Welcome to the Q&A Forum

Hi there,

In any case I would not expect a lot of traffic to posts that are syndicated. Generally people use syndicated content to supplement what's already there. So if the idea is to get more traffic through these posts, I would suggest finding another way.

Having said that, my understanding is that syndication-source still works, but canonical will override it. If you have something that says they no longer recognize it I have never seen it. I've used syndication-source when I don't know the URL or where a CMS/API doesn't allow me to get the URL, as you can point syndication-source back to the root domain URL.

I've also seen sites just post something like "Originally published on http://example.com/page" without incurring any penalties.

To do it by the book, I'd suggest canonical first, syndication-source second, and a source link third. And in this case I'd suggest doing it by the book because syndicated content is generally not going to give you a lot of organic search benefit. Hopefully that helps.

Penalties from third-party spam are rare, but it happens. Just to make sure we're on the right path, I'd first take some top pages and fetch/render them. Submit 4-5 of the top pages to the index. If everything looks fine in the render and submitting them to the index works, we eliminate 99% of technical issues.

If you're seeing spam links coming in you should definitely start with the disavow. Do you have a manual penalty notice in Webmaster tools? If so, I'd submit a reconsideration request explaining that your client is not doing this link building AFTER the disavow is in.

Beyond keeping the disavow updated and trying to get Google to care about your reconsideration request, there's not a lot you can do. In one case I was able to get an IP address from a legitimate site that had spam comments and trace that back to a competitor's office. In most cases you just have to try keeping up with the garbage.

If it's just an indexation issue you can try the "fetch as google" feature under the crawl menu. You can google the site:URL to see whether this is necessary.

Your PA might have dropped on the home page if you changed the URL. For pages that Moz hasn't been able to look at and compute the default score is 1. All new pages are always 1, and will update when the new index comes out (roughly every 1-2 months). Also make sure you do a 301 redirect if you've changed the URLs.

Sometimes it requires a little more time and effort depending on which keywords you're targeting. Without seeing the site and looking at the links I can't say much more about it, but the first thing I'd look at is your backlink profile. If it looks like a lot of links anyone can get by posting a comment or something, there may be the problem.

Just to add on to what Kevin is saying, a reconsideration request is generally not the way to go unless you've received a manual warning in Search Traffic > Manual Warnings. It's more likely that the site doesn't rank due to limited backlinks and content. Make sure you look at the keywords you want to rank for and ask yourself whether you're providing as much value to the users as those on page 1.

TL;DR: You're right to be skeptical that this is an urgent issue (in my opinion), but it is something worth fixing at some point for several reasons.

I was far more concerned by search results, but I see you've added those to noindex/disallow in robots.txt, which is great. Not many people know that works!

I think it's very possible that Google understands the difference between a classified ad and an editorial content piece. They definitely treat products and content differently. That said, it's generally a good idea to avoid relying on Google's intelligence, as many have been let down by Google's failure to understand.

Duplicate content is generally something SEOs are overly-concerned with. More often than not it triggers a filter - not a "penalty." I don't see it as the most dangerous thing you could be doing by any stretch of the imagination. That said, I've seen several classified sites do the following, which I'd recommend as a "best practice" approach. At one time Craigslist did this, and may still be doing it.

• Accept non-spam ads with a pending status
• Check against listings in a given period of time for duplicates. This happens even if the ad is changed slightly, so there's some kind of semantic+image analysis going on.
• If a duplicate is found under the same user name, inform them that they've already posted the ad. From here the rules are up to you. Many sites say the ad can't be posted again for 7 days (if the old ad is deleted) or 30 days (if not). They then encourage users to buy a featured listing that shows up higher than others.
• If duplicates are found under different user names, give a warning that it's against your terms of service (make sure it is) to post duplicate ads from multiple accounts, that accounts can be banned, and have them certify the post is not the same.

You don't need to follow this exactly, but it's here to give you some ideas on having your users prevent duplicate content for you. Given the general positive architecture I've seen on the site it looks like you know what to do with the site better than I would.

