Welcome to the Q&A Forum

I have been trying to find any case studies of people who have optimized images for SEO for their e-commerce website, but haven't been able to find any case study, indicating obtained results. I am wondering how much increase in Google Image search traffic others have been able to obtain when optimizing their e-commerce images for image SEO.

I need this information to justify development resources needed for for example an image sitemap, changes to file names and alt texts, title tags and possibly EXIF data. File size is already ok.

Hope someone has experience with this and can share some results.  Also, would be great if Moz would do a Whiteboard Friday about this (hint!).

Hi Dmitrii,

Thanks for your reply. At first, I also thought that using multiple domains in one sitemap was not possible, however I came across this article from Google that mentions this possibility:

https://support.google.com/webmasters/answer/75712?hl=en

Does anyone have experience with this?

Hi everyone,

I've been searching about a problem, but haven't been able to find an answer.

We would like to generate a XML site map for an international web shop. This shop has one domain for Dutch visitors (.nl) and another domain for visitors of other countries (Germany, France, Belgium etc.) (.com). The website on the 2 domains looks the same, has the same template and same pages, but as it is targeted to other countries, the pages are in different languages and the urls are also in different languages (see example below for a category bags).

Example Netherlands:
Dutch domain: www.client.nl 
Example Dutch bags category page: www.client.nl/tassen

Example France:
International domain: www.client.com 
Example French bags category page: www.client.com/sacs

When a visitor is on the Dutch domain (.nl) which shows the Dutch content, he can switch country to for example France in the country switch and then gets redirected to the other, international .com domain. Also the other way round.

Now we want to generate a XML sitemap for these 2 domains. As it is the same site, but on 2 domains, development wants to make 1 sitemap, where we take the Dutch version with Dutch domain as basis and in the alternates we specify the other language versions on the other domain (see example below).

<loc>http://www.client.nl/tassen</loc>
<xhtml:link<br>rel="alternate"
hreflang="fr"
href="http://www.client.com/sacs"
/></xhtml:link<br>

Is this the best way to do this? Or would we need to make 2 site maps, as it are 2 domains?

Hey Sean,

Thanks for your response! The pages are all category landing pages, and with a mega dropdown for navigation, faceted navigation on the left and a footer, the total number of links adds up quickly. They're all internal links though, and it was my understanding that a high number of internal links does not directly harm the value of a page?

I used the SEO Workers website to count the links by the way. According to them, the lingerie page has 342 links (they call it URLs but I guess they mean links), the swimwear page has 199 links and the nightwear page has 190 links. So, the number of links for the swimwear and nightwear pages are roughly the same, yet the swimwear page does not currently rank for swimwear, and the nightwear page is ranked #5 for nightwear.

What's strange though, is when I copy the source of each of the three pages into Notepad++ and count hrefs, I find 203 occurances in total for each of the pages. That's strange in the sense that SEO Workers apparently does something different, but not strange since the page templates are exactly alike, but the content is different so the number of subcategories, filter options etc will be a little different.

I'll PM you the links, just in case you're curious and want to see for yourself

One of our clients is a lingerie retailer, that also sells swimwear and nightwear. They have a webshop, with landing pages such as /lingerie, /swimwear and /nightwear (only in Dutch). The nightwear page is doing ok (#5 in Google NL), but for the swimwear and lingerie pages, something strange has happened and I have no clue what it is.

On May 7, the /swimwear page ranked 24 for swimwear in Google NL. In that period, the homepage was ranking at #40-50. One week later, on May 14, the homepage was ranked 24 for swimwear, and the /swimwear page was nowhere to be found. The homepage is currently ranking 17/18 for swimwear. Something similar seems to have happened for lingerie. The homepage was ranking at around 11-13 for a few weeks in April/May, the /lingerie page was ranking at 8 in May, and then after the 4th of June, both the homepage and the /lingerie page no longer seem to have any decent rank for lingerie.

