Nielsen vs Clark - they're both wrong

I'm glad somebody is finally saying this. Almost all of the arguments against Nielsen's statements have been fully of hand-wavey speculation, emotional arguments and a billion other logical fallacies. Whilst Nielsen himself (as you pointed out) fell into the trap of declaring that a separate mobile site is always appropriate.

It's naive to assume that the right information architecture for desktop is also right for mobiles. It sure would be nice if IA could cleanly be treated as completely independent pages, but that's not what IA is. If your data and other research tells you that the majority of users on mobile have different goals (for the most part) to desktop users, you're doing them a disservice by deciding that a responsive layout is enough.

Jason, great points. I think you make a solid case for letting data drive decisions. I do fall in the responsive camp after nightmarish experiences both developing mobile sites, and using mobile sites.

So, let me ask a couple of questions ( not challenges) about the examples you brought up:
The Bank site:
- What percentage of desktop users go directly for the log in? I bet it is high, and my bank has a log in on the home page, wouldn't all bank sites?
- Why not use responsive to make the log in more prominent for mobile users?

If it is important enough to mobile users to have a link to the "full" site, why then should the user then be faced with a monstrosity that is difficult to use on a mobile?

Additionally, in site switching logic, how do you decide which version of the site the user receives? Is it based on User Agent? Because on my 10" tablet, I am constantly served a mobile site because of the Android User Agent, even though it has the same size screen and better resolution than my 10" netbook. Often times, when I click the "Visit the Full Site" link, I am directed to the home page of the full site, losing my place, which may be difficult to find again. Adding to my frustration, many times, clicking deeper into the full site, I am once again redirected to a mobile site.

As mobile devices continue to improve, and change size, targeting based on User Agent is becoming more useless, and provides a worse user experience. Not to mention the continuous dev work of trying to figure out what a device is capable of, and tailor the experience to that.

I do appreciate the idea that there may not be a "right" answer, and that both camps seem to be falling into the religiosity of their technology. Maybe, however, responsive is not just a technology solution, but a maturing of how we consider user experience and front-end design.

Good conversation!

Scott, great questions!
Almost all of the mobile users on the bank sites go RIGHT to login.
> Why not use responsive to make the log in more prominent for mobile users?

They run different ads to their mobile users, so now we have a responsive site where some things appear on the page ONLY for

In addition, banks (by their nature) have multiple tables on their website. Data tables with rates and things. Formatting this VERY IMPORTANT data to be accurate is cumbersome with a responsive site. We discussed making it a responsive site and making all rate tables PDFs, but that didn't seem right either.

In this solution we were able to make an affordable site that looks good and functions the way we want for 95% of all users. Responsive would have taken a lot more work, and I don't think it would have been a *better* experience for anyone.

RE: User Agent
I know, I wish there was a better way. Right now well below 1% of their site traffic is Android tablets, so we opted to make it as easy for the 99% as possible, knowing that in the future if people start using Android tablets we're going to have to change what we're doing.

Again, didn't seem worth investing a lot of time (and money) in a "technically" perfect solution which didn't actually make life easier for a significant number of users...

In my mind part of usability is having the balls to say "we're going to prioritize this user over another user". If you try to make something for everyone, it won't ever be great for anyone. If you can make it great for *most* of your users today, I always opt for that over making it "OK" for everyone...

Interesting and funny fact is that .net magazine exactly does what Nielsen suggest and Clark opposes. Access .net from smartphone and you'll get redirected to dedicated mobile link and website.

But what about mobile apps? If Clark were correct nobody would and nobody should create mobile apps, they should all just give us their desktop website. He's right, though, that most apps lack the features of their desktop counterpart.

Just today I updated the "meetup" app on my Android and looked up one of my groups. Actually I couldn't because you can only see and find actual events but not any group's details, member, browse discussions or whatever. We all know the Facebook apps sucks big time in comparison to the desktop, Twitter as an absolute minimalism website isn't capable to provide all its feature to their app. The list goes on and on.

Totally agree that decisions about how to prioritize content should be made based on real data. I hear too many people making assumptions about what they think the "mobile user" will want.

That said, organizations should still aim for providing more content rather than less:

* Content that is accessed by only a small number of readers can still be INCREDIBLY important to those people. Analytics data is a scalpel, not a machete.

* You can't generalize what people want based on the current bad experience offered by most mobile websites. You're never going to get accurate data about how to prioritize UNTIL you get all your content up.

* Many CMSes don't support multi-channel publishing, so having two versions means you're doubling your workload. Forking your content is inherently, categorically bad. You're better off publishing exactly the same content as on the desktop site until you can figure out how to make your content adaptive on the backend.

* If the content really isn't useful, why do you have it on the desktop site? Better to have a holistic process for pruning outdated or irrelevant content.

