Welcome to the Q&A Forum

Hi,

From a code management point of view - as Peter says it's very common practice to split your CSS into different files as they are then much easier to manage and maintain. You can use a tool like Yahoo's YUI compressor to minify - as Bradley says - and aggregate (merge) these files.

From a web performance point of view, less files does not always mean better performance. Web browsers used to only download up to 2 files per domain, but now it's pretty standard for them to support 6 or more. See a browser breakdown for Max Connections and Connections per hostname here: http://www.browserscope.org/?category=network&v=top. I wouldn't recommend trying to split across 6 files, but you might find that if you have one massive CSS file it will download quicker when split up.

There is another disadvantage to having a single, CSS file in that you're not making the most of web browser caching. Every time you change any CSS, all users will have to download the entire file again. Again this may not be a problem for you, but something to bear in mind.

My advice would be to point Google Pagespeed at your website's key pages and act on as much as the feedback as possible: https://developers.google.com/speed/pagespeed/. It is a fantastic resource and presents its findings very clearly.

George
@methodicalweb

Hi,

The good news is I've worked on very similar projects before and taking a look at your examples, you're configuring it almost  by the book except you shouldn't use rel=next/previous and a canonical. It's either/or, so you're probably going to need to ditch your canonical.

i.e. http://googlewebmastercentral.blogspot.co.uk/2011/09/pagination-with-relnext-and-relprev.html

However, I've configured sites almost exactly as you have and found that Google has just randomly chosen different (and multiple) combinations of page and sort order to rank in different sections. Once they get added to the index, it's a real chore to get them removed.

I've learnt that if you genuinely don't want your sorted pages to appear in SERPs, you should use AJAX (and not have AJAX crawling turned on) e.g. "/?pg=1#dir=desc&order=price". Everything after the hash won't get crawled by Google.  If you can't do AJAX, then you can add noindex to sorted pages and (at your own peril) nofollow / robots.txt to stop some pages being crawled. Using nofollow / robots starts to move into crafting page rank though and IMO is to be avoided.

Another approach to avoid pagination and the performance impact of very long lists is to create more subcategories to break the inventory up. Might not be possible for your inventory, but worth considering a complete side step.

George

Hi Chiaryn,

Thanks - you've been really helpful! I had assumed that as the referrer wasn't in the Web UI (per WMT), it wasn't available anywhere. I'd also assumed it was a copywriting issue and not a product data issue.

Need to readdress my assumptions

George

Hi Sika,

What you're seeing isn't anything to be concerned about, and Moosa has already answered the cannibalisation part. I'd advise against tracking positions for individual keywords using a single tool from one week to the next. If you imagine there's a natural fluctuation of rankings and Moz is taking individual snapshots that might be up one day and down the next so they're only really meaningful when tracked over a longer period of time. You're also not taking into account the long tail - bunches of keywords that are similar which may not be fluctuating nearly as much as the one you're tracking.

The real indicator of where you rank should be the organic traffic to the page. I doubt that traffic will be fluctuating even nearly as much as the positions for the few keywords you are looking at.

As for algorithmic negative impact - you would probably see significant drops across multiple tracked keywords if this was the case - and those drops would be sustained until you diagnosed and fixed the problem.

Regards,

George

Hi Aaron,

The search experience on the website is a bit unconventional in that you search for a company name and it returns pages of results alphabetically listed with the name you are searching for hopefully in there somewhere!

You could make changes to the pagination using rel=next/previous, but what you're displaying isn't really "true" results pagination. I would therefore be cautious about changing it if the site is ranking well.

Canonicals would only be required if you were showing the same content on different URLs. A quick "site:" search like the below only returns one result, so either Google isn't showing the duplicate URLs (very likely given your question) or it isn't a problem for you:

site:www.formationsdirect.com inurl:companysearchlist.aspx?name=AMNA+CONSTRUCTION+LTD

You can look in webmaster tools to see which query string parameters it is picking up and configure the behaviour you want GoogleBot to take. You can also get some sense of the duplication if it is an issue.

Regarding the company page URL you gave, anything after the # in the URL won't get crawled so you don't need to worry about canonicalising those.

Again, if it's ranking well, be very careful about trying to solve a problem that doesn't exist. If you can find duplicate content then definitely redirect or canonicalise it and see what kind of impact it has. I would do this before taking on anything more significant like the website information architecture and navigation.

George

Hi,

I took a quick look at your site, sitemap and index status and only 25 urls in Google, but very many more in the sitemap.

What I couldn't work out is where your /item-details/ urls in the sitemap are linked to from your website? I can't get to them through buying -> catalogue. It won't help indexing status if they aren't being linked to from anywhere.

