Why We Shouldn’t Make Separate Mobile Websites


There has been a long-running war going on over the mobile Web: it can be summarized with the following question: “Is there a mobile Web?” That is, is the mobile device so fundamentally different that you should make different websites for it, or is there only one Web that we access using a variety of different devices? Acclaimed usability pundit Jakob Nielsen thinks that you should make separate mobile websites. I disagree.

Jakob Nielsen, the usability expert, recently published his latest mobile usability guidelines. He summarizes:

“Good mobile user experience requires a different design than what’s needed to satisfy desktop users. Two designs, two sites, and cross-linking to make it all work.”

I disagree (mostly) with the idea that people need different content because they’re using different types of devices.

Firstly, because we’ve been here before, in the early years of this century. Around 2002, the huge UK supermarket chain Tesco launched Tesco Access—a website that was designed so that disabled people could browse the Tesco website and buy groceries that would be delivered to their homes.

It was a great success—heavily stripped down, all server-generated (as in, those days screen readers couldn’t handle much JavaScript) and it was highly usable. One design goal was “to allow customers to purchase an average of 30 items in just 15 minutes from login to checkout.” In fact, from a contemporary report, (cited by Mike Davis), “many non-disabled customers are switching from the main Tesco site to the Tesco Access site, because they find it easier and faster to use!” It also made Tesco a lot of money: “Work undertaken by Tesco.com to make their home grocery service more accessible to blind customers has resulted in revenue in excess of £13m per annum, revenue that simply wasn’t available to the company when the website was inaccessible to blind customers.”

However, some blind users weren’t happy. There were special offers on the “normal” Tesco website that weren’t available on the access website. There were advertisements that were similarly unavailable—which was a surprise; whereas most people hate advertisements, here was a community complaining that it wasn’t getting them.

The vital point is that you never know better than your users what content they want. When Nielsen writes that mobile websites should “cut features, to eliminate things that are not core to the mobile use case; [and] cut content, to reduce word count and defer secondary information to secondary pages,” he forgets this fact.

Tesco learned this:

“We have completely redesigned Access so that it is no longer separate from our main website but is now right at the center of it, enabling our Access customers to enjoy the same features and functionality available on the standard grocery website. As part of this work we have had to retire the old Access website.”

Nielsen writes:

“Build a separate mobile-optimized site (or mobile site) if you can afford it … Good mobile user experience requires a different design than what’s needed to satisfy desktop users. Two designs, two websites, and cross-linking to make it all work.”

From talking to people in the industry, and from my own experience of leading a dev team, I’ve found that building a separate mobile website is considered to be a cheaper option in some circumstances—there may be time or budgetary constraints. Sometimes teams don’t have another option but creating a separate website due to factors beyond their control.

I believe that this is not ideal, but for many it’s a reality. Re-factoring a whole website with responsive design requires auditing content. And changing a production website with all the attendant risks, then testing the whole website to ensure it works on mobile devices (while introducing no regressions in the desktop website)—all this is a huge task. If the website is powered by a CMS, it’s often cheaper and easier to leave the “desktop website” alone, and implement a parallel URL structure so that www.example.com/foo is mirrored by m.example.com/foo, and www.example.com/bar is mirrored by m.example.com/bar (with the CMS simply outputting the information into a highly simplified template for the mobile website).

The problem with this approach is Nielsen’s suggestion: “If mobile users arrive at your full website’s URL, auto-redirect them to your mobile website.” The question here is how can you reliably detect mobile browsers in order to redirect them? The fact is: you can’t. Most people attempt to do this with browser sniffing—checking the User Agent string that the browser sends to the server with every request. However, these are easily spoofed in browsers, so they can’t be relied upon, and they don’t tell the truth, anyways. “Browser sniffing” has a justifiably bad reputation, so is often renamed “device detection” these days, but it’s the same flawed concept.

On mobile, Twitter.com automatically forwards users to a separate mobile website.

More troublesome is that there are literally hundreds of UA strings that your detection script needs to be aware of in order to send the visitor to the “right” page. The list is ever-growing, so you need to constantly check and update your detection scripts. And of course, you only know about a new User Agent string after it turns up in your analytics—so there will be a period between the first visitor arriving with an unknown UA and your adding it to your detection scripts (in which visitors will be sent to the wrong website).