So, trying to find the problem:

• I checked the Moz On-Page Grader for all three pages/keywords. All pages get an A.
• Robots.txt hardly contains anything, and there's no robots meta tag to block the page.
• Checked Google Webmaster Tools just in case, nothing out of the ordinary.
• The page contains 342 URL's which is a lot, but that shouldn't be too problematic, I thought?
• I used the Moz Keyword Difficulty Tool to analyze the current top 10 for lingerie, and compared it to our client's stats. Page Authority in the top 10 is 37-68, our clients is 33. Domain Authority in the top 10 is 27-100 (wiki page), our clients is 43. MozRank for the top 10 is 6,6-8,0, our clients is 6,7. MozTrust for the top 10 is 7,0-7,1, our clients is 7,0. The basics all look fine!

Now, the page could use some link building, the number of links is a little low compared to the top 10 entries. We're also in the process of redesigning the page, adding more products but also more content (text and images) to inspire visitors a little more. But being low on links doesn't seem a good explanation for such a relatively sudden change in ranking. As far as we can tell, looking back at commits by our developers, nothing changed around that time that could have this impact. Of course it's possible that we missed something, we're still looking.

Also, it's strange that these sudden changes took place for swimwear and especially lingerie, but not for nightwear.

I'd really appreciate any tips, hints etc. to find out what's going on here. Thanks in advance!

Michel

The point I'm trying to get across is this:

"I asked the question of why these pages are considered duplicate, the answer appears to be : because textually they are even if visually they are not."

I don't think that's the complete answer, or even the most important part of the answer. Surely having mostly similar content across pages won't help, but as I've tried to point out, there are other factors that come into play here. It's not just about the content, but putting the content into context for the search engines. In order for them to understand what it is they're looking it, there's more that's important than just the content.

Michel

Hi Simon, Thanks for your response. So you do not think that duplicate content will be a problem when using hreflang? And also, I guess that you feel that Google does not do an awesome job matching languages to countries when using only a language variable on a .com domain? Michel

I do not agree. I see these kinds of pages on e-commerce websites on a daily basis. For webshops that sell only a certain kind of product, almost all product pages will look alike.

In this case, the H1 is different, the page title is different, and the description is different. This is only a small portion of the page but that's not uncommon, so I would argue that it cannot be just that.

I would look into URLs, marking up your data using http://schema.org/Product, possibly making small changes to accomodate the tags. For instance splitting up brand, color etc. so that you can mark them accordingly.

I wonder if the details.aspx has something to do with it?

www.1010direct.com/TDV-019-GOLD-50/details.aspx
www.1010direct.com/DGV-DD1165-970-53/details.aspx

Basically, both pages are called details.aspx. Depending on how you look at it, you have 2 pages that are named the same (with mostly similar content, though not unusual for e-commerce websites) in different subfolders. I'm not sure if there's some kind of difference in the way Moz works, and if that's part of why Moz marks this as duplicate content?

Are you unable to create 'prettier' URL's? Such as:

www.1010direct.com/tim-dilsen-019-gold-50-glasses.aspx
www.1010direct.com/dolce-gabbana-dd1165-970-53-glasses.aspx

With or without the aspx of course.

I'm trying to wrap my head around various options when it comes to international SEO, specifically how to rank well in countries that share a language, and the risk of duplicate content in these cases. We have a chance to start from scratch because we're switching to a new e-commerce platform, and we were looking into using hreflang.

Let's assume an example of a .com webshop that targets both Austria and Germany.

One option is to include both language and region in the URL, and mark these as such using hreflang:

webshop.com/de-de/german-language-content (with hreflang de-de)
webshop.com/de-at/german-language-content  (with hreflang de-at)

Another option would be to only include the language in the URL, not the region, and let Google figure out the rest:

webshop.com/de/german-language-content (with hreflang de)

Which would be better?

The risk of inserting a country, of course, is that you're introducing duplicate content, especially since for webshops there are usually only minor differences in content (pricing, currency, a word here and there). If hreflang is an effective means to make sure that visitors from each country get the correct URL from the search engines, I don't see any reason not to use this way. But if search engines get it wrong, users will end up in the wrong page and will have to switch country, which could result in conversion loss.

Also, if you only use language in the URL, is it useful at all to use hreflang? Aren't engines perfectly able to recognize language already?

I don't mention ccTLDs here because most of the time we're required to use a .com domain owned by our customer. But if we did, would that be much better? And would it still be useful to use hreflang then?

webshop.de/german-language-content (with hreflang de-de)
webshop.at/german-language-content (with hreflang de-at)

Michel Hendriks
Docdata Commerce

Thanks Oleg and Irving!