Now I don't think 250 out of 10k is bad. Having consulted with a few local classified sites that's actually quite low. But I do think there's something to be gained by detecting duplicates to prevent users from gaining an unfair advantage over those playing by the rules. And if you sell featured listings this is an excellent way to help those who are most desparate to sell while increasing revenue.

I hope that helps.

Obligatory disclaimer: This is merely free advice for your consideration, and not the Moz official stance. The consequences of any changes you do or don't make are ultimately your responsibility.

I'm trying to understand WHY they're using hidden H1s at all? I see no good reason to put H1s in a hidden class if it's not displayed on the page. Is there some sort of functionality mapped to the H1s? If so, it should be done some other way if you don't actually want it to show. I don't have the full context, but it really sounds like a bad attempt to get keywords on the page. It won't work, and it could do some harm.

Looking into the code it seems like there's an excessive amount of markup. Each schema type has various rules about what it should apply to, so it's not hard to violate one of the rules if you're not consciously applying each rule. Just a quick look and I see 13 of these:

itemscope="itemscope"

Not sure what that's supposed to be doing, but I don't think it's correct syntax. Unfortunately looking deeply into a page with so much markup (101 itemprops!) to fix every possible violation is WAY beyond the scope of a question that I can answer here. Generally I'd suggest only using markup where you really want the snippet to appear rather than tagging everything. There's just too much to go wrong.

Beyond that, use Google's structured data testing tool and the structured data section in Webmaster Tools to make sure what you have is working. There are probably some Magento plugins that could handle this for you. Really I think all you need to tag are the product and review-centered markup.

Sorry that didn't help as much as you'd like, but there are dozens of potential minor technical problems to fix. There's enough markup that I can see why Google thinks its spam. In this case, less is probably more. I'd consolidate way down to review markup, make sure it's accurate, and only expand markup carefully once that's working.

Honestly, Google is pretty smart about these spam links. They see hacked sites all the time, and chances are good that those spam links are not related to the content you're actually building the site around. That means the real risk is likely very low. That's especially true if the spam pages on your site created by the hacking are now 404ing. It's clear you're not trying to manipulate rankings.

You can feel free to add spam links to the disavow pre-emptively.

https://searchenginewatch.com/sew/news/2309486/matt-cutts-use-the-link-disavow-tool-even-if-your-site-hasnt-been-penalized

Just make sure you only blacklist a site if the entire site is pure spam. Otherwise use the page disavow feature. I'd say you may as well do this, but it wouldn't be my top priority. The exception to this is if you were hacked and the spam links are relevant.

Example: If the hacker created doorway pages about viagra, and your site is about sports memorabilia, I'd not be too concerned. Just make sure all the pages about viagra are dead. You can still submit a disavow. I probably would just to be on the safe side, but only once I got around to it.

If the hacker created doorway pages about sports jerseys, and your site is about sports memorabilia, I'd be concerned. This could look like an attempt to manipulate Google's rankings, so I'd step on the disavow. And again, make sure any pages created by the spammers are 404ing or 410ing (even better).

Generally I'd not worry too much about duplicate warnings like this, especially if the duplicate pages are being created by parameters. Though Moz won't be able to see it, check Google's search console (AKA webmaster tools) under parameters to see how many portfolio IDs are being crawled. You can even set the parameter manually, though I find "Let Googlebot decide" usually works just fine.

The canonical tag is a good idea. Definitely use a canonical tag that to point back to the original version of the page. Without seeing the site I can't give you more specific info than that.

I'd probably not use robots.txt, just because I'd worry your content would be harder to crawl once a session is tagged with profile ID. It all depends on where users/bots pick up the profile ID variable from.

Very hard to respond to this, but I'll do my best. First, there's no real point in starting a new site if it's going to be thin content all over again. If you just post videos from other sites with some text, you're just a re-aggregator of content. I'd think about what you can do to make your site actually more useful. Think about things like this:

• Are there search,filter, rating, or navigational features your competitors are using that your site lacks?
• Are your titles accurate and specific, or vague and targeted towards similar terms? You want to be accurate and specific. You don't want to mislead anyone with promised value only to disappoint them.
• How can you add useful content at scale?

