Welcome to the Q&A Forum

Agree with the others here, two or three sites definitely are a lot harder to maintain, and a properly-structured website can cater to different types of audiences, even when they vary a lot. You'll want to be very careful with the UX and perhaps work with some conversion rate optimisation people as well to segregate the audiences and not confuse the sales process for either, but this should be possible, especially if you focus your marketing for each audience to drive traffic to pages where there is no initial cross-over (shared navigation, etc.). I am guessing since I have not seen either product or preliminary site (if there is one), but I believe that one site would be better given that there will be shared pages, such as company information.

If you were to create two sites, I suggest canonicalising the duplicate content to one version of the site, most likely the one with the larger user base or higher potential return. E.g., www.site.com/company-information and www.b2bsite.com/company-information are canonicalised to www.site.com/company-information if www.site.com is the primary business interest.

As Silkstream says, the practice itself isn't spammy but it has been thoroughly abused by "SEOs" in the past to the extent that it should be done with care and for non-SEO purposes only. Firstly, it's not really likely to help your SEO anymore. Most of the links you will gain will be nofollowed, and while there is some reason to believe that a LOT of high quality nofollow activity (like you might get from a very successful viral campaign) is a good signal, it's not technically an important metric for SEO and shouldn't be treated as such. You should be bookmarking where it would make sense to share / save content for other marketing purposes, with SEO taking a very far back seat as a reason why.

Hi Nils,

If you aren't producing translated content, but are just going to put some posts out in English, I do not believe you will see a problem. You could consider putting the English content into a subfolder like /en/ and including a header such as http-equiv="Content-Language" content="en"> in the head. You could also target that subfolder to an English speaking country in Webmaster Tools, but you can't pick more than one country, so that doesn't seem the right option here (i.e. you'd have to choose the UK, US, or another place, and if you're trying to get visitors from all over the world, that could hinder your rankings in all locations bar the country you choose).

If your entire website is geo-targeted to Sweden, either by a .se domain name or geo-targeting in WMT already, you might have a harder time ranking in foreign countries as Google has been told that the entirety of the site is for the Swedish market. However, on a generic TLD like .com, you should be able to pick up a good amount of search traffic from other markets with English content.

Hi Christopher,

Can you elaborate on where you've seen a meta refresh reported?

Cheers

PageRank is a pretty outdated way of looking at a page's worth, especially because it's updated irregularly, and often fudged / inaccurate. Google realised a while ago that people used PageRank as a metric to buy and sell links, meaning that them making it very accurate was helping fund an industry they very much didn't like.

That said, 2.5 years with no internal PR showing up on any URL seems slightly odd. Do the pages show decent Moz metrics, like Page Authority?

Does the site sit on HTTPS URLs exclusively? At least a little while ago, that used to stop toolbar PR from showing up but it didn't mean that pages wouldn't rank.

It's also possible that the site receives few or no links from elsewhere, meaning that there is not much authority to be passed through the site from the home page or any other target page. Does this seem likely to you, looking at the site's links? As others have said, there could also be internal linking problems that halt the flow of authority, but 2.5 years is a while to show no internal tbPR whatsoever.

If there is a need to write these from a customer's point of view, no matter how they are going to find the content, then the posts should be written. You don't need to cultivate everything on your site for search - there should be a reason behind why you create and publish content, however, and EGOL is right that a content plan is an important start if you feel that you're writing / publishing "in the dark" with no structure.

Short updates might be very useful for your current readership, or for historical purposes, so that future readers can go back and research what constituted a "record price" in 2014. They may never arrive at that content through a search engine, but it's useful to them and they find it regardless.

EGOL and Samuel raise some very good points here. I feel like you're also somewhat at risk of looking like you're trying to deceive the audience if you're deliberately divorcing the community from the brand site and image, but using it to drive the community to commercial action. I am not sure how subtle they are planning on making that connection, but subtlety in that manner rarely works: it's a lot more "honest" and appreciated to cultivate a community and openly market a product, as Moz has done.

Using the Moz example, this site did not start out as a tool / analytics provider, so the community actually came first. When I was a full time employee, we were largely an SEO consultancy. The blog was a labour of love as far as Rand and the staff were concerned. In 2007 when we first introduced Premium (now Pro) accounts, we clearly used the community (which was pretty substantial, even back then) to market the service. To all but a very few negative folks, this went down fine.

