Welcome to WebmasterWorld Guest from 126.96.36.199
Forum Moderators: incrediBILL
I'm Jakob's worst nightmare. Time and time again, my sites have proven to me that purposefully counter-intuitive web design can -and will- produce excellent results. As always, this depends on the audience and the product or service. For instance, on one very large network of high-content sites I sell sponsorships. To ensure that traffic sees the sponsors, I do not have site search options on the main pages.
( Coming soon to a thread near you: "Ugly as a brand" )
I tend to agree with most of the things he says but I am not convinced that everything he says should be taken as gospel or the only way to do something. Like RC I have some sites that go directly against what he says and do very well. I have others that incorporate many of his ideas that do well also. Basically I think it comes down to the fact that there is no "right" way to build a website. It can be done many different ways and still be very effective.
I'm very curious about those specific reports that stlouislouis mentioned - "207 Design Guidelines for E-commerce" - for that exact reason. In fact, I almost bought the set last week and I still may.
By the way, I certainly agree that useit.com is far from a good looking site. It's very obvious that there's a LACK of visual design style here. The Nielsen Norman Group does not do design - they do usability studies - and I assume that his website (and his company's site) look the way they do to emphasize that. They probably waste NO time at all on calls about doing design, LOL!
Two years ago, when the Nielsen Norman Group e-commerce studies first came out, Jakob picked Amazon as the site that adhered to the highest percentage of his company's guidelines. In June 2002 in a new 15 site survey [news.com.com] which didn't include Amazon, LL Bean took the top spot at 66% compliance, and House and Garden was at the bottom with 38%.
If we remember how these guidelines are established - user testing - the fact that the best site missed a third of them is notable.
A highly usable site can still be mighty ugly - that's a different area of concern. But a beautful site can be nearly inscrutable, and that's the issue for Jakob.
All design requires tradeoffs at many different levels. When the tradeoff is between style and substance, between eye candy and usability, which way do you go? I think it's essential to recognize when you come to those critical choices. Otherwise, aesthetics may unconsciously trump usability and hurt the bottom line.
That's where these NNG reports are valuable - recognizing what the usability issues are.
In a one month followup study, only 0.9% of first time visitors (determined by cookies) used the search function.
We feel that's extremely good, because only 12% of those first time visitors were one hit wonders. 88% explored the site in some way beyond the Home Page - for an average of 14 page views per new visitor. The main links for "new visitor" information were chosen 57% of the time.
This contrasted dramatically with the previous version of the home page. For the old version, 45% of new visitors were one hit wonders, and less than 10% of those who stayed around visited the new visitor information that was so laboriously prepared. The "page views per new visitor" stat was only around 3.
That's the kind of results that make me want to take in more of Nielsen's recommendations.
Nielsen's thinking is more flexible than that.
The perfect website might score 90%, because specific websites always have valid reasons for deviating from a few of the established guidelines.
A judicious use of graphics, layout, colour and design, if it makes a web page easier to look at, if it makes it easier to find the important elements, it aids usability.
But I also thing there are times when you are allowed to break the rules. You have to know exactly what you're doing, of course, but depending on the actual site and your intended audience, you might be able to do something rather more... innovative. Clearly, if you're going to create a gaming site, you don't want to make it look like Google or Amazon.
For example, the BBC website has a section aimed at very young children (aided, presumably, by their parents). It is very brightly coloured indeed, with almost no text on it at all. The navigation buttons are Flash, and when you mouse over them an inanely cheerful voice "reads out" the button text. Normally, this would be a horrendous and intensely annoying technique, but this is a site aimed at young children who can't read and certainly won't have an aural browser at their disposal. (It does beg the question of whether children of that age ought to be surfing the net, but that wasn't the concern of the designer.) Go to the BBC's website, scroll down to the bottom and click on CBeebies.
With this kind of input, just one idea can make a big difference. So I'm pretty sure the cost will be paid back many times over.
Some of the items I see look like "no brainers". But in my experience, that can be exactly the kind of issue we tend to overlook. So having it spelled out is valuable, and knowing that actual test groups found it important also means a lot.
Certainly, his points on testing, useability and accessibility are competely on target, but I strongly disagree with his points on design and presentation (which are taking us back 6 - 8 years). I'll be the first to argue that design, useability and accessibility are inherently linked, but I feel that a good designer/UI developer can create a nice design while staying within the inherent limitations of useability and accessibility.
That's probably true.