The biggest issue you have however is the way canonicals are set up on the problem pages. If you go to this page:

https://www.wilkinsons-auctioneers.co.uk/item-details/?ID=2710

It has the following canonical (without the id):

rel='canonical' href='https://www.wilkinsons-auctioneers.co.uk/item-details/' />

If you search on Google, that canonical URL is indexed, so if you fix this by adding the id to the canonical they should start to appear in SERPS.

You have exactly the same problem on your auctions pages. e.g. https://www.wilkinsons-auctioneers.co.uk/auction-items/?id=13&pagenum=51

Another point that will help you rank is to use friendlier / more descriptive URLs for the items.

Hope that helps

George

SearchMetrics would be a good place to start - you won't get individual keyword historic performance but it will show your website's overall SEO visibility over time. Particularly useful for tying in with Google algorithmic updates.

George

Hi Aaron,

First off, since your rankings haven't been affected I would definitely hold off changing anything in WMT unless you're sure as it might cause more harm than good. If you paginate what looks like potentially thousands of pages I'm not convince Google will look on this fondly. The URLs will probably also change regularly as more companies are incorporated because the pages are set to show fixed list lengths.

Resolving the duplicate content onsite is definitely the best course of action. The fact that Moz is crawling these duplicate pages indicates that it's picking up links from somewhere on your site. If you are able to stop exposing these links and only linking to the "preferred version" i.e. canonical then this will give you some control and a better understanding of the site's information architecture.

Regarding setting up of canonicals, I suspect that this will be a harder job as of the 3 duplicate URLs you provide, it's not immediately clear which one would be the canonical. There are probably also thousands of instances similar to this duplicate group across other company lists and Google will have picked at random which one it sees as the canonical on each one. Marking another URL in the group as the canonical stands to (at least temporarily) cause a drop in rankings and SEO visibility if done across thousands of pages simultaneously.

If I was you and I felt compelled to address the issue I would pick a sample ~10% of the duplicate groups, set a canonical on each of them and see what happens in terms of rankings over 3-6 weeks. I would also add the canonicals to a sitemap and try update any links on your website to make sure only the canonical is referenced.

It's risky though, as your rankings are good even though I understand the principle of what you're trying to achieve. When I've tended to do things like this it's when a website has had nothing to lose.

George

I would throw HTTP 410s for them all if they don't get traffic. 410 carries a little bit more weight than 404s and we're not talking about a small number of pages here. I wouldn't redirect them to the homepage as you'll almost certainly get a ton of "soft 404s" in WMT if done all at once.

Matt Cutts on 404 vs 410: https://www.youtube.com/watch?v=xp5Nf8ANfOw

If they are getting traffic, then it'll be a harder job to unpick the pages that have value.

George

I've read conflicting studies about use of Trust badges. Sometimes they're a good idea as it instills a feeling of trust in a nervous/cautious customer but other times it can have a negative effect by scaring off a user who hadn't considered security/privacy to be an issue until you mentioned it! It depends on the level of technical/online shopping experience your customers have.

As Gregory says - test it - get them to give you a month free trial and see if it impacts your conversion.

George

@methodicalweb

Hi Rafal,

The key part of that statement is "we might still find and index information about disallowed URLs...". If you read the next sentence it says: "As a result, the URL address and, potentially, other publicly available information such as anchor text in links to the site can still appear in Google search results".

If you look at moz.com/robots.txt you'll see an entry for:

Disallow: /pages/search_results*

But if you search this on Google:

site:moz.com/pages/search_results

You'll find there are 20 results in the index.

I used to agree with you, until I found out the hard way that if Google finds a link, regardless of whether it's in robots.txt or not it can put it in the index and it will remain there until you remove the nofollow restriction and noindex it, or remove it from the index using webmaster tools.

George

Hi Glenn,

Assuming you're on IIS7, I think these are the steps you are looking for:

http://technet.microsoft.com/en-us/library/cc732930(v=ws.10).aspx.

I strongly recommend you test the redirect configuration out on a test server first if you don't have a great deal of IIS experience.

George

@methodicalweb

Hi there,

You're describing quite a common frustration with Google Seller Ratings not appearing. The official line from Google can be found here: https://support.google.com/merchants/answer/190657?hl=en. The excerpt that is relevant to you is as follows:

My seller ratings aren't appearing

Reviews are not added to Google Shopping results in real-time, so you may notice a delay between receiving a new review and its addition to your rating in the Google Shopping results. The same is true when a review is removed from a seller rating website. If you notice your reviews have stopped appearing, please make sure that the store name and registered domain match in your Google Merchant Center account and third party seller rating websites. Learn more about how to update the store name and website URL in the Google Merchant Center account settings.

For Google Shopping star ratings to appear, typically your business needs at least 30 unique seller reviews, each from the past 12 months. However, we may show ratings for merchants with fewer than 30 reviews if we have sufficient data from other sources to determine an accurate rating.

George

@methodicalweb

Yes it's the worst possible scenario that they basically get trapped in SERPs. Google won't then crawl them until you allow the crawling, then set noindex (to remove from SERPS) and then add nofollow,noindex back on to keep them out of SERPs and to stop Google following any links on them.

