Welcome to the Q&A Forum

Hi Jordan,

Yes, these are absolutely duplicate content. The vast majority of the text is the same, with some small additions and relevant city names changed. Google has been able to determine this sort of duplication for a long time. You do see some websites get away with this sort of content, but it's generally not something we'd recommend because Panda (and / or manual review) would pick this off as manipulative very quickly.

You would likely be a lot better off creating one page or one section of the website for brain injury rehabilitation in "the north east" with one copy of the relevant information. Additional pages can detail anything in particular about individual clinics, including unique contact information, testimonials, photographs, etc.

I don't believe that this content would come close to passing manual review as they are right now.

Hi there,

I would really like to do something about links, but have been told this will have no effect until the next "Penguin refresh".

This is only true if you are currently under a penguin-related penalty. That penalty won't be lifted until the algorithm update is rolled out, but if you are not currently under this type of penalty, link development will help your site's authority and, most likely, rankings. As long ago as 2007, you'd need to wait a few months to see improvements from any type of SEO work you did, but results usually come in a little quicker nowadays. That said, when I worked for an agency, we used three months as the minimum time frame to judge progress and make adjustments. Still, you'd see changes sooner than that in most cases.

If you do not have a good idea of your target keywords and which pages should be ranking for those keywords, I would put that in place, i.e. the matrix. I would have really thought the SEO agency would have done this early on, to be honest, even just for reporting purposes.

If you think you have thin pages, this can be an issue with Google's Panda algorithm, whose purpose is in part to devalue websites with too much thin, "useless" content. Unfortunately, it hit real estate websites particularly hard in some cases due to the nature of the industry. If you have a feed of listings, the content of that feed doesn't differ much property-to-property, and is duplicated extensively if those properties are also shared via a feed with aggregators or partner agents. As such, fleshing out these pages and sections can be very useful, so I would put this in as an important set as far as on-page work goes, after you have identified your primary keywords and developed a content plan to optimise for them.

Make sure you're really creating good content for those primary keywords, however - not "doorway pages" or content with no purpose besides attracting search bots and clicks.

The user interface testing for conversions is absolutely essential but is not an SEO item per se. Conversion rate optimisation can increase sales by hundreds of percentage points very quickly and if you have good internal or external resource to do it, do it now! That said, if SEOs are telling you to do this because "it's too hard to improve rankings" and not because it's an important part of their overall service, I would take this as somewhat of a red flag. It's not too difficult or impossible to improve rankings unless you are penalised or can't afford to invest in SEO (meaning you'd be unlikely to afford good CRO too). I am a big fan of using good SEO and CRO together - the results in terms of improved revenue can be quite astounding.

Cheers,

Jane

Hi there,

Although tied to the parent domain as a subdomain, this subdomain on its own has very few inbound links from third parties. OSE sees only 7 linking domains that are not Your Next Shoes. For a subdomain (which does not traditionally inherit terribly much authority from its parent domain), this is quite low. Although it does not explain a sudden drop (unless Google considered it to be ranking unfairly before, e.g. because links from the parent were counting as impartial inbound links when they should not be), general poor performance might be because of a low number of linking sites when compared to competitors.

Depending on how extensive your quoting of the government content is, you might just be able to link, or you might be better off canonicalising. A simple quote on an otherwise unique page is not reason to canonicalise, just as if you had quoted from a newspaper website in an article about a subject. There is no way you'd need to canonicalise your own article to that subject.

An entire page, lifted and republished for legitimate reasons, you could canonicalise to avoid any duplication confusion (even though a link was the proper way to go about identifying the original source of the content in the past).

Hi there,

Do you have access to what the .htaccess file looks like at the moment, i.e. how have the 302s been set up?

Cheers,

Jane

I would say that obsessing over nofollows or no nofollows is over-complicating things much more than simply linking to more information about a subject. The vast majority of content on the internet that links to informational resources, such as the example you've given (even though you own both resources) is not written or linked to with nofollow / SEO in mind. This is what I mean by it being sad if no one can create content and link between properties, whether they belong to one or more parties, without considering Google, SEO and nofollow.

It's shortsighted to claim that links between owned properties should be nofollowed. This is far from a set rule. Google does not consider linking to your own properties to be spam in and of itself. It would consider deliberate link manipulation via link networks to be spam, but is it spam if amazon.com links to amazon.co.uk? If Moz.com links to Opensiteexplorer.org? If Virgin corporation links to its health club chain? Or if I link from my blog to my consulting site? Hell no, it's not. It would be manipulative of me to create 800 websites promoting SEO consulting and link them all to my own website, however, just as it would be spammy for Virgin to write a new blog post every other day on sites they own, linking to their health clubs with the anchor text "gyms in London".

