Speed matters

More proof that speed as perceived by the end user matters. This time from a Google Research paper (PDF). Google's experiments demonstrate that increasing web search latency 100 to 400 ms reduces the daily number of searches per user by 0.2% to 0.6%. Furthermore -- and this is where it gets really interesting -- users do fewer searches the longer they are exposed. In other words, the cost of slower performance increases over time and persists.

To use Peter Van Dijck’s words:

In other words, if your website is a little slower, users will use it less (we knew that), but they'll also use it less and less over time, and when it speeds up again, they’ll still use it less than before the slowdown.

Based on their observations, Google suggests site builders to think twice about adding a feature that hurts performance if the benefit of the feature is unproven.


Alex UA (not verified):

Based on their observations, Google suggests site builders to think twice about adding a feature that hurts performance if the benefit of the feature is unproven

So, does this mean you're going to pull the Overlay from D7 core? (Sorry, I couldn't help myself)

December 18, 2009

I agree that the Overlay in Drupal 7 is a bit slow. However, it does improve usability so it is probably the right trade-off to make. It might not be the right trade-off to make for everyone, which is why the Overlay can be disabled. That said, hopefully we can continue to improve performance of the overlay ... feel free to help! (Sorry, I couldn't help myself either. ;))

December 18, 2009
Alex UA (not verified):

It's not just the technical performance of the overlay that's a big problem, it's the idea/the way it was implemented. I realize that some usability people said that this was better, but I completely disagree. In my opinion it's complete overkill and is an attempt to re-create the browser. I don't think that there's anything "proving" that this is a usability gain, and I think testing would reveal that this implementation is a problem for many different groups of users. Can you point to another similar implementation on the internet? If not, then why wouldn't this cause *more* user confusion?

Here's what I don't get: there was a ton of work done on the download/update manager, and the module page on d.o. is about to be redesigned, so why is it so hard to have people download this from contrib if they want to/if it's needed?

If I thought the overlay was a worthy cause I'd happily help out (I, and Zivtech, help out where/when we can, as long as we see the value of the task), but I disagree fundamentally with the Overlay's current implementation and I'm about 99% sure we'd never use it.

December 19, 2009
BryanSD (not verified):

Another advantage of speed...let's those spambots go through your site quicker so they can move onto the next.

Personally, I don't think a slightly slower site prevents me from visiting or staying at a site any less. Slower sites though do cause me to contribute less. Nothing is worse than submitting a comment and seeing it disappear due to a timeout. The end result is either you wasted several minutes with no comment left behind or you end up submitting the same comment several times. Not a good experience for people wanting to take part in the conversation.

December 18, 2009
Wim Leers (not verified):

For those who are interested in page loading performance, and more specifically, measuring and monitoring it for their actual visitors and not simulated ones (in real time if desired), please take a look and comment on my master thesis proposal: "Web Performance Optimization: Analytics".

December 18, 2009
Coornail (not verified):

Google is also started ranking pages by their load speed.
You can check out yours at: Google webmaster tools > Labs > Site performance

December 19, 2009
Jan-Nicolas Van... (not verified):

Google has a very useful Firefox plugin available for page speed testing and optimization: http://code.google.com/speed/page-speed/download.html.

December 22, 2009


Updates from Dries straight to your mailbox