Thursday, 13 May 2010

Page download time - usability findings

Fascinating findings about user perceptions of page download speed and how it's connected to a website's ease of use. The data is pretty old now, but interesting nonetheless. Technology changes quickly but people don't.

In this podcast, Jared Spool talks about some really interesting findings his company User Interface Engineering almost stumbled across in the process of researching something else. They found that users perceived some sites to be slower than others, even when in reality, the complete opposite was true. There was however, a strong correlation between a user's success with a site and their perception of how fast pages were loading.

When things are going well time flies; when you're not enjoying yourself, time is more likely to seem like it's dragging.

The truth about page download time - UIE podcast to download or listen online

The truth about download time - article by Christine Perfetti and Lori Landesman

Why speed matters - Jakob Nielsen
So while Jared and co are telling us that if the user experience is good, download times aren't so critical, Jakob Nielsen counters with the impact of slow page load times on our attention and ability to focus.

I don't think Jared and Jakob are necesarily contradicting each other here. They both agree slow is bad. It's just that Jared is talking about the user's perception of download speed, while Jakob covers the reality of page load time.

Short-Term Memory and Web Usability - The human brain is not optimized for the abstract thinking and data memorization that websites often demand. Many usability guidelines are dictated by cognitive limitations.

Website Response Times - Slow page rendering today is typically caused by server delays or overly fancy page widgets, not by big images. Users still hate slow sites and don't hesitate telling us.

Response Times: The 3 Important Limits (excerpt from Usability Engineering 1993 - Jakob Nielsen)
  •  0.1 seconds gives the feeling of instantaneous response
  •  1 second keeps the user's flow of thought seamless.
  •  10 seconds keeps the user's attention.
  •  After 10 seconds, they start thinking about other things, making it harder to get their brains back on track once the computer finally does respond.

Follow up
A funny coincidence - just week or so after writing this.

We recently implemented some changes to the content management system interface to make orientation easier and one critical task quicker to do.

A colleague was telling me how much she liked the interface tweaks and how great it was that the system performance had been tuned up. Only thing was, the system is operating at the same speed it always has...

1 comment:

  1. Not to dispute the fact that user perception can differ from an objective measure, but there is no doubt that page speed, measured objectively, matters to users. This blog post describes the results of a live traffic experiment on the Google search site.

    But Google has also done usability lab studies looking at the speed question. Of course, in a lab setting one doesn't expect users to consciously observe a change of a few hundred milliseconds. Also in a lab setting, the larger sample sizes needed to detect a response to relatively small changes are infeasible. Nonetheless, the lab study does confirm a difference in user response to speed differences of just a few seconds.