If Rand had started a separate company and website hosting and selling SEO tools and subscriptions, and had dropped hints and posts every now and again here to promote it, it would not have been met with nearly as much favour.

The SEO points about two sites being twice as hard to maintain, market and rank than one also apply, but I would question whether a smaller brand can build a totally separate community that is a) big enough, and b) independent enough not to annoy users if they figure out the commercial intent behind it. If there's one thing that's slightly worse than no community, it's an empty one.

HI Diego,

You don't need to worry too much about whether the pages are tags or categories, but could you give me examples of the URLs you plan to use?

Hi Louise,

If you can ID the parameters, you can also look at blocking these in Webmaster Tools. This page explains more. As with any blocking of URLs, of course, proceed with caution.

Hi Christopher,

You have added canonical tags to http://blog.fdmgroup.com/uk/blog/, pointing to http://www.fdmgroup.com/uk/blog/ so this takes care of the duplicate content problem between those two versions of the blog being the same.

For pages like http://www.fdmgroup.com/ca/blog/page/8/, you could look at different pagination options to ensure you receive benefit from these pages. Option 1 in that post (noindex, follow) is one of the simpler ways of doing this, but have a look at the others too.

As others have said above, combating scrapers is a big reason, but you're relying on lazy scrapers not removing the tags.

Another reason is to prevent rogue content management systems from attaching unnecessary query strings to URLs, creating pages that can end up loading infinite times under different URLs. A canonical tag in the source file would mean that any number of duplicate pages point back to the original. The same reason goes for sites that have issues with redirecting www / non-www URLs to the correct version, or who deal with the same problem regarding secure / insecure URLs. In all these cases, the canonical tag is a bandaid, not a cure - it would be better to fix the underlying problem of the rogue CMS, incorrect redirection, etc. but the canonical tag (self-referencing) is there if you need it.

Google doesn't seem to have an issue with it, but if you're at all concerned about the other engines, use with care.

Hi Eduardo,

If you're suffering from a Penguin issue, you may see the site's results improve over the next few days or weeks. Google confirmed last night that they are rolling out Penguin 4.0, starting yesterday. Most people have seen more movement in US results so far, but there have certainly been some international changes, and it's no longer 2011 when non-US results took a very long time to update after Google first released Panda.

If nothing changes for the site in the next few weeks, either Penguin is not the issue or you have not done enough to remove the bad links. Have you tried removing them, not just submitting a disavow report? Removal is recommended; if removal turns out to be impossible, that's when you need to use the disavow tool.

The site's issues could stem from on-site, meaning that you'll need to look through the ideas presented here and conduct a full on-site audit.

"I still want the category pages to distribute their link authority to the individual pages, then I should leave the category pages in the site map and meta noindex them, rather than robots.txt them. Is that correct?"

This will achieve the goal, yes. You would ideally include noindex, follow (as opposed to nofollow) in the meta tag of the page you want to exclude. This means that Google crawls the page in full and allows PageRank to flow from that page to the pages it links to, but doesn't include any of the page's content or its URL in the index.

If you exclude the page via robots.txt, Google never crawls the page at all. You sometimes see URLs whose pages have been excluded via robots.txt showing up in Google's index, because robots.txt doesn't say "don't index this URL"; it simply says "don't crawl it." That's also why excluding a page in robots.txt and putting a noindex meta tag on the page would be redundant - Google would never see the noindex tag because it would never crawl the page.

Hi Christopher,

Google has definitely become a lot better in recent years at identifying this sort of duplication and dealing with it (largely because this has to be one of the most common accidental / non-malicious duplication causes - categories on blogs and news sites). That said, cleaning it up is for the best. I have been meaning to clean up a blog belonging to a family member in this manner for months (years...) because the version of each piece of content Google has chosen is wrong. Pages marked up with dates, e.g. example.com/2014/01 kept ranking better than the original posts in that date range. So even when Google makes the decision for you, it won't necessarily make the right one. You're risking visitors coming to a page they didn't expect, or a page that doesn't answer their query as succinctly as the "best" version would have, and if you are in e-commerce of any sort or focusing on conversions, this can make a big difference to how optimised your traffic's on-site experience is.

