Welcome to the Q&A Forum

Hi there,

Just to round off this thread, I would say that Robert is right - often low rankings aren't about what you are doing "wrong" but are about what your competitors have done right that you have not considered. If your SEO practices are out of date (I am guessing here, but things like seeking links from directories, focusing on the minutiae of markup rather than robust site structure and quality, unique content, pushing out small daily "blog" updates that are repetitive rather than writing two or three longer, quality pieces per week, etc.) then your rankings will suffer if your competitors are more up to date.

If you are still seeking links from third party sites with solely "optimised" anchor text (e.g. links saying "Car insurance" pointing to the car insurance page) you can end up with a penalty or poorer rankings as a result, usually at the hands of the Penguin algorithm or a manual action (although if you are still ranking 8th, you do not have a penalty.

Hi there,

Most definitely seconded to what Casey said. Linkscape doesn't filter based on spam, so the authority you see passed here is skewed in terms of real world effectiveness in Google. Squidoo really is one of the most awful dens of spam online. Interestingly though, I was just taking a look at the full list spreadsheet of sites hit by the Farmer update (according to Sistrix's data) and Squidoo isn't mentioned.

But it really is awful. Stay well away

Hi Remko,

If the entire site sits on HTTPS, definitely ensure that you have a valid certificate.

As far as it being necessary to have an entire site sitting on HTTPS, opinion on this has changed over the years. It previously was not a great idea; it's now much more common and many well-ranked websites use HTTPS as standard. I would not change your entire site including pages that do not need to be secure based solely on this session at SMX, especially as it was indicated that there's opposition to making this a ranking factor. I'd wait for more evidence before I did something like that. However, if making a site load solely on HTTPS URLs becomes absolutely necessary, then I'd be more comfortable doing so now than ever before.

Hi Peter,

I understand that the platform only allows for subdomains. From a purely SEO point of view, subfolders or pages are preferable to subdomains because authority does not appear to pass between a parent domain and its subdomains in the same way as it does between subfolders and parent domains. If your landing page sites are only one pagers, they may be seen as quite thin as well.

However, there is no reason why you can't build quality content like this - it just may take more link building to establish the authority for the subdomains than it would for pages on the same site. You will need to ensure that as much unique content as possible is placed on the landing pages to increase their 'worth' in Google's eyes, given that they are separate from each other on subdomains.

Hi Asif,

The numbers you see in WMT will be fairly accurate: the numbers for site: searches are rarely accurate, so I wouldn't pay them too much attention. You can usually trust what you see in WMT though.

Cheers,

Jane

Hi there,

Which structure is best depends on the structure of the website from an information architecture point of view more than anything else. For example, do you have a main California page, to which you link to a San Francisco page, on which all the services in SF are listed? Or do you have a landing page for professions such as plumbing, to which you link to different areas, i.e. California and San Francisco?

If California is the parent and SF / plumbing are under that, then /ca/sfo/plumbing would make sense. And vice versa.

I would say, however, that you should only go through the process of 301 redirecting existing URLs if you feel that the site is at a clear disadvantage because of the current URL structure. You can lose a little authority when you 301 one URL to another, and rankings can suffer / fluctuate as a result. Most go back to roughly where they were before, but this is not guaranteed. We generally advise avoiding 301s / URL changes unless it is undeniably necessary.

Cheers,

Jane

Hi again,

Sorry for the misunderstanding.

The links being from identical content should not be harmful to you - this sort of citation is fairly common, e.g. when a site has shared a piece of viral content like a video, infographic, tool (calculator etc.) or cited case study that others reproduce and link back to. I'd be far more concerned about the quality of the linking sites.

I think the duplication question will also be even less of an issue if you remove the poor-quality links: identical citations from high-quality sources will pass as normal and not malicious liking activity.

Hi there,

Unfortunately, Blogger is very outdated when it comes to professional blogging. It lacks a lot of the independence and functionality. It is bad to the extent that you cannot 301 redirect a blog on a URL like myblog.blogspot.com if you decide to move over to a domain of your won. If you client is just using the Blogger software within their own domain already, I would seriously consider having them move to Wordpress. Google shows no sign of putting the sort of effort into Blogger that would elevate it to the level of Wordpress in terms of what it can do, and as you undoubtedly know, Wordpress is virtually limitless in what you can do with it.