It's very subjective whether nofollow should be used, but I really do not agree that there is a hard-and-fast rule that any link using "optimised" (i.e. descriptive) anchor text should be nofollowed, just as much as I disagree that links to things you own need to be automatically nofollowed. At which point is a link natural if every link on the internet that meets these criteria is "unnatural"?

I agree that if the link does not add value, it should not be there, but in your example, you mention a person and link to more information about that person. Since a quote from that person is the crux of the page's subject, it absolutely adds value to link to more information about that person, no matter how well-known that person might be to the website's audience. I find it hard to argue that the link does not add value.

Yep - this is the purpose of the tag and by all accounts, it works well. It has been supported since late-2011, I believe, and I've both had no bad experiences with using it, nor have I heard stories of it not working properly.

Hi Ben,

It's absolutely possible to out-rank the exact match domain, although without seeing the actual examples, it's tough to say what will be needed. Even the DA / PA numbers don't tell me much, unfortunately.

Exact match domains have an advantage of sorts, simply because their content tends to be highly relevant to the query / subject matter. You will find, however, that if there are better options available for Google to rank, Google is not married to the idea of always ranking exact match domains first. Far from it. It used to be pretty easy to spam using this tactic, but that clearly didn't provide a good user experience for anyone. For example, if I owned carinsurance.co.uk because I bought it in 1999 and threw up a thin affiliate website on it in 2008, it's not going to rank well for [car insurance]. As it turns out, no domains like this do rank well for [car insurance]. (I don't know who is behind http://www.carinsurance.co.uk/ and I used it as an example before looking at it, I promise.)

Better overall SEO - link development, branding, content, social media work (if applicable to your industry and done correctly), etc. all contribute to out-ranking an exact-match domain, no matter how long that domain has ranked well for the query in question.

Hi Devanur,

Just one more thing to add (and I am definitely not trying to insinuate anything here!) but one thing to keep in mind with third-party tools like Moz, etc. is that they have access to all the publicly available data about your site and your competitors' sites, such as how many links they have and from where, but they do not have access to Google's filters or "spam" / quality measures. As such, Google may be devaluing some links that other services are discounting. Moz and other providers work very hard to ensure that they count / discount links in the same way as Google probably does. And for the most part, they get this right. However, when all of your metrics across different tools show that you should be ranking better, and you are not, it is a good idea to manually analyse your links and content in comparison to your competitors' and see if you can spot any issues that are not showing up in automated analyses (e.g., do you have a very high number of links from one particular type of site, whereas your competitors do not, etc.).

Pixelbypixel is correct - the metrics measure the authority of a site based on the number of links it has and where those links come from. Moz has some basic trust measurements built into these metrics so it is not a case of "the more links the better", although Moz clearly also does not have access to the sort of spam / anti-manipulation algorithms Google uses for weeding out bad links when it comes to assessing authority.

In general, the more high quality links you build to a site, the better its DA should be, and the better the PA of the pages receiving those links. The above point is also important too though: just because your DA / PA is higher, your site may not rank better automatically. Relevancy to the particular query, age of site and a lot of other metrics are going to influence which site actually ranks best for any one query. DA and PA are a good metric on which to base your progress, however.

By the way, the redirect tool I mention in my other response is throwing a 500 for that page right now!

http://i.imgur.com/mrpd4lK.png

Hi Daniel,

You can get a long way with a good free tool like this. My former agency got a lot of good press and social love for a redirect checker they developed last year: http://www.ayima.com/seo-knowledge/redirect-checker.html

This operates as a Chrome plug-in, which I would highly recommend, seeing the reaction the tool got. It's super easy to use, which people absolutely love.

I would say that this is absolutely not an instance where you would want to use nofollow. There is a huge difference between this and linking to a insurance company's commercial car insurance page with the anchor text "car insurance". It's sad that Google and the SEO community have jointly scared everyone to the extent that we are afraid of linking to information sources about non-commercial terms (e.g. "Brigham Young" linking to a Wiki page about Brigham Young). Nofollow is meant to indicate that you do not wish to vouch for the source of the information or that you have been paid to include the link and thus don't want to indicate that the link is purely editorial. This use is still true, eight years after nofollow's creation and it would be sad if we reached the stage where people are basically hesitant to link without it in almost every circumstance.