Where you'll be "penalised" for duplicate content, especially by Panda, is as you cite above: when the duplication looks like it has been done for spam purposes. This has happened accidentally to people when their content management systems have gone mad with infinite duplication, but it likely won't happen with simple blog categories.

In short, Google sees this sort of duplication all day, every day and will choose its favourite version to rank. However, if you can guide its choice, you're in control of what your visitors see.

You mention country-based categories in your original question. If internationalisation and duplicate content are a concern, you might want to check out the href lang tag (also called rel="alternate" tag - it gets called either by the community). Could be useful if you're publishing the same thing in different countries.

Hi there,

While looking at metrics like this is important, you're approaching modern link development in a slightly misguided, old-fashioned way. This is the way we used to link about link dev three or four years ago.

For one, there is little use finding an older article (anything published more than a few days or weeks ago) and trying to get a link on it. This was a very popular, effective tactic up until about 2010. However, figuring out if a link was included as an original citation or added after the fact is incredibly easy. Thus, it's easy to discount links-added-later, and you can be fairly sure that Google is suspicious of these links given how popular this link development tactic was pre-2010.

If a site has good Moz metics (DA, etc.) and no or very low toolbar PR, this can be an indication that Google finds something wrong with the site... but toolbar PR is a relatively outdated metric now, so take its figures fairly lightly. If you can potentially get a link from a site, look at all of these metrics and then look at both the backlink profile of the potential linking site and which other sites it links out to. Does it look like it sells links to unrelated, spammy websites / industries? Don't go for a link with them. Do they advertise a text-link service where anyone can buy a link on their site? Definitely stay away.

If you come across a less-popular site in your niche with original, quality content, don't turn a link down on that site because it has low metrics.

"2. If a site is a PR2 DA 30/ PA 32 with 14 root domains, 250 total links.... Would a link like this give me any benefit or should I skip links like this? Why?"

To answer this, you're going to need to look further than these metrics. Those numbers are good enough that if the site is otherwise a good match for yours, the link could be worthwhile.

Brilliant, glad to hear it!

Honestly, doing site-wide redirects can be very simple if done correctly in .htaccess, avoiding anyone (bot or browser) encountering 404s. Redirecting properly avoids a 404 bloat in Webmaster Tools, allowing the webmaster to identify "real" 404s and fix them as well. I'd go with redirecting anything you plan to move.

As Mike has said, the fact that you are often linked to, and the content is translated, would put this far down my priority list as far as SEO goes, in that it's not going to hurt you (unless the site are extremely poor quality and linking to you a high number of times, which doesn't sound like the case here).

I take it they're just lifting the content, not scraping the HTML, given that the content is translated. If they were taking the HTML, you can place the canonical tag in the source code and make it quite likely that they'll publish their version with a canonical tag, pointing to your site.

From a legal point of view, you probably have good claim to go after this behaviour, which is up to you. This should not hurt your SEO, however.

Hi there,

I would firstly ensure that you are 301ing the 404ing pages to an appropriate location - it's best if the redirect points to a location with "similar" content to what used to exist on the 404ing page, if that makes sense. Blindly redirecting everything to a home page, for instance, isn't technically what a 301 is meant to do (i.e. indicate that the resource has "moved permanently").

"* Implemented endless scrolling to replace previous pagination"

Can you expand on this one at all? This is a bit of a red-flag one for me, as it would be necessary to know how you did this in terms of on-page technology and whether endless scrolling is search-engine friendly. Good pagination can be very useful for proper indexation. A good guide to pagination was written by one of my former colleagues at Ayima here: http://www.ayima.com/seo-knowledge/conquering-pagination-guide.html

Hi Matt,

If you 301 one URL to another, the majority of the metrics previously afforded to the old URL will be transferred to the new URL. Various evidence, some of it quite anecdotal, says that not all the authority passes from the old to the new, but it's certainly the best option you have if you want to change URLs.

If you are moving to a new domain, you can write a rule in your .htaccess file (if using an Apache server) or in your control panel (if using an IIS server) that transfers every URL to its new location. Even if URLs have no external links, this will simply direct any traffic to the old URLs to their new locations.

Hi Berner,

"So when you are looking at your sites most popular keywords, do you make the most popular keyword your homepage?"

