The Customer Analytics Blog
Actionable advice on optimizing the customer lifecycle
Woopra vs. The Competition (Technically Speaking)
I would love to recruit some of our Wooprites to do a little testing if I could.
John P’s been running Woopra over on his blog along with a slew of other tracking scripts (don’t know why he hasn’t removed the others yet – we’ll give him a hard time!), and he’s been asking me why Woopra performs so much better than Google Analytics. This topic has come up in Forum discussion’s as well, and I thought I’d do a little more testing to try and verify a hypothesis.
In that discussion, it has been noted in reports from 3 separate people that Woopra is catching a much larger number of visitors than Google Analytics (please go add your results too). No one seems to know why though. So I did some research using the Firebug plugin along with FireFox to gather some real performance data on Woopra vs. Google Analytics, P-Metrics, and MyBlogLog – all of which are installed on John P’s blog.
Here is what I found:
If you look at the full size version of that image you’ll see that Firebug is reporting the following response times for the various scripts associated with the tracking:
| Woopra | 125ms for 3kb |
| Woopra | 62ms for 95b |
| 219 ms for 10 kb (75% slower than Woopra) | |
| 62ms for 35b (same as Woopra) | |
| PMetrics | 360ms for 4kb (188% slower than Woopra) |
| PMetrics | 125ms for 43b (100% slower than Woopra) |
| MyBlogLog | 250ms for 3kb (100% slower than Woopra) |
| MyBlogLog | 187ms for ? (200% slower than Woopra) |
In addition to this data, I manually downloaded the Woopra javascript and noted that its filesize is 7.7kb, and also the Google Analytics script which weighs in at a hefty 22.2kb. So, in addition to having 75% slower response time for the download request, the script is actually triple the size as well!
No wonder Woopra is so much faster and captures more visitors!
And by the way, I cleared my cache and repeated the test, and Google actually performed even worse (but I didn’t have the heart to use this data for the comparison). So this was not a one time event.
John’s server is located in the US, but these results should be similar for anyone, anywhere around the globe because of our distributed network. So I’d really appreciate it if anyone else could spend a few minutes doing some testing and either provide your results in the comments, or a link to a blog post with them. We’re very interested to see results from other people.
