, then 4 on second page, 0 on third page, and 2 on fourth page
b) top 3 on first page, then 3 on second page, 6 on third page, and 1 on fourth page

So these appear to be shifting, with overall count going down, at least for this search, following just the particular domain we're paying attention to. On other searches, I've seen the count for one domain go up.

I think this may be the equivalent... for evaluating pages on domains that have received large user preference in other tests... of using, say, multiple sitelinks to test user preferences (which is at least one of the things I believe Google has been doing in other tests). Instead of sitelinks, though, Google has got so many pages to choose among on various mega-sites that they perhaps need to try multiple listings rather than multiple sitelinks. I can't imagine what else they've got in mind. I feel that Google does believe in following the numbers, and it may be that the numbers their engineers see are challenging what we all feel is useful and fair.

I should note that it may also be screwing up usefulness in big ways. I tried a [repair brandname widget] search today, and useful results were almost impossible to find, not only on Google, but also on Bing. Ultimately I needed to drop the brandname and just use [repair widget], without the brand. All that said, I'm convinced that this is another of those queries where there actually weren't very many good pages satisfying it. Cleanest results initially appeared to come from Blekko ;) ...by quite a bit... but Google had the largest index and ultimately that's where I found the marginally most helpful content. It took a lot of digging. It's not a search I'd tried before, so I can't compare.

[Robert Charlton]

Rasputin - Re the TA results you mention...

A related (and perhaps even worse because it's more extreme) issue relates to the same 'brand' site but with lots of different tlds occupying all the results.

We did discuss TA, with specifics (let's not do it again), back in October 2011, in this thread...

How Google is Showing the Results
http://www.webmasterworld.com/google/4381552.htm [webmasterworld.com]

On the 'kolkata' search we discussed where I'd seen 14 results on the first 3 pages, I'm now seeing 20. Before, there was host crowding, so I only saw two pages per cctld. It appears now that there is some brand crowding (ie, more than two results on some cctlds), and also, it strikes me, perhaps more cctlds... but life is too short and I haven't counted them all.

I was surprised, btw, re TA, that in a Q&A session at one of the recent conferences, I believe with both Matt Cutts and Duane Forrester, a question was asked that IMO was clearly complaining about TA and multiple cctlds. Matt was apparently genuinely unaware of the situation... and in the context of that session there was no time to clarify. Now, in the large scheme of things, TA type results may not even be unusual. Now, in the large scheme of things, TA type results may not even be unusual. I sure hope this is a test.

[atlrus]

I wonder how would this relate to SEO in terms of "link juice".

If you previously had 2 pages from Website A ranking for "blue widgets" and now you have 16 pages from Website A ranking for "blue widgets", would those 16 pages pass the link juice equal to the 2 pages before or did you just get 8x the link juice available for "blue widgets" from Website A?

Also, is anyone seeing more than one (or two) websites having those multiple page returns for a given keyword? Because what I am seeing is just one website per keyword returning multiple pages, at least for the first 10 Google pages.

[diberry]

Instead of sitelinks, though, Google has got so many pages to choose among on various mega-sites that they perhaps need to try multiple listings rather than multiple sitelinks. I can't imagine what else they've got in mind. I feel that Google does believe in following the numbers, and it may be that the numbers their engineers see are challenging what we all feel is useful and fair.

That's the best theory I've heard so far as to why Google could be doing this intentionally and think it's a good idea.

But since this started, I've had a surprising number of non-techies complaining to me that Google is just useless now and they're preferring Bing. They actually seem only to want ONE result from each site, because once Google points them to BigGiantBrandSite.com, they're savvy enough to know it might be worthwhile to try their search term in the site's engine. If that doesn't work and they hit "back" to Google and see nothing but the exact stuff they just found by searching the site, that's pretty annoying. I also get the sense they feel like it's insulting their intelligence, sort of "Duh, I KNOW Amazon has a hundred of these."