ZenCart, Magento and poor SEO – solving the issues

Pixel Ovenmitts from JustMustard.com

Pixel Ovenmitts from JustMustard.com

We’ve been working with JustMustard.com on a very special project involving ZenCart eCommerce software. JustMustard is an online retailer and wholesaler of a variety of very strange and very awesome gifts and presents. We really enjoy working with them. Plus we really want some Pixel Oven-gloves.

However, ZenCart is viewed with both praise and disdain from programmers. It is an excellent off-the-shelf system for an online shopping experience, very secure, and very customisable. However, like most off-the-shelf systems, it does require a lot of work to give the results wanted.

As an aside, if you are looking for a new shopping system, we do recommend looking into both off-the-shelf systems such as osCommerce and Magento against having a custom solution being built. A custom solution can mean better SEO and should be more in needs with your business, and although the costs may appear high and a require a good long-term development team, on the other hand, a generic site needing custom changes later can be expensive or impossible, and is more likely to be attacked and spammed. In JustMustard’s case, we’re finding that ZenCart (after some work) is working very well; but in other sites using a generic site could be a costly mistake..

ZenCart, along with many other eCommerce sites are not set out for great SEO opportunities. ZenCart’s biggest mistake is having duplicate content, since a product can be accessed by many URLs (via the cPath); and often different URLs for the same product is used across the site. ZenCart also doesn’t make any use of the URL for keywords, often making the common mistake of using the querystring for page navigation.

We’ve been able to work with the ZenCart system and rewrite all the necessary URLs, and also automatically use the product title (and optional description) as part of the keywords in the URL. This is simply done using some fairly logical PHP techniques alongside some custom htaccess rules and regular expression matching.

It means a URL which previously looked like this:

/index.php?main_page=product_info&cPath=1_4&products_id=39

Is now redirected, alongside any similar-looking URL, to the much neater and unambiguous;

/product-39/cactooph-toothpick/

The same technique has been applied to product categories, and has been to static product pages too.

If the URL visited isn’t exactly the right one, the script tries to find the product, category or static page, and redirects the customer to the correct URL. The URL includes the latest keywords via the product title; ensuring customers have an informative URL, and search engines have valuable keywords and non-duplicate content.

Scripts like this can really clean up a working website, and really help to achieve both better keyword position and make it easier for your customers to navigate your site; without requiring a complete re-write. Our version will work just as easily for any future product; and doesn’t require any additional effort from product administration.

list of shortcodes on derp.co.uk that come from rewriterules

derp.co.uk: we love rewrite rules!

We’ve done this on every site we’ve set up, and found it helps bundles. Often this is just a few pages, but we’ve integrated this same system on far more advanced websites too – including a complex laptop, battery and power browser system on the laptopshop.co.uk website – a total of over 300,000 pages. We’ve also used rewrite rules to a ridiculous degree on the image site, derp.co.uk

Certainly ensuring non-duplicate content and keywords are important to SEO. Changing all the URLs on your site can be a lot of work to achieve, and it should never be taken lightly, it can sometimes be better to improve slowly and carefully. At the very least, it is essential to ensure all old URLs redirect to the new address and include a sitemap. For a number of reasons, this was a good time for JustMustard.com, and we’ll be keeping a close eye on it and tuning accordingly..

Despite our technique being relatively simple, it requires quite some work before it will automatically with all ZenCart installs, and static pages have had to be arranged using some of our other tools. JustMustard.com was a perfect candidate for this work, but every site we’ve seen has different needs and requirements.

For this reason, we’re not able to provide source code, though we may provide some of our techniques if there’s demand; and we’re always happy to help if you contact us. We hope this is useful and motivating to get those site issues fixed!

Instead, we are happy to discuss similar installations and assistance with SEO and eCommerce.

Best Practice for Next/Previous Pagination in Search Engines

Most developers will come across a need for paged content; perhaps for products on a website, or for long articles; but how is this best done for search?

Google have recently announced more support for previous/next pages, and have offered two options;

  • Using the <link> tag in the <head> container, to ‘Canonicalise‘ the individual pages to a ‘View All’ page.
  • Use the same <link> tag in <head>, with the ‘rel’ attribute to point to the previous and next pages.

 

The decision of which method to choose depends on your needs – if you’d prefer visitors to arrive at your ‘View All’ page, or perhaps you use a ‘Configurator’, such as one for choosing the make and model of your laptop; you may wish your customers to land there.

Although; you may prefer also customers to land on the page the search keywords the user arrived on; rather than having to look through tens of products.

However, there’s a few other points to consider. With a view-all page, particularly one with a simple list of all products; this may not be the best experience for your visitors. It may also mean a higher latency and page load time, causing visitors to get bored – and also isn’t best for search engine ranking.

But, if customers land on a particular page number, the spidered content may be outdated, and mean the customer lands on the wrong page; and has to search for their product. This commonly happens on forums, but may equally indicate slow spidering.

It should be noted that these only act as ‘hints’ to Google; but particularly if your site is struggling with many paginated pages, especially with outdated content, it may be worth considering these alternatives.