But what is even truer is an old IT adage (that dates way back before the WWW): If it ain't been tested, it doesn't work.
You need evidence that the design works for the intended audience, on the equipment they use, and in ways they will be comfortable with. Enthusiasm from a designer can be a good starting point, but it's only step one.
Every audience is different, so subtle changes in a theoretically great design may make the difference between being a visitor magnet and a ghost site.
There's another old adage that could do with being dusted off: you can't improve what you don't measure.
Tedster's results are in line with both these adages....
* gather evidence for improvements;
* make the changes;
* measure the difference;
* keep the changes if they worked;
* repeat until happy.
Anything else (such as arguments that Frames or splash pages or Flash or Java are good or bad) is just handwaving.
I view design as the way content is presented. It should not just be functional, it should be clear, visually pleasing, appropriate and, if you can manage it, refreshing but without actually dominating and taking over. The design should have a purpose other than saying, "Hey, look, I have Photoshop and Fireworks!"
I don't think you can get away from the importance of testing. The guys who have sat down and recorded how people use sites have scientific data that we don't. They know the average user far better than we do and I value that information greatly.
But he never said whether that increase in searches is valuable to his business purpose. Or if it increased sales, or site stickiness, or repeat visits, or anything important like that. He seems to assume that an increase in searches is inherently a good thing. I know for sure that it ain't necessarily so.
We had a good thread about Making Site Search an Asset [webmasterworld.com] - and the MANY pitfalls involved.
If Jakob has a blind spot, I'd say it's around the need for a highly visible search box. He's making some uninspected assumptions here that come, I suspect, from his concentration on big clients.
Going further, many of his suggestions about improving search functionality have a lot of value. For instance - you've got a search function, but how does it deal with either A)too many results or B) no results?
Again, I really dig the guy for all the reasons stated in this thread. By the way, the reports are excellent and have already caused me several moments filled with wincing and wanting to hide. And it's not HIS comments that did it, it's the quotes from his user panel.
Those "average jane and joe" panel members nailed more than few issuess that I know I will need to fix. And it's not all about HTML and page layout, either. A lot of it is about what kind of content you need to make the visitor comfortable enough to buy.
The performance aspect is basically a measure of how easy (or hard) it is for a user to complete their task - where a task is, for example, information retrieval or making a purchase. Performance is relatively easy to measure - does the user complete their task? How hard was it? We can measure time, we can measure the number of clicks etc. We can easily complete user tests that give us a pretty good indication of what aspects of our sites we need to fix up. Performance is where IA and search fits in.... if IA & content is well structured, then a search facility may be less of a necessity. Yet, a _good_ search function could work wonders, if it HELPS users achieve their task. And putting the search box on the front page... if it helps minimise the number of clicks a user has to make, it must have its benefits. The necessity for a search function also depends largely on the purpose, and the size of the site. If the site's primary purpose is to provide users with information, then it is often worthwhile to offer an easy to access search function. It is also of benefit to provide a search function if the site is particularly large, or has a complicated navigation structure.
The preference aspect is subjective, and hence, much harder to measure. The only way to measure user preference is to... well, ask the user what they prefer! Preference is where the aesthetics of a page come into play.
Preference and performance are related. Show a picture of a few sites to a subject, and they'll tell you which they prefer based on aesthetics. Let them play for a little while, and which one they say they prefer might change - because they've been able to see how the site performs. This indicates that performance impacts preference, but not really vise versa. It is much harder to prove that preference impacts performance.
Neilsen seems to focus on performance, sometime to the detriment of aesthetics (such as on his own site). IMHO, to make a site as usable as possible, we should try to maximise both aspects - make a site as pretty as possible, without detracting from ease of use. Make a site as easy to use as we can, without making it ugly!
Oh and one more thing - never forget the _purpose_ of the site, and what you want your users to be able to achieve. That's my rant anyway.
Thought I'd report in now I've had the report for a while.
Firstly, it has already paid for itself. A point on homepage design led to the alteration of one of our site's homepage which has made a significant change to sales. The fact that anyone with half a brain could have spotted it is not important ;)
I would say that overall, the book has 2 main benefits:
1) It reminds you of all the basics you may have forgotten about since your site was built (see above!)
2) The quoted comments from the study participants reminds you of the way Joe and Josephine Bloggs actually surf, and their level of knowledge when it comes to the web. Very, extremely, mind-numbingly low.
Oh, and it has lots of lovely glossy colour piccies. :)