I'd generally lean towards using the existing site, but it can take a while for Google to reconsider it. It's your call, but the important thing is to make sure your pages aren't thin and you're delivering value.

Given that you're using an iframe on the link you gave, Google will just see the second domain as a type of scraper and probably disregard it. If the "best" domain is your primary domain it's probably not a big deal. The iframed site will just never rank for anything. It's not ideal, though, because any links on the second site will not help you.

Out of curiousity - why are you doing this? It's at least a strange experience for the user to see the site with a title and logo that doesn't match the domain. If it's for tracking, there has to be a better way. If it's for domain name, I think you can create a better brand and get that domain. This is partly unsolicited advice, but "bestup" doesn't really make sense to most people. You'd probably get a lot further with something like "John D. Makeup" and the matching domain (in my example johnpmakeup) won't require these strange concatenations.

I agree with the responses here: having two sites is generally not the way forward. I just want to add one other option. Most people who received link-based penalties were building links to 2 or 3 pages. If that's the case you could 301 redirect most of the site on a page level and let the pages with bad links 404.

If you have cleared the penalty it might not even be an issue to 301 redirect the whole domain. Just make sure that you update your disavow on the new domain.

Thanks for the update, and good luck!

So there were no external links pointing to the subdomain admin.site.com ? If that's the case you could probably just noindex/nofollow the thing or let it 404. You could write an .htaccess rule to rewrite the domain name, but it's actually probably not worth it now that I think about it. The exception, of course, is if the subdomain had external links pointed to it.

It's hard to say without the URL or the keyword. What I can say is that Google's always testing new ranking orders and seeing which one users like more. We believe they measure that "like" by how many people choose a different result or try a new search vs. the other results. If you legitimately think your result is best, it might bounce back after a test period.

Long story short, make sure you're looking at the usefulness of your page and how well it solved the implicit question in the keyword. Definitely look at things like linking root domains, keywords in the title, and page keywords - but that's not all there is to it anymore.

If it is Panda then cleaning up the site to make it more useful is your best bet.

• Make sure the design is clean.
• Make sure the writing is answering the user's question, and that it's well written.
• Make sure that the function of the page matches the search query.

Fixing these things should result in an eventual improvement in rankings.

If you were cloaking, that could be a penalty all by itself. Make sure to fix anything shady, and consider whether non-standard implementations or architecture could have confused Googlebot or made it look like you were trying to manipulate rankings. Check Webmaster tools for any warnings or manual penalties for sure.

You can also cross-reference drops in traffic to what others have seen on similar dates to make sure it was Panda.

Okay, so the situation here is a little unclear, but the solution should be pretty straightforward.

If the admin.site.com was different from the original site domain, simply noindex/nofollow all of the pages on that domain. I recommend this over a robots.txt rule because it will actually remove them from the index. You can add a disallow all rule in robots.txt later once the site is completely noindexed.

If the admin.site.com was the same domain, I'd recommend redirecting all of those pages to the new URLs again and then launching a noindex/nofollow version blocked with robots.txt, though I'm not sure why it needs to exist for reference. If the subdomain was different from the old site you could also probably just noindex/nofollow all of it without the redirect. It's not best practice, but it's not that big a deal.

Hope this helps to answer.

Hi Sarah,

Somehow I answered this and I must have forgotten to post the answer! Arg, it was a long one, too. Let me try to summarize what I'd do:

-If possible, noindex any page that doesn't display content while not logged in. Wait for those pages to drop out of the index, and monitor for errors.

• If not possible, skip straight to blocking pages behind a login wall with robots.txt. For example, to block anything in the login folder:

Disallow: /login

Or to block anything with a login variable:

Disallow: /*?login

This should prevent bots from crawling those URLs where you don't have any content to show them. Make sure to use this carefully.

I do apologize for the delay. If you have additional questions please feel free to PM me. I'd be happy to do a quick consult online or over the phone, as I feel bad that I never actually answered, and I can give you more specific ideas if we look at the site. If this answers your question that's fine too.

