Page load time analysis

by Ravi Pathak on December 27, 2008

in Blog

Brian Clifton provide great idea and a script to track page load time using google analytics event tracking. Immediately we started using it and realized that its invaluable from actionable analytics perspective. We worked with a client for implementing this quickly. Basically, here is how event tracker works.

trackEvent(category, action, optional_label, optional_value)

We started passing data of the time taken to load the page in the “optional_value” parameter measuring from the start of the page load to the time when load is finished. We also passed this data get maximum granularity of milliseconds.

So here is what we have got. You see that #1 item takes about 900% more time to load. So, this is insight. Actionable as well. See how it becomes actionable.

As of now we have figured that there is one page on the website which takes huge amount of time to load. ok great. Now we used charles proxy to understand what kind of time taken for different server requests on this page. Below is the output that we got from the Charles.(have eliminated 1 to 12 as they weren’t significant).

If you notice #34,#40 takes both about 16 and 9 seconds relatively. That’s huge for any given page. Although these where async requests so that didn’t deter users from clicking on any other link , however this is huge.

Unfortunately, I am not in position to comment on what we did after-wards, its obvious that we could reach to the root of the problem very quickly. We realized and found one call to the server which has creating a problem, we modified that and we are all set to go.

 

Another piece of extreme learning that we had gotten is that suddenly few pages started taking about 10 times more time to load than last month’s average. We tried and investigated then we realize that it was due to the link /internet cable got broken in the middle of the sea somewhere which passes the bytes. It was terrifying to see it however news got us relaxed a bit.

Sounds easier said than done? What do you think ?

Liked this? Get more by Signing up for our free newsletter!

Ravi Pathak

Ravi Pathak

Ravi Pathak is a Co-Founder of Tatvic and expert at Managing different web analytics tool. Ravi's actively works on conversion optimization projects to improve conversion rate and test newer hypothesis with e-commerce companies. He regularly tweets via @tatvic or @ravipathak and is easy to have conversation with. Google Plus Profile:

Website - Twitter - Facebook - More Posts

  • http://discrepant Aman

    Great post, its a beautiful use of