How the Winners Do Mobile SEO [Guest Post]

**Note from Stephanie: This is a guest blog post by Bryson Meunier, a Chicago-based SEO and Director of Content Solutions at Resolution Media, an Omnicom Media Group Company. His career at Resolution Media has given him the opportunity to help some of the world’s top brands achieve success in mobile and SEO including: Lowe’s, LeapFrog, Apple, Groupon, FedEx, and others. 

There are a lot of opinions on how to do mobile SEO right these days. Do you need to build a separate mobile site, or is it enough to make your website responsive, so that it reformats for the device that’s accessing the content? Or can you skip all of these entirely and simply make your web site accessible to and readable on devices of all kinds, like Apple.com has? I’ve been researching mobile SEO best practices since that meant making your site accessible to PDAs, and I’ll be giving my own professional opinions on the subject at SMX Advanced in Seattle this year during the iSEO: Mobile Search Engine Optimization Done Right panel with Cindy Krum and Google’s Pierre Far. But before I gave my considered opinion, I wanted to see what the most successful web sites are doing today, if anything.

To that end I looked at the sites getting the most traffic from Google on the Internet according to SEMRush, and crawled them with Google’s smartphone Googlebot user agent. 

Smartphone Googlebot, Meet the Traffic Leaders

If you’re not yet aware of smartphone Googlebot, it’s a relatively new crawler introduced in December 2011 to skip redirects and display URLs that the smartphone users ultimately see. For example, if I’m Citysearch and I want smartphone searchers to get the mobile site rather than the desktop site because it converts better, I can serve the mobile site to smartphone searchers and smartphone Googlebot so that when someone is searching for a local restaurant my mobile URL rather than the WWW appears:

When I did this using a Firefox User Agent Switcher and the Smartphone Googlebot User Agent there were 3 basic actions that the sites took:

  1. They did nothing in response to smartphone Googlebot. This is either because they have a site that they think provides a good experience on smartphones, or they have a mobile site or responsive site and it simply didn’t recognize smartphone Googlebot as a mobile user agent that required a mobile experience.
  2. They redirected smartphone Googlebot to a mobile URL. This was usually an m dot subdomain (e.g. m.yahoo.com, en.m.wikipedia.org), but was sometimes only a parameter or a separate URL that indicated somehow it was intended for access from mobile devices. Exactly one time it was a separate dotMobi TLD.
  3. They didn’t change the URL, but did reformat the site for easier mobile browsing through responsive web design.

It’s worth noting that I could not crawl 7% of the most popular sites. Six weren’t crawled because they weren’t safe for work (and not really that relevant to what most white hat SEOs are going to be doing anyway) and one because it returned a 404. 

Responsive Sites or Mobile Redirects?

Given the uproar in the web design and SEO community recently over Jakob Nielsen’s recommendation to build both a mobile site and a full site, I half expected the majority of the sites crawled to return responsive web sites rather than mobile sites on separate URLs. There’s a popular theory among SEOs these days that mobile URLs dilute link equity, making the site less competitive for search; and if that was the case I would expect the majority of these sites to be responsive, since that’s apparently more competitive than sites with mobile URLs since responsive design uses the same URLs and doesn’t split link equity.

I also wasn’t sure if any of the sites would be prepared to welcome smartphone Googlebot, given that it’s only been around for about 5 months and the results are currently spotty at best. Knowing how slow large websites (especially those run by large corporations) can often move, I thought that adoption would be low, if anything.

Our hypothesis, then, was that these 100 sites that get the most organic traffic from Google would not serve mobile URLs, as the split link equity from mobile URLs would make them less competitive in organic search, making it difficult for them to get enough organic traffic to be in such an elite category. More likely the sites would use responsive design, as for whatever shortcomings it has it uses the same URLs as desktop and does not split link equity. Furthermore, we hypothesized that not many sites would do anything special for smartphone Googlebot, as it has only been around for 6 months at this point, and adoption seems to be low given that the smartphone search results are still largely populated by desktop web sites.

What I found was just the opposite. 

Of the 100 sites, 83% of them took smartphone Googlebot to a separate mobile user experience, and only 10% did nothing. Of the 10 sites that didn’t take smartphone Googlebot to mobile content, more than half of them have mobile or responsive sites and are likely not redirecting the bot because they’re unaware of its existence or have yet to implement the change. 

This is much higher than I was expecting and suggests that the great majority of sites that get the most traffic from Google are looking for a way to make their mobile content visible in Google smartphone results today, rather than waiting for next year or more budget to make simple fixes that will result in more mobile search traffic. 