Good luck!

Firstly, please don't assume that you've been hit by Panda. Find out. Indexation count is generally not a good basis for assuming a penalty.

• Was there a traffic drop around the date of a known Panda update? Check this list. https://moz.com/google-algorithm-change . If the date of traffic drop lines up, you might have a problem. Otherwise it could easily be something else.
• How many links does your site have? Google indexes and crawls based on your authority. It's one area where it doesn't really matter where the links go: just having more links seems to increase the amount your site is crawled. Obviously the links should be non-spammy.
• Do you have a site map? Are you linking to all of these pages? It could be an architecture issue unrelated to penalty.

If it is a Panda issue: generally I think people take the wrong approach to Panda. It's NOT a matter of page count. I run sites with hundreds of thousands of URLs indexed, useful pages with relatively few links and no problems. It's a matter of usefulness. So you can decrease your Panda risk by cutting out useless pages - or you can increase the usefulness of those pages.

When consulting I had good luck helping people recover from penalties, and with Panda I'd go through a whole process of figuring out what the user wanted (surveys, interviews, user testing, click maps, etc.), looking at what the competition was doing through that lens, and then re-ordering pages, adjusting layout, adding content, and improving functionality toward that end.

Hope that helps.

1. Just to add to this, I'd look at the backlink profile and see the ratio of good to bad. If it's nothing but terrible links, I'd kill the 301 redirect. Though honestly at this point any damage that's done has probably already been done. Do you see any notice of penalty in WMT? If so definitely kill the 301. If not, and there are some good links, probably just move forward. If there are any pages with particularly good/bad ratios, you can do page-level redirects rather than domain-level redirects.

If you have no reason to believe you're under penalty, it's probably best to simply move forward. Try to have the worst of the worst removed, build or re-build links to the .com, and work on getting legitimate links going forward. It's all very hard to say without looking at the site or the links, but hopefully this is helpful.

2. Keep the disavow on the .com as long as spammy links are still more-or-less pointing (albeit through a 301) to your site.

Are the links bad? If not, don't worry so much about where the links are pointing. The link distribution models to even out linking page, anchor text, etc. aren't necessary unless the links come from spam sites.

Building (legit) links to deeper pages is generally a good strategy since the terms are more specific and often less competitive. If you think you're under penalty, though, let me know and my advice would be very different.

I've been watching some large sites, and I've seen some pretty crazy things around the 14-17th, with most sites seeing the biggest impact in ranking drops on the 16th. It was clearly an update, but I can't find reliable info on what exactly it was - probably because it affected a small niche of sites. Does your site receive a lot of traffic for local terms by chance? My working theory is another update to local search results. (Trying to show more companies that work in that area, possibly with more local packs.)

The reason I'm thinking local is because it was the biggest shift in rankings I've seen in a long long time, and tools like Mozcast didn't pick up a big change. There are so many possible combinations of local terms that it's nearly impossible to track even a small set of keyword + city terms.

Anyway, I've seen about the same as you - a big drop for some key local terms, followed by a slow and steady recovery. The results looked pretty bad when the update was first made, so maybe Google realized it wasn't a good update and has rolled some of it back.

Of course this is all just speculation. I'd love to hear what other types of terms people have seen ranking decreases on.

So Google Analytics uses what it (sometimes) calls "Last Non-Direct Click" attribution by default. You can change this, but it basically just means the last channel clicked by the customer gets the credit everywhere but the Multi-Channel Funnels reports where you can change the model.

First, check that your total goal figure align across models.

Generally the multi-funnel conversion reports are off a bit. I'm not sure why, but the standard reports are more accurate than a report in multi-channel reports or the Model Comparison Tool. The goal value in other reports should be close to the "Last Non-Direct Click" model. Paid search tends to have less variance. All of my other channels are less than 1% off. Are you seeing more variance?

Hope that helps! It's just one of the strange ways where Google comes to a number in a different way, and it doesn't always align perfectly.

