Should We Abandon Category Navigation?

[100up]

Our site was Panderized and we're currently embroiled in the thin content cull.

One of the realizations we've had is that our category navigation was buggered up to the point where we had so many superfluous pages because of pagination and sorting issues that we're talking about making some pretty radical changes.

Our site is essentially product based, and navigation is possible via search, alphabetical merchant listings and category navigation. So to free up some crawl space, we're fixing the category pagination problem. But looking at our analytics, it's clear the vast majority of our users find what they need using search or the alpha listings.

I just want to work towards cutting the category navigation altogether, but that makes people nervous. From an SEO standpoint, how much can the 100's of category pages be doing to hold the whole thing together?

[tedster]

Big question - and ultimately something only you can analyze. Have you checked your analytics to see how often your visitors use those category pages? Before I bow down to any theoretical improvement for the sake of Google, I'd want to know that information.

You might consider using noindex for the category pages, to see if that helps. Then any link equity still circulates through them. Although my current impression is that unless shallow content pages are removed, only a little improvement is possible.

But very few "pandalized" websites are reporting anything near a full recovery of their lost traffic, and there's not enough hard data and testing available so far to say anything definite.

[walkman]

A week ago I removed them from my product pages totally. Left a link to a page where they are listed. But I left the a b c ...z navigation though. I did it just in case, so I'm designing for Panda

[helpnow]

@tedster What is the difference between removing "shallow content pages", and noindexing them?

[tedster]

Removing means the request for that URL returns a 410 or 404 status. Noindex is just an instruction to search engines, but the URL still returns a 200 status.

[helpnow]

@tedster Right, but what is the difference vis-a-vis google and shallow content? If a category page is shallow, but provides an important step in navigation for users, it needs to stay, but you can noindex it to make sure google doesn't add it to the SERPs. Are you saying that a noindex is not as good a 404/410? I was under the impression that a 404/410 = a noindex in terms of getting it out of the SERPs and not falling into disfavor with google for causing shallow pages to get into the SERPs?

[tedster]

No one knows for sure yet - because no one has tested how things work and actually recovered their lost traffic.

From reading the answers given by John Mueller, it seems as though noindex should be used for content you intend to improve. If you don't intend to improve shallow content, it sounds like he advises that you remove it.

My guess is that Google still applies some kind of site-wide demotion if they know your site still has shallow content available. In other words, they prefer to send traffic when they feel the entire SITE is a good destination. Just my current impression. As I said, we don't have a lot to work with.

The one site I know about that did recover traffic did NOT use noindex, they did drastic surgery and removed or re-wrote the pages that were hit the hardest.

[bramley]

Info site. Many search results lead to detailed multiple-pages. I no-indexed the pages beyond page 1, (btwn 2-10 more) and improvement in SERPs was of the order of position 5 to 4. As Tedster said, gives a marginal kind of improvement; not as I expected.

But the gripe here, real, is that single page 'thin content' can serp better than multi-page detailed and higher quality multi-page results (the extra pages (generally a marker of more in-depth, quality) can count against you.

Where multi-page sections can be usefully divided by text / subheadings i do that. But in some cases the following pages are just images, and large ones. That's a big attraction and uniqueness to my site, so I don't plan to change this just for Google.

Yet in both cases, including long text over multiple pages, with correct headings etc, can be beaten by single-page thinner sites.

There is an upside, however, in the case of text on multiple pages when it comes to long-tail searches.

IF the comments by John Mueller do mean what we think, then this a bit 'shallow'; I see the point, but life is a bit more complex than that - the algo should be more nuanced (noindex is not necessarily hiding anything, just keeping the index to chapter headings, not listing every page, so to speak - it can be good sense / legitemate).

The instruction no-index does not imply no-read. If G is wise, they will not simply tally no-indexed page count but tally which are actually thin, or not - and I hope take into account non-text content (which I believe they can do).

So, if you noindex 80% of your site (I did so only because surprised it WAS indexed), one can expect a circa 20% improvemnt in serps (Actaul scoring might be more dramatic of course but a serp increase is no mean feat, and even a small increase welcome).

[edited by: bramley at 6:47 am (utc) on Apr 2, 2011]

[bramley]

Strange thing - wmt complained about a page with no title - this was just a typical blank page in a JS or PHP library to prevent directory listing. There are no links to it, so I wonder if G is now looking at all points along a url path, besides following actual links ...

Seen some other examples which made me think - how did G find that !

[edited by: bramley at 6:43 am (utc) on Apr 2, 2011]

[bramley]

In one case I am beaten in the serps by a wikipedia image - yes, that;s right, a single image beats my carefully designed page with introduction text, history, best things to do and see etc, plus multiple images and several videos (not plucked from elsewhere but self-made).

[bramley]

The only reason I have taken to reading these forums in the past weeks is a page that got a penalty and i just hapened to noticed it. No idea why even now, kept changing it every couple of days and eventually it was lifted. So sad that one cannot just make pages for visitors - the G algo is just too paranoid, too hair-triggered and the not telling you what the issue is just bizzare and non-sensical.

I always thought, erroneously it seems, that G itself decided what to index - could tell that if you had say Galley of X and then Gallery of X part 2 etc that it need only index the first page, and note that there was more in a good light.

Still, I have learned some useful things by looking again over all the details - and only hope G does the same; it's not there yet ...

Sorry for the post flood, but I will disppear now i had my say ;)

[Sgt_Kickaxe]

It's just not shallow content, I've tracked some mighty shallow sites that got by unscathed. Those sites, though extremely shallow, do not have very much duplicate content. Their product descriptions are all unique and their page titles are not cookie cutter or extremely keyword based.