Put this in a commercial context and multiply the rate at which the target page or linking website receives / links out with high-value terms, and you have more of a problem. I have had clients ask me about ratios for years - "can we safely build links with 30% commercial anchor text?" - to which we'd have to say that there is no "safe ratio" for any particular keyword, niche or industry.

Google looks at far more than the anchor text when deciding on what is natural and what should be penalised / filtered. A page about a person or a product might use that person's or product's name nearly 100% of the time and be perfectly natural. I have also personally seen pages with 80%+ brand anchor text be penalised (not by Penguin but manually) because the links were clearly part of a sophisticated but fairly uniform paid link scheme, despite using anchor text links "Brand.com" and "visit their website". A high ratio of commercial anchor text is the icing on the cake for some of these penalties but there is no need to nofollow every link or even a selection of links just because it happens to be exact-match in terms of its destination.

Hi Jennifer,

I am wondering if perhaps the drop in traffic was not related to the removal of the pages. You say the drop happened just a day or two after the pages were removed. While Google works very fast with indexing new content, it should take a little longer than a day to process the removal / redirection of a large number of unpopular pages. It doesn't crawl rarely-updated / rarely-visited URLs on a regular basis (you can check on the last date a page was cached quite easily: http://i.imgur.com/NPmTF5S.png

Does analytics give you a good idea of where the missing traffic used to come from, i.e. which pages are not receiving traffic that did before?

Egol is right - pages like this can persist in quieter niches, especially if there are not a huge number of other results to take the thin results' place. I notice that this guy's page for London has a lot more content, probably indicative of the fact that the competition for rankings and business is a lot higher in London than it is in surrounding areas.

Hi Louie,

It seems unlikely to me (which doesn't mean it's impossible) that they have a city-specific penalty. Can't say I have ever heard of that before either, and I would say it would be hard for Google to decide on something like that, manually or algorithmically. The idea that it could be a large number of links from the Sydney region honestly seems like a stretch: that should really just result in a link-based penalty that affects the site nation-wide. Where this could have happened, however, is if someone abused directory links with way too many links using "sydney" in anchor text but I'm sure you'd have seen something like that already. Even then, I haven't heard of that hurting all rankings in that region alone, but rather just for queries including "sydney".

Have you tested this searching from IPs in each region, i.e. I just wanted to make sure it's 100% confirmed that the site does rank well and receives traffic from Brisbane, Melbourne, Perth etc. and that you are not seeing personalised results in those locations and not in Sydney. I know it seems like grasping at straws but it's a really weird case for a non-geo-specific query.

Hi there,

Just to add to what Martijn has said, you should be able to specify the pages that target the US with this tag and leave the pages for international users unspecified. This means creating different pages / URLs of course, rather than using IP detection which isn't advised for changing content.

Hi Louie,

Has there been any improvement to this? If not, could it be possible given the query and the sites that do rank in the Sydney region, that the query is particularly competitive in Sydney as compared to other cities, and that the ranking sites have put any extra effort into that location?

I understand that it isn't a city-specific query, but is the service / product something that might be considered a more profitable query to rank for in Sydney?

Google doesn't support city-specific geo-targeting so it won't be something they've set up incorrectly in Webmaster Tools.

Is there any address or contact information on the client's website that might specifically indicate that they serve areas that are not Sydney?

Cheers,

Jane

Hi there,

I think what William is saying is that site B will receive the benefit of a new linking domain, but that linking domain (site A) is many times stronger than site B because of its extra backlinks. As a result, site B still gains a good amount from receiving the redirect.

There are different opinions on this, i.e. on whether a redirect seems to help a lot more than this, and every case in which this takes place seems to have different results, largely because of all the other metrics affecting both sites. It's obviously hard to test this in a controlled environment.

There is debate around how much of a boost a cross-domain redirect actually has on the destination website - certainly in days gone by, the site at which the redirect is pointed will receive the vast majority, if not all, of the redirected site's rankings. This can still happen if all of the content on site A's redirected pages is reflected on site B, or if site B's content is sufficiently the same. However, if you were just redirecting for the purpose of making site B stronger and there was no real cross-over between the content of either page / site, this process shouldn't have a hugely beneficial effect on site B.

Hi Luke,

Noindex, follow will work fine for the duplicated tag pages, although you could consider canonicalising them or redirecting them to a more useful resource - if you can do this en masse in an automated way, or if you only do this manually for tags / topics of high importance.

