Welcome to the Q&A Forum

The new pagination advice is really tough to navigate. I have mixed feelings about rel=prev/next (hard to implement, doesn't work on Bing, etc.) but it seems generally reliable. If you have pagination AND parameters that impact pagination (like sorts), then you need to use prev/next and canonical tags. See the post Alan cited.

I actually do think NOINDEX works fine in many cases, if the paginated search (pages 2+) have little or no search value. It really depends on the situation and the scope, though. This can range from no big deal at all to a huge problem, depending on the site in question, so it's tough to give general advice.

I'm not having great luck with GWT parameter handling lately (as Alan said), especially on big sites. It just doesn't seem to work in certain situations, and I have no idea why Google ignores some settings and honors others. That one's driving me crazy, actually. It's easy to set up and you can try it, but I wouldn't count on it working.

Cross-domain canonical is the most viable option here. As Mike and Chris said, it is possible for Google to ignore the tag in some cases, but it's a fairly strong suggestion. There are two main reasons I'd recommend it:

(1) Syndicated content is the entire reason Google allowed the use of rel=canonical across domains. SEOs I know at large publishers have used it very effectively. While your situation may not be entirely the same, it sounds similar to a syndicated content scenario.

(2) It's really your only viable option. While a 301-redirect is almost always honored by Google, as Chris suggested, it's also very different. A 301 will take the visitors on the partner site page directly to your page, and that's not your intent. Rel=canonical will leave visitors on the partner page, but tell search engines to credit that page to the source. Google experimented with a content syndication tag, but that tag's been deprecated, so in most cases rel=canonical is the best choice we have left.

The W3C standards might allow for no protocol, but you would never just put "//" - that's part of the protocol ("http://", "https://", "ftp://", etc.). This usage is technically incorrect. It could cause minor issues on some browsers (although probably not on newer ones).

Does it matter for SEO? Well, that's a bit trickier. Google tend to ignore base href unless there are ambiguous relative URLs, like canonical tags that have no base URL and are unclear. Practically speaking, it's probably not a huge problem, but it is possible for it to cause issues down the road.

Either way, if it's on a sitewide template, it's a 5-minute job, and what they have is wrong. I'm not one to knock devs (I've been a dev and I've managed devs), but they need to stop arguing and just fix it.

I'm not sure what's tripping our warning (although that can occasionally be hyperactive), but I do see one problem. You're using the non-www version (which is fine, in theory), but your "Home" link goes to the "www..." version. You're sending Google a mixed signal with this. It doesn't seem to be causing major problems (no "www..." pages are in Google's index for your site), but I'd change the internal link to conform to the non-www default.

The basic problem is that Google and other crawlers (including ours) will keep visiting that "www..." and then get redirects and/or hit the canonical. It creates an odd loop that could waste crawler bandwidth and sends a mixed signal about your intent.

Wow, that's a tricky one. I haven't seen a lot of good data on rel="alternate" hreflang="x" yet, other than general suggestions that it's worth trying and seems to sometimes work. Technically, Google would say that the green/red/yellow versions aren't true duplicates, but practically, I think the canonical here is a good approach.

My gut feeling is that the canonical will overpower the hreflang - adding the alternate language versions to the color variants won't hurt you, but it probably won't do anything. If it were me, I think I'd go with your first approach - just add the hreflang tag to the canonical version and leave it off the variants. Monitor that - see how it impacts your international rankings, and go from there.

I tend to agree with Andrew that blocking all 1,000 pages really removes you from a lot of potential ranking opportunities, but two issues come into play:

(1) You may be stuck behind the safe-search wall, which can really diminish the ability of those pages to rank.

(2) The existence of these photos on other sites is definitely going to increase the chances of something like a Panda penalty, or, at the very least, aggressive filtering of that content.

Long-term, I think what Andrew said about making the content richer is critical - you're going to have to provide a clear value-add that Google can see. In the mean-time, it doesn't have to be all-or-none. Maybe instead of opening up all 1,000 photos to ranking you could open up just your most popular category, try to beef up that content, and see how it goes?

It depends a bit on the site structure, but I'd actually be wary of setting the category page canonicals back up to the main blog. These aren't really duplicates, and that could send an odd signal (and potentially negative) to Google, especially if there are a lot of them.