It causes "duplicate content issues" in the sense that only one is likely to appear in search results if you do both. Given that Linkedin is probably more authoritative it's more likely to show up, but it just depends. It won't be a "penalty" necessarily - especially if your author name is on both. If the goal is just a lot of views, go for it. If you're trying to build the site you should probably prioritize the site.

Generally I'd recommend posting first to the site, making sure it's indexed, and then cross-publishing or cross-promoting if you don't see a lot of traction with it. But that's just because I'm generally trying to build my sites, and not LinkedIn. As I said, there are probably more views to be had on LI, but it may not accomplish site goals.

It probably won't hurt your rankings, but we would need a URL to say more. Do you see links to the fonts in your source code? If it's not immediately obvious you could try running a crawl to see where these are being linked from. It's probably a simple matter of cleaning up some CSS.

Interesting - thanks for coming back to update.

You could simply remove your name from the pages The sitelink example from Apple is at Google's discretion. They usually only do it for sites with high authority or where they're sure of your intent and have identified an entity associated with your search.

I would not mess with your robots.txt. Just decide which pages you do or don't want to show up for your name. You could try something like {Type of photos) photos by {Your Name} for pages you DO want to show up.

Generally lots of search presence is great, so I wouldn't worry. You COULD de-optimize, but it seems unnecessary.

You will see no benefit from mass-301 redirects to the home page, and it might even hurt. The only way to get value is if you redirect it to identical or similar content. I wouldn't even count on that adding value or traffic now that the site has changed hands and content has been gone for a while.

Still, if you can recover (and have the right to use) the old content, it might be worth 301 redirecting - or better yet, putting on the same URL so no redirects are necessary. Only do this if the content was good and the links were quality links. Otherwise let them 404.

If you can't put up the same or similar content, don't bother. Focus on building the site naturally.

Please tell me that you're using rel="alternate" tags for your desktop site and rel="canoncial" tags on the mobile site.

https://developers.google.com/webmasters/mobile-sites/mobile-seo/configurations/separate-urls?hl=en

If you're not, you need to implement ASAP. Make sure you're following all of the guidelines in the linked document.

Issues like this are why I typically recommend responsive sites for all but the best-staffed mega e-com sites. You need to redirect page to page - not from a category to the home page or something. You also need to make sure your canonical and alternate tags are absolutely perfect. I'd also recommending annotating the mobile site as directed above in the sitemap, especially if you've seen issues in the past.

Out of curiousity, what kind of redirects are you using? They should be 301, 302, or JavaScript redirects, because that's what Google supports. If these are done on the server side (and both versions have analytics installed) you should preserve the landing page info. Note that all redirects should redirect to the page indicated in rel="alternate" lest you give Google the wrong signal.

Read over the guidelines and let me know if you have any questions. If the guidelines are implemented correctly Google will rank your mobile site for mobile users, which will cut down on the redirects. This thread may also be helpful on how to implement redirects:

https://productforums.google.com/forum/#!topic/analytics/KQ2RPsLR35Y

Alright, well my advice runs counter to common wisdom: don't worry so much about whether it's on a sub-domain.

https://www.youtube.com/watch?v=_MswMYk05tk

https://www.mattcutts.com/blog/subdomains-and-subdirectories/

My feeling is that you should worry a lot more about what the blog is linking to.

I've personally had access to look into 2 of the most-cited subdomain examples. In one of them the company moved into a subdomain, but broke a massive document up into about 7 pieces - and each piece started getting organic traffic. This happened at the same time, so I don't think it can be attributed to the sub-domain movement. In another case the blog logo linked back to the blog home, and there were almost no links to the sales pages. Changing the header resulted in more links to the root domain, so yes - rankings improved.

I've heard other anecdotes of people moving to sub-folders and seeing big boosts in their non-blog rankings due to the links from the blog. For the last 5 years, though, I've seen no credible evidence. I've been in a good position to see the opposite. One more story: a friend of mine just moved a MASSIVELY-cited blog (blog PA 75) off the sub-domain, but his nav was already mirroring the www version (which has 1/15th the links). Impact on rankings: absolutely none.

I'd still install on the same server if I could, just for the sake of maintenance, but if you can't find a technical solution just make sure you link architecture makes sense.