This may be why they get more organic traffic from Google than most of us.

How they actually responded to Googlebot might surprise some SEOs, as a full 60% of the sites redirected to a mobile URL, while only 11% of the sites were responsive. 

When we account for the sites that may have mobile content but didn’t redirect it, we see that 83% of the top sites either have mobile sites or responsive sites, and 71% of them offer mobile sites at mobile URLs.

You may have heard the “one URL to rule them all” line or read Duane Forrester of Bing’s plea to use one URL for each piece of content and thought that mobile URLs would be detrimental to link equity. It’s reasonable to think that duplicate content would cause problems with canonicalization and ranking as it does in desktop search for things like printable URLs and indexed session IDs. Some have recommended responsive design as the best way to build a mobile web site in order to avoid this split link equity. There are good reasons to use responsive web design, and 12 of the top 100 sites have made their sites responsive. However, as 71 of the top 100 traffic generating sites in Google use mobile URLs, it’s unlikely that this is a serious issue for ranking in Google.

True, this is only a sample of 100 sites, but they are the sites that get the most organic search traffic on the Internet. They could clearly be doing more to optimize for mobile search traffic, in my opinion, but the fact that they’re getting more overall traffic than most of us means they can’t be doing everything wrong. This study is also consistent with findings from Mongoose Metrics that showed the majority of sites in the Quantcast Top Million redirect to mobile URLs instead of using responsive web design. At around 52%, however, their majority is not as clear as the 71% of top performing sites in this study that display mobile URLs. 

Many methods can work for mobile SEO as there are responsive sites and sites without mobile content in this list of top performers as well; but if we’re following the leaders toward best practices we must consider the option the majority of them use, which is redirecting to a mobile URL.

In conclusion, contrary to much of what you read these days from my mobile SEO colleagues and other advocates of responsive web design in all cases, we found that most of the sites that get the most organic traffic didn’t use responsive design. This could be a positive indication that certain sites should continue to build a separate mobile site because the split link equity is either recognized and properly handled by Google with the Skip Redirect update, or it’s not enough of an issue to make a site less competitive in search.

There’s more to mobile SEO than redirecting or reformatting, of course. There are a number of best practices to consider, some of which are true, and some of which aren’t. I will be going over a study of the smartphone search results at SMX Advanced that tries to get at what is most important in terms of mobile search ranking factors. We released a part of it earlier this year that demonstrates an unusually high number of sites with mobile versions ranking in Google smartphone search, but will be discussing even more ranking factors at the conference. If you have tickets already, be sure to join us June 6, 2012 in Seattle at iSEO: Mobile Search Engine Optimization Done Right and at the Local University Advanced workshop, where I’ll be talking about mobile search ranking factors in detail.

If you’re interested in mobile search and you can’t make it to SMX, be sure to read my blog and/or the mobile search column in Search Engine Land, where my fellow columnists and I write weekly about all things mobile SEO.

Stephanie Chang

Stephanie Chang

Stephanie helped open Distilled’s New York office in June 2011 after working for a year at a New York-based full-service agency. She oversaw the SEO and social media execution for a variety of clients including B2B, B2C, e-commerce and international...   read more

Get blog posts via email

16 Comments

  1. Hi Stephanie

    First very nicely done study and presentation.

    I've always assumed rightly or wrongly that the dilution of link equity would be one of the first items dealt with at the algorithm level as it would be in the best interest of those searching for content or companies.

    For the last 2 plus years, I've been using an iPad for most of my off work browsing. It has been just in the last four months that I've been using a 'true' mobile device a Galaxy Nexus with ICS 4.0.4

    During this time I have surprised myself that the GN is now my device of choice and my habits mirror what I've been seeing in the analytics of the many sites that I have access to.

    Again, from an antidotal observations, I find it very frustrating to browse sites that redirect you to their mobile site or have a plugin that reformats the sites.

    A good example is Danny's <a href="http://www.searchengineland.com" Search Engine Land which seems to use a Wordpress plugin to reformat their content for a mobile device.

    Many times it is difficult to 'share' the page with Google Plus or other social sites. Yes, you can go and turn off the mobile version on some sites but when you return you often have to again go to the bottom of the page and in many will only let you see the mobile version.

    I find most if not all a poor user experience much as if I was browsing a WAP site.

    My belief is that at least for right now a responsive or fluid design is the best way for most SMB website to show their content. In my opinion the majority of times the user experience is of higher quality and maybe just as important the company doesn't have to be burden with another website as an expense.

    reply >
  2. Thank you for your research, it's interesting to see what the big boys are doing with regards to mobile SEO.