There are only very minor differences between the countries that speak the same languages. Sometimes there will be price differences, though not always, and we use callcenters that have different phone numbers for different countries. That's all though, and so that's not much.

So, we don't need the country in the URL. We would be able to determine country by using Geo IP and preselect that. I guess the most important question for is, if there's any reason you think we should include the country in the URL anyway?

We have made the decision to start using a new ecommerce platform, which means we will have to migrate our existing webshops. Some of our new customers will be launched on the new platform straight away. Some limitations we used to have when it comes to URL structure have mostly been lifted, so I've been thinking what the perfect URL would be in terms of SEO.

Since we mostly work for pan European customers, the multi regional and multi lingual aspect is a very important one, as it's important to rank well in all countries. I've always figured that even though it would be good to integrate country into the URL somehow to indicate to the engines that this content is meant for a certain country (either by using local TLDs or indicating using Webmaster Tools that a certain subdirectory or subdomain is targeting a country specifically), there are multiple countries with the same language (for instance, they speak French in France but also in Belgium), which could cause duplicate content issues:

www.webshop.com/be/fr/french-product-name
www.webshop.com/fr/fr/french-product-name

I guess it won't matter much whether you use fr.webshop.com, www.webshop.com/fr or www.webshop.fr, it's mostly the decision IF you want to include country somehow. What do you all think, is this important? Or is the multi lingual component enough for pages to rank well in several countries? For instance, if we were to use the language component only:

www.webshop.com/fr/french-product-name

Would this have the potential to rank well in both the French speaking part of Belgium, as well as France?

Michel

We have a website that displays a number of products. The product has variations (sizes) and unfortunately every size has its own URL (for now anyway). Needless to say, this causes duplicate content issues. (And of course, we are looking to change the URL's for our site as soon as possible)

However, even though these duplicate URL's exist, you should not be able to land on them by navigating through the site. In theory, the site should always display the link to the smallest size. It seems that there is a flaw in our system somewhere, as these links are now found in our campaign here on SEOmoz.

My question: is there any way to find the crawl path that lead to the URL's that shouldn't have been found, so we can locate the problem?

Indeed John, using men-shirts would certainly be an option as a solution to the double names. If I understand you correctly, this would result in something like this:

www.onlinestore.com/men/
www.onlinestore.com/men-shirts/
www.onlinestore.com/men-denim-shirts/
www.onlinestore.com/levis/levis-stone-wash-denim-shirt/

So I wonder if a flat URL structure, supported by proper internal linking of course, could be a viable option, or could perhaps even the prefered option?

Michel

One of the first things I remember reading about SEO and URLs, a long time ago, is that keywords are important, and hierarchy is important, for search engines and for users. Hierarchy in URLs would give the search engines an idea of the structure of the site, and users would be able to edit the URLs to continue navigating.

I'm wondering about URLs, hierarchy and usability lately, since I've seen that ASOS uses a new URL structure on their site. At first glance, I thought it was brilliant, so I would like to get all of your opinions as well.

For those of you that haven't seen the URLs: for categories, ASOS uses a structure as you would expect it, but for products they don't insert the category in the URL. Instead they insert the brand name as the first part of the URL, followed by the product title. Some examples:

Category:
www.asos.com/women/dresses/...

Product:
www.asos.com/french-connection/french-connection-tie-waist-pocket-stripe-dress/...

I can see the importance of brand name for a site like ASOS, and like how they stressed this by inserting not the category but the brand for products. I don't know how much ASOS still relies on organic non-ASOS related keyword traffic, but still.

Now, for hierarchy, I guess a good internal linking structure will tell the search engines about the hierarchy of a site as well, right? So perhaps hierarchy in the URL isn't that important? Perhaps something like this would be just as good as anything, given a good internal link structure?

www.onlinestore.com/category/
www.onlinestore.com/subcategory/
www.onlinestore.com/brand/product-title/

Now, I understand that if you use this structure, you wouldn't be able to have men/shirts and women/shirts, but let's say that you don't have subcategories that use the same names. In this case, how important is hierarchy?

And, what do you think about this URL structure for an ecommerce site for which brands are important?

There is certainly a need to communicate that the products are made of leather. We want to be found when people search for leather billfolds, and I feel it's only natural that we also communicate this on the website. Especially since not everyone knows the brand, and since leather is not in the company name.