Despite all this work to set up a second parallel website, you will still find that some visitors are sent to the wrong place, so here I agree with Nielsen:

“Offer a clear link from your full site to your mobile site for users who end up at the full site despite the redirect … Offer a clear link from your mobile site to your full site for those (few) users who need special features that are found only on the full site.”

Missing out features and content on mobile devices perpetuates the digital divide. As Josh Clark points out in his rebuttal:

“First, a growing number of people are using mobile as the only way they access the Web. A pair of studies late last year from Pew and from On Device Research showed that over 25% of people in the US who browse the Web on smartphones almost never use any other platform. That’s north of 11% of adults in the US, or about 25 million people, who only see the Web on small screens. There’s a digital-divide issue here. People who can afford only one screen or internet connection are choosing the phone. If you want to reach them at all, you have to reach them on mobile. We can’t settle for serving such a huge audience a stripped-down experience or force them to swim through a desktop layout in a small screen.”

The number of people only using mobile devices to access the Web is even higher in emerging economies. Why exclude them?

Mobile Usability

I also agree with Nielsen when he writes:

“When people access sites using mobile devices, their measured usability is much higher for mobile sites than for full sites.”

But from this he draws the wrong conclusion, that we should continue making special mobile websites. I believe that special mobile websites is like sticking plaster over the problem; we generally shouldn’t have separate mobile websites, anymore than we should have separate screen reader websites. The reason many “full websites” are unusable on mobile phones is because many full websites are unusable on any device. It’s often said that your expenditure rises as your income does, and that the amount of clutter you own expands to fill your house however many times you move to a bigger one. In the same way, website owners have long proved incontinent in keeping desktop websites focussed, simply because they have so much room. This is perfectly illustrated by the xkcd comic:

A Venn diagram showing 'Things on the front page of a university website' and 'Things people go to the site looking for.' Only one item is in the intersection: 'Full name of school.'
A Venn diagram showing “Things on the front page of a university website” and “Things people go to the site looking for.” Only one item is in the intersection: “Full name of school.” Image source: xkcd.

As I wrote on the website The Pastry Box on April 13th:

“The mobile pundits got it right: sites should be minimal, functional, with everything designed to help the user complete a task, and then go. But that doesn’t mean that you need to make a separate mobile site from your normal site. If your normal site isn’t minimal, functional, with everything designed to help the user complete a task, it’s time to rethink your whole site.

“And once you’ve done that, serve it to everyone, whatever the device.”

In a previous article, Nielsen wrote in September 2011 that he dropped testing usability with featurephones:

“Our first research found that feature phone usability is so miserable when accessing the Web that we recommend that most companies don’t bother supporting feature phones.

“Empirically, websites see very little traffic from feature phones, partly because people rarely go on the Web when their experience is so bad, and partly because the higher classes of phones have seen a dramatic uplift in market share since our earlier research.”

This is a highly westernized view. Many people can’t afford smartphones, so they use feature phones running proxy browsers (such as Opera Mini), which move the heavy lifting to servers. This is often the only way that underpowered featurephones can browse the Web. Statistics from Opera’s monthly State of the Mobile Web report (disclosure: Opera is my employer) shows that lower-end feature phones still dominate the market in Eastern Europe, Africa and other emerging economies—see the top 20 handsets worldwide for 2011 that accessed Opera Mini. Since February 2011, the number of unique users of Opera Mini has increased 78.17% and data traffic is up 142.79%.

A caveat about those statistics: not every user of Opera Mini is a featurephone user in developing countries. They’re widely used on high-end smartphones in the West, too, as we know that they are much faster than built-in browsers, and users really want speed.

Nielsen’s dismissal of feature phones reminds me of some attitudes to Web accessibility in the early 2000′s. His assertion that companies shouldn’t support feature phones because they see little traffic from feature phones is the classic accessibility chicken and egg situation: we don’t need to bother with making our website accessible, as no-one who visits us needs it. This is analogous to the owner of a restaurant that is up a flight of stairs saying he doesn’t need to add a wheelchair ramp as no-one with a wheelchair ever comes to his restaurant. It’s flawed logic.