If you're talking about a few category pages, leave it alone. Use rel=prev/next for pagination and make sure you're handling and search filters (and not spinning out URLs), but just let these pages get crawled normally. They're an important path on the site.

If you've got a ton of categories, sub-categories, and tags, then I'd go with META NOINDEX. Important note, though: in most cases, you'd use NOINDEX, FOLLOW (not NOFOLLOW) - you don't want to cut the path for crawlers to reach your individual posts. Again, this does depend a bit on the site architecture and whether you have other crawl paths.

Did you 301-redirect or canonical from the old site that was optimized previously for this term? You could be seeing an indirect impact of back-links to that old site (which wouldn't show as links to the current site, depending on how you redirected them).

Patrick's explanation is definitely plausible, but it can be really tough to tell without a lot more data. Did you see a corresponding drop in rankings across keywords along with the organic traffic drop? Have you lost traffic only for specific pages or across the board? Did your total traffic drop as well, or did the organic traffic fall into other buckets (like direct or referral, which could indicate an attribution issue)?

The more you can segment this loss, the better. If you're seeing ranking drops across many keywords that mirror the organic traffic loss, and you've got a lot of link profile issues, Penguin is certainly a reasonable possibility.

Are these just snippets (link + paragraph) or are you displaying large portions of the posts on the home/category pages?

Alan's absolutely right about how canonical works, but I just want to clarify something - what about these pages is duplicated? In other words, are these regular searches (like product searches) with duplicate URLs, are these paginated searches (with page 2, 3, etc. that appear thin), or are these user-generated searches spinning out into new search pages (not exact duplicates but overlapping)? The solutions can vary a bit with the problem, and internal search is tricky.

I'd tend to agree. It won't harm you to go over, but since no one's going to see it, what's the point, generally. Now, if you feel thinks are being cut off arbitrarily (like the name of a product) and those keywords should be included somehow, fine - go over the limit. You can potentially rank for those terms and they may show up in other places. Some directories, social networks, etc. read the title tags, too.

If it's just ego, though, like a long brand name, I'd suggest letting it go. Unless you're a really big brand, it's not going to matter that much, and you're likely to rank for your brand anyway. If the brand is getting cut off, then name recognition won't help your CTR anyway.

Usability research definitely shows that people pay much more attention to the front of a headline, and even just the first two words, so even if a longer title "makes sense", people may not care or even see it. So, sometimes, I think we obsess over getting something just right that may not end up mattering that much, practically.

Personally, I'd pick a different battle with the client, and maybe just keep a few long that are critically important.

I honestly don't think that's a big deal - as long as you aren't creating tags or adding categories in a way that this could spin out of control. You've basically got 20-ish search result pages. They aren't high value, but they are useful paths to the blog content and they could rank for category keywords. I think it's a balancing act, and in many cases internal search can spin out of control and harm a site. My gut reaction, though, is that you're not in that situation, and cutting off these pages might do more harm than good.

Yeah, normally I'd say to NOINDEX those user-generated search URLs, but since they're collecting traffic, I'd have to side with Alan - a canonical may be your best bet here. Technically, they aren't "true" duplicates, but you don't want the 1K pages in the index, you don't want to lose the traffic (which NOINDEX would do), and you don't want to kill those pages for users (which a 301 would do).

Only thing I'd add is that, if some of these pages are generating most of the traffic (e.g. 10 pages = 90% of the traffic for these internal searches), you might want to make those permanent pages, like categories in your site architecture, and then 301 the custom URLs to those permanent pages.

Has anyone been doing link-building on your behalf? The pattern of links definitely looks intentional, and 99% of the time, that ends up being someone (an employee, an agency, a consultant, etc.) working on behalf of the site. If you have any other parties involved in the search efforts on your behalf, I'd make doubly sure to get a status report one exactly what they've been doing.

Without seeing the actual site in question, that's my opinion, yes.

Technically, the trailing slash version is the "correct" version. Almost all modern browsers automatically add it, so the practical implications are pretty small, but I think your consultant's essentially correct. As @walrus said, you're probably talking about in the ballpark of 1%, so I wouldn't obsess about it.

On the CRO side, usability research actually suggests that (on the web) the first 2 words of a headline carry most of the weight. In SERPs, I think that's even more at play - people scan, they don't read full TITLEs. The natural assumption is that long TITLEs carry more information, but in many cases, what's good for SEO is good for users with titles and headlines.

