All About Website Usability Blog – Holly Phillips

The huge difference between .1 and 1 second, part 2
November 21, 2009, 4:29 am
Filed under: confidence, visual design

My last post referenced an astounding finding by Jakob Nielsen about the difference in perception between a 0.1-second page response time and a 1-second response time.  So why do I think this is such a big deal? 

Let’s put this together with a few other things that we know about website visitors.  We’ve found in other usability research that visitors (at least ours) like to be in control.  They want to be the ones to decide when to initiate a chat session, or when to talk to a salesperson.  We also know that they like to see all their options, and not feel that any information is being hidden from them.

So, let’s say that we have two different tabbed page designs — Design A has a .1-second page response time when switching between tabs and Design B has a 1-second response time.  According to Nielsen, our visitor will feel that he’s in control with Design A, and that he can access all of the information because it’s all “on the page”.  On the other hand, with Design B, he’ll feel that the computer is in control and is meting out the information to him.  The user is much more likely to prefer Design A because of these factors. 

This insight can help us make coding decisions, such as whether to load the information on every tab upon first entry to a page, or to load it only when the tab is clicked.  This can thus make the difference between a satisfying and unsatisfying final design, even if they both look identical.  Food for thought…


The huge difference between .1 and 1 second, part 1
November 14, 2009, 4:09 am
Filed under: confidence, visual design

Jakob Nielsen’s Oct 5 Alertbox describes the difference in user expectations about timeframes on the web.  While he discusses instances where users are willing to wait 1 hour, 1 day, 1 week, and even longer for specific types of results, the really interesting thing is his finding about the perception of response times of 0.1 second vs 1 second.

On the surface, it would seem that there shouldn’t be much difference in user satisfaction to have to wait 0.1 second vs 1 second for a page to respond.  It’s not like that extra 9/10ths of a second will significantly contribute to the time available to read a page.  But the interesting finding is not that one seems appreciably faster than the other.  Instead, it’s the subjective impression given by the time:  Nielsen found that a page that responds in less than 0.1 seconds feels as if you made something happen, whereas a page that takes 1 second to respond feels as if the computer is doing something to create the response.

Why is this so interesting?  We’ve watched users interact with webpages that have very similar designs at different sites.  In one case, both pages had a tabbed-structure on the lower half of the page, where clicking on one of the tabs kept the upper-half of the page constant but changed the lower-half of the page.  On the first site, the new tab was brought up immediately (less than 0.1 seconds), and on the second it was brought up in about 1 second (we of course alternated which sites the users saw first).  Users didn’t comment on the speed, but instead commented about how they preferred the first site because “you don’t have to go to a whole new page for each tab, like on the second site”.  In reality, neither design was taking the user to a new page.  But in the perceptions of the user, the second site was slower specifically because it had to bring up a new page.  Ie, the slower page implied that the computer was doing something, whereas the faster page implied the user was controlling the action.

This perception of having to bring up a new page then led to comments about the second site “being harder to navigate, since you have to go to all these different pages”.  The first site was deemed easier to navigate since “all the information is right there on the same page so I never had to navigate around” — even though they actually did navigate to the different tabs.

To me, this finding is incredibly insightful and helps explain how seeminly tiny time differences can have profoundly different cognitive impacts on website visitors.

Can layout alone influence satisfaction?
November 7, 2009, 4:24 am
Filed under: visual design

In customer usability studies we often watch customers struggle to find what they’re looking for, to figure out the navigation, or to wade through irrelevant search results.  But understanding the impact of page layout on customer behavior and satisfaction is tough to do unless you focus specifically on that and conduct task-based usability testing using a few different page layout designs.

Chaparro, Shaikh, and Baker conducted a study to help quantify the impact of page layout on usability.  Their finding: 

“Layout on a web page (whitespace and and advanced layout of headers, indentation, and figures) may not measurably influence performance, but it does influence satisfaction.” Chaparro, Shaikh, & Baker

Could this be one of those elusive design characteristics that help explain why customers who can easily find their desired information or complete their desired task still only rate their satisfaction as 8 out fo 10?  We’ve seen in our studies that satisfaction is strongly correlated with task completion up to a point.  After all, if you can’t complete your task, it’s pretty unlikely you’ll be very satisfied.  But that correlation tends to break down at the high-end; perfectly successful visitors with nothing but positive comments often don’t give us the maximum satisfaction ratings.  There’s clearly something else that goes into their mental evaluation in addition to successful task completion.  Improving page layout may help crack this nut.