Developing Usable Websites For All Devices

The W3C Mobile Web best practices say:

“One Web means making, as far as is reasonable, the same information and services available to users irrespective of the device they are using. However, it does not mean that exactly the same information is available in exactly the same representation across all devices. The context of mobile use, device capability variations, bandwidth issues and mobile network capabilities all affect the representation. Furthermore, some services and information are more suitable for and targeted at particular user contexts.”

There will always be edge cases when separate, mobile-specific websites will be a better user experience, but this shouldn’t be your default when approaching the mobile Web. For a maintainable, future-friendly development methodology, I recommend that your default approach to mobile be to design one website that can adapt to different devices with viewport, Media Queries and other technologies that are often buzzworded “Responsive Design.”

Combining these techniques in a smart way with progressive enhancement allows your content to be viewed on any device (and with richer experiences available on more sophisticated devices), with the possibility of accessing device APIs such as geolocation, or the shiny new getUserMedia for camera access.

Although many other resources are available, I’ve written “Mobile-friendly: The mobile web optimization guide” which you’ll hopefully find a useful starting point.

Further Reading

(jvb) (il)

In your experience, what kind of “mobile websites” do you create most often?

↑ Back to top

Bruce Lawson evangelises open web technologies for Opera. He co-authored Introducing HTML5, the best-selling book on HTML5 that has just been published in its second edition. He blogs at brucelawson.co.uk.

  1. 1

    This is a very productive debate on technical organization and capabilities of competing approaches. It’s also interesting to consider the business implications of single site vs. segregated sites. For better or worse, organizations tend to budget and spend in alignment with site organization, and in planning these expenditures, the benefit represented by an audience is a true consideration. Functional segregation along the lines of mobile and desktop “enables” budgetary segregation by audience. When the features are prioritized separately, this can have the very unintended effect of denying discounts to blind people, as cited in the original post.

    I think it’s the right thing to support a unified site, because it puts all the stakeholders into the same investment bucket, for better or for worse. We are already at a point where the ROI represented by user audiences is tipping in favor of mobility. That said, the desktop is set to become the next usability slum, as investments in mobility begin to eclipse those made against desktop oriented sites. Is a blind person more likely to be a desktop or a mobile user? The answer might change often. The answer is not relevant, but it does bring to light the complexity of feature prioritization in the face of the resource, budgetary and schedule constraints we are all juggling. It makes sense to favor device agnosticism and unity, if only to avoid unfavorable impact of unintentionally segregating and denying benefit to an audience, which may also very well be one that is a constitutionally protected class of people.

  2. 2

    Love this article; you captured all the reasons we decided to go responsive with intridea.com. Creating and maintaining multiple mobile solutions just wasn’t pragmatic.

    Though redesigning the site to be responsive was no small undertaking we certainly saved a lot of time and money in the long run. Additionally, I’ve found everyone who sees our site is struck by the novelty of the responsive design!

  3. 3

    Who is the person trying to reach your web world?

    First Timer? Are they a regular to or, have been to your website before?

    Are the expectations / experiences of mobile device users different for repeat traffic -vs- new traffic? (to your web world. mobile or otherwise)

    Will you lose current website traffic if mobile web expectations aren’t met?

    This post is less than 3 months into the future since the article was posted and analytics suggest YES, businesses ARE losing customer base due to poor mobile web performance.

    Are ALL mobile users needs the same for ALL businesses?

    Does a business use the mobile web to attract a prospect to later come back on laptop or work station for more details?

    So many questions. I think the answer is that there is no ONE answer.

    … and as luck would have it, once we figure out how to plug and chug away with mobile web applications, our portable devices will project a hologram screen and we will be back full circle asking… “how did we used to maximize full screen websites?” lol.

  4. 4

    One should go for responsive design iff speed & performance does’t matter… Mobile devices are comparatively slower than computer. Seperate design proves to be faster and more usable & easy to implement…

  5. 5

    …”Earlier in this century…” Trying to impart historical grandeur to your opinion?

  6. 6

    Again…I find it strange that some that disagree are the ones that are missing the point about having critical information RIGHT in front of your EYES..on ANY platform.

    Example….how often do you want to know about the company, contact names. address and a real phone number or persons name to call?…..In the real world that is in the high 90% range..

    Remember the Yellow pages….learn from the past ..people are still people not binary code

  7. 7

    Just voted for responsive design. Even more than just for usability purposes this is definitely the ideal approach to take for a site in terms of ability to achieve quality placement in the major search indices (whether on a desktop or digital device query).

