On Attending Jakob Nielsen's Seminar, "Designing Web Usability"
(part of the Montclare Speaker Series)
Just because it's impossible does not change the fact that it's necessary.
-- Jakob Nielsen
This was an amazingly helpful seminar; I'm really glad to have attended. I'll start by saying that it wasn't what I expected. I've delved deeply into Jakob's web site, and have read quite a bit of his book, Designing Web Usability : The Practice of Simplicity. I was often left with the impression that the author was a bit preachy; that he somehow put himself on this pedestal, taking pot shots at the web, telling us what needed to be fixed but not revealing to us how to do it. I sort of expected the seminar to be dry, dull and humorless.
I could not have been more wrong. Jakob turns out to be a dynamic speaker with a great sense of humor. Seeing him in person absolutely convinced me that he has truly earned the guru status that has been bestowed on him. I came to realize that Jakob is a scientist focusing on very scientific usability issues...he is not an HTML specialist to tell us how to fix our code! Here are some highlights from my seminar notes:
One way to think about the success of your site is to determine the number of successful users you have as a percentage of your unique visitors stats (your "conversion rate") If it's e-commerce, this is the percentage of visitors who have actually bought something from you. For other sites, it may be the percentage of users who actually register.
(unique visitors) x (conversion rate) = level of success
So, to increase your success you could work really hard to get that unique visitors number up...which is what people usually try to do. But when you do that, you're only increasing the number of people who look at your site once and leave in disgust. If you focus on that conversion rate by making your pages more usable, your actual level of success will increase much more dramatically.
User skepticism is rampant. On your site, tell them your business model, tell them how you are making money off of them. This will (presuming you are an honest company) build their trust in you. Validating credibility is essential.
The browser wars could be won if the browser companies focused on value-added features like reputation managers -- imagine a trusted 3rd-party source that would pop up a message when you visited a site, saying, "We've found that this company provides poor customer service...for this type of product, we recommend these sites instead..."
Results of user survey on factors that encourage repeat business (HOME):
75% High quality content
54% Often updated (frequency depends on type of site...how often do you realistically expect them to visit?)
58% Minimal download time
66% Ease of use
All other factors, like customizability, were significantly lower.
Usability studies (not web-specific) have shown that response times by the system to user interaction have the following limits (this is where he made the comment quoted at the top):
For the user to feel they are directing the action (and not feel uncomfortable waiting): less than 1 second
For the user to feel an actual physical connection between their actions and the response of the system: 1/10th of a second.
To keep a user at your pages, the page *must* load within 10 seconds.
Guidelines for success (RUN - combined with the above, HOME RUN):
Relevant to the user's needs
Unique use of online - what you're doing with your web site cannot be done off the web (why re-purposed content does not work)
Net-centric corporate culture - the people in your company must understand the web and how it is used, why brick-and-mortar companies often have terrible web sites
Another useful metric: The percentage of pixels on the page that are dedicated to what users actually want. Eg. "Welcome!" pages with pictures of company principals and text messages are useless.
Another quote: "The average user experience on the internet is 'I'm on the wrong page!'"
He gave the 7 steps to better usability, the process of usability testing as a looping constant process of repeated improvement. Important point: develop parallel prototypes of your design; 3-5 alternative mockups is about right. Each iteration of improving and refining your prototype will tend result in a 38% gain in usability.
Rules for usability testing: Real users, real tasks, SHUT UP (let the user do the talking, do not lead them.
When starting a usability testing session, be exploratory but do not lead them. Keep this part fairly short (as it is in real life)...ask them how they would use the site and what they would expect to be able to do. Then follow with actual tasks for them to complete. DON'T use the terminology used in the interface itself -- give them scenarios to work through. Ask them to talk out loud about what they're doing, why, and how they feel about the results as they go.
Focus groups and user surveys are not usability testing -- these are useful, but usability testing must focus on the users.
You only need to test with 5 users...here's why.
On content: page titles and headlines must make sense when they are viewed out of context -- eg. search engine results, etc. People tend scan, not read, on the web -- content must be scannable (use bullet lists, etc.), objective (no marketing-speak) and concise (about 50% shorter than writing for print).
If you have a chance to see Jakob speak, don't miss it. It's not just a listing of all these important points, he really delves into the background and supports each one completely with the results of real research. Very impressive!
Back to Terrie's Production Docs