Of course, you don't want to be too spammy, or just string keywords together - that might get someone's attention but then cost you the click. Overall, though, I wouldn't be afraid of short TITLEs.

More info here in a post from 2009 by Jakob Nielsen:

http://www.useit.com/alertbox/nanocontent.html

Yeah, I think that's relatively safe, although it depends a bit on the scope relative to your overall site index and link profile (I wouldn't set up 500 affiliate URLs with a canonical on a site that only had 600 indexed URLs and a few dozen non-affiliate links). Keep in mind that Google may still choose to devalue the affiliate link, but the canonical tag will keep these landing pages from looking like duplicates and should prevent anything harsher.

I think that situation's a bit different - if you aren't interlinking and the sites are very different (your site vs. customer sites), there's no harm in shared hosting. If you share the IP and one site is hit with a severe penalty, there's a small chance of bleedover, but we don't even see that much these days. Now that we're running out of IPv4 addresses, shared IPs are a lot more common (by necessity).

I've been so buried in Private Q&A, I didn't realize you'd resurfaced over here. Long time, no see

Ah, understood. So, yes, in theory cross-domain canonical does handle this. I know major newspapers that use it for true syndication. There is risk, though, depending on the sites and content, and there is a chance Google will ignore it (moreso than in-domain canonical). So, I mostly wanted you to be aware of those risks.

META NOINDEX is safer, in some respects (Google is more likely to honor it), but if people start linking to multiple versions of the content, then you may lose the value of those inbound links on the NOINDEX'ed content. Since it's not showing up in search results, that's less likely (in other words, people are going to be most inclined to link to the canonical version), but it's a consideration.

It's really tough to give a recommendation without understanding the business model, but if you absolutely have to have separate sites and you feel that this content is valuable to the visitors of all three sites, then cross-domain canonical is an option. It's just not risk-free. Personally, I'd probably start with unique content across the three domains, then phase in the most useful pieces as duplicates with canonical. Measure and see how it goes. Don't launch 1,000 duplicates on three sites in one day.

Since this issue can occur site-wide, I do tend to agree with Anton that 301-redirects are a better solution for this particular problem (although canonical tags will work, if that's your only feasible option). It is important, as implied in the comments, to make sure hat your internal links are consistent and you aren't using both versions in your site (although, with "www" vs. non-www, that's pretty rare).

Practically, it depends a lot on the size of your site, whether you have links to both versions, and whether Google has indexed both version. This is a problem in theory, but it may not currently be a problem on your site. You can check the indexed pages of both the root domain and www subdomain separately in Google with these commands:

site:mysite.com inurl:www

site:mysite.com -inurl:www

(the first pulls up anything with "www", and the second only pages without it).

If you're seeing both in play, then sorting out how to do the 301-redirects is a good bet. If you're not, then it's still a solid preventive measure, but you don't need to panic.

It's kind of tough right now, because the SEs are touting Schema.org and they're claiming it's the future, but it seems like support of it today is still lackluster. Even a Bing rep hinted that at MozCon - for now, the best advice is probably to use Schemas plus one of the older formats. Whether RDFa or microformats probably depends on your niche. The best advice I can give is find a rich SERP in the niche you're looking for and see what those sites are using. If you can see it in a SERP, you know for a fact Google is honoring it in some cases.

I think I'd start slowly in that case. Keep the relationship aspect in mind, too. Even if all three companies know the writer/client and are aware of the relationship, sooner or later one of these articles is going to take off. If one site gets the SEO credit and the other two sites aren't ranking, there may be friction. Even if the work is spread out evenly and all high-quality, you don't control (ultimately) what content finally sticks and is successful. I just think things could get weird all-around if you send every article three places and only one gets credit.

Irving's correct - that tag is basically nonsense, as best I can tell. It's just a comment. There's no approved tag to block part of a page from Google. Google has a directive that looks similar to this, but it's for Google Search Appliance (internal site search) and doesn't impact organic search (we actually just tested it recently, coincidentally).

I disagree a little bit about cross-domain canonical. I have some friends at major media properties who have had good luck using cross-domain canonical for syndicated content, and it does seem to pass link-juice when used appropriately. It is true that Google doesn't always honor the cross-domain use, so it's very situational.