    More on this here: http://www.seo.com/blog/the-best-way-to-develop-an-optimized-mobile-website/

  8. 8

    My view is also that there is no right or wrong answer to this question. It will largely depend on what your website is about, what its purpose is, the information architecture of the site etc..

    In some cases, responsive design will be the best move, but in other cases, a mobile specific site will be the far better approach.

    My own site is a good case in point. It is a portfolio site designed to show off my work, with a horizontal scrolling JS slideshow that is perfect for desktop and laptop screen sizes and browsers.
    It still functions well on mobile browsers, but the design required a new approach for a better mobile experience. The best approach was to create a mobile specific portfolio site, that was vertical in design, simpler, and without using any JS.

  9. 9

    It just goes to show that you can’t please everyone and sometimes site authors try and force feed their way of accomplishing a particular task. While I appreciate all the articles I read–because they do provide great information–let’s not forget that my audience isn’t necessarily your audience and we are probably both trying to achieve a different goal.

    I do like these kinds of articles however, because they often turn the lights on in our heads and help us build upon our thought processes. I guess what I’m trying to say is what works for you may not work for me and so on. You may like to build “responsive” pages and I might like to build a separate mobile site. Why? Because we’re all in different situations. Not all of us work for design and developer firms where we have to listen to someone telling us what to do. Some of us do this for fun – like me ;)

    Information is good and helping each other is good, but remember we’re all working on something different and what you’re working on in no way shape or form has any bearing on whether I build a separate site or not as long as I’m achieving my end goal.

  10. 10

    One thing that seems to go ignored is the most important aspect of site design – lead conversion. When you build a full site, with all your calls to action at the top and on the side – this can disappear with responsive designs. Thus your primary tactics to get people to submit their info can literally be non-existent. This means you have to put those tools within the content. I’ve been doing this more and more but it does create a situation where I have multiple lead gen forms on every page in an effort to ensure every visitor knows how to take action or contact me.

    Keep in mind, I don’t use Adwords so my type of focus is different than most. My goal is to find people who’re seriously interested in receiving access to something or someone that costs money. Thus, every visitor could be a handsome reward. I can’t afford to lose that person’s information because my site was shrunk on their phone. LOL.. I’m NOT necessarily saying responsive is bad but this site is a great example. You’ve got a crap load of paid advertisements on your right sidebar. The problem is most people aren’t seeing that side bar because it squishes down in certain resolutions.

    My personal choice is to continue to design for desktop, but then add in tools and techniques that cater more toward mobile users. I have to build for both, but more importantly, I have to convert both!

  11. 11

    I agree about mobile site’s what is the point!

    Basically design one site, optimize it for all devices which is easy by the way just validate your site you website will and should run fine on most devices long as you do not have big image files on your website or any un-optimized elements, instead of worry about devices what your sites should really be thinking is user Interactivity, we are now living in a touch screen age on many devises soooo….

    Shouldn’t we be worry about sites that are more touch friendly rather than devices friendly?

    • 12

      What about design features that work great with a PC/desktop /mouse interface but poorly on the mobile, and vice-versa? Should you throw away a great bit of software just because not everybody can use it?

  12. 13

    I love how this was posted on a terribly designed mobile site which doesn’t use the correct width in landscape or portrait. I think that mobile web sites shouldn’t exist we are in an era where our smart phones out power some desktop computers so f*ck off mobile site builders, and stop FAILING at tring to make your site eaiser to use my making an unfinished site where you cannot zoom, has very buggy code, and a bad CSS template…

  13. 14

    mobile sites should die. my site is completely readable on any phone I’ve tried. For one, I keep simple. Tables mostly. Also, I keep the max width at 900 pixels. bagtoss . com

    I’m actually trying to find out if when viewing on a phone, if the network you are on or the browser manipulates the code for some reason. I have a Samsung Galaxy S4 and there is an option for viewing the desktop version. I only have one version and when I select this option, the picture viewer I have starts to work correctly. It is possible the picture view has code for mobile versions but the developer claims it does not.

    From the developer FAQ page: “Why can’t I view my gallery on a 3G mobile connection?