Technically yes, but be smart about it. One reason why the puppies query is so much more popular is not just because people want information about the puppies or want to adopt them. You're literally getting searches in there from people who just want to see cute pictures and videos of Bernese Mountain Dogs (because who wouldn't?!). If you are interested in this traffic, go for it. If you are more focused on people who may convert into customers (whatever that means for your particular website), pick home page title tags and headings that better reflect the home page's goals.

I am not sure what you mean in this instance by "more link juice". "Link juice" is a colloquial term for PageRank, or the authority that flows into and around a website from inbound links.

"When you have keywords that have parts of other keywords, is it safe to use the longer keyword because it has parts of the other phrases?"

Yes - this is pretty common. If I have a site where I sell accessories for swimming, I can have a page about "Speedo competitive racing suits" that covers the term "racing suits" and likely also "competitive suits" and to a degree, competitive swim suit terms related to the Speedo brand. I will have mentioned "swim" or "swimming" on the page, so Google is aware of what type of suits I sell.

For question three, bolding and italicising terms is absolutely not necessary anymore. It's questionable when this last worked, but I'm guessing it was before I started working in SEO in 2006 Placing your primary keywords in URLs and as H1 tags, however, is common and expected. It would be odd to have a page about Bernese Mountain Dog grooming that did not use that phrase as its H1 / main heading, and it wouldn't be at all unusual for that page's URL to be example.com/bernese-mountain-dog-grooming. The H2s, if there are any, should be used for sub-headings throughout the page. So to take the grooming example, perhaps there are there H2 tags (you can have more than one) regarding the best brushes, coat clipping it applicable and shampoo.

I hope this makes sense, let me know if you have any issues.

Best,

Jane

Hi there,

I would say that, taking William's point into account, canonicals might work in order to remove any possibility that Google would see the new site as copying the old one. That said, I can't guarantee that they could not either manually or automatically (manually would be much easier) note that the two sites are owned by the same person and that the domain change is a measure taken to avoid a penalty. The truly safest thing to do is to re-write the content and start afresh. The next safest is to remove the content from the old site, force a re-crawl / wait for Google to update its cache of the old site excluding the content, and then re-publish on the new site.

Canonicals will make this process quicker, but I don't believe it can be guaranteed that they won't result in Google making a stronger connection between the two sites, which might not go well. Again, this is only if there are enough similarities for Google to understand that this is not a scraper / scrapee situation but a situation where one entity owns both sites.

I'm sorry not to give a definitive answer.

Hi Stephen,

Great that you've probably found the cause - this will absolutely cause mass de-indexation. I had a client a year ago canonicalise their entire site (two sites, actually) to the home page. All their rankings and indexed pages dropped off over a matter of about six days (we spotted the tag immediately but the fix went into a "queue" - ugh!).

The bad news is that it took them a long time to get properly re-indexed and regain their rankings (I am talking months, not weeks). Having said this, the sites were nearly brand new - they had very few backlinks and were both less than six months old. I do not believe that an older site would have had as much problem regaining rankings, but I can't be sure and I have only seen that situation take place first-hand once.

Hi Stephen,

That's part of the syntax used to transfer a site URL-by-URL - what the entire snippet of code means is that a browser or search engine will come across www.oldblog.com/post123.html and immediately be taken to blog.newdomaincom/post123.html. If you are not using the same URLs on the new blog, you won't need to or be able to use this particular function in your .htaccess file.

You say that "the new blog has a different wordpress theme & categories than the old blog" which makes me think you'll need to redirect category URLs separately at least (I must have missed that line when I read the question at first; apologies).

Once you've redirected everything, you will need to keep the old domain. If you let it lapse or sell it, the redirects will no longer work unfortunately.

Hi there,

Ryan is correct - high numbers of subfolders like this aren't ideal. That takes into account real and virtual subfolders (i.e. "subfolders" that have no content on them but are generated by a CMS versus subfolders that contain landing pages).

Ideally, these would be rewritten to www.example.com/women/denim/product1.html, etc. You will need to check with the developers why the CMS creates these subfolders and what can be done about it.

Google is much better at indexing / ranking URLs like this than it was a few years ago, and it's not exactly a deathblow if you're told that this can't be changed. It's still not ideal though, so check whether it can be simplified.