Good answers here - did you get this taken care of? I'd say choose one domain and redirect or forward the others that have the same stuff. To explain it to my boss, I'd say.

• It confuses customers to have the same content on two domains. They might not know which company they're dealing with.
• You probably don't want half the traffic going to one site and half going to the other, especially if their content and user intent is similar. Every live domain is another analytics profiles I have to check on and watch for issues.
• Don't expect any ranking bonus from multiple domains, because when content is duplicate Google will just choose 1 page to rank.
• Maintaining multiple sites is more work than it's worth. We can get more done by focusing on our core domain unless there's a strong case for creating a new brand. (I wouldn't create a new site unless it was for a distinct brand).

Again, probably want to 301 redirect to the primary/canonical domain. If the domains have no links and no traffic (as I'd expect) forwarding through the registrar is fine, too.

I generally recommend meta noindex without blocking via robots.txt for reasons like this.

First, click secondary dimension and check out source / medium. Are they coming from Google? If all you did was use the remove tool on Google Webmaster Tools, Bing/Yahoo don't get the message yet.

Search engines cannot see a noindex tag  on the page if they're blocked from crawling the page. They can't access the page to read the tag. So it can sit around in the index despite not having been crawled for a while. (Though usually it's removed eventually).

Also keep in mind that you see some landing page traffic from instances where GA fails to fire the first time. It's usually a VERY tiny percentage, but I often see traffic to some (virtual pageview) popups that can't even load without entering info on our site (e.g. it's not even a possible landing page).

Might I also ask why you removed the job listing from the index? I was thinking this might be a good time to use rather than blocking crawlers outright. That assumes, of course, that you keep listing up for a fixed time. If you know when the job listing is going to expire, you can just tell the search engine. They might even send some traffic to your individual listings while live.

Unfortunately, Google has been scaling back reviews quite a bit. We have several sites where we used to show up with star ratings on every page. Now we're lucky to get them to show up at all.

I have noticed it relies a lot on the query. If you are trying to rank for local queries where Google pulls in a local pack, for example, they're really not likely to show your stars. Maybe this is to prevent confusion. Maybe it's to help themselves on Google Maps.

Try doing a site: search with the exact URL. In my experience it's most likely to show up in low-competition searches, and nothing is less competitive than a site: search with your own site. For example https://www.google.com/search?q=site%3Ahttp%3A%2F%2Fwww.shopperapproved.com does seem to pull in stars.

The good news: what you have is working.

The bad news: you're at Google's mercy as to when your snippets will pull in. Page and domain authority (in Google's eyes) does seem to play a part, and the query will change things up. My advice is always to implement it, continue marketing, and wait for Google to realize you're really an awesome company/site with reviews worth highlighting.

Unfortunately, there's no magic number of reviews or count of links. It just depends on competition in the SERP.

The errors are appearing because users are immediately redirected to the login page. It doesn't actually matter whether these are 301 or 302 redirects because you're not looking to pass link equity anyway. The easiest way to deal with this is to tell crawlers (like Moz's and Google's) to ignore stuff they couldn't see anyway because it's only shown to logged-in users.

I generally block the /customer/* folder in robots.txt to avoid issues like this. And if you have a wishlist in the primary nav that's worth blocking, too. You could also just remove the Wish List link when users aren't logged in. The idea is to tell the crawlers to avoid any portion of the site where a login is required. So you could just upload a text file to canyonos.com/robots.txt that looks like this:

user-agent: *

disallow: /customer/

disallow: /wishlist/

These lines will block the entire folder, so you should only do it if there's nothing in the folder that search engines can see. In other words, use this line if everything in the folder is behind a login (it usually is in Magento).

Google's info is pretty good here:

https://developers.google.com/webmasters/control-crawl-index/docs/robots_txt?csw=1

First make sure what you're dealing with.

1. Go into Google Analytics >> Acquisition >> Organic and find the week (or day, if you can) of the drop in organic traffic.

If you don't have GA, you can try something like Searchmetrics.

