Designing For The Exception
November 08, 2004 | 
30 Comments
Gerry McGovern writes about how designing for the exception is an obvious Web design mistake. In general I tend to agree with him. He points out breadcrumbs, for example, as being a design detail that only serves a few users:
We spent a lot of time designing breadcrumb trail navigation, which tells the reader where a particular page sits in the overall classification…Nobody actually questioned whether breadcrumb trail navigation was genuinely useful. It just seemed like a great idea. We were sure that everybody would love it. It turned out that it was a feature that the vast majority of people ignored. Everybody I have talked to since, and every report I have read, has stated that breadcrumb trail is, at best, a like-to-have feature.
I’ve seen this same behavior with users. In fact, I’ve never once seen a user acknowledge, in any way, the presence of a breadcrumb trail. I don’t think breadcrumb trails hurt, but I’m pretty sure their usefulness is limited to a small few, at least on the sites I’ve tested. These are the exceptions Gerry is talking about.
This could be said for many design “features” and I do think it’s important, when making design decisions, to sit down and really think about where you’re really going to add value across the board, as opposed to designing for a minority.
Filed under: News
Comments
1. JD said:
I came accross this myself during a recent development stint for my own blog. I had been looking at a large number of sites, looking at how the “breadcrum” navigation had been implemented, only to realize that I don’t think I’ve ever actually used it. After asking many friends and co-workers and hearing similar responses, I’ve opted to drop the breadcrum navigation and go for the cleaner look.
Posted on November 8, 2004 08:50 AM | #
2. Paul Griffin said:
Not to get bogged down in the example at the expense of missing the bigger point, but I have personally always found breadcrumbs to be rather useful. Even if I don’t click on them or really acknowledge their presence, I can always orient myself in a large site at a glance (assuming that they are logically organized). I find their usefulness to be subconscious, rather than active. To me, they are just another “you are here” sort of feature, not unlike a visually separated tab or menu item to denote the page you are on.
That said, I agree wholeheartedly that we often bog ourselves down worrying about the exceptions, rather than the rules.
One has to wonder, however, whether these sorts of thoughts might eventually lead to some wondering about those of us in the standards community who obsess about getting a site to look right in an obscure or underused browser. Should we instead be concerned mainly with whether our site looks right in IE (and lately, according to my stats, Firefox), and let support for other browsers fall by the wayside as exceptions, merely “nice to have”?
Just a playing the devil’s advocate on a monday morning ;)
Posted on November 8, 2004 09:10 AM | #
3. Jason Santa Maria said:
When I started out as a designer (before I knew how to do anything resembling website building), I didn’t really use them. I find that since I now build sites (and have designed breadcrumb navs into certain projects) I use them a lot more frequently. Maybe it’s because my mind now works in a more architectural way, needing to understand the entire structure and where I am in it. With all that in mind, I think they still work for certain projects. To be honest, sometimes I would rather have more, redundant navigation, than strand a user.
Posted on November 8, 2004 09:12 AM | #
4. Josh S said:
I find myself using breadcrumbs all the time. But maybe, like Jason, I use them only because I know how they work and I build them myself. But I do think not having breadcrumbs (or some way of letting the user know where they are in your site structure) leaves a user helpless and confused, and is therefore poor web design.
Posted on November 8, 2004 09:21 AM | #
5. Roger Johansson said:
I did usability testing on a site I’m currently working on just this afternoon, and out of four participants, three used the breadcrumb trail, mainly to navigate back to the home page. When asked about it, the fourth person knew what a breadcrumb trail is, but hadn’t seen the need to use it during testing.
Posted on November 8, 2004 09:21 AM | #
6. Small Paul said:
Yeah, the old Apple Mac theory of design: get rid of absolutely everything, except those things that you absolutely can’t do without.
I’m currently of the opinion that the navigation system should tell you where you are, as well as where you can go. So I’m a big fan of stuff like:
I reckon it conveys the information with the markup, and gives you enough hooks for nice styling to make it even clearer.
Posted on November 8, 2004 09:22 AM | #
7. matthijs said:
If the research about people NOT using breadcrumbs is correct, i’m apparently an exception.. On big sites, with a lot of sections and pages, i often take a quick look at the breadcrumb trail. especially when i arive at the site via a google-search or link. On big sites with navigation at the top of the site, at the left, at the right, etc, it helps a lot to see were i am.
Matthijs
Posted on November 8, 2004 09:25 AM | #
8. Gabriel Mihalache said:
There are 2 kinds of bread-crumbs:
Type 1 will show where’s the current page in the hierarchy of pages of the site.
Type 2 will act like a history, listing pages previously visited (bread-crumbs fit this definition better).
Both are kind of useless.