Hi Ryan,

It's going to be hard to do this without creating duplicates - if they aren't commissioning re-writes of descriptions but just pulling from the database, identical content like this is far from ideal.

One school of thought is that there really isn't any such thing as a "duplicate content penalty" unless you have some huge, gratuitous problem that results in a Panda issue. Google simply chooses the version of the content it favours and drops the other. The local site would still be much more relevant for a query like "fridge sale san antonio".

An example of a big retailer that has a similar(ish) site at the moment is Marks & Spencer Outlet here in the UK (outlet.marksandspencer.com). M&S is probably the most recognisable high street brand in the UK, to give you a perspective on size.

Looking at what they're doing, they're listing pages like this: http://outlet.marksandspencer.com/Limited-Edition-Jacquard-Textured-T69-1604J-S/dp/B00IIP7GY2?field_availability=-1&field_browse=1698309031&id=Limited+Edition+Jacquard+Textured+T69-1604J-S&ie=UTF8&refinementHistory=subjectbin%2Csize_name%2Ccolor_map%2Cbrandtextbin%2Cprice&searchNodeID=1698309031&searchPage=1&searchRank=-product_site_launch_date&searchSize=12

This is the same product as this: http://www.marksandspencer.com/jacquard-textured-coat-with-wool/p/p60056127. I love it that the "outlet" version is more expensive... anyway...

The product details, which are all included in the HTML of the main site, are not included in the Outlet page. The Outlet URL is indexed (what queries it ranks for / could potentially rank for are unknown) - but I would be keen to hypothesise / experiment with the idea that if that product was on a page about it only being available at M&S Moorgate, and looking for coats at M&S Moorgate was as popular a query as [fridge sale location], the Outlet page would rank.

You will never get an SEO to say that you should "copy and paste" descriptions across domains or within them, but essentially the pages have to provide a service / information that makes them worth ranking for relevant queries.

Hi there,

Unfortunately the site seems to be down right now, so I can't have a look (I'm getting a 500 unavailable error), but I would say that while copy is important to give Google an idea about the page's topic, copy for copy's sake is not necessary. In other words, as long as the page says enough to get the message across about what it's meant to convey, you don't need to add another 200 words for no reason.

That said, each page should contain a good paragraph or so of unique text in most cases. I don't love putting minimums on the number of words that should be used, but keep in mind that text presented as images, etc. should use a technique like CSS image replacement in order to show that text in a search engine readable format, i.e. HTML.

I would not be confident that this has caused your 30% drop in traffic unless the pages are now extremely void of unique content and could be seem as "duplicates" or near-dupes of each other as a result, or simply as not very useful. On the other hand, if there are SEO errors with the site already, a big drop in content will certainly not help.

Get that 500 error checked out!

Really good point. Taking that into account, I might guess that an anti-manipulation method Google might employ is to grab registration details, hosting data, analytics codes, etc. and other identifying factors to determine whether the canonicalised content is owned by the same person. That is, canonicals between tightly-linked sites where the "duplicate" is penalised could hurt the canonical source, stopping people using this in place of the old 301 trick. If the scraper site has nothing in common with the source, Google does not pass on any negative metric from the duplicate.

This is just a theory too of course! I'd be confident assuming that they're taking precautions to stop this becoming a common trick. Awesome point!

How are the indexed numbers looking in WMT today? I see 3,370 results for a site: search on the domain, but those can be iffy in terms of up to date accuracy: https://www.google.co.uk/search?q=site%3Agoautohub.com&oq=site%3Agoautohub.com&aqs=chrome..69i57j69i58.798j0j4&sourceid=chrome&es_sm=119&ie=UTF-8

Not that this should matter too much if you are submitting a sitemap through WMT but your robots.txt file specifies sitemap.xml. There is a duplciate sitemap on that URL (http://goautohub.com/sitemap.xml) - are you using sitemap.php, which you mention here, in WMT? .php can be used for sitemaps, but I would update the robots.txt file to reflect the correct URL - http://i.imgur.com/uSB1P1g.png, whichever is meant to be right. I am not aware of problems with having duplicate sitemaps, as long as they are identical, but I'd use just one if it were me.