302ing to the search pages isn't good for two reasons: one, search engines traditionally don't follow or pass PR through a 302, and they also prefer you don't include your own "search pages" in their indexes. The "easy" way around this is exactly as you describe: produce quality category pages in place of what was a search results page. You can probably get away with having search pages indexed, and many companies get them to rank, but what Google wants to avoid is the indexation of hundreds or thousands of random search results pages from a website, often with complex query strings that can result in almost infinite numbers of pages being created.

Cheers,

Jane

Hi Joanne,

Sorry for not totally expanding on what I meant as the outcome using javascript! Yes, you're "hiding" portions of the mega menu. There is speculation that search engines might execute JS commands to find out if there is content behind them, but at the moment this still seems a valid way to "remove" links from Google's total link count. Keep and eye on whether this changes though (and acknowledge that it might not be totally reliable, even now).

Perhaps check out this tutorial for JS menus: http://www.siteground.com/tutorials/menu/javascript.htm

Cheers,

Jane

Hi there, glad to have been of help!

As far as how long to wait goes, it really is hard to say but if you see nothing at all in a month, I'd consider other options like canonicalisation (or coming back here, showing what you've done and seeing what people say - as I am only one opinion! ).

Cheers,

Jane

Hi,

It should do - the link is gone, the page is gone and deindexed. There should be no reason why the PageRank distribution changes to pass a higher percentage to the remaining links.

You should even be able to achieve this by simply removing the link to Page4 without removing the page from the server or using the URL removal tool.

How long this takes is not something I can say, but as soon as Google re-crawls the updated page, it should take note of the missing link and update how PR is passed from the page accordingly.

Hi Stephan,

I agree with Scott that this is possible but question how much more trouble it would be to run two installations for the two blogs?

The canonical tag will take care of the duplicate content. I don't see much wrong with duplicating using this approach if the duplicated full article is better for users on the non-canonical site, or using the snippet + link approach if that's fine for usability / reader interest.

This is a pretty old, annoying phenomenon called keyword cannibalisation. As DJ123 has said, Google can't work out which page it should rank for the generic term. It's incredibly frustrating when the signals are otherwise perfect to point out the right page.

I would add the "shoes" link--pointing to the main page--to all pages, not just the Sports Shoes page, e.g. from Hiking, Dancing, etc. as well, as is the case in this graphic. This graphic is from a very old blog post on keyword cannibalisation, which will read as pretty basic now but might be worth a glance.

This could be a risky idea, but if the metrics for the main shoes page are significantly stronger than the Sports page, you could consider canonicalising the Sports page to the main page for as long as it takes for Google to pick up the main sports page and "drop" the Sports page. Once that happens, remove the canonical tag and fetch as Googlebot from Webmaster Tools, waiting for the Sports page to be re-indexed. I doubt that the Sports page would take precedence over the main page again after this, but this is risky because it will force the Sports page out, and that page is currently ranking. You're taking a fairly confident gamble that the main page will take its place, but it's a gamble nonetheless.

I have seen some companies commission interesting studies that are vaguely related to their niches, supposedly for research purposes, but if these are good enough and interesting enough, they are often cited in news sources.

Take insurance company Aviva - insurance is pretty boring, right? The company compiled a study on the amount of money new parents spend on their babies a couple of months ago - the article Aviva published is here: http://www.aviva.co.uk/media-centre/story/17298/first-time-parents-spend-492-million-preparing-for/

Here is some of the press that study got:

http://www.thesundaytimes.co.uk/sto/business/money/Consumer/article1411345.ece (also in print edition, hence brevity of online snippet)

http://www.dailymail.co.uk/news/article-2609904/Do-really-need-Moses-basket-newborn-Parents-splash-1-600-baby-wasted-unnecessary-items.html

http://www.mindfulmoney.co.uk/financial-planning/first-time-parents-spending-almost-500m-preparing-for-their-babys-arrival/

http://www.covermagazine.co.uk/cover/news/2340859/just-one-in-five-new-parents-have-life-cover

https://www.healthinsurancedaily.com/health-insurance/product-area/life-critical-illness/article440449.ece

http://www.futureproofinsurance.co.uk/blog/?p=397

Unfortunately, these articles don't link to Aviva's website - either the home page, the release or anywhere else. This is an action that should be followed up by the SEOs (disclaimer: I have worked with Aviva on projects in the past, but neither myself nor anyone I know was involved in this piece of content creation or distribution - I just liked the study and appreciated the attention it got, minus the links. I read about it in the newspaper).

