Conversion ratio and page parse time study. Is it worth to improve performance?

[data_digger]

Hello,

We all know that parse time is very important to conversion ratio. Customers get frustrated when they have to wait 10 seconds for page to load, and often abandon such site.

I was sitting today drinking coffee wondering about how important page generation time is to conversion ratio and I couldn't answer that question honestly. Why? Well, let's say that We have two similar shops, one with average parse times of 1 sec, and second with avg. 2 sec. How much more orders will shop 1 get compared to shop 2? What if avg. parse times would change to 1 and 4 sec? Maybe shop 1 would get 9x times more orders then shop 2? Or maybe only 1.2 times more orders? (20% increase). Is it worth to spend hundreds or thousands of dollars on dedicated server just to improve performance a little bit?

I started googling around and amazingly I couldn't find anything on the subject. Maybe I ask wrong questions, but it seems that there's no such study.

Is anyone aware of any papers on the subject?

[stevesh]

It appears to make a lot more difference than I would have thought:

http://www.svennerberg.com/2008/12/p...version-rates/

http://blog.mindvalleylabs.com/gone-...ed-matters/176/


Lots more links when you Google 'page load speed conversion'.

I hope all this will make the Flash fans do some thinking.