I believe that the golden rule is usability. If the web site is usable (including aesthetically pleasing). it will make it easy for the users to achieve their goals. Isn't that what a web site is for?

Hi Jason,

Many thanks for the thoughtful critique and useful advice.

I fear that my essay made me seem absolutist or dogmatic, which I am not. In a moment where there's so much churn and change in both technology and culture, I feel there's little place for dogma or religious fervor.

In particular, I regret if I gave the impression that I'm in "the responsive design camp." I am not. In fact, I did not realize that such a camp exists. Responsive design is a tool, not a philosophy. It's a means, not an end. If there are folks who are saying you must always use this tool, that all sites must be responsive, then that's not right, either.

I've written many times before that building a separate mobile website vs a responsive website depends on the project at hand -- and yes, ABSOLUTELY that should be data-driven.

If I may, I'd like to recommend this article I wrote some time back: "Responsive Web Design or Separate Mobile Site? Eh. It Depends."
http://globalmoxie.com/blog/mobile-web-responsive-design.shtml

While I strongly believe there's no place for dogma, I do think it's important to locate guiding stars -- the assumptions and beliefs that gird us in our research and data-gathering and design principles. An especially bright guiding star for me is equal access to content, no matter what your platform. Nielsen's guidelines riled me because they default to denying content to certain platforms. My simple notion is that the default should instead be to allow access to all content from any platform. What can I say, I'm a glass-is-half-full guy.

For the general case, all the latest data and research points to the fact that phones are becoming a primary (in many cases, exclusive) browser for many users... 25 million users in the US and rising. In less than two years, if current trends continue, more people will access the web via mobile than via the desktop. This indicates that in very short order, the desktop will no longer be the primary web platform (and in fact, the days of a single primary platform may be over).

I suggest our default assumption should be that content should be accessible and well formatted on all platforms. That assumption may shift according to the project at hand, but I believe that should always be our starting assumption.

I do not believe that the information architecture or presentation of a site should necessarily (or even usually) be the same for a mobile version -- or a TV version or a speech-driven version or for any other platform -- as for the desktop. Some use cases may be more popular for mobile than for desktop, and vice versa, and the priority and presentation of content should reflect that. For your university or bank, yes, OPTIMIZE the experience for the use cases you mention. Prioritize those tasks.

But cutting content outright, making it inaccessible (or inscrutable) to mobile devices, is likely to be a mistake in most cases. Jason, do I have it right that your mobile-only banking customers can't access a huge swath of content about loan information, exchange rates, or other services unless they swim through the desktop version? I honestly don't believe that's ideal; I suspect that in a very few months your user stats will bear that out. We've already seen poor results for a number of commerce sites which prevent the ability to purchase certain product lines, denying a purchase to customers:
http://globalmoxie.com/blog/mobile-commerce-belly-flop.shtml

I do take your book counterpoints and appreciate them. One would hope, though, that the reader of the Cliff's notes or the abbreviated paperback would also have access to the full version of the book somewhere, somehow. Mobile-only users, however, would find themselves wholly cut off from content if it were not available to them via their device -- or at the very least would experience it wholly differently if forced to read it in a format not appropriate to the device.

A clarification: while it's true that I'm an author and a blogger, those roles are only incidental to the way I fill my days: doing design work and design strategy for my clients. I do reams of user research and usability testing. I assure you my views are not based on hand-waving or speculation or assumptions, but real data and observation, supplemented by the wealth of survey data about general use trends. But then again, so are Jakob Nielsen's.

Data is only as good as the assumptions that underly its analysis and gird its conclusions. While I agree with the symptoms Mr. Nielsen saw (that people are easily put off by overwhelming amounts of content on the small screen), I simply disagree with his cure for those symptoms.

Mr. Nielsen's cure is to cut limbs off the patient, cutting content for the mobile platform. I would suggest that it's better to simplify before suppressing. We should not use a customer's device context as an excuse to avoid difficult design decisions. And basing decisions on the generally poor state of mobile experiences isn't a good guide, either. As Karen McGrane writes above, "You can't generalize what people want based on the current bad experience offered by most mobile websites." Well put.

Siloing and suppressing content should be our last resort. Separate websites may very well be needed for technical reasons and of course when building very separate apps. But if you build a separate website with the express reason of eliminating content or features, I suggest that in most cases that will be the wrong approach. This is the advice I give to my clients, but of course, your mileage may vary.

Thanks again, and all the very best.
Josh Clark

Great post.