As Marie says, doing this isn't likely to throw a penalty your way due to the way Google handles duplication (and you're unlikely to be doing this on page after page after page of your website). You always want to link to the source of the quote, with the (very likely) assumption that the source is not linking back to you (and if they are via a trackback "comment", that link is probably nofollowed).

If you're quoting something like the author is in this example, you've taken quite a hefty chunk of text from the source, but you're providing a LOT of original commentary, linking back and providing a blockquote tag (which is not "necessary" but probably sends a good signal).

I like what David and Samuel have to say here. There is and always will be room for press releases when it comes to spreading PR and information. Businesses will continue to put releases out there and often link back to themselves, but they should be sure to link with brand terms, generic words ("find out more", etc.),  or URLs (similarly to the way people generally build links now). You wouldn't put out a press release linking back to yourself with "car insurance" and spread it to 150 different sources now, if you knew what was good for you.

News stories and releases are always going to get picked up and spread, but what Google is looking for when it comes to actually hurting sites with links like this is a lack of a natural pattern. Does the site receive next to no media attention, but suddenly has 500 links from an identical piece of copy, whilst also receiving no new social media attention, no additional coverage (e.g. no one has taken the press release and written their own story about it nor conducted an interview with a company representative)? The pattern there is unnatural and warrants further investigation.

Is the company regularly being cited, mentioned and written about? Does it put out a release about a real new product or development and have that release picked up by real news sources, some of whom put their own thoughts on the web about the company development? This is natural-looking.

I hope this makes some sense. Essentially the goal is to spread information in the way you would if Google was not an issue, with the resulting coverage being beneficial to your SEO efforts nonetheless. Putting out press releases about nothing and expecting links back from newswires, etc. isn't a brilliant idea but using press releases for PR can be very beneficial for SEO when done properly.

It really depends on what you define as blackhat. On-page trickery (cloaking, redirects for search engines bots, etc.) can be discovered by browsing as a search bot, digging into code, viewing caches, etc. Danny Sullivan and Rand uncovered a large amount of cloaked (and stolen) content on stage at SMX Sydney a few years ago. It was quite entertaining at the time

Some people are basic enough to use tactics like hidden, white-on-white text, as Martijn says. I'm yet to see that tactic actually working post-2004 though

If it's links they're using, the easiest way is to use a tool like Open Site Explorer, Ahrefs or similar to check the links out. Sneaky people can block the OSE / Ahrefs / MajesticSEO bots from crawling the sources of their backhat links if they have access to the linking sites. You can block the bots either in robots.txt or by rejecting the visits to stop the bots from noting that the links exist. That way, the backlink analysis tools will never see that blackhatsite.com links to rankingsite.com, and so forth. It takes a big network that the spammer controls to block link research tools' bots' access to every link you build, however, so this isn't too common.

Whether all big brands / well ranked sites are using blackhat tactics pretty much depends on your definition of blackhat, but it's certainly true that it is very hard if not impossible to rank top 3 for competitive terms (car insurance, poker, credit cards) without parting with money that results in links being built. This doesn't mean that they're all buying links, but they're definitely investing in marketing that results in links, and the whitest of the whitehats will say that this is technically not organic, natural link development. It is, however, what we do - marketing.

The idea of the "sandbox" is an old-fashioned way of looking at a site that doesn't have enough authority to rank well yet. In the past, Google employed different ways of making sure brand new sites were worth ranking before they ranked them, but those methods have changed and the term doesn't apply anymore.

Is it possible that the site simply has too few backlinks to be competitive yet? It's often easier to rank well in Bing than it is in Google (although it's surprising you mention that it is not ranking in Yahoo as Bing powers Yahoo's search engine now).

As Marie and Gary have said, there are a lot of legitimate reasons why a site would acquire a high number of new inbound links naturally. Even if those sorts of links prompted Google to manually have a look at the backlink profile, this sort of activity isn't going to hurt.

What Google is looking for with backlink acquisition is unnatural patterns, so if you were to add 1,000 links to a backlink profile of (say) 1,500 links, with no discernible media attention, content creation or seasonality, you'd be breaking a pattern of link acquisition (or the lack thereof) and possibly trigger attention. Google is incredibly good at pattern recognition, for obvious reasons. If a pattern changes, there has to be a reason for the change. If they determine that the change was manipulative an unnatural, that's where you run into trouble.