2. Correlate to this:

https://moz.com/google-algorithm-change

3. Look for recovery in organic traffic.

If you think Penguin is still impacting you, I'd recommend requesting removal. In my experience it's not very effective to seek the removal of links, but it's worth a try. If you don't think it's impacting you or still affecting you, just do good stuff going forward. In either case, submit a disavow on the bad links. Don't get caught up fixing things that aren't impacting your site.

If there is an organic impact, this is what I'd suggest.

-Start with a good form letter. Include the URL where the link is located, write up a professional, well-phrased request, and send it out. If there are a ton of sites, you might want to have someone who knows a good link from a bad link sort it for you and gather contact info.

-Send a follow-up to sites that don't respond.

-Submit the disavow and wait for an update.

You could hire someone to do link removal, but I don't think it's often a good way to spend money. Many of these sites are abandoned, and no one is reading the email.

If it's Panda after all, submit a new question and we'll talk

Google is fickle when it comes to sitelinks. There's not a lot you can do to control them aside from telling them which pages NOT to show for certain URLs in Webmaster Tools. There's definitely a correlation between your highest-authority pages and pages that appear prominently on your site by default.

The only thing you can really do to get site links is to prove to Google that you're a brand and that you appear more prominently in a search for something like your brand name. It's less a technical matter and more a matter of spreading the word about your brand. Links, shares, online mentions, and real-world awareness branding actions are all likely to help.

Hi there,

Impressions data is always estimated, but Google does two different calculations for the total and each impression. For example, I'm looking at a site right now with 157,358 impressions for one day in the chart and 170,000 in the graph. Google's taking a guess at the number of impressions for the whole site, then for each keyword. The guess in the chart is just a sum for each keyword.

I've seen discrepancies much higher, especially for smaller sites. Which one's more accurate? They're both such overblown estimates it's hard to say. Google Analytics impressions data is intentionally awful. They don't want to make it too easy to parse for those who want to manipulate Google's algorithms.

If you see the same discrepancy in actual sessions, though, let me know.

I don't see any reason it wouldn't be pulling the page in. Maybe check on the spelling of the keyword? If you have the right URL and the right keyword, it could be a bug, but hard to say at this point. We might be able to help more with a screenshot.

Yeah, Kate's right. I don't think you can go wrong, and unless you have redirects that disagree with your preferred version, go with the non-www.

If you're using standard www and non-www redirects use the one that shows up in the search bar for the site or area of the site that you're submitting. We have a site in Google News that is www. and the news is non-www. For no reason, but we haven't bothered fixing it. We were included on the non-www. So it's possible and likely that both would work fine, but I know for sure that the one that shows up in the URL of your feed works best.

Yeah, so I always thought that Google hating on affiliate links was a myth. Then I had an affiliate site started losing rankings. They went back up when I nofollowed the links. That was years ago, but it's still my preferred solution.

I suspect Google is already smart enough to understand what's happening, but they avoid doing anything about it because it's not impacting rankings and some people use it for legitimate reasons. Maybe some site owners don't trust the site they're sending traffic to, and don't want them to know which content is working best. Maybe others have a poor method of tracking that involves redirect.

I'm familiar with Yoast's solution, but I don't think it's helpful. It might make it look like you have more internal links, but I sincerely doubt Google is going to reward you for that. I recommend nofollow in any case, and direct links unless you have some other reason to avoid them.

Magento will generally canonicalize correctly by default. What I'm guessing you're seeing are category pages that have variables for faceted navigation - again, a pretty common issue in Magento. I've seen and even "fixed" this issue for past clients with no noticeable results.

If you ARE having problems getting your pages into Google's index (check indexation status in Webmaster tools and look for exact product/category URLs to check) there is a plugin that will remove the variables for faceted navigation on category pages. The main downside is that someone who emials/shares a link will not share it with the filters/sorting attached.

Generally I recommend leaving it alone. If you're not having indexation problems, though, it should be fine to leave alone and let Google figure it out.

If you're really worried about indexation I think that's a fine solution. It's definitely easier to manage, and it'll also be easier to track pageviews in most analytics platforms. The only downside is that if someone emails or links to a category page with filters applied the recipient won't see it. But generally people share products and not category pages, so it's not a big deal. I'd probably go that route.