Posted on November 8, 2004 09:28 AM | #
9. Dan BOwling said:
Jason has it right, better safe than sorry. But then again, you have to be sensible about it. The Titanic needed lots of life boats, but no one ever put them in a home bathtub as a “better safe than sorry” measure. Breadcrumbs are a great tool for large sites, orienting users even if they don’t click on them. They don’t require much screen space either, and they provide an alternate (usualy less confusing, therefore more accessible) method of navigation.
Ok, now out of the tree and into the forest. I believe that designing for the exception is a good idea, except when in detracts from the target users ecperience. For instance, adding a “Skip to Content” feature helps out with a small minority of users that veiw the site without CSS, or through a screen reader (ect.); however, the to vast majority of users fail to recieve any benifit at all. Same with a style switching feature.
Posted on November 8, 2004 09:48 AM | #
10. Kim Siever said:
I am in support of using breadcrumb trails. I use them frequently.
I have to wonder if one of the reasons they are allegedly not sued much is because in the history of web design, they are relatively new.
Posted on November 8, 2004 09:55 AM | #
11. Keith said:
re; breadcrumbs – I’m pretty sure most of you are the exception. I’ve seen many, many users on sites with breadcrumbs and have tested specifically their usefulness and again, never once seen a user pay any attention to them.
I think tech savvy folks would be more like to use them. But I guess it really could differ site to site. Then again, they can be a titanic (sorry) pain to manage on some sites.
So the question is do they add value?
Just to clairify, I’m not against breadcrumbs and that’s not really the point, just an example. They are one of many examples of desgining for the exception, if that applies.
I will say this, if you rely on breadcrubs to orient your user and as navigation you might have problems.
Posted on November 8, 2004 10:24 AM | #
12. Olivier Travers said:
What Gerry McGovern completely misses is that on many sites a (small) minority of users account for the majority of usage, be it in page views, time spent, or sales. I don’t care that the people who come from a badly written search engine query come to my sites, spend 30 seconds there to understand their mistake, and hit the back button. I’m not going to make money (or whatever is my goal) with those guys anyway. (Of course I’ll try to make sure they don’t leave before understanding whether what I have to offer is relevant to them or not.)
I do care however, that the people actually targeted by, and interested in, my content, engage with the site and discover more of it which happens over several sessions. That’s what a breadcrumb can be useful for, to support both a sense of scope and place, as well as provide contextual navigation within large database-driven sites.
Look at most sites’ traffic stats. You’ll see a majority of users (especially first comers) are in and out quickly. Even with the best site it’s still going to happen, because there are plenty of stupid or hurried people out there, and because web search engines are still mediocre. Just because a breadcrumb didn’t make sense on McGovern’s CMS (doh) doesn’t mean they’re not useful in other contexts. And just because many drivers don’t use their turn signals doesn’t mean they’re useless.
Posted on November 8, 2004 10:33 AM | #
13. Jeroen Mulder said:
We did usability testing on a four-level deep website some weeks ago. Beforehand we decided to include a breadcrumb trail displaying the location of the current page in the site’s hierarchy for the simple reason to increase efficiency.
Since we didn’t want to overload the user with information that belonged on the second level of the hierarchy on in the primary navigation, we decided to allow shortcuts to this level (as well as the third level if you’re on the fourth.. etc.) through breadcrumb trails.
I think breadcrumb trails, though depending on the targetgroup, are indeed more designed at exceptional people. They are often used by above novice users or by disabled users, who learned to use them to their advantage.
I also haven’t made any observations that users acknowledge that the row of links is indeed a breadcrumb trail. Since users often skim through a document looking for relevant links, I have observed that the breadcrumb trail is used a lot ,if present, by people who know how to use one.
But to draw up a conclusion about the usefullness of breadcrumbs trails is a bit early, in my opinion. The term ‘super user’ is a bit misleading in my opinion, as one could consider a returning visitor who is familiar with your website a ‘super user’ as well. What super users differentiate from normal users is their deeper knowledge of the way websites work. In a perfect world, you want your original first-time user to become a super user on your own website as well, so the user will use it in the most efficient, effective and pleasing way the website allows him to.
Posted on November 8, 2004 10:33 AM | #
14. Chris Platts said:
Breadcrumbs (a misnomer, but it seems to have stuck) often look like a bit of an afterthought, tucked away somewhere above or below the masthead in
However on a rich, deep site (like many corporate intranets) they can be a useful navigation aid. On our intranet we have integrated them into a local navigation scheme (left column - in addition to tab-like top-level global navigation), consisting of a vertical ‘breadcrumb’ trail back to the home page, followed by the highlighted current page, indented child links, and outdented(!) sibling links. All links are of similar size, but are given different visual weight using color, font and border style.