We’re always happy to help anyone who would like some assistance; and Google have also put together some information on both the ‘Rel=Next’ and ‘Rel=Prev’ attributes, and well as using a ‘View All’ page.

Googles changes the operators

If you’re a regular user of Google, I’m sure you’re familiar with the plus operator. It’s used as a tool by power users to narrow down results to ensure words will be in the search results.

If not, you’re probably familiar with Google rewriting your query to try and help your results; most commonly this is seen with misspelling; but often a word or two may be dropped from your query to assist with search terms you may have meant.

As Eric Schmidt once said, “We know where you are. We know where you’ve been. We can more or less know what you’re thinking about”. They also know you can’t come up with the best search terms for the results you want.

Sometimes things go wrong – often, the wrong word is dropped. Previously, adding the + operator; e.g. [winlirc +audiocapture] would solve this, now it requires quote marks – e.g. [winlirc "audiocapture"] to give useful results.

If you make a search on Google using the + operator to ensure a word is used, you’ll now receive the following message;

Showing results for google +replaced plus with quotes
The + operator has been replaced.
To search for an exact word or phrase, use double quotation marks: google “replaced” plus with quotes

Here’s a screenshot:

Screenshot of Google error - replaced the plus operator with quote marks

Google have replaced the plus operator with quote marks

So what’s going on? Google are changing some of the operators; they’ve updated their documentation on their help page. The negation operator, ‘-‘, still works as expected; as does the double-quote marks for both phrases, but now also for words too.

We’ve certainly noticed that over the last few weeks, for longer search queries some words are dropped that are seen as, in Googles own words; ‘evidence might come from language analysis that Google has done or many other sources’. However, in our experience, the search results have been mainly to give popular results, rather than more specific results that the search terms imply.

As an aside – as always, the underscore is treated as a special case – and is another reason why it should be avoided in URL structures, always use hyphens instead!

So what does the ‘+’ operator do now? For now, it seems very little.

Why has it been replaced? Nobody seems to know, although both the plus and quote operators were probably very similar in code.

What’s the problem? Well, it’s a minor gripe, but it now takes at least four key presses (and navigating both sides of the word or phrase); instead of just two. Google also seems to being less ‘choosey’ on keywords entered by the user; instead intending to match more popular results.

Of course, real expert users can continue to use the nfpr=1 querystring to disable most of Google’s corrections.

Since Google haven’t made an announcement, if anyone else has any more information about this change, please let us know in the comments.

SSL Google won’t be sending you queries

In a fairly unexpected report, Google have announced that SSL will be enabled for a number of users by default. Along with GMail, secured by default last year, and Twitter and Facebook providing secure versions as an option, Google will be introducing secure searches to users over the coming weeks.

Google have provided a secure search at https://encrypted.google.com/ for a while, more information on it is provided in their help, but it ensures that no third-party can see your searches.

It also means that webmasters will not receive the query that the user has searched for. This was done by parsing the referer* in the HTTP request. For example, url= was the url clicked on, the q= represented the query, and cd= as the position the link was in. An example script, in PHP, can be found on this site; but could apply on any landing page in any language.

This is a major blow to webmaster who use this data to analyse keywords; and it will break those sites that automatically generate content based on your search query. This was often used for analysis – but also for link farms and advertising sites for gathering keyword information and misleading users. This hopefully will reduce some of the link farms that use this technique; but some webmasters do use this to highlight search terms on the landing page.

So what does this mean? Webmasters will still know users came from Google, but not any of the information about the search. Users who encrypt their search will not be passing on their keyword terms via the HTTP header.

You can however, receive a list of the top 1,000 search queries through Google Webmaster Tools, and presumably will still continue to receive keyword and other information from Google Analytics.

The official announcement can be found on the Official Google Blog.

* referer is spelt here with one ‘R’ as defined in the HTTP spec.

 

SEO Exploit via DNS “piggybacking” found in the wild

A recent report out from the ISC (the ‘Internet Storm Center’, a program of the SANS Technology Institute) warns domain owners to check their DNS records.

We offer a DNS Records tool on frag.co.uk/tools/ that may be some help to our customers.

The report details around 50 organizations that have had new machine names added to their DNS zone information. These were then pointed to sites used to boost the search engine positions of pharmaceuticals, personals, and adult web sites.

A good explanation, and some other sites that are affected can be found on the ISC Diary – What’s In a Name blog post.

For example, the Federal Commission of Taxation in Argentina at www.cfi.gov.ar have a subdomain they presumably were unaware of, at “buy-viagra.cfi.gov.ar”, thankfully now suspended.

We run our own internal DNS servers here at frag.co.uk, and can proudly announce that all our domain owners are unaffected, secured and safe from this attack. Please feel free to contact us for more details.

Do you outsource your DNS? Would you ever catch something like this?

We may be able to help. One of our totally free tools on frag.co.uk/tools/ offers a “DNS Records” tab that allows you to keep an eye on your DNS records. Do let us know how you get on.

eMail: web@frag.co.uk
Phone: 07739 315821

Meta


frag.co.uk
on Google+