    I'd like to mention one thing that may throw the study off a little bit--responsive design, as it is understood now (using media queries to determine dimensions of the viewing area and respond appropriately), is a pretty new idea (the term was coined in May of 2010. See http://www.alistapart.com/articles/responsive-web-design/). Mobile sites, on the other hand, have been around for years (and the big boys could afford to create them). Therefore, judging effectiveness by prevalence could be misleading.

    What I'd like to see is an analysis of paired sites--two very similar sites, one of which used responsive design and another which used a desktop and a mobile site--to see which performs better in normal and mobile SEO. It would probably be fairly difficult to come up with a large sample size, but it could be a more accurate comparison. If I ever see such a thing, I will try to remember to come back and share it here.

    reply >
  3. Bryson,

    Great post, always good to see people posting real world examples.

    There’s a popular theory among SEOs these days that mobile URLs dilute link equity, making the site less competitive for search

    When I helped Cindy with her recent Googlebot-Mobile Smartphone Case Study, as best as I could see at the time having the SERPs optimsed as a result of Googlebot-Mobile didn't change the rankings, it simply sent the user directly to the appropriate mobile URL.

    As such, we're not really getting more or less traffic as a result of having an optimsed SERP. We or more specifically our visitors are getting a better user experience by skipping the desktop URL/redirect and going directly to the mobile URL.

    If a business provides an m.domain.com and the content is identical or substantially similar to the desktop content, they should be using rel="canonical" tags to address the duplicate content issues. As a result of that action, they are also helping to address the link equity dilution as the PageRank from the mobile URLs will be transferred to the appropriate desktop URL nominated via the rel="canonical" tag.

    That doesn't help if the content you're delivering to the mobile website is deliberately different as a result of mobile device keyword research. In this scenario, you either leave the mobile URLs indexed because you want them to rank on their own merits (with or without the SERP optimisation) or you rel="canonical" them back to the desktop URL regardless.

    If you're content is based around mobile device keyword research, since the SERPs rankings aren't affected (at least not that I could tell) by a Googlebot-Mobile optimised SERP, the mobile optimised content isn't going to rank on its own steam even in a mildly competitive environment without link building.

    Which is why we come full circle about one URL to rule them all. Long term I think the best course of action is a single URL, you're not splitting link equity (even if you're managing it using rel="canonical") but more importantly you get the benefit of leveraging the rankings of your desktop URL that will invariably be much stronger than your mobile URLs. The challenge then becomes writing great content for your site that hits the mark for both desktop and mobile device users - which in certain industries varies quite a lot.

    The one URL to rule them all approach is still relatively new and I think it poses a lot of obstacles for many businesses. For example, depending on your industry - simply reformating your desktop content into a mobile template might not be enough to meet the users needs, they might want/require/desire tailored content. The other real problem is the technology component, a lot of content management systems on the market are simply not capable of delivering differ... [continued below]

    reply >
  4. [continued from above] ...ent templates based on user agent, let a lone different content based on user agent over the same URL.

    I also wasn’t sure if any of the sites would be prepared to welcome smartphone Googlebot ... I thought that adoption would be low, if anything.

    I think you'll find that it isn't that they were prepared to 'welcome' smartphone Googlebot-Mobile or that they moved fast after the announcement, it was that they implemented their mobile website according to best practice.

    When we put out mobile website together, it was before Google announced the smartphone version of Googlebot-Mobile. Our implementation used 302 redirects on a per URL basis because it makes the most sense from a users perspective.

    As it turned out, Googlebot-Mobile crawls with an Apple iPhone-esque user agent string. The vast majority of businesses that have invested in a mobile website will have made sure that prominent devices such as an iPhone are catered for well.

    When Googlebot-Mobile arrived with an iPhone based user agent, businesses we already prepared for it because they were doing it right already & Google reinforced that by saying not to do anything special for smartphone Googlebot-Mobile.

    I can't wait to read the findings of your upcoming mobile research, it's fascinating stuff given the incredible growth happening in this space around the world.

    Regards,
    Al.

    reply >
    • Al, thanks for leaving such a thoughtful comment and I’m sorry I didn’t get to it earlier. It’s difficult to keep up with comments unfortunately when you’re travelling and writing as much as I am these days. At any rate, I appreciate it and I’ll try to give you the thoughtful response you deserve.
      This point was especially interesting to me:

      “If a business provides an m.domain.com and the content is identical or substantially similar to the desktop content, they should be using rel=”canonical” tags to address the duplicate content issues. As a result of that action, they are also helping to address the link equity dilution as the PageRank from the mobile URLs will be transferred to the appropriate desktop URL nominated via the rel=”canonical” tag.