301-redirects are certainly more reliable, but they are different in the sense that a canonical will only remove the source page from search, whereas a 301 will remove it for users as well. So, it really depends on what you're trying to accomplish.

It's tough without seeing the sites, but MozRank is a bit more of a pure quantity score than DA/PA (which are built to predict ranking power). The fact that you have a lower MozRank but significantly higher PA suggest to me that the competitor is playing a numbers game. Your top links may have a higher quality, but they may just have more overall links, multiple links from some domains (site-wide maybe), etc.

I'd rather be higher on PA, buy our numbers. The picture I'm seeing here suggests your overall quality is better than your competitor's, so I don't think I'd go chasing after their link profile.

This isn't a typical application of rel=prev/next, and I'm finding Google's treatment of those tags is inconsistent, but the logic of what you're doing makes sense, and the tags seem to be properly implemented. Google is showing all of the pages indexed, but rel=prev/next doesn't generally de-index paginated content (like a canonical tag can).

Where is GWT showing them as duplicates (i.e. title, META description, etc.)?

Long-term, there are two viable solutions:

(1) Only index the main gallery (NOINDEX the rest). This will focus your ranking power, but you'll lose long-tail content.

(2) Put in the time to write at least a paragraph for each gallery page. It'll take some time, but it's doable.

Given the scope (you're talking dozens of pages, not 1000s), I'd lean toward (2). These pages are somewhat unique and do potentially have value, but you need to translate more of that uniqueness into copy Google can index.

Dr. Matt Peters presented some preliminary data at SMX Advanced today (6/11) and will be presenting more data at MozCon next month. I'm not sure when the full report will go out, but it should happen yet this summer.

Sorry, I should've said "I'm not sure how / how much we measure..." - it's been a while since I've dug into mR and mT, and I don't know how much an internal link structure could tip the balance. Like PageRank, it's iterative, so you'd have to crunch the numbers - it's not a simple, linear equation I could spell out here (even if knew/remembered it).

Truthfully, it's rarely worth getting hung up on, especially if PA/DA are solid and the site is ranking well. Knowing exactly why the competitor has a higher mR would rarely provide much or any advantage. It's generally better to track their strongest links and see if you're missing anything.

I agree with Phil that I'd canonical that home-page - I don't think it's the whole problem, but parameter blocking in GWT is inconsistent, in my experience. For something that important, I'd want a stronger signal. I'm also not clear on why your doing a 301-redirect to ref=collegexpress.com when a crawler hits the site - it just seems like an unnecessary step that could weaken your home-page.

Although careersandcolleges.com isn't targeting "collegexpress" on-page, you do seem to be using it in inbound anchor text. For example:

http://home.epix.net/~captclint/schools.html

This sets up conflicting cues for Google, and since the Careers and Colleges domain is a bit stronger, it could be enough to cause you some problems. I'd make sure that brand anchor text is isolated, as much as possible.

In some rare cases we'll see Google canonicalize two domains, if they think the content is too similar. I'm not seeing any evidence of that here, though. Both domains are ranking, so that suggests that either the Collegexpress domain is being weakened somehow or Collegesandcareers is overpowering it.

I strongly doubt that checklist is enough to tip you over the edge - it could just be that the page is code-heavy relative to any unique content (we including the template text in our duplicate calculations, so sometimes it can be over-sensitive).

If there's no unique information at all, then I'm with @MyHolidayMarketing - the time to add unique content to 30-40 products is time very well spent, for both SEO and conversion. If the problem is that you're spinning 30-40 products out into 1000s of URL variations, then you may want to consider not indexing all of those or canonicalizing them. Having every color and size of every product in your Google index can start to devalue your core pages after a while. That really depends on the scope.

This is a pretty common problem with event-oriented sites, and there's no easy solution. It's a trade-off - if you keep creating new URLs every time a new event is listed, you risking producing a lot of near duplicates and eventually diluting your index. At best, you could have dozens or hundreds of pages competing for the same keywords.

You could canonical or 301-redirect to the most recent event, but that has trade-offs, too. For one, a huge number of either can look odd to Google. Also, the latest event may not always be an appropriate target page, especially if more than just the data is changing. Unfortunately, without seeing the content, it's really tough to tell.