Initial usability studies suggest that users appreciate the consistency in presenting the local navigation at any level of the site. It also scales to any navigation depth (not that I am suggesting it should substitute for good IA), although there is sometimes some initial confusion indicating that it is not so good for public sites with many first time or occasional visitors.
We’ll see how it goes when the site goes live - hopefully next week!
Posted on November 8, 2004 12:00 PM | #
15. Jason Santa Maria said:
Yes, I do think they add value, as long as, like anything you put in your designs, you can justify their existence. On large “drill-down” style navigational sites, redundant navigation (and I can be a big fan of redundancy) can really aid a user where the site’s structure IS the navigation (like e-com and bulletin boards). Exposing the structure may work to your advantage. I don’t really believe there to be a yay or nay call on this one, because it is dependent on the type of site you have and the audience you are targeting. In the same way we target marketing and messages, we should be able to target our sites and their structure.
So, roundabout answer, the more information you have to bring the user through, the more it may behoove you to have a few options to make that information navigable in a variety of ways (keyword search, breadcrumbs, etc.)
Posted on November 8, 2004 12:19 PM | #
16. Aaron Schaefer said:
I agree with what Dan said above:
Isn’t designing for the exception really just a way to say inclusion of all, even if it’s the minority of your visitors? Isn’t that the point of accessibility?
Keith is right in saying that if a user has to rely on breadcrumb links to establish navigational structure, perhaps your navigation should be improved. But likewise, I doubt anyone would rely on something like breadcrumbs as their primary navigation…it’s merely something supplemental to help those that prefer to use them.
If we didn’t pay attention to the minority of internet users, we would all be making designs that only worked in IE, couldn’t be used by the disabled, and wouldn’t be viewable on an alternative browsing device such as a PDA.
As long as adding a design detail doesn’t take away from anything else, I think it shows that you’ve already covered the majority of users and are willing to put in the extra work to make things even better for everyone else.
Posted on November 8, 2004 02:04 PM | #
17. Adrian said:
I have only ever found a use for breadcrumbs on photo gallery pages. They can be quite useful on large sites when looking for obscure drivers and the like but in most cases I find them ugly and impractical.
Posted on November 8, 2004 04:02 PM | #
18. Lachlan Hunt said:
Breadcrumbs are useless. I never use them, and in fact I hardly ever see them anymore (at least not on the sites I visit regularly). If a site is designed well, there’s little need for breadcrumbs. When they’re used to show the heirachy, there’s usually nothing that can’t be figured out by looking at the path in the URI (assuming the site uses nice, clean URIs). If they’re used to show the history of where the user’s been, then that’s nothing that my back button can’t provide.
Most, if not all of the information provided by breadcrumbs should be available from a variety of sources, including the site navigation, the page title, page headings, etc. So, basically, breadcrumbs are only useful for poorly designed sites.
Posted on November 8, 2004 04:29 PM | #
19. eric scheid said:
IIRC, the research done with breadcrumb trails has relied on the user clicking the links to be counted as using the breadcrumb trail.
One simple experiment which could be done, and hasn’t as far as i know, is an A:B test where one site has no bread crumbs, and the other has breadcrumbs but they are not linked. Then measure the user’s ability to find stuff.
Posted on November 8, 2004 04:29 PM | #
20. Keith said:
Everyone – I feel the main point is being missed here. Breadcrumbs are an example that, in a particular case, support the main idea – that usually designing for the exception is a mistake.
Breadcrumbs was a good example because they don’t add a ton of value. Meaning if you need breadcrumbs you’ve got bigger fish to fry. I’m not against them, but when Gerry calls them “nice to have” I think he’s right on – for his case if nothing else.
Jason brings up a good point about their value being related to the goals, audience, etc. of the site. Just like almost everything else in Web design.
(Anyway, we’ve already gone over that one. A few times.)
As well I don’t think Gerry is talking about exception in a usability or accessibility frame of mind. More like the “CEO doesn’t like it, so let’s make a knee jerk design decision to make him happy” kind of way, even though the doesn’t explicitly say it that way.
Posted on November 8, 2004 04:45 PM | #
21. Cameron Moll said:
Of the few times I’ve used breadcrumb trails in recent navigation (as a user, not a developer), it’s been because the primary or secondary nav failed. Meaning, I couldn’t find my way back to or across a category using the provided navigation structure and had to use a link in the breadcrumb trail.
A safety net? Yes. But if the primary/secondary navigation fails, breadcrumb trails won’t fix the (larger) problem.
My 2c.
Posted on November 8, 2004 08:33 PM | #
22. Jason Santa Maria said:
Alright people, read Keith’s last comment. Breadcrumbs are just an example, he doesn’t want our opinions on ‘em. The question here is; what do the design details (like but not limited to breadcrumbs), which are potentially used by only a select few, add to a design?