Hi Peter,

In some ways, subdirectories seems even more sensible when you're dealing with single landing pages, as they'll work together somewhat to look like a fuller site from Google's perspective, rather than just a collection of subdomains happening to exist on the same domain.

Hi Conrad,

Sorry for the late reply on this. I hope it is still helpful.

In general, this is always what we recommend, unless the ccTLDs (or whatever is being redirected) has a substantial history that makes it so strong that it can survive well on its own. If you're languishing in the SERPs or need a boost, this can certainly provide it. Keep in mind that you don't take 100% of the link equity from one site to another during a 301, but you take a large portion, and the added strength from the domain should more than make up for this.

There is absolutely no guarantee that you will get positive results, or won't get negative results - the most likely scenario is that you'll see a dip in rankings and traffic as Google acknowledges the redirects, but that original rankings will return within a matter or days or weeks. After that, your ascent through the SERPs should be faster than when the content was on a ccTLD as opposed to a subfolder.

Where I would not do this, again, is where substantial effort has been put into the ccTLDs and they are climbing steadily on their own. This, to my mind, is always a move to be made when other options do not appear to be making a lot of difference.

Hi,

Is your target market international or targeting one particular country? I would only recommend a .fr domain if you are targeting a French market: if you're looking at the UK, see if you can get a .co.uk domain. For the United States or an international market, go with .com. It is harder to get a geo-targeted domain, i.e. one on a .fr, to rank in a country that is _not _France with a language that is not English. You will see .es domains ranking with Spanish language content in South America, but you rarely see situations like the one you're proposing.

From the UK, I see some .fr domains ranking in local search for this term, but no sites ranking in "regular" results: http://i.imgur.com/ZhKTxk9.png

Hi there,

You have two good answers here already from Andy and Dennis. If possible, could you provide examples of the site and the terms for which the internal pages are ranking? Feel free to PM them to me if you would rather they not appear on the thread. This should help me look at whether the internal pages are just more relevant to the queries, or whether it's truly strange that the home page doesn't rank.

Cheers,

Jane

Hi Anthony,

It looks like a simple error on Google's part, especially since your other pages are displayed correctly, but do try the actions listed by Chris like Fetch as Googlebot, perhaps after also making some minor changes to the title tag to spur a new title to be indexed (nothing drastic, try "PicoSure Tattoo Removal UK | Fastest Treatment in Manchester, Liverpool, Birmingham" perhaps).

I don't believe this will be the problem in this case, but beware of copying pages' content on other websites like Yell:

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

http://i.imgur.com/e94blAB.png

It's a bad idea to place your content on other websites, especially authoritative sites. Google heavily filters (and sometimes penalises for) duplicate content and the last thing you want is Yell or another review / directory site being considered more relevant for your text than you are.

Hi Matthew,

Although the metrics of the two sites (number of backlinks, etc) are quite similar, domain / site age are going to be playing a part here. The years of indexation are going to put the .co.uk site at a big advantage over the .com. Although exact match domains are not a guarantee of ranking, you do also have an exact-match with the .co.uk, but not with the .com (with "usa" added). Do you have plans to acquire http://www.blindbolt.com/?

Hi there,

If you 301 to https://www.mysite.com/, the authority of **http://**www.mysite.com/ will be largely attributed to the HTTPS version. That is, all the links (anchor text, etc.) pointing to the non-HTTPS version will help the HTTPS version rank well. There may be some dilution along the way, but you will not lose the benefit of those links simply because the HTTPS version is canonical and a 301 is in place.

You can feel free to specify the https version in WMT as a result. In fact, there have been murmurings that several Google employees favour HTTPS sites, but we are a long way from that being a recommendation. It's just now much more safe to say that using HTTPS is safe (it used to present some ranking / crawling issues). http://searchengineland.com/ranking-benefit-making-site-ssl-yet-googles-cutts-like-make-happen-186810

The next time Google crawls the previously-linking page, it should note that the link is gone, as it will cache the page. Any benefit or detriment that the links was passing on should technically be nullified. Moz's crawl will update when it re-visits the page as well, but the frequency of Moz, Googlebot and the updating WMT is really hard to predict. If the link is on a highly-traffic, well linked-to page, G and Moz should pick it up quickly. If not, it can take a long time.