Aviva has conducted similar studies in the past, as have a bunch of similar companies whose core product is otherwise both competitive and not a perfect candidate for "content marketing". The trick is to spin the market such that you find an interesting facet to research and write about.

If that market happened to be shoes, a study could focus on how many Americans suffer unnecessary injuries due to wearing the wrong footwear. For instance, running shoes that don't suit their running styles, high heels in jobs where people spend a lot of time on their feet, shoes that are too small, whether Vibram shoes are beneficial to foot and leg health or whether they're a gimmick, how much money per year Americans spend in health costs associated with foot / leg health that could be avoided by wearing better or correct footwear. I've made this up quickly, so please excuse the rough idea!

The key is to find something that resonates with a lot of people, not just your target market. Aviva doesn't really mind that the majority of the people who consume their new parents article are not going to take out an Aviva health insurance or life insurance policy. Whilst they're definitely raising awareness for these products, they're also raising basic company awareness and they should gain links because of that press.

The more obscure or dull the industry, the harder this can become - and consumer insurance does have direct ties to a large portion of the population's everyday life. Again, you're looking for an angle that outweighs the fact that "nobody links to the content/information with rare exceptions." The more you can tie your angle to both your basic niche and the general public's interest, the better.

You can also get more specific than "the general public" and target a group of potential linkers who would otherwise not be interested in your market - say you are a shipping / international logistics company. You might find that the music industry is very liberal with their blogging and link-giving. Research and write about the amount of money lost per year in damages shipping music equipment, providing insurance options and developing guidelines for safety in that specific market. Again, a very off-the-top-of-my-head example (and I doubt the music industry links like this at all, but you get what I mean).

Hope this helps!

Jane

This should not cause a problem. As soralsokal says, internal anchor text linking isn't a regular problem for over-optimisation; the only way you'd be in danger of attracting unwanted search engine attention is if you were churning out hundreds of low-quality blog posts, linking internally over and over again to commercial pages with spammy anchor text.

Google is used to seeing navigational links on ecommerce sites, the vast majority of which are "optimised". The optimisation is fine; it's for users. You wouldn't link to your wine page with any other anchor than "wine" because it tells people what the page will be about Google knows this; navigational elements like this are fine as a result.

Hi Mark,

Redirection and an updated XML sitemap will be enough to ensure this change is acknowledged properly by Google.

Cheers,

Jane

Hi Robert,

Jonathan has said that nothing will be passed as parameters, but even if they were, I don't see that there would ever be a _penalty _for this (in the true sense of a penalty - algorithmic or manual penalisation for spam). You could flood Google with a million query strings and no canonicalisation if you did it badly which could get dangerous, is the only thing I can think of and even then, this would be easy to fix with canonicalisation on your own sites.

Yep, it's another "it depends" - if Moz links to Search Engine Land in multiple blog posts over many years (as it has done), this is going to count for more than one vote. Those links also undoubtedly go to different pages on SEL's site (new posts, etc.). But if I write one blog post every other day, linking to my affiliate site in every post, this really won't help the affiliate site at all

Hi Israr,

Since your site is properly targeted to Pakistan in Webmaster Tools, don't worry about having a .com domain as opposed to .com.pk. The WMT configuration does that job perfectly.

I would be tempted to guess that one of a number of things is happening: It could very well be that the US results are being affected by recent updates and those updates haven't rolled out to the rest of the world yet (the US usually sees updates first; in the first Panda roll-out, it took Google a long time to extend the roll-out past the US. I remember being at a search conference in Poland over a month after Panda was released in the US and people kept bugging the Google rep about when it would come to Europe!). In this case, you might be in for a good ranking improvement in Pakistan soon.

It also might be the case that there is more competition for the term you want to rank for in Pakistan than there is in the US. If the query pertains especially to Pakistani audiences (mentions location, is about a product or service only available in Pakistan, etc.) then other competitors might have fared better in providing adequate localised signals, whereas your site might be appealing to a more general audience (as far as the signals it gives Google are concerned).

Is your website hosted in Pakistan? The location of a website's hosting should not be a major factor at all, but it can make a small difference if Google has trouble figuring out what the intended target country is. Your WMT geo-location should have taken care of this though, even if the site is hosted in the US or elsewhere.

Also, do you have a good number of links from other Pakistani websites, whether they be .com.pk domains, sites hosted in Pakistan, and sites with a lot of Pakistani links themselves? Again, not a massive ranking factor, but if most of your links come from the US or other foreign countries, this isn't as good of a location signal as if you have a lot of links from other local websites.