Configuring URL parameters again is just a directive regarding the crawl and doesn't affect indexing status to the best of my knowledge.

In my experience, noindex is bulletproof but nofollow / robots.txt is very often misunderstood and can lead to a lot of problems as a result. Some SEOs think they can be clever in crafting the flow of PageRank through a site. The unsurprising reality is that Google just does what it wants.

George

Hi,

I'm looking to revamp the fortunes of an ailing Fashion ECommerce blog, which once had an impact on SEO for the site which it linked to but now has fallen by the wayside.

Blog sits here: www.mydomain.com/blog and links to products and categories on the ECommerce site www.mydomain.com.

The blog has about 2000 posts on it written over the past 5 years, which are almost all rewritten content about existing stories, events or embedded youtube videos related to fashion on the Web. None of the blog topics are unique, but the posts have been rewritten well and in an entertaining way - i.e. it's not just a copy and paste.

The blog is written on an old, proprietary platform and only has basic Social sharing. You can't comment on posts, or see "most popular" posts or tag clouds etc. It is optimised for SEO though, with fashion category tags, date archives and friendly URLs.

The company badly needs a shot in the arm for its content marketing efforts - so we're looking into the creation of infographics and other types of high quality, sharable content with an outreach effort. Ideally I want this content to be hosted on the Ecommerce site, but am faced with a few options which I'd appreciate the community's view on:

How I should handle the mix of the legacy content on /blog and the addition of new, "high quality" content?

1. (Pimp v1) Leave the /blog exactly as is and add the new, high quality content as new posts to it. Invest in pimping the /blog UI so that it has features such as commenting/tag clouds etc. They could migrate the blog to Wordpress, but leave it on the same URL.
2. (Cut loose) Leave the /blog alone, and start afresh with a new Wordpress blog for the new, high quality content. e.g. /News or news.mydomain.com. The old blog posts probably aren't worth bothering about, but it might be risky to delete them as there are a lot and are better off with them than without.
3. (Pimp v2) Set up a new Wordpress blog (e.g. /News or news.mydomain.com) for the new content and move the old /blog content to it. 301 the old /blog posts to the new location. The depth of old content that exists will add weight to the new content from a user's perspective, but will seem sparse if published on its own. Not sure why I would do this, but it's an option...
4. (Kill) Kill the old /blog content, start a new one for the new, high quality content.
5. Maybe there's another option I haven't considered.

Thanks in advance,

George

I agree with the 2 responses above.

Your blog is probably ranking because it has more links/shares (or at least more recent links/shares) and potentially more relevant content. You should try improving the content on the laptops products page.

You should also make it very easy for someone visiting your blog to get to your products to purchase them. If the blog is well written and useful/engaging, this might be a good opportunity for your content marketing to form a key part of your customer journey.

I bumped SEO-Buzz's answer

In practice, sometimes it's impossible to write unique product descriptions. I've worked with websites that have over 15K products, with around 5K changing every quarter. To keep on top of that you'd need an army of copywriters.

In that situation I would recommend doing the following:

1. Make sure your category / hub page content is awesome / unique / relevant. These will then be your main landing pages.

2. Pick key product sections - based on high margin / good stock availability / competitive pricing and update the product descriptions for them. If you client sees improved sales as a result, they will probably roll out this strategy to the rest of the site.

Hope that helps

George

Hi Graeme,

For old product pages - your solution is good regarding showing users alternatives to the out of stock products. No need for an "out of stock page" as there's no value in that for crawlers or users. Regarding point 2 - if you redirect discontinued product pages to category pages that should be fine although Google may regard that as a soft 404. If there are loads of products like this and you 301 them in one go then the chances are it will flag up in Google WMT. If there are a small number and you introduce them gradually then you'll probably be fine.

For the crawl errors question, adding value to the pages in terms of related products is a good solution if that's viable and the pages will be different enough from each other (i.e. no duplicate content). One thing that isn't clear at the moment is if you're redirecting empty category pages all to the homepage - or if it's possible to redirect or canonical them to their parent category.

e.g. For home -> clothing -> men's clothing -> shoes

If all the men's shoes are discontinued, then redirect that page to men's clothing rather than to the homepage. This reduces your chances of getting a soft 404, and is also arguably a better user experience.

Hope that helps,

George

Yes this is a good idea as it's a catch all for URLs that might include tracking URL parameters, or other parameters that don't affect the page content. When there are no tracking parameters, it's going to be more development and testing work to hide the canonical, when having it there doesn't cause any issues. It's also quite a brutal but effective catch all if your page was accidentally accessible via other URLs - e.g. non-www or https.

George

No need to be concerned. Aside from all the really well documented best practices on canonicals, in your original question you've spotted at least one big site that does this. They pay the SEO big bucks and rank well.