These shallow sites also discuss the same product on various pages, they aren't full of "one page per keyword set but never mentioned on site again" type content. I'm leaning towards unique being more important than shallow right now.

[walkman]

"their page titles are not cookie cutter or extremely keyword based. "

This, I can confirm (at least by by itself) does nothing bad to your rankings.

We need some sites to come back so we can see. Have you seen if they had internal dupes, like too many tag pages?

[tedster]

If Google's Panda has hurt site-wide rankings because of navigational pages (and that's a big IF in my book) then that is an area Google should be tweaking in Panda.

It's one thing not to rank a category page, preferring to rank the actual content pages instead. But there's no reason to allow that decision (don't rank category pages) to hurt traffic across the whole site. I'm assuming here that the site has significant and deep content on the great majority of its pages - like 95% or better.

If a site is, say, 50% navigation pages (categories and so on) then I would lean toward a major simplification. Either that or get more content pages online.

[g1smd]

On a related note, I am of the opinion that "category in URL of product page" is a very bad idea, especially on sites with many ways to get to any particular product page.

Removing the category from the URL reduces Duplicate Content issues at the product page level. Usability is improved when those product pages also link back to all the relevant category pages.

[Shatner]

>>From reading the answers given by John Mueller, it seems as though noindex should be used for content you intend to improve. If you don't intend to improve shallow content, it sounds like he advises that you remove it.

What if you use disallow in robots.txt? Is that as good as removing it?

If it doesn't... what's the point of disallow then?

[Shatner]

>>So, if you noindex 80% of your site (I did so only because surprised it WAS indexed), one can expect a circa 20% improvemnt in serps (Actaul scoring might be more dramatic of course but a serp increase is no mean feat, and even a small increase welcome).

I did that as well, but haven't noticed any improvement. You have? Were you hit by Panda?

[SEOPTI]

Disallowing your navigation in robots.txt will kill your site. No link juice will flow.

[crobb305]

In one case I am beaten in the serps by a wikipedia image - yes, that;s right, a single image beats my carefully designed page with introduction text, history, best things to do and see etc, plus multiple images and several videos (not plucked from elsewhere but self-made).

Number 1 on a few big phrases in my industry: Yahoo Answers. The user-submitted responses provide links to sites of their choosing.

Scammer's haven. Anyone who has posted their own answer and included a link to their site can get traffic now that Google values it over all other content.

I think this largely refutes any theories of user-generated content being targeted by Panda.

[tedster]

A request - let's stick with this thread's topic which is category page navigation and Panda. There are plenty of other threads here where we can discuss more general Panda Update topics. Thanks.

[crobb305]

Funny you pose the question about eliminating categories, because I was thinking about adding them. My Pandalized site has always used a simple navigation, no category pages, all pages reside in the root. I have a sitemap page and top navigation for organization, and internal pages are just linked throughout the site --as needed-- from relevant pages. I have seen some top-ranking sites doing very well with categories.

let's stick with this thread's topic which is category page navigation and Panda.

It's easy to get sidetracked when reading the posts, and you see something you want to reply to. By the time you've read halfway through them, you can forget what the topic was.

Disallowing your navigation in robots.txt will kill your site. No link juice will flow.

I'm starting to feel like a navigation that appears on every page is part of my problem. Some of the top-ranking sites I have examined lack a common navigation area (not good for users, IMO), use images for links (reducing anchor text occurrence), or use nofollow on some/all of the nav links. When I look at my WMT data, I can see that my 5 hardest-hit pages tend to be the ones I link to from every page via top or side navigation, with an order-of-magnitude more internal links to those pages than to the others. I fear this is a problem, but I do not intend to remove, disable, or nofollow my navigation.

[tedster]

my 5 hardest-hit pages tend to be the ones I link to from every page via top or side navigation

First thoughts about that. Such heavily linked pages are the ones your site signals as the most important.

Panda may well be using a kind of "reverse PR flow" when it allows the pages it identifies as shallow to affect the entire site. In other words, the "shallowness" devaluation flows backwards to the pages that link TO the presumed shallow page.

I agree with you that removing the navigation itself is a weak idea, especially top navigation. I've seen major enterprise sites try that approach with recent redesigns, and it really kills their search traffic, even with millions of backlinks working for them.

If I'm right on this, the essential fix would be addressing why that target page is evaluated so poorly, as shallow content. Disrupting the navigation would be attacking a symptom and not the cause.

[tedster]

Taking this thinking one step further, those category pages that the OP asked about would only be problematic because they allow a few "shallow" pages to spread their curse back through the rest of the website in a more intensified way. And again, the fix might well be to fix those main problem pages rather than the category navigation.

[Shatner]

>>Disallowing your navigation in robots.txt will kill your site. No link juice will flow.

Didn't say I disallowed ALL my navigation, just the large number of category pages. Most of these were basically generated by tags, so I guess you could also call them tag pages, but I was using them to create useful categories.

Google had indexed them but they never brought in any traffic or had any ranking in the SERPs before Panda, and I never worried about it because they weren't there for traffic they were there to help direct users to specific topics.

So I don't want to remove them, but I disallowed them, and left the links to them there for users. Is that enough?

[bramley]

Update: seen a bit more improvement feed through today, so a typical page has gone from position 5 to 3.

This was after noindexing all follow on pages. My site is roughly place A, page 1,2,3 place B page 1,2,3,4,5 etc So I noindexed all pages other than the page 1s. It seems to have helped me quite a bit. At first it seemed marginal but has become better as the nondex is increasingly picked up and the recalcs fed through the G network.

Not sure those pages 2+ are shallow exactly (only so perhaps compared to the first, which I make strongest) but there was only a need to index the first pages (as one does for the chapters in a book).

There is a menu of all pages on all pages for any place, plus one back to the places index.