To be honest, I think that the first scenario is more likely and that a movement upwards in the US results is a good sign...

Hi Alex,

When it comes to bad links pointing to your site, definitely remove them as an when you can (i.e. immediately, acknowledging that it will take time to contact every linking webmaster and negotiate the link removal). If you are going to suffer any ill effects in rankings from removing bad links (which does happen if Google was passing PageRank through those links), it's still best to get this done sooner rather than later. What you do not want is bad links that haven't been noticed to suddenly be picked up upon, resulting in a manual penalty.

Hi Jonathan,

I think I understand what you're asking - you're asking if it matters for social signals if a link out from a social media website (be it Twitter, Facebook, Pinterest, etc.) goes through a 301 or a 302 redirect, e.g. a bit.ly link goes through one 301 to get to the target page, right?

Unfortunately, I don't know whether a 301 or 302 has any influence over how Google treats links from social media (which are usually also nofollowed, but we're talking purely about social signals here). I can only speculate that if I were Google and I wanted to look at social signals as opposed to SEO ranking signals, I would not take into account whether the redirect was a 301 or a 302.

If I were building my own tool to do this, I'd absolutely use 301s just because they are best practice for SEO, so you're going to get the benefit of the 301 if the link comes from somewhere other than a nofollowed social media site. There would be no reason to penalise a URL shortener or its target URLs for this.

Third vote for using the site and not redirecting it here (pending a re-brand of sorts to suit the fact that it's now owned by your company).

You have the opportunity to take up a lot more real estate in the SERPs than you do now, whereas redirection at worst removes the old site from the results. At best, you might see a small improvement in your rankings, but you're already ranking better than that site to begin with. As EGOL says, PageRank is logarithmic so if you're dealing with what is a weaker website (whether in actual PR or toolbar PR - the latter of which is a fairly outdated metric but still gets used for basic comparison purposes), it may actually be many times weaker than your stronger domain.

At best, using both site might allow you to target different sections of the same audience, appeal to different needs from the audience base (very large clients versus smaller businesses, etc.), and for results where you rank top 10 with both sites, you're about 20% of the results, not 10%

Hi Marcel,

Not sure that this can be put down to Panda: it could be related since Panda deals solely with on-page, but Google works on so many changes at once that it might not be. There is a team dedicated to working on SERP display, optimising click-through from SERPs, etc. so it would be very difficult to say exactly what this relates to. It's definitely been happening (a higher number of people asking in here about why their descriptions and title tags are changing in SERPs) for a few months now.

Hi Damien,

So if I'm reading this correctly, the consultant is saying that due to the size of the site (tens of thousands of pages), and the need to categorise its content, that subdomains are the best choice.

I would say that there are far bigger websites using categories within subfolders, notably big retailers, e.g.

http://www.marksandspencer.com/c/beauty, http://www.marksandspencer.com/c/food-and-wine, http://www.marksandspencer.com/c/mands-bank

http://www.waitrose.com/home/inspiration.html, http://www.waitrose.com/home/wine.html, http://www.waitrose.com/content/waitrose/en/home/tv/highlights.html (<-- the last one being a crappy version, but a subdomain nonetheless)

and so do websites that deal with providing content for very different audiences:

http://www.ncaa.com/schools/tampa, http://www.ncaa.com/championships/lacrosse-men/d1/tickets, http://www.ncaa.com/news/swimming-men/article/2014-03-29/golden-bears-and-coach-david-durden-earn-third-national-title, http://www.ncaa.com/stats/football/fbs

Has the consultant provided examples of other websites doing this that would take on the same structure?

There are hundreds of examples of websites whose structure / categories are properly understood despite existing in subdirectories, so I'm still sceptical that this is a necessity.

This is not to say that a subdomain approach wouldn't work and is definitively bad or anything, I'm just not really convinced that the reasoning is strong enough to move content away from the root domain.

I disagree about user experience - from a user's perspective, the only difference between subfolders and subdomains is the URL they can see in the address bar. The rest is aesthetic. You can do or not do everything you'd do with the design of a website using subdirectories that you'd do with a website(s) employing subdomains. For example, just because content sits on www.icy-veins.com/wow/, its navigation wouldn't have to link to www.icy-veins.com/hearthstone/ or mention the other brand in any way if you don't want to. You can still have separate conversion funnels, newsletter sign-ups, advertising pages, etc.