Posted on November 9, 2004 03:42 AM | #
23. Kev said:
Without wanting to focus too much on breadcrumb navigation, I think design elements that serve a valid purpose are a good idea, even if they are only used by 1% of the site audience.
Adding value to design is always a good idea and as long as the design element isn’t an afterthought or useless or hampers more productive areas of the design then where’s the harm?
re: breadcrumb navigation. Lots of accessibility advocates say that breadcrumb navigation is a good idea for deep sites.
Posted on November 9, 2004 06:10 AM | #
24. Keith said:
Kev – the question, to me, is do these elements add value in the larger scope of things? In my case, with breadcrumbs, they were a bit of a maintenance drain. As well, when they get to long they create screen real estate and display problems. So I’ve got to weigh that against the potential value, if any, that they have. In my case it might not be worth it to keep them.
See where I’m going with this?
The added value isn’t there.
Another example. If I am asked to reorganize a site based on an internal org chart that suits the needs of an internal group (the 1%) that is probably going to have a negative effect on the other 99% of users. There is no added value (to the majority) in that.
It’s an extreme example, but proves the point.
I like to talk to my Golden Triangle when it comes to this stuff. You’ve got business goals, user goals and site/maintenance goals. These things should be as balanced as possible.
Breadcrumbs (to go back to that) might help user goals a tad, but if they have a high “cost” as far as maintenance goals go they might not be worth it. Or they may take valuable screen real estate, thus cutting into the business goals. The cost in relation to the value is too high.
See what I mean?
Also, to your point about accessibility advocates (which I consider myself to be one, by the way) – many also have questioned the value of breadcrumb navigation. My advice is to test with your own site and see if the value is there for you.
Posted on November 9, 2004 08:22 AM | #
25. Kev said:
Sure, then its your call as to whether they’re important to you or not. I’m merely suggesting that if they pose no harm (and what poses harm is, I entirely agree with your implication, yours to judge as the designer) and have some tangible benefit, or potential benefit then I see no harm in designing not for the minority exactly but for as wide an audience as possible.
Indeed and I completely agree with your advice. I’m not meaning to suggest that ‘everything plus the kitchen sink’ is a good approach to design but if something fits in with the overall design and serves a valid purpose - why not use it?
Posted on November 9, 2004 10:16 AM | #
26. Kim Siever said:
After thinking about this some more, I thought I would qualify my previous comment.
I do not use breadcrumb navigation to the exclusion of any other method of navigation. Breadcrumb trails is only one navigation I try to offer users. I also provide menus and link tags (for Opera navigation bar). I also offer a site map and where possible a search engine.
I do not wan to tell our users that they only have one option for navigating our site (say, a single, simple, styled list). I want various optionsā€“so long as they do not overwhelm the pageā€“available so that the user’s preference can be accommodated.
Posted on November 9, 2004 12:22 PM | #
27. The Wolf said:
Life without breadcrumb? No way!
When browsing a forum or deeply categorized websites, this is often the best way to go back to the page you want.
Posted on November 10, 2004 11:13 PM | #
28. Justin said:
It’s not that this breadcrumb discussion isn’t kinda interesting, but I think the most important part of the article was this:
I probably like this because it applies to me a lot, but I suspect that a lot of designers and programmers are also guilty of this.
Posted on November 11, 2004 07:35 AM | #
29. @mber said:
I use breadcrumbs all the time. I find them helpful in orienting myself logically within a site. From taking a quick survey among friends, several said that they use breadcrumbs, and those that don’t usually use them said they do at least look at them as a way of orienting themselves and not feeling stranded, or to see where a particular page has been “categorized” on a site. I think breadcrumbs are a useful feature, and since they’re not difficult to implement, what’s the harm of having them, even if a minority of a site’s users use them? Those who don’t want to use them, won’t; and those who do, will. Better than pissing off those who do want to use them, since those who don’t won’t be affected either way.
Posted on November 11, 2004 10:26 AM | #
30. Jonathan Snook said:
Okay, this conversation is a few days old and I didn’t really have anything to add at the time but now I do. Alas, my point may be lost by this time.
If something provides a modicum of usability, even if only used by 1% of the population, then it’s worthwhile. “Hey, don’t use ALT text… it’s only used by 1% of the population” or “Hey, screw readable URLs, nobody uses em! article.asp?id=123&p=123&c=cookie is perfectly fine!”
I guess more importantly there needs to be a priority list of the features that you wish to add to your site and that due to cost or time, the lower priority features (readable URL’s and breadcrumbs being just examples) can be dropped. Because in the end they are helpful but certainly not critical to the success of your site.
Posted on November 12, 2004 12:32 PM | #
Comments are now closed