    Some mobile networks perform ‘Content Modification’ on HTML content viewed over a 3G connection. This may cause Juicebox to break. Most networks add an option to ‘Request desktop site’ which bypasses the issue. ‘Content Modification’ by mobile networks not only breaks Juicebox, it breaks many websites that uses JavaScript. We are aware of this issue and are investigating possible solutions.”

    If this is truly the case, then is there some code I can add to the page to tell phones to stop it! Just render it in desktop view.



  14. 15

    Nothing irks me more on the web when a mobile site is dumbed-down and stripped of content, and to get the information I need I have to click on “go to full site”.

    Users are impatient, they don’t have time to hunt through a clumsy mobile interface. If a user can’t get to the content they want on a device in the same time as on a desktop, than it’s a design failure.

    Why pay so much for these smartphones and cell plans when you often get a lesser version of the web with one?

    Responsive design is the way to go.

  15. 16

    I spoof my web browser’s user agent (Chrome) as a means to avoid awful redesigns of blogs, such as Ars Technica, there’s this sickness going about, fixing what’s not broken, Gawker’s done it several times, too bad they don’t have an alternative mobile site, now I just don’t visit any of their sites (Lifehacker, Gizmodo, etc…) because their design layouts came out of their asses. I also spoof Google Search, so it’ll serve their mobile CSS instead, it just looks neat and well organized.

  16. 17

    Oh, the irony of having to type this comment after having had to scroll through 300 miles of ridiculously-oversized text, on a phone with a full browser and broadband connection that would be perfectly capable of showing your full website if only you hadn’t listened to some idiot ‘UX Designer’ who convinced you that forcing this special needs version of your content was a good idea for some reason.

  17. 18

    Abdalhamid AlAttar

    January 5, 2014 4:16 am

    You can’t create a ruler and make it work for everything. It’s not about do it or don’t do it. Part of analyzing your needs is to decide which approach to take. Some websites better to be separated, some better to be responsive. Others better to have mobile apps. It’s all dependable variables, you can’t just say Yes or No before you study what you want to do.

  18. 19

    This particular web site is an example of why the author might be wrong.

    The text is rather large on a PC monitor, and the amount that fits on one screen is a bit minimal. I feel like I’m reading something written in a children’s book with big friendly letters for kids.

    The same layout would be OK if the text just resized itself a little better and made use of more of the screen. But there’s no question that in trying to make something for the mobile, the desktop experience has suffered greatly.

    A paint vendor’s site used to have a fantastic colour pallete to allow to to quickly find any shade you wanted. That excellent idea has been binned to make it mobile friendly, you have to navigate back and forth between loads of pages and you can’t see all of the colours together.

    The desktop is the rich experience, the mobile is the common one. Don’t try to make them the same at the cost of losing the slick stuff.

  19. 20

    If you cannot find what you want on a website, regardless of whether it is a desktop or mobile version, that is purely down to the way in which that particular site has been structured. I would say that it is far easier to ‘drill-down’ into the required content of mobile sites that I have seen (done).

  20. 21

    I think you kinda missing the the point here, yes I too prefer to see a desktop site on a tablet because of the screen size! But sites for mobile users is a must,,, if only for load time! A link to the desktop for further info is a good idea! x

  21. 22

    That’s right Jeanna, Asking the right questions, but to whom? The user, of course. So I’m with Jacob; useability studies rock. You can sit in your office testing all the browsers and screen sizes under the sun. You think you’ve got it right until you let that baby out into the real world.
    Then you find out it’s back to the drawing board.

  22. 23

    True, future will be different hopefully better. But what about the billion devices out there now ?

  23. 24

    So enjoyed your response. Especially this part.

    “I also have scant evidence that people care about this. They care about the experience, regardless of how you deliver it. They care that it works on their handsets. They care that it allows them to achieve their tasks or goals. They care whether the contact options suit their budget. They care what your page load time is like. They don’t care if it’s a separate mobile website, as long as it is built for their needs.”

    Thank you for your wisdom and insight regarding this matter.


Leave a Comment

Yay! You've decided to leave a comment. That's fantastic! Please keep in mind that comments are moderated and rel="nofollow" is in use. So, please do not use a spammy keyword or a domain as your name, or else it will be deleted. Let's have a personal and meaningful conversation instead. Thanks for dropping by!

↑ Back to top