We at tagitup.net have been suggesting to our clients for the past 18 months basically exactly what Mr. Nielsen suggests, not as an extreme approach but as a sound starting point. We feel mobile users are best served by providing them with what they seek and only what they seek. For our service this often starts with a QR code found in the real world. From where we place that code and how we frame it, we can anticipate what a user may be seeking when they scan it. "Scan this for yada yada". Sure that means that we need to be fortune tellers in a sense. But that is the game of understanding your clients needs. If we say you'll get yada yada, they we deliver exactly that, and a little more just for good measure.

Now the data may not support this. I mean that if we start looking at things like "time on site" etc, we will obviously see less time on site for minimized mobile sites. But we feel this is a good thing! If they just bounce, we failed to provide what they expected. If they drill down then we succeeded. tap-tap-like, tap-tap-buy, tap-tap-call. We do provide links to a full site, but most often those links are ignored when browsing on mobile, as we would expect.

We try to serve the people's needs first. Our stance is that happy browsers who feel they haven't been fed a "bait and switch" are more likely to view the company favorably, and therefore more likely to engage and return.

Though the data is clearly essential to understanding the broad scale utility of the sites, it doesn't always translate directly to user experience.

Thanks for a great discussion.

I got so bored writing a rather entertaining response that would have had you take a magical journey through the gum drop forest on a unicorn that I gave up, much like this debate. There is no answer it really does just depend on the service, the objectives of the customer and sadly but most critically the outlook of the business it is related to. As leisa reichel said behind every shit user experience is a shit company.

I'd like to hear the unicorn story actually. bummed.

I can't help but think that this article missed the point in two key ways.

First, Jason Mark summarizes Clark's argument as "Always build a single responsive site." In reality, Clark's article advocated the same sort of analytics-driven decision making that led the author of this article to build separate mobile sites for his client. The key take-away from Clark's article, in my opinion, was the fact that there is no single "mobile use case" all companies can target.

Figure out how your visitors are using your mobile site (or how they'd *like* to) and tailor it to that. If responsive design techniques work, use them. If a separate site with pared down content is what they want, that's the right way to go. The danger is in *assuming* that a mobile visitor has a particular task in mind simply because of her user-agent.

Which brings us to the second problem: the Cliff's Notes comparison. The danger isn't that somewhere, someone is presenting truncated content. As Jason Mark noted, many a harried student relies on Cliff's Notes to absorb a work quickly. Assuming that a visitor with a mobile device wants a limited-functionality site, however, is like assuming every college student wants the Cliff's Notes version of a book simply because they're wearing school colors. Sometimes, it's a perfect match -- often, though, it will result in annoyance and confusion.

Josh,

Thanks for taking the time to respond and thanks for the link to your last article. Nielson certainly touched a few nerves with people and I was feeling like many of the responses to his article (which had some good points) were... um polarizing.

RE: the bank with the loan rates and such
Right now mobile users who want loan rates and such do have to "swim" through a desktop version of the site in order to get loan rates and such, but a responsive site didn't fix those problems because the data tables. Instead of spending significant time and money on it, we just left that as a know issue, knowing that mobile users that went anywhere other than banking login made up much less than 1% of all site visitors in a month.

I'm sure there's ways to solve those data-table problems with a responsive design (some sort of responsive table for example), but to do it in a way that's easy to update CORRECTLY and still be in compliance was looking tricky.

Any people don't have to "swim" through the bank sites any more than they have to "swim" through Apple's non-responsive website.

I don't usually mind if I have to go through a non-mobile site on my iPhone... in a few years my standards will probably change.

But I do share your concern that we shouldn't just cut off content from one class of user...

(we'll continue to monitor)

Eaton,

Interesting point about the cliff's notes, but I might have been a little unclear with my point.

> Assuming that a visitor with a mobile device wants a limited-functionality site, however, is like assuming every college student wants the Cliff's Notes version of a book simply because they're wearing school colors.

Usability is about tradeoffs.

You can't have everything for everyone every time. You need to look at who your users are, how they're behaving, how you want them to behave. Analytics aren't the be-all-end-all, but they're a great starting point.

If you can say that 90% of the visitors to your site behave a certain way, good usability dictates that you prioritize their needs over the remaining 10% (unless there's something special you know about the remaining 10%).

To change your analogy a little:

Assuming that a visitor with a mobile device wants SPECIFIC CONTENT THAT'S SIMILAR TO THE CONTENT OTHER MOBILE DEVICE USERS WANT, however, is like assuming every college student wants the Cliff's Notes version of a book simply because they're wearing school colors AND THE LAST 20 PEOPLE WHO WERE WEARING THOSE SCHOOL COLORS ASKED FOR THE CLIFF'S NOTES.

At some point that won't be a safe assumption, but today, that's a good assumption. If tomorrow people wearing school colors look at your funny and say "I don't want the Clif's notes", it's time to change again.