Echoing what others have said about link building using directories - I can't endorse that as a link building tactic at all. Get listed where appropriate for reasons other than SEO (are you a local hairdresser in Everett, WA? Getting listed on a site about service providers in Everett is absolutely fine). The sort of directory-based link building that was common and fairly useful for "volume" in ~2009 is no longer a good idea. You'll show up on sites like these without effort (people seeking to fill out directories they run), and shouldn't worry too much about those sorts of links, but it's not advisable to seek them yourself.

The idea of using canonicals interests me, but I am not 100% sure it is risk-free. It used to be the case that you could 301 penalised websites and remove the penalty (we're talking 2010 and earlier here). Google is very keen on transferring penalties these days, so I would be surprised if they are leaving a loophole for canonical tags open like this, or if they will keep that loophole open for long.

You would ideally leave the site live and remove its content as William says - once you see that the cached version of the site no longer contains the content you want to move, you can feel free to take the old site down and put the content up on the new site.

We don't know what lengths Google is going to or will go to to avoid people being able to re-use previously penalised content (including good content from penalised websites) but the safest thing you can do whilst using this old content right now is ensure the old content has been deindexed before putting it up again elsewhere.

The actual safest thing you can do is re-write the content, but I realise this might not be possible.

Check out the Moz recommended list - they're mostly SEO companies but some will offer copywriting / content services, or will have partners they can recommend.

Hi,

Marie is right - we can't provide you with numbers like this. The last time that this was possible was 2009 / 2010 (in my experience). You knew that if a well-ranked website had 1,000 inbound links and 50% of those had commercial anchor text, 5% of which was from a certain type of website, then those were good numbers. Not necessarily to copy, but successful numbers nonetheless and a framework for what was working at the time.

Now, as Marie says, Google is far better at determining what is real and what isn't, so even if you stick to someone else's numbers in terms of where they get their links from, your attempt at copying that may not have the same results because it does not look nearly as natural or high-quality.

I also have to point out that you have mentioned directories and social media: social media links are almost invariably nofollowed, so if you are acquiring them, you should be thinking about these links for other reasons besides SEO. A natural, high quality backlink profile will include social links, so don't forget about them, but they are not technically a source of inbound followed links anymore (for the most part).

Directories are often low-quality / spammy and often require payment for inclusion. I would only have a site listed in a directory if I was doing it for other reasons than SEO nowadays too.

Hi Simon,

Like Andy says, is there a possibility that you can have these windows open in a new tab with a nofollow attribute if necessary? Whether to keep people on site or send them away has been a problem / debate in marketing for a long time (well before we had to worry about penalisation / outbound link issues to go with it). There's a fine line between keeping people on a site they no longer want to be on and accidentally allowing them to forget about where they were before.

You might want to test different ways of sending visitors away to see which results in the best conversion metrics (whatever you choose those metrics to be). As a user, I almost invariably open links in new tabs myself, but different people have different ways of handling this themselves and I'd test what works best for your user base in terms of visitor retention / other goals.

Hi Stephen,

Checking in to see if you had checked the points William has raised above. Do you see anything that could have resulted in the drop? Also, are you comfortable sharing the site here? We might be able to have a look too (feel free to PM if you are not comfortable sharing publicly).

Cheers,

Jane

Hi Nick,

There still could be ways to harness the client base - you have to be careful how you go about doing things like asking for links, and while tactics like this one are now a bit outdated, looking at innovative ways to get links from the client base is still worthwhile. What the client actually does will depend on the resources of the client base and how they can inspire some sort of action from them, but it could still be doable. As with most aspects of SEO however, it's a bit more difficult than it was back in 2008 / 9!

Hi Rohit,

Just to follow up and add to what others have said: if you built links from third-party websites, pointing to your site, Google should crawl the third-party websites and see the links pointing to you within 15 days. If Google is not crawling the site more regularly than once every 15 days, the websites are probably lower in quality and not great sources of backlinks to begin with.

If you are using Moz / Ahrefs, etc. are slower in crawling and indexing than Google. Moz often updates once every 30 days, so could pick up these news links in its next crawl. Ahrefs tends to be slightly faster, but is still not instantaneous in picking up new links.