      That doesn’t help if the content you’re delivering to the mobile website is deliberately different as a result of mobile device keyword research. In this scenario, you either leave the mobile URLs indexed because you want them to rank on their own merits (with or without the SERP optimisation) or you rel=”canonical” them back to the desktop URL regardless.
      If you’re content is based around mobile device keyword research, since the SERPs rankings aren’t affected (at least not that I could tell) by a Googlebot-Mobile optimised SERP, the mobile optimised content isn’t going to rank on its own steam even in a mildly competitive environment without link building.

      Which is why we come full circle about one URL to rule them all. Long term I think the best course of action is a single URL, you’re not splitting link equity (even if you’re managing it using rel=”canonical”) but more importantly you get the benefit of leveraging the rankings of your desktop URL that will invariably be much stronger than your mobile URLs. The challenge then becomes writing great content for your site that hits the mark for both desktop and mobile device users – which in certain industries varies quite a lot.”


      When I wrote this post you may have had a point, as we didn’t have any guidance at that point what to do with equivalent URLs that were not exact duplicates. I might have argued that in the case of a mobile home page or any specific pages that we create for mobile searchers but not for desktop searchers we’re really trying to rank for concepts and keywords that are most relevant to mobile users. So when the searcher is trying to find the brand keywords, they could go to the desktop site, which would redirect to the smartphone site; but if they’re looking for the brand keywords plus modifiers that indicate they’re looking for a mobile site (which many of them are) they will get a good user experience in search results, which they wouldn’t get if they just built a responsive site. And if they were looking for nonbrand keywords we would simply redirect them appropria... [continued below]

    • [continued from above] ...tely based on the user agent, and they would get the content they’re looking for, regardless of device.

      However, given that Google gave guidance on June 6 of this year on how to properly configure your mobile URLs in this case (https://developers.google.com/webmasters/smartphone-sites/details), it’s kind of a moot point. As it is, Google does recommend adding canonical tags for equivalent URLs, even if they’re not exact duplicates. What’s more, Pierre Far said that Google should be able to rank the proper content for the proper user, even if the canonical tag is on the mobile URL. So given that there’s really no advantage to using responsive web design to create an experience that is better for users on two URLs.

      And I think with their recent announcement Google is acknowledging this as well. Responsive web design, as you say later, might not be enough to meet the users’ needs. As I've said many times in the past on Search Engine Land and elsewhere, different search behavior sometimes requires different content, and if Google, who has historically bent over backwards to ensure their search results provide a stellar user experience, suddenly penalizes those who want to provide the best user experience even if it’s more work for them, will be setting a precedent that won’t ultimately be beneficial for users of their product. As it is they stated a preference, but also left it up to the site owners on whether or not that preference best serves their users.

      So I’m ultimately less optimistic about the long term prospects of responsive web design than you are, I think. It works great for duplicate pages, and it’s a lot less maintenance for webmasters, but I completely agree that there are industries for whom responsive web design is not the right answer, and I don’t see that changing any time soon. It seems to me that Google had the right idea by saying use responsive web design if it works for your users, but if it doesn't we'll still be able to handle your content appropriately if you tag it properly.

      Thanks for the clarification on smartphone Googlebot. You may be right that these sites are just treating it like an iPhone and sending it to iPhone content rather than doing anything special after the announcement. Great point.

      And thank you, sir, for the discussion. It’s nice when I encounter others who think critically about these new issues rather than approaching them with preconceived notions about what’s right and wrong for SEO. It’s really only through interactions like this that we can reach any meaningful consensus as an industry and best serve our businesses and our users.

      Best,
      Bryson

  5. I agree with Mike Elbert above that judging effectiveness by prevalence could be misleading.
    I would really want to see a bigger set of data of more recent sites that use both types of mobile effectively.
    But this is very helpful and makes me think once more which route to take with mobile.

    reply >
  6. I too agree with Mike Elbert too. I've have found Google's John Mueller http://goo.gl/a22cN to be a very good resource for questions regarding the crawling of sites by the different Google Bots

    reply >
  7. I asked John Mueller about separate mobile site vs responsive designed and this was his answer
    http://goo.gl/6kubk

    reply >
  8. Thanks, all, for your comments.