The other option is to create a static URL for every pairing and update the content on that page (maybe creating archival URLs for the old content, that are lower priority in the site architecture). That way, the most current URL never changes. Again, this depends a lot on the site and the scope.

If you'r just talking about a couple of URLs for a handful of events, I wouldn't worry too much about it. I probably wouldn't reverse the URL ("A vs. B" --> "B vs. A"), as it doesn't gain you much, but I also wouldn't lose sleep over it. If each pairing can generate dozens of URLs, though, I think you may want to consider a change in your site architecture.

Keep in mind that our metrics (including MozRank, DA, and PA) don't measure the risk of Google devaluing or penalizing your link profile. By no-following these links, your raw inbound link equity will go down, and it will impact our metrics if those links made up a large portion of your profile.

On the other hand, that doesn't mean that it's the wrong thing to do. Taken to extreme, these kinds of footer links are getting increasingly risky. Unfortunately, mitigating that risk does mean losing any positive impact of those links. It's not an easy decision, and it may be painful in the short-term, but you're betting that short-term pain against the risk of long-term catastrophe.

Whether that's the right decision for your site today is a tough call (and certainly one we can't make without details), and I'm assuming you had good reason to make it. I wouldn't worry too much about our metrics, and I'd focus instead on the impact on your actual rankings and traffic (especially traffic that converts). Don't reverse the decision too lightly, because switching back and forth can also send negative signals to Google.

Just the home-page, or the entire site? GWT is usually the fastest and most reliable way. I'd block the entire site in Robots.txt, too - sometimes, Google wants to see that prior to a GWT removal (although, usually, NOINDEX Is enough).

I assume they still want the page active for visitors (and just want it off of Google)?

The problem with (2) is that, if you cut the crawl path, Google can't process any on-page directives, like 301s, canonicals, etc. Now, eventually, they might try to re-crawl from the index (knowing the URL used to exist), but that can take a long time. So, while canonical is probably appropriate here, you may have to leave the old event/URL active long enough for Google to process the tag.

If these are really isolated cases, I wouldn't worry too much. Maybe rel=canonical them, and eventually Google will flush out the old URL. If this starts happening a lot, I'd really consider some kind of permanent URL for certain match-ups and events.

There's no easy answer. This stuff is very site-specific and can be tricky.

To add to what Tom said (and I endorsed his answer) - if Google honors the cross-domain canonical (and it's not a guarantee), then the non-canonical domain would fall out of the rankings (or, at least, certain pages would). Clients don't like to hear that, but the alternative is two domains competing with each other. At best, Google is likely to filter out one of the domains anyway (even without canonicals or redirects). At worst, there's potential for a Panda-scale penalty. There's very little to gain from this scenario, and a lot to lose, IMO.

I'm not a legal expert, but are they still using the trademark on the home-page? If they're not using it at all, but are only ranking on it from past efforts, there's not much of a case. They can't control the algorithm - they can only remove references on the site. Of course, it depends on how pervasive those references are and how much you want to fight it.

I just worry that de-indexing the home-page is going to have much broader, long-term consequences.

If you want to make the party threatening action happy at any cost to your site, you could rel-canonical your home-page to their home-page. I wouldn't do it, but it's another possibility.

Ultimately, though, all these things take time to process. If you, in good faith, have removed all trademark references and have requested removal with Google, you can't control the rest and I doubt you're liable for it.

If Page.fr is in French, I'd use rel=alternate/hreflang and not the canonical tag. It should allow both versions to rank for the appropriate audience without causing issues with duplication:

http://support.google.com/webmasters/bin/answer.py?hl=en&answer=189077

These aren't "true" duplicates and rel=canonical could disrupt the ranking ability of the French site.

Here's the problem - these things never happen in isolation. Technical problems, a weak link profile, possible negative SEO - they all compound each other. If this site had a strong historical link profile, throwing a few dozen (or even a few hundred) bad links at it wouldn't have an impact. Big sites can absorb that. The problem here is that the entire link profile looks too much like spam. Now, add in technical issues, and you're just crippling your efforts all around.