The Moz index is updated about every 3 - 4 weeks; again, the removal of one individual link requires the latest crawl to have noticed that it has gone, which is likely but not a certainty.

Hi there,

For the question of how long rankings will take to come back, unfortunately that's impossible to answer for sure. I had a client once mistakenly point all of their canonical tags at their home page, effectively de-indexing the whole site. The fixed the error within a day, but it took five weeks to get their rankings back. This was a very new website that had very little authority to begin with, but it was still a huge hit to take.

If your site is a little older and more established, your recover should be a lot quicker. Unfortunately it's still very hard to put a time frame on this.

Hi John,

In the case where you did receive a warning, feel free to aggressively try to remove the links and / or make a disavowal submission. For generic warnings, I don't believe example links are given - I have seen example links provided in follow-up emails to the webspam team though.

If the rest of your profile is, to your knowledge, clean then I would start with these links. It's disappointing on Google's part because they used to be quite adept at ignoring links like this that should be discounted - their spam trigger finger is a lot itchier now than in days past and a lot of legitimate sites are getting caught in the mix.

Hi Kanmin,

This does not mean that anything has been approved, just that Google has received the disavow file.

This links to information on how to submit a reconsideration request. You will hear back from Google after you submit this (within a few weeks, most likely) if your penalty was manual and that penalty has been lifted. If the penalty was algorithmic, you will not hear back - you would only see the site's ranking and traffic improve.

Cheers,

Jane

Hi Paul,

I'm really not a fan of providing entire sites in HTTPS and really recommend only using secure URLs on pages through which sensitive information is passed. As Tom says, ensure that every URL is only allowed to load with either HTTP or HTTPS and that the other version 301 redirects to the canonical version. If this isn't technically possible, ensure that the canonical version of the correct URL (the one you use to link within the site, etc) is referenced in a canonical tag on the page.

Cheers!

Hi Sally,

This could have to do with Google's desire to present the most diverse selection of results for terms, and since there are a range of other websites relevant for the [david lerner] query, Google has dropped the subdomain in favour of keeping the main site in first position, and showing a potentially relevant result for a totally different business in second position (the leggings result). This can happen even when a result has ranked well for years.

To solve this, it would be good if the site displayed "sitelinks" for the brand query (i.e. the six or more links that appear below a branded search, usually linking to other key pages within the website). Obtaining these links usually requires strength and authority that far outweighs other companies with similar names. It's not very common to see subdomains appearing as sitelinks, but it does happen sometimes, when a brand and the subdomain are particularly strong, e.g. https://www.google.co.uk/search?q=ycombinator&oq=ycombinator&aqs=chrome.0.69i59j0l5.1467j0j9&sourceid=chrome&espv=210&es_sm=119&ie=UTF-8 (look at the Hacker News and Y Combinator Posthaven links).

You have a partial duplication issue with both http://news.davidlerner.com/ and http://news.davidlerner.com/news_index.php displaying the snippets for largely the same stories, but this should not be enough to damage the subdomain from a Panda point of view. You may consider restricting bots' access to the /news_index.php page unless it is performing a different task from an SEO perspective for you. Again, I do not believe this has caused the site to drop out of the rankings (and it is still perfectly indexed, ranking for its URL and full name) but it's something to consider tidying up.

Hi there,

If you are using the canonical tag and it is displayed properly on the UTM URLs, pointing to the canonical URL, then these links will pass value if they are followed. There are a range of reasons why someone might use UTM tags - a lot of services tag outbound links with this for tracking purposes. I would definitely try to get links nofollowed if the UTM (or another metric) clearly identifies that they are paid links and could be picked up by Google either manually or algorithmically, but the fact that the link contains the tracking code won't absolutely determine it as paid, and you can still gain authority from these links with correct canonicalisation.

Cheers,

Jane

Hi there,

Ah, I see what you mean.

You can certainly add these sitemap links to the .com's robots.txt file. I don't believe there is syntax to specify language for this though - if you look at large multinational sites like Adobe.com, you can see their multitude of sitemaps below their disallowed files: http://www.adobe.com/robots.txt

Microsoft's is similar: http://www.microsoft.com/robots.txt

I'd say you're safe enough to link like this - the sitemap will be accessed, plus if you have used hreflang tags and geo-targeted the subfolders in Webmaster Tools, Google will understand that it's looking at maps / content for a specific language.