In retail that's called "good customer service" or "attentive sales staff". On the web that's called "good usability".

Good stuff, Jason, thanks. FWIW, Filament has some good thinking and techniques around complex data tables -- no doubt about it, one of the more challenging aspects of a responsive design (along with fixed ad sizes, I'd say). For more on their table technique:
http://filamentgroup.com/lab/responsive_design_approach_for_complex_mult...

Josh

I think the analogy of the Cliff Notes is off the mark - people buy those specifically whereas we cannot determine the intent of the mobile user unless we perform use analysis.

I think that there is a middle ground somewhere though - I try to determine that with each mobile site that I develop based on the customer's customers if there is no hard data. Once we gather some data (as quickly as we can) then we can make adjustments.

Well, a "middle ground" article was certainly forthcoming, so thanks for putting it out here.

I was bemused by the ferocity of the outrage at Jacob Nielsen's mobile-specific site suggestion and the emotional (nee political) reactions to his antiquated model of thinking. Surely an #OWD (Occupy Web Design) hashtag would sprout up claiming eminent domain over all that is true and just in the world (of web design). Well, the sentiment was surely there if not the hashtag.

But the outrage seem to come mostly from those (and their "followers") who have the most to gain by #RWD being "right" versus "wrong," as if those were our only two options. Frankly, not one project I have worked on in the last year-and-change has been so cut and dry, but I digress...

Some of the reactions I read by certain self-appointed #RWD gurus have actually made me less endeared to their (supposedly 'our') cause. Not that I question #RWD, just the fundamentalist attitude of many of its proponents.

Nielsen's research appears to be sound and logical and the freak-outs appear to have been caused by his suggestion to build a mobile-specific site. I get that some folks felt he needed to get his hand slapped for suggesting an implementation path, where he clearly has no subject matter expertise. But the intent of much of the backlash appeared to attempt to discredit the entirety of his research and relegate him to being part of the "old" (meaning "wrong") way of thinking.

The "real world" that I work in is one where it's ALL middle-ground and gray area. I glean what I can from the research, implement the way that makes the best business sense for my clients' needs and leave the emotional responses to those who clearly have too much time on their hands.

I think you have to be careful about using Analytics to make development decisions. If you have a shop with a step at the front door you will find that no one in a wheel chair comes to visit, this doesn't mean that they don't want visit!

If you have a site that only supports desktop users you will find that the vast majority of your users are desktop users. Looking at which pages a mobile user is looking at on your desktop site is a start but it's not the full picture.

Most sense I've heard on this topic yet. We spend hundreds of hours with consumers, 5 rounds of iterative builds and loads of research and analytics work. We designed what they wanted - not what our cherished notions, egos or opinions told us.

We operate a glass repair and replacement business and the key functions, content and features that people clearly wanted are represented in the final build. The consumers steered the product design towards favourable outcomes for both themselves and us, as a business.

I find a great deal of small businesses who simply wish to have a small mobile site, with some limited features so they can interact with consumers (rather than having no presence). We should be helping build experiences, not design methdodologies - and simple focused around the consumer rather than dogmatism.

What works for a news site is not going to be the same for a food delivery service, bill payment, online banking, glass repair, airline services etc. etc.

I see, like you, elements of arguments of both camps that make sense - but you're right - these are extreme points of view from people, in my opinion, who are not builders. Us builders have a say in this debate too and the answer is always in the customer.

I would be annoyed if an extreme viewpoint prevented many companies from just HAVING some sort of mobile presence around consumer needs, rather than design wan*ery or frippery - sometimes the user needs are subsumed by people pushing their viewpoint. It's not about us - it's for them.

Many businesses are struggling to get investment or effort put into mobile (or don't even have a presence yet) - sometimes this debate overshadows that problem, which is a shame.

@Josh, Interesting stuff at Filiment although that sounds *pretty close* to what Nielsen proposes, in terms of only showing subset of information to mobile and giving them an OPTION to see more.

@logicalthings #OWD tag cracked me up.

"The truth is when working on mobile you should always look at the site analytics and make smart decisions based on what you find. If you find that your mobile users use the site in a significantly different way than desktop users, check out Nielsen's guidelines and see which apply to you. And if you find that your mobile users are accessing the same content as desktop users then consider a responsive design."

This approach misses the possibility that your current mobile visitors' behavior is influenced by your site's strengths and faults. In other words, they may not be consuming content x or y because it's poorly formatted or difficult to access - rather than because of a lack of interest.

A website page featuring a Flash-based interactive map would be useless to most mobile visitors. Mobile traffic and engagement on that page will be abysmal. Create an HTML (or even a static) version of the same map and you might see a huge surge in mobile traffic, perhaps even exceeding desktop traffic.

Data can be misleading without the right insights.