I guess the answer to your question is "nothing", at least when it comes to category names. However, would you go as far as to also use it in product names? This would mean that on a category page, you would see a list of products (18 on each page at the moment), all with leather in the name.

In a recent WBF about over optimization, Cyrus said it seemed that especially the combination of the title tag, URL and incoming anchor text seemed to trigger a decrease in rankings. I'm not too worried about that happening here, I think we can safely optimize both title tags and URL's at least. Now to decide what to do with the rest.

This may be a very basic question, but with all this talk about overoptimization I just want to make sure we get this right.

We run a webshop for a manufacturer of leather products. Billfolds, iPhone sleeves, briefcases etc. Their company name (also the domain name at which the webshop is active) does not include 'leather'.

Obviously, leather is an important keyword for these products, but having a category page with 'leather X', 'leather Y', 'leather Z' not only looks weird, it might even look spammy. The same, though to a lesser extent, is true for the category names. Do we really want to have 'leather billfolds', 'leather ipad sleeves' etc. at the top of every category?

Can anyone give some tips, pointers, best practices perhaps for when an important keyword is basically true for every category/product/page of your site? How do you include it without overoptimizing?

Basically, you are asking 2 questions. Is it important to redirect, and what do we redirect to?

1. It is important to redirect old URL's, but only if those URL's are indexed. You see, people may still click on them, as they will still be in the search engines for a while. Of course, if the URL's are not indexed, you don't have to redirect anything. You are correct to use 301 redirects, and you will lose some link juice but it's better than nothing!

2. The best URL is an URL with relevant keywords. Relevant, so 'catalog' does not need to be in the URL. Your example of http://www.k9electronics.com/designer-dog-beds/ is the best URL. It's good because people will understand what your page is about, and search engines like that.

Good luck!

I guess that is the only solution then. I don't quite understand why Google picks that information to show in the SERP text (as well as the 170+ items) but we'll try disabling the Smarty debugging when we're not actively using it. I hope it helps!

Ah thanks Alan! It looks like there is a problem in the code that generates the breadcrumb URLs. We will get that fixed asap, whicih should lower the number of duplicate content warnings considerably.

But if the ads are unique (just not unique enough) why would you want to use a canonical tag? You'd use that to show Google which is the prefered page but in this case, is there such a thing?

Making the pages unique will be difficult, your best bet is to try and add and/or emphasize the information that is unique.

Maybe you could change the way people place ads. Not just have the same kind of information such as make, model etc., but ask people to write a little something about how they got the car, what kind of maintenance they've had done, why they are selling it. If they write all this in their own words, you will have more unique content. Plus you can tell people that adding this information helps sell their car quicker.

We have an international webshop with languages in the URLs. Our URLs are now set up as follows:

http://thermalunderwear.eu/eng/category/product

Now, we know that there's some kind of strange redirect problem causing problems with our indexation, this is a technical issue that should be fixed soon. But whether this is the cause of some other strange problems, I do not know. I'd be happy with any help/advice/tips.

1. The SEOmoz site crawler starts at http://thermalunderwear.eu. This currently does not yet redirect to http://thermalunderwear.eu/eng like we want it to, but all the links on the page do include the default language code. So all links on the page are http://thermalunderwear.eu/eng/category etc. However, apart from those URLs, the site crawler finds many URLs in the form http://thermalunderwear.eu/category/product etc., so not including the language variable. Where it gets these I do not know, and since these URLs dont exist and the webshop simply shows the homepage, these URLs all have 50+ duplicate titles/content. Why oh why?

2. If I do a Google search for indexed URL's with English as language, I get many results formatted like this:

Coldpruf Enthusiast mens thermal shirt - Thermal wear for men ...
thermalunderwear.eu/eng/men/coldpruf-enthusiast-mens-thermal-shirt 170+ items – Fine-ribbed longsleeve thermal shirt men from Enthusiast ... {$SCRIPT_NAME} eng/men/coldpruf-enthusiast-mens-the {$ajax_url} http://thermalunderwear.eu/ajax

What are those variables doing there? It looks like it's taking something from our Smarty debug console, which is hidden but still active in the source code, but also the ajax URL which is in a completely different location. What is Google trying to show here?