Cheers,

Jane

Hi there,

Have you been able to remove the old urchin code and is analytics showing the traffic to the new domain now?

Cheers,

Jane

Hi Carl,

Dana has a brilliant answer here, but I'll add a couple more things:

Links are still a heavy part of the ranking algorithm, but it's absolutely clear that Google wants to move away from this reliance, simply because link building has been scaled so effectively by so many people that it is not a true measure of impartial authority anymore. I'm sure they would love to rely on social signals as this is people talking about what they like, don't like and do in a candid fashion, but Google doesn't have the ability to infiltrate the most valuable social network for this (Facebook) yet, and they probably never will.

As such, they still need to rely on what is posted on the rest of the web, and citations / "mentions" seems a good place to start. It has long been true that they seem to look at "links" that are not link but simply cited URLs (e.g. if I write out http://www.site.com/, this CMS doesn't link to it but Google crawls this page and sees the URL). This is plainly obvious if you are checking a site's 404 errors through Webmaster Tools - I have seen a lot of "404s" that aren't links: Google reports a broken link in WMT and shows you where the link is placed on the third party site; you visit that site and see that it is not a link, but simply a URL written out like the example here.

Does PageRank pass "through" that "link"? Impossible to tell exactly how Google deals with those from an authority point of view, but they see it. And they would clearly be wrong to ignore data like this, so we can expect that such citations are at least factored into how a site appears to Google.

It's also likely that they take into account mentions of a brand name in a similar way, and links that are nofollowed. It's impossible to say how, but I would certainly like to guess that 100 mentions from authoritative sources (including many newspapers that do not link out, as a rule) are more likely to increase a site's authority than 100 links from questionable or lower-quality sources.

Hi there,

I'll go through the questions one by one:

1. Does Changing the Anchor Text Bad? Does Google have a new way of detecting anchor text changes.

Google will note the changed anchor text when it re-indexes the pages the links sit on. There is no hard evidence that this is viewed as "good" or "bad", although if the links are otherwise "bad" (from spammy sources), changing the anchor text won't help - the link would be better off being removed. However, if the link is neutral or good, there is no evidence that changing the anchor text will hurt you.

2. Can I change the Title/Anchor to something more generic, so that the anchor text profile is more natural.

Definitely - if you are going to change the anchor, it is best to change it to something related to your brand, e.g. the name of the company or URL of the website.

**3. Since this is a Directory is http://obln.org/ - Does anybody know being on this directory bad? **

Free web directories are not advisable places to build links, so despite this conversation about anchor text, I would remove the link in this case. Yes, they have a PR of 5, but it is the best rule for modern SEO to not use free directories as link building tools because they have been abused so badly for spam in the past.

Hi,

CleverPhD has a great response, but I wanted to add specifically that if you do this move, you will not lose the authority of those 13k links / 1,200 linking domains if you do the 301 redirection process properly. The guides linked to in CleverPhD's comment will help.

301 redirects pass the majority of authority from one URL to another, but you may also want to create a list of the top linking websites from that 1,200 and contact them after you have completed the move, asking them to update their links to you. This will ensure that you retain your best links on top of properly redirecting all the others.

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.

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.

Hi again David,

I haven't heard of this particular company - could be good if the 40 penalties stat is true, but get your research in and you should be fine.

Also check out the Recommended Providers list - there could be some good options on there too.

Hi,

Technically this is duplicate content, but it is most likely something search engines understand to ignore (especially if no links point to the upper case / incorrect URLs). If it particularly bothers you, place a self-referencing canonical tag in the page's file so that www.examples.com/pAGe/PROducT references www.examples.com/page/product in its tag. Either that or a 301.

Cheers.

Hi there,

My former agency has a good post on pagination that you might find useful: http://www.ayima.com/seo-knowledge/conquering-pagination-guide.html

You definitely want to cut down on duplicate content as much as possible - let me know if that post does the trick for the ecommerce question!

Cheers

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.

You should definitely be able to ignore these as they're quite common (not specifically these websites, but these types of links). That said, if you are under penalty and can't work out which links are causing it, having cleaned up all the rest, I'd go for a disavowal. If you are 100% certain that the penalty is due to Penguin, approaching the manual webspam team isn't likely to help you at all. But getting that disavowal in before the next Penguin update (which, despite Google's updates of this week - Panda and "payday loans" - can't be far away) might help.