"Would I get more of a boost in rankings to pages of my site by placing links in the text of my articles on my site to other pages on my site?"

As Spencer says, you're talking about internal linking with generally comes with a slightly different set of rules to inbound / outbound links to and from third-party websites.

Feel free to link to relevant parts of your site from articles, but do so where and when it appears natural, and don't overdo it. That is, don't link to a product every time you use the product's name or mention something to do with the product's function. Imagine if Moz linked to its blog home page from every page on the website that mentioned a blog in text. It would look manipulative and spammy.

Hi there,

The points others raise here are very good. Look also at the total number of links coming into the website. Links are still very important for rankings - Google may be trying to steer away from having inbound links influence rankings as much as they have in the past, but they are still a big part of the ranking algorithm and it's difficult to rank well without them.

Open Site Explorer only shows the site receiving links from 13 separate domains. Cross-check this with your competitors - how many do they have? And as David mentions, do these links point to the pages you want to have rank?

Hi Vince,

Being publicly outed for selling links (or buying them) can result in harsh sanctions that seem to go above and beyond fair, although every case of this happening is anecdotal. Google has been known to suspend or ban people's entire networks for publicly flouting their guidelines. They're certainly not above making an example, and being publicly noticed on Twitter is a tough way to go.

If all outbound links that could be considered commercial have been nofollowed, I see no reason why you need to worry about outbound linking anymore. Andy may have a good point about ads on the page, but the ads I see are fairly nonintrusive. You may want to try serving them as banner ads with JavaScript, like the "links" created by ad networks (was this how you did it before, when it wasn't good enough for Google?).

Hi Adam,

Not to point out something that is likely well taken-care of, but did the GA / Analytics code populate across the site?

Also, is there any heavy JavaScript on the site, especially above analytics code, that might prevent analytics code from loading properly. We had this happen with a client a few years ago. We built custom analytics for this client (they did not want to run GA). Client placed our code in the footer. Client placed slow-loading CRO code in the header. CRO code took so long to load that people had often clicked away from the page they landed on before our code had had a chance to record their visit, as JavaScript generally loads in the same order as it's placed on the page. We had them move our little piece of code up to the top of the page. Problem was solved (in the mean time, we were recording a 20,000 visit loss each week!).

I'm just wondering if this is a tracking issue since all search traffic, not just Google has been affected. It would be quite rare to find an issue that has the same effect at the same time to both Bing and Google's algos. They're similar, but they're not identical and Bing generally tends to take longer to respond to change than Google as well.

Any chance you have raw server logs to compare analytics stats to?

As someone who used to work at a company on the Recommended list (and who was in charge of the Contact form) - we did get leads. They were about 80% appropriate for the business model, although a lot were too small or wanting short-term projects (my agency generally took on corporate-level accounts for long periods of time). I would say that this sort of linking is different to just linking to clients, e.g. if we then went on to link to an insurance company that came to us through a Moz Recommended referral. But if your SEO company partners with a PPC agency or design firm, provides some services to the other and sends out a relevant link, that seems a little more relevant.

On the other hand, I hate these discussions about what we're "allowed" to link to

I hope not, for your sake! 13 hours later - do you see any new downturn?

"But if they are already in the html, would that be considered cloaking?"

There are times when presenting something in HTML that is otherwise not visible, but having other features on the page that allow people to read / click / access that content in other ways, is fine. Linking is a tricky because links are so inherently valuable in terms of SEO. You don't can't really be too careful.

I'd be wary of presenting links (a subset or full set) in HTML if there is a form process to actually arrive at the links' targets. Essentially you'll be linking to products X, Y and Z on a page, for search engines but requiring a specific input from a user to see X, Y or Z - an input that only very few overall visitors are actually likely to make. I would say this qualifies as showing different content for SEO's sake and not providing a UX alternative that is pretty much the same thing. Others may disagree with me on that - I'm being wary here

I would very much like to see the HTML if you are still active in this thread when it is produced, but you may be left with a situation where the pages need to be linked to elsewhere throughout the site to ensure they are crawled.

Always good to get a blog post or case study out of things like this, at least! I also seriously hope (and suspect) that you will take less time to recover than my client did - the fact that their sites were less than six months old when they made this error will certainly not have helped their chances.

Cheers,

Jane