Also make sure that your category pages still update the URL when you go to page 2, or that page 2 is somehow also being indexed. You don't want products that don't get indexed because categories can't be crawled.

Yes, the bots will crawl the pages, but they will not INDEX them.

There is a concern there, but mostly if the bots get caught in some kind of crawl trap - where they're trying out a near-infinite set of variables and getting stuck in a loop. Otherwise the spiders should understand the variables. You can actually check it in Webmaster tools to make sure Google understands. Instructions for that here:

https://support.google.com/webmasters/answer/6080550?hl=en

Ultimately Google will definitely not penalize you for having lots of duplicate content on URLs through variables, but it might be an issue with Googlebot not finding all your pages. You can make sure that doesn't happen by checking the indexation of your sitemap.

You could also try to block any URLs with the URL parameter in robots.txt. Make sure you get some help on the RegEx if you plan to do this. My advice is that blocking the variables in robots.txt is not worth it, as Google should have no problems with the variables - especially if the canonical tags are working.

Googlebot at least is smart enough these days to know when to stop crawling variable pages, so I think there are more important on-site things to worry about. Make sure your categories are linked to and optimized, for example.

Very odd, then, that they're being removed from the index. Do you think it's possible that the images have different URLs depending on which server they're cached on? That could definitely do it. I'd have a friend across the country pull them up and see if the image URL changes.

I'm assuming  that the image has some dynamic characters on it, which is pretty common with CDNs under certain configurations. Unfortunately, I've never used MaxCDN. If the image is just cdn.site.com/image.png - I'm afraid I have absolutely no idea why they wouldn't be re-indexed. I have similar CDN images that pull in fine.

If I'm not mistaken Magento has canonical tags on category pages by default, so you might be trying to solve an issue that doesn't exist. Take a look at the source code on faceted navigation to confirm. Or you can send me the site and I'll look over it.

Issue 1:

I think your intended ranking page is not indexed.

https://www.google.com/?gws_rd=ssl#q=site:http:%2F%2Fwww.localsearch.com.au

It's probably because, as Donna indicated, you have so many pages. This happens when you have what are essentially search pages that are indexed. Stuff happens like having a page for plumbing and plumbers in the same city, for example.

In the short term, you can make sure that non-indexed pages are linked to across the site. Long-term you're going to want to think of a way to organize your site to make sure Google and users can find the most important pages. For example, add breadcrumbs back to the city page, and have the city page linking to your most important types of pages (even if they're still searches) for the city. Right now your city pages are just more search pages, which is a big wasted opportunity to layout which pages you most want people to find. Also make sure  you figure out what's going on between these two "types" of the exact same page. There should only be one for the same results where possible:

http://www.localsearch.com.au/Gosford,NSW

http://www.localsearch.com.au/Search?where=Gosford,NSW

Issue 2:

Look at the "linked from" and figure out where these bad pages are linked to on the site. Google wouldn't make up a URL if someone wasn't linking to them, and my guess is your site is causing them. With a highly-dynamic site like yours it's usually either a crawl trap or a combination of dynamic URLs through a particular path that the server wasn't expecting.

Alternatively, and maybe more likely, Google has been trying to parse Javascript lately, and doing a rather poor job of it. I've seen Google try to find links in Javascript that were never intended to be links. You can either ignore these errors and wait for Google to get better, or you can dig into the JS with a dev and see what's causing Google to interpret something as a link. There's usually another way to put the code together where Google understands.

Hi there,

Could you tell me whether the URLs on your images are static on the CDN sub-domain? Or do they change regularly?

I was actually trying to give you the tools to figure out what's cached and indexed. You can just run a site search for the content and look at the cache, though. For example:

http://webcache.googleusercontent.com/search?q=cache:wIHvFryFPo4J:www.kingskolacheny.com/%23!cateringandcontact/c1d94+&cd=1&hl=en&ct=clnk&gl=us

If nothing shows up it's probably not indexed.