Hey Joost,

That's a tough one because it probably should be subjective and depend on other factors about the linking site, the site it links to and how it links. If the BBC were to link to me twice, once to a new product page on my website and once to my home page, I'm not going to be concerned that the link is only worth the first link in the HTML code, and freak out if that's not to the page I'm interested in. Same thing goes for a lesser site to the BBC, but that would be a highly authoritative example.

That said, if you're counting links from c-classes or IPs, which is a very common way to assess backlinks, that page on the BBC is going to count as one "vote".

If I see a sidebar linking out twice to the same domain, I'm not going to be all that comfortable claiming that both those links are going to be any more useful than one would have been.

I don't believe Google would be simplistic enough to treat two links from one URL on bbc.co.uk to two different pages on one website the same way as it would treat two links on a blogspot blog to two pages on another website, if that makes sense.

Hi Christopher,

The set-up you have at the moment is actually fine for SEO - http://www.fdmgroup.com/ resolves the website. http://fdmgroup.com/ employs a 301 redirect to take users and search engines to http://www.fdmgroup.com/: http://i.imgur.com/et4AKYV.png

This is correct, http://fdmgroup.com/ would be seen as a separate "version" of the home page if it was allowed to load without sending everyone on to the "www" version via the 301 redirect. Furthermore, every page within the http://fdmgroup.com/ "version" of the website would be duplicated from the "www" version of each page.

If you want to change to use http://fdmgroup.com/ instead, using that both on advertising and having that be the version of the website that resolves for both users and search engines, you will need to reverse that 301 redirect. This will mean that when people try to visit http://**www.**fdmgroup.com/, they are redirected to http://fdmgroup.com/. This is a simple process.

However, it is inadvisable to go through redirection like this unless you really, really have to. When you redirect a URL with a 301 redirect, a large portion of the URL's authority is passed on to the new URL. Not all of the authority is passed though. As a result, your rankings and traffic can take a little hit for a short while. This is not usually a big problem, and usually resolves itself quickly but it is best avoided unless the redirection really has to take place.

I am tempted to say that "it looks cleaner" is not the best reason to go through this change when your current set-up is totally fine and correct for SEO purposes.

That said, you absolutely could reference fdmgroup.com in offline advertising for stylistic purposes. When people type that URL in, they'll be redirected to the www version just as they are now. This is pretty common because of the stylistic benefits of not including www in TV / print advertising.

Hi Alfredo,

Some interesting discussion here. Coming back to the root of the question, Matt's response of "Google has a lot of experience in returning relevant web pages, regardless of the top-level domain (TLD)" is very important. It cancels out a lot of the question about the new TLDs.

I would definitely not class this as one of the statements he makes in order to spin a line for Google either - this should be taken as their stance, in my view.

So with that said, the question is largely down to EMDs.

Google targeted low quality EMDs a couple of years ago: http://searchengineland.com/low-quality-exact-match-domains-are-googles-next-target-134889. Low quality is the key point here; this doesn't really address whether a high quality site called SciFiBooks.company (or .com) has inherently more chance of ranking for [scifi books] than a very similar website (in terms of backlinks and content quality) with a different name. Debate about that is ongoing, as you can see here.

Hi Damien,

There are cases where subdomains are very necessary or inevitable, usually because of technical limitations (and even then, they can usually be worked around via practices like reverse proxy). When you see subdomains in the wild and aren't sure why they're being used, they will often just be legacies - old set-ups that no one wants to change because it would require redirecting old URLs, which is inadvisable if those URLs don't need to be redirected and if they rank well.

In this case, I'd be really interested to know why the SEO was adamant that the new structure should use subdomains and not subdirectories. Google is much better at working with new subdomains now than it was in years past, but if there is no good reason to use them, subdirectories are still the safer option for SEO purposes, and the content housed on subdirectories should automatically inherit authority from the parent domain. New subdomains seem to be far less likely to inherit this authority, as other responders have said above.

Find out exactly why the SEO wanted subdomains - if their reasoning isn't solid, you may want to place this content in subdirectories and place 301 redirects from the subdomains to the subdirectories. If you are going to do these redirects, doing them sooner rather than later is advisable as redirection usually comes with a short period of lower rankings / traffic.

On that note, redirection does usually result with that short period of traffic loss, but that should happen quite quickly and be fixing itself in 2+ weeks, not getting worse.

Hi Joost,

I don't have hard data on this at all; this is a what-I-know plus gut feeling answer.

Gary is right - multiple links from one page to another target should be treated more like one link to the target domain, but this might not be a uniform rule. In effect, two links from one page whether those links be site-wide or individual shouldn't have much more or less effect on the target website than just one.