He can certainly buy both domains and that would be a good option (301 one to the other), but he will likely need to serve different stores or dynamically change pricing etc. due to the different currencies and possible shipping / returns differences. It's also worth noting that if he's targeting North America, some of their spelling is different to Australia (US vs. British English).

He can use the href lang tag to show Google that even though both stores are in English, one is English for the US and one is English for Australia, one is Spanish for Spain and one is Spanish for Mexico, etc.

The two responses above are good - we advise not doing redirects unless they are truly necessary due to some people's experience with ranking problems that are hard to fix (i.e. redirection shouldn't be carried out "just because", for aesthetic purposes that aren't important, etc.). However, if the redirection is necessary and it is done properly (check out those resources TrentonM11 linked to) you should be able to complete it with very little problem, as Anirban says.

Hi Denise,

I have to say that the rebrand strategy worries me too. There was a good need to have a domain name like this in 2006 / 7 or so, but it's very unnecessary now and verges on dangerous if it is seen to be a spammy move. It also reflects quite poorly from a usability point of view - if a brand-name regional insurance company like Pemco (http://www.pemco.com/) for example, and you change your domain to www.seattlecarinsurance.com, it doesn't look particularly professional.

I would definitely avoid making this change across social media as well. Is there any other reason besides SEO that this domain change was made in the first place? If not, I strongly suggest changing this back to the original name because there is unlikely to be any SEO benefit in doing this at all.

Hi there,

This is pretty much exactly what the rel="alternate" tag is for. Have you looked into that option? This allows you to use the .co.uk for the UK and the .com elsewhere: https://support.google.com/webmasters/answer/189077?hl=en

You will have trouble ranking a .co.uk in countries besides the UK, although moving everything to a .com in the future might be an option as well. However, it should be perfectly feasible to rank both sites with proper geo-targeting like this.

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.

Hi Edmond,

Sure thing - it's actually very easy. I'll show you here. Log into Webmaster Tools and go to the account you want to geo-target. Click on the "settings" button where the arrow is in this screenshot:

<a>http://i.imgur.com/ziFanER.png</a>

Click "site settings": <a>http://i.imgur.com/j65txrk.png</a>

Choose "Australia" (or wherever) from the drop-down: http://i.imgur.com/6KNNOVb.png

You can only do this with non-country specific domains, so you can't geo-target a .com.au domain to the UK, etc.

You can then add site.com/nz/ as a new project in Webmaster Tools so it shows up in the initial list with all your domains. That subfolder, you target to New Zealand and so forth.

Cheers,

Jane

Hi again,

Sorry it has taken a few days to get back to you. I replied in the other thread about ccTLDs versus using one site. Some additional info: in general, you will have an easier time using the subfolder structure recommended in the other question (again, as long as there are no factors which make it important to have country-specific domains). The Singapore / Indonesian sections of the website will naturally inherit authority because they sit on the strong .com. Just being linked to by the .com isn't enough to give them such a large boost.

Apple uses this strategy for internationalisation: http://www.apple.com/uk/ for the UK, http://www.apple.com/nz/ for New Zealand, http://www.apple.com/sg/ for Singapore and so forth.

On the other hand, BlackBerry uses subdomains: http://uk.blackberry.com/ and http://sg.blackberry.com/.

Amazon obviously uses ccTLDs.

All of these domains are hellishly strong in their own rights; traditionally it has been thought of as best to use one site like Apple does if you are no a mammoth already. However, you can make other options work with good link development. I think in your case, one domain is something to seriously consider.

Hi Denise,

You should be safe to redirect back to the original domain - two months is long enough for Google to have completely indexed the new site and assigned it rankings, but a 301 redirect generally takes care of this and reverses the situation back to its original state. It's not optimal - optimal would have been never doing this in the first place unfortunately - but you do need to complete the redirect back again because the original site is now thought of as the new site, as far as Google is concerned. You need the 301 to say "hey Google, we're not on pmd.com anymore, we're over on brand.com!" just as you did the first time, when you changed the pmd.com two months ago.

Again, it's not optimal because they fewer redirects a domain / site can go through the better, but I'd class this sort of redirect for keywords as damaging and not something you want to keep around.

