March 23, 2004 | View Comments (7) | Category: Design
Summary: Designing websites in increments to fully expose the user experience.
The Web's Best Kept Secret, aka. D. Keith Robinson (okay not really a secret), has posted an entry (The People-Centric Web) that focuses on some of the issues that were discussed yesterday. One of the issues that Keith discusses is watching users actually use your site to get an idea of how they actually use the site. These observations over time should help you improve the quality of your site. Very rarely do we get the right design the first time, especially when our sites begin to grow. So this is where the not so new concept of design betas comes in.
With the advent of the web also came the new software development trend of hosting open betas for software. No longer were you constricted to a small user group of QA testers, but you could open your software up to whatever size group you chose, and they would do all the testing for free. From this you are able to spot holes in your software and hopefully correct them.
A downside to this is that sometimes you can fall into the trap of expecting your users to handle most of the work for you. Open betas do not mean you can avoid planning the software ahead of time or taking care that you have the beta working to the best of your ability at the time of its release.
I have an extremely bad (or good, depending how you look at it) habit of releasing websites before they are finished. I do this because: 1) I get anxious and 2) as long as I have the proof of concept out there everything tends to work itself out. This strategy has proved to be moderately successful for me. Admittedly, you will probably never get the ooh-ahh type of designs out of me, but you do get an organic style website that grows with the users needs.
For people who have been here a while, you remember the redesign process that I went through for this site. The key factor in what made that process successful was I had an end goal in mind (the architecture) and I was rewarded with great user feedback. For a recap here are the entries (next time I need to remember to take screenshots of each phase):
In essence, I see all of my websites in constant beta stages as is the way I think all websites should be viewed. Obviously commercial websites would follow a more point release type of concept, but still the underlying theme is the same. Web design is different than any other type of medium in that for it to remain successful it has to evolve.
Slashdot has had the same design for as long as I can remember. They may have made some minor tweaks here and there, but nothing significant enough to warrant any discussion. It can not be argued that they are popular, yet they also kept the same design. Kind of goes against what I said in the paragraph above huh? Well Slashdot has the advantage of having a fanatical user community which will always keep the site alive. However, running and keeping a site alive of that size is not a cheap thing to do.
From my knowledge, Slashdot has two major sources of revenue: subscriptions and advertisements. Now maybe a redesign would not help increase subscriptions, but I do think it would do wonders for ad revenue. I am sure Slashdot gets an insane amount of pageviews, but imagine the increase they would get if their site actually became more browseable. Meaning you can actually go back into the archives without any issues and just spend the day strolling through the past. Sure I could do that now after some minutes of figuring things out, but the whole experience just is not fun like getting lost in a huge library. Instead I see Slashdot as just having a homepage and with RSS becoming more and more popular it is possible that pageviews will decrease.
Even though a magazine may be laid out by a print designer, the same design (or at least structure of the magazine can remain the same) can be kept the same for years or even decades. Users generally tend to use magazines the same way over a large period of time.
Software grows only when new features are introduced into the package. Websites can grow and change on a daily basis.
A really tough concept for me to grasp while doing websites for clients was that I was not always going to be in control of the website. When the project was done, I was done with it. Even if I looked at the site a month later and noticed a billion and one ways that it could be improved, I was helpless. This is frustrating. Bowman laments about the same issues on the Wired Redesign in his interview with Digital-Web.
Understanding that your website works more like a garden that needs occassional pruning instead of a brochureware site (unless you really are just running a brochureware site) is a major step in getting back to the people-centric web. Thinking like this also helps to bring the fun back in design for me. I can spend less time sitting there planning a whole website and more fun designing and tweaking the web site.
Sam Royama laments on the fact designers no longer are given the opportunity to develop relationships with clients or their customers. I think this type of design is how it might happen, especially since there really is no concept of time on the Web. It is either now or past.
Trackback URL: http://9rules.com/cgi-bin/mt/mt-tb.cgi/184
Hopefully it is obvious in my writings that I agree with you. But I think a bigger distinction needs to be made between Web Sites (categorized as information-based, non-transaction, non-processing) and Web Applications (whether it is ecommerce for customers, or distributed computing solutions for business).