    Mike Ebert, I had actually contemplated doing that study before we did this one, but thought it would be nearly impossible to form an apples to apples comparison since web sites are going to have history and link equity and hundreds of other factors that are going to affect ranking. If we found two sites that were redesigning for mobile, they would have to compete for the same keyword(s), have the same link equity and social shares, use the same content, etc in order for us to attribute their success or failure to responsive design or mobile URLs. If you find a study like that, please share it, but I have my doubts about whether a meaningful study like that can be done.

    Your point is well taken about judging effectiveness by prevalence being misleading. My point in doing the study was not to say responsive design can't be effective for SEO, but only to show that mobile URLs can. The fact that responsive design is used by even 12 sites in the sample is enough in my mind to show that sites using responsive design can generate a lot of organic traffic. But the fact that the majority of the sites used mobile URLs should also demonstrate that mobile URLs can be effective for SEO as well. There are many in our industry who still believe that responsive design in all cases is the way to go for SEO, but it becomes more difficult to defend that position when so many sites that get so much organic traffic from Google use mobile URLs. No?

    Runner2009, thanks for sharing. I know John Mueller of Google has said both approaches can work in the past as well, and that's no different than what I'm saying here. Mostly objecting to the idea that mobile URLs result in a loss of organic traffic due to diluted link equity as a result of duplicate content, which doesn't appear to be the case for most of the sites in this sample of top performers.

    For what it's worth, I advocate a hybrid approach that uses responsive design for duplicate URLs and mobile content on mobile URLs for the homepage and other pages intended for mobile searchers. Seems to solve issues with both responsive design and mobile URLs, as neither method is perfect by itself.

    reply >
  9. cara

    Is it best to serve a 301 or 302 redirect to the smartphone Googlebot?

    reply >
  10. cara,

    Google say it doesn't matter if you use a user agent based 301 or 302 redirect, so long as it's consistent.

    I prefer to use a 302 in case another bot like Bing happens to discover the user agent based redirect and treat it differently than Google do; in which case you wouldn't be hurting your rankings in Bing if they handled it differently.

    Al.

    reply >
  11. Bryson/Stephanie,

    Thanks for the mobile SEO guidelines in your article. The merits of responsive vs separate mobile site are clearly discussed and evaluated, but what about one other alternative:

    Same URL ("one URL to rule them all" approach), but not merely displaying different content when viewing a page from a mobile device by utilising a different style sheet for a whole lot of display:none, but rather:


    detect user-agent at a http request/response level
    if mobile user-agent, return only a subset of the content that is more suitable for the user experience on a mobile device (E.g. strip out bloaty header/footer/sidebar/stuff) - i.e. only show the core product/textual content


    This is the approach we'd like to use for the mobile version of our car rental website.

    Any pros/cons to consider if we take this approach?

    My main concern lies around the fact that Google might somehow mix up the "full" version of a page with the "lite" version of a page - which could have some serious negative impact on our current rankings in desktop search - especially if the mobile pages have:


    much lighter content
    much less internal linking between pages


    Any other thoughts or things to be aware of?

    reply >
    • Andre Van Kets - this is where Google compares the page's content based across its traditional, feature phone and smartphone crawlers. Personally I'd have no issue with that.

      Another thing that we need to ensure, when talking about separate URLs is the correct use of canonicals and rel="alternate".

      Lee

  12. Andre,

    What you're referring to is adaptive rendering, where the template actually changes based on different criteria such as a user agent. You see this regularly these days on WordPress sites that use one of the mobile plugins; same URLs & it delivers the content via a different structural layout (ie, less of everything - not just hidden with CSS).

    I think your concern about Google getting confused is legitimate, however with the announcement that they recommend one URL to rule them all & responsive design, they've added another tool for webmasters to 'hint' to Googlebot what is going on with their mobile implementation - the Vary HTTP response header.

    In the link I referenced above you'll see Google recommend that if you're doing conditional redirects based on user agent, that you also include a Vary: User-agent in your HTTP response header to notify Google that the site/user experience is changing based on the device being used.

    While I haven't implemented the HTTP Vary response header yet, I would recommend that you apply that header in the response when you adapt the design based on user agent. That way Googlebot-Mobile using a smartphone user agent will get the message correctly & I expect you shouldn't have any issues with Google getting confused about it.

    Cheers,
    Al

    reply >
  13. Hey ,

    I am agree with Mike Elbert as well. Most of the experts forget to ping the website for mobiles as well. It has been seen lately that most of the users and Internet marketers or Idea searchers are using Mobile phones and most of the websites are not properly optimized for Mobile SEO.

    Thanks a lot writing this guest post. A lot of points got cleared here.

    Thanks

    Saif

    reply >

Leave a Reply

Your email address will not be published. Required fields are marked *

*
*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>