Neither of us can prove where the links came from, but it is my strong gut feeling that these links look a lot more like something a past SEO company or past employee would've done than any attacker. They're not blatantly bad - they're just uniformly low quality. The bad redirects looks like exactly something a company that moved sites would do - not an attacker. I think you need to dig deep, because either someone isn't telling you something, or someone just doesn't know. It's amazing how often past efforts never get documents or just disappear.

Majestic will index the links a bit faster than us, but nothing will do it as soon as they get posted. Right now, if this is entirely negative SEO, your only tools are disavow and legal action.

It is my sincere belief, though, that there are many fundamentals missing that could shield your site, and that you're going to end up chasing a problem that doesn't exist and wasting a lot of time and money.

I would either go with parameter blocking or META NOINDEX this page (that's probably a bit more effective). It would be better to block it from Google than to 404 100s of variants, as you could see a spike in 404s and that can cause some problems.

Sorry, edited this - you don't want to 404 the login pages, because that's going to return a 404 for visitors as well, and the pages won't function properly. You want to just keep this away from the bots.

If the canonical version of your URLs is secure (HTTPS), then yes - you should use absolute paths with "https://" in the them for your canonical tags.

If you are fighting a war, I'm not sure you'll win this war this particular way. There is not tool (including Google) that can give you link data in real-time (or even in a couple of days), so you'll end up acting late and always in a reactionary way. You'll have to cut as you go, and you'll always be a step behind, if they keep up the attack.

Google will often discount these links over time (unfortunately, it just takes them time to catch up), but your client's efforts may be better spent building a stronger link profile that can withstand attacks like this, even if it means some short-term pain. Otherwise, I fear this is a battle you'll keep fighting for months.

Unfortunately, the keyword-heavy brand name makes life a little tougher, because Google is more likely to see branded anchor text as suspicious (as opposed to having a brand name that isn't a general search term, like "Moz"). It may help to make sure the client's brand is strongly established in social channels and other profiles (including Google+). Let Google know this is associated with a brand, and you're not just keyword-stuffing (which may be how it appears to them, because of these targeted links).

Agreed - the "canonical" shouldn't just be a tag - it should reflect the actual link structure of your site, or it's not going to be as effective. Adding the additional keywords to the URL could slightly boost the profile pages, but there's always some risk to making a massive change like that. See my post from last year:

http://www.seomoz.org/blog/should-i-change-my-urls-for-seo

Yeah, it gets messy. I think I'm on the same page as Dan, but just some clarification. The two "approved" options are:

(1) Canonical all of the paginated series, including filters, to the "View All" version. One warning - Google only recommends this if the view all page loads fast and isn't too huge (and that's just good advice for usability, too).

(2) Using rel=prev/next for the paginated series, but then rel=canonical to point the filtered version to the current page. It's a bit mess.

Here are two good posts on the subject, but I'm afraid they reveal just how messy it is and how much SEOs differ in opinion. No real-world solution is perfect, that I've found:

http://www.ayima.com/seo-knowledge/conquering-pagination-guide.html

http://searchengineland.com/the-latest-greatest-on-seo-pagination-114284

Yeah, these are the newish attributed answer boxes (that come from the index/web), and here's what we know so far. Basically, if Google interprets a query as question-like (which is an ability that may have come out of Hummingbird or been increased by that update), then they look at the sites ranking on page 1 and try to see who's title/snippet/content best match the question.

In other words, you have to rank on page 1 organically, and then you have to provide a decent answer. This second part still seems pretty crude right now - Google seems to be trying to match against question/answer keywords in a fairly literal way. So, if you see a competitor ranking, you may want to tweak your page title, Meta Description, and on-page content to better match the intent of the question or to sound more like a succinct answer.

Canonicalization is a bit of an "it depends" answer in a lot of cases. If the domains are really identical and if you want to discontinue using the two duplicate domains, then 301-redirects are probably the best long-term bet.

If you need to continue using the other two domains (as landing pages for other marketing campaigns or paid search, for example), then 301s aren't going to be viable, and cross-domain canonicals should be a decent choice. I'd also say that if you just have a small number of duplicates pages (not the entire sites), then canonical tags are a good bet.

The only time I'd be cautious with 301s is if: (1) one of the domains has a bad history or questionable link profile, or (2) you're 301'ing a large number of domains - two domains is probably not an issue. In that cases, canonical tags may be a little safer (people aren't gaming them as much as 301s - yet - so Google isn't quite as suspicious).