I think Web Sites have more an opportunity to grow organically as you say. Especially sites with good content that gets updated often. People will forgive some discrepancies in IA of visual design as long as they can still get to the content they want.
But for Web Apps I think the story is much different. For ecommerce, you have to be as right as possible from day one, or be the only app on the block that offers your particular content. Otherwise you will not convert browsers into customers as fast as you want, and may even kill your opportunity to do so in the future. For distributed computing, each error (however you define error) is a loss of money for the company.
I guess in this lies the importance of user profiles, task analysis, and usability objectives.
I hope this topic of being People-Centric falls under the heading of "well, duh" for everyone. I know it won't, and I guess like many of the conversations of late on Web standards, it is something that will continuously have to be talked about.
Matthew Oliphant (http://usabilityworks.typepad.com)
Okay, I reread the post and I guess I missed this line:
Obviously commercial websites would follow a more point release type of concept, but still the underlying theme is the same.
Sorry.:)
Matthew Oliphant (http://usabilityworks.typepad.com)
The Human-Computer Interaction (HCI) field uses tools (hardware and software) to study how people use software or surf web sites. They can track eye movement, clicks with timestamps, etc. There is a camera recording everything. It is possible for testers to be told to do a play-by-play of what is going on in their mind.
Zelnox
We signed up for one of those free 1and1 hosting plans and use the web space as a test-bed for new sites we are working on.The clients seem to love the fact that they can come and see how things are going and give us heaps of feedback where changes are needed. Because their servers let you run scripts anywhere, it is easy to put a complete site in a folder containing its own cgi-bin etc. When everybody is happy, it is a simple process to transfer the whole site from our server to theirs with only some permissions to change. I guess that we are lucky in that our situation allows us to develop relationships with clients (My partner/wife has even turned down work because she didn't like the people!). Once the site is up and running, any ideas I might dream up can be implimented on this test version and we also use it for training the client on how to make best use of their piece of cyberspace. Without a doubt, the best thing for me is watching people, some who can hardly use an email, use a new site - more educational than any design manual.
Adrian Rinehart-Balfe (http://www.boogenstein.com/)
Once again, a little late to the party...I would agree with you, but I would add the caveat that constant change can be *worse* than improving features.
But only if those constant changes are happening at extremely short intervals.
It can be extremely frustrating to a user to get used to one behavior, only to find the next day that that behavior is changed, and then they have to go through the learning process all over again. It's not so bad if these are subtle changes, mind you, but let me give you an example (web app):
at my previous job, I was the primary developer for this beast of an online ordering application. This app has been in existance for 3 years or so, and has mutated (via poorly thought out business rules) into a rambling, un-user-friendly beast. But not without a few decent learned behaviors.
In the last 4 months of my short employment there, I was asked to change 4 things that while on the surface didn't seem like large changes, ended up being major UI process ones (i.e. all the users had to relearn the interface 3 times). All within a 4 month period.
So...I would say that while yes, a web site/app can (and should) be treated like a garden, the reality is you still have to create the decent arrangment first and then tweak. Can't be ripping up the beds and moving them every month or so.
Tom Trenka (http://www.dept-z.com)
Keep track of comments to all entries with the Comments Feed
©2003-06 9rules, Inc. | Thoughts Feed | Notes Feed
Amazon Stores: Powertools | Jewelry | Books | Health | Landscaping | Kitchen | Movies
#1
I think it would be really cool to watch users working with my website. I've spent an awful lot of time trying to make things useable and pleasant - to make the experience rewarding. I will continue to evolve the site (on an almost daily basis).
But, despite all my effort, I'll never know if I'm making right decisions or wrong decisions until I either a) get feedback, or b) watch a user at work.
I completely agree that user-testing is an important part of any UI development (whether it be for web, print, a chunk of hardware, etc, etc). Such testing should be high on any developer or designer's list of priorities.
DarkBlue (http://urbanmainframe.com/)