That said, if Google felt that site-wide link or text blocks were being used manipulatively, there is no reason why they would not discount the value of those links altogether. It's interesting that you may have seen a correlation between multiple links from site-wide areas and poorer performance. It would also be interesting to see the data - you could put together a good blog post about it with enough data, for sure.

Exactly as Anirban says. We're seeing a lot of directory SEO questions lately, and it's worrying. Both the exact anchor text and the fact that directory links are more often than not junk is why you should remove them.

If they were directories that were HIGHLY relevant to your business and frequented by people looking for your type of services, they might be fine. For instance, if you found a directory listing gardeners, plumbers, electricians and builders in Los Angeles and you happen to be a gardener in Los Angeles, there is some sense here.

However, the vast majority of directory submissions we see are made-for-SEO and should be removed ASAP.

Yep, I would guess that your Panda issue was far more likely to have been the URLs than the menu - that's pretty standard Panda fare so that's great that you've sorted that out.

I am leaning towards really doubting that the menu is hurting you at all, but of course experimenting with removing non-essential menu items is possible. It's hard to draw real conclusions from the results you see after you modify something like a menu because other factors will influence rankings during that time, e.g. you gain or lose links, Google changes its treatment of unrelated metrics, your competitors gain or lose links, etc.

I'm assuming these internal keyword links are a lot more effective if they're only competing against 100 links instead of 400+

In theory, kind of - one link out of 100 that uses the anchor text "natural skin care" will receive a higher percentage of the available passable PageRank than one link out of 400 using the same anchor text. The link from a page with just 99 other links should not be seen as any more relevant for the phrase "natural skin care" than the link from the 400-link page though. But it should receive more of a boost for that keyphrase than if it had to share the passable authority with three times the number of targets.

I am making an assumption here, so if anyone else knows that relevancy in internal anchor text goes up along with the division of passable PageRank, please chime in!

Google doesn't have the same reverence for anchor text that it had a few years ago, but most of its efforts in not relying so much on anchor text have been focused on inbound links from other websites, rather than internal links.

WMT and GA data rarely matches up perfectly, but William's point about the GA code not working for whatever reason is important.

My former agency wrote our own analytics code for one client site. It was a small piece of JavaScript. The client placed it below a very slow-loading piece of conversion rate optimisation JS that took a very long time to load. Most people coming to the website knew what they wanted from the home page quite quickly and we had about 50% of our SEO traffic clicking away from the home page before our code had loaded. As a result, a huge amount of our traffic was missing from our weekly reports: we could only report on those people who'd stayed on the home page for long enough to have our code load, or who had landed on other pages and taken longer to figure out what they wanted or who had come to the page they wanted immediately.

We had our code moved up in the page code. JavaScript processes commands as they appear in the code, so our code loaded "immediately" after it was placed high in the code, and the CRO code did its thing afterwards. Our traffic went back up (after we had been through hours of explaining that rankings and interest had not dropped, so it was absolutely a tracking problem and not a real traffic problem).

Not saying this is happening here, but take a look at the page load times, JS on the page and placement of the GA code, plus raw server logs if you have access to them, and see if there is possibly a tracking problem with GA.

Hi Joanne,

The "fewer than 100 links" quote is a little bit outdated in terms of what Google can actually handle. This article from late last year makes clear that the limit is long gone. The guideline is an okay benchmark when it comes to ensuring that you aren't overloading a page with useless or irrelevant content and links, but it's definitely not a requirement for crawling / indexing / ranking success anymore.

Your drop-down looks pretty standard to me. Some sites choose to present menus like this using JavaScript to call lower-level URLs (e.g. Hair, Face & Body + Natural Skin Care might be links in the HTML but then a JS call is required to bring up Certified Organic Skin Care, Cleansers, etc. - Google and other search engines do not traditionally execute JavaScript and thus won't see the links) but this is an increasingly unnecessary tactic, and could be detrimental if those URLs are not linked to in a crawlable manner elsewhere within the website.

Having a lot of links on a page has one "detriment" - a page does not "leak" PageRank / authority if it links out a lot, but it does mean that less PageRank passes through each link if there are hundreds of links, as opposed to 50 or 100 links. The total amount of PR that can be passed on is divided amongst more links. Again, it's important that links to all your pages be crawled so that those pages are indexed and receive authority, so I don't think you have to worry about this.

Further example: http://www.marksandspencer.com/ uses a JS-powered navigation, http://www.debenhams.com/ does not.