I hope this helps!

Best,

Jane

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 Eduardo,

Ah - I understand. I don't see this as being a problem - many internal links like this have many thousands of links pointing to them internally and Google understands that this is the nature of internal linking. If you crawl a few other larger sites, including this one, with a service like ScreamingFrog (you'd need the paid version to see the full site crawl), you'll see that resources like this blog are linked to thousands of times, since there are thousands of blog posts from years ago, articles, tool pages, indexed Q&A pages etc.

Cheers,

Jane

Hi Claudio,

I would echo the guys above in saying that it sounds like you could do more to revoke the penalty on the original site. If you begin anew, I would definitely not 301 the old domain (not that it sounds like you were going to), but I'd also invest in completely new content, rather than duplicating the old content. Google's ability to track duplicate content is amazingly good, so even a noindex on the old content could still have G draw a connection between content it penalised in the past and the new site.

Moosa is absolutely correct that it is better (and unfortunately much harder) to remove bad links than it is to just disavow them. Google's spam team often appreciate genuine effort to remove links - disavowal appears to work best if you have been unsuccessful in your link removal and can prove that you got in touch with as many sites as possible (screenshots of emails unanswered or answered unfavourably, for instance).

The other very good thing about removing links is that they can never hurt you again in the future if Google one day decides to change the way it views previous disavowals... which we certainly can't count on it not doing.

Hosting these pages as subfolders instead of subdomains should be fine, and is the "traditional" way of presenting new content with the highest likelihood that the content will inherit authority from its parent domain from an SEO point of view. They should rank just as well as if they were on subdomains or separate domains.

Separate installations and themes should not make a difference; Google will be absolutely clear in its understanding that these sub-brands are related to the main brand and to each other, so assuming you are fine with this, different design and CMS installations will not make a difference.

Hi there,

For the guest blogging point, I would say that the absolute best type of guest blogging achieves two goals:

1. It looks like a guest author, who is an authority of a subject close to the heart of the publishing website, being commissioned to write an article for that website. My father runs a blog about swimming. The website has been online for eleven years. If a swimming coach from a successful program writes a post for him about a new method of training and links either to his / her own website and to a product that helps with the new training method, neither my father's blog, the coach's website nor the product page are ever going to be hurt from doing this. This is a standard practice that happens every day (usually without giving thought to SEO).

How many people sometimes write for big magazines who aren't permanently on staff there? Moz was told that Youmoz was fine. What you need to avoid here are the crappy, "made for low-quality sharing" **posts that are easy to spot.**It can be really obvious posts come with a generic author bio that links to a commercial site, are poorly written, or are totally at odds to the topic of the site. Using the swimming blog example, if he were to publish a piece by a "freelance journalist" about the same new product, link to that product and include a crappy author bio about how "Peter is a freelance journalist who enjoys hiking, traveling and indie music. He is often found working at [COMMERCIALSITE.COM]", then you're in much more dangerous territory.

2. Because the goal is to provide something that is of relatively high quality, both you and Google get what they want: a web with high quality content that provides Google with the signals it needs to rank good content above bad. Essentially, Google has created an environment around guest blogging where you have to do it particularly well in order to succeed. Which is pretty much fair enough. Just a lot more work The harder you work to "hide" the fact that you're guest blogging, the more Google's problem with low quality guest blogs goes away.

So how does this tie together with keyword optimisation? I'd definitely agree that you no longer want to seek links to your car insurance site with the anchor text "car insurance". Some of these are going to pop up; let them appear on their own if they're going to, but don't seek them.

In that regard, on-page targeting is more important than the basic off-page metric of anchor text.

Regarding off page: we have enough reason to believe that Google understands the content of a page that is being linked from and how that relates to the page that is being linked to: a blog post on an 11 year old swimming blog about swimming that links to a swimming accessory with the anchor text "this product here" will still mention most likely ALL of the product's primary keywords. Google indexes the blog post, sees the product's name and keywords about what it does, follows the un-optimised link to the product website and sees those terms (including commercial keywords) again in the product page's title tag and description. Google relied on anchor text in the past, likely because this sort of semantic processing was too hard for a primitive search engine. This requires a much more sophisticated approach to analysing content, but makes it much more likely that the picture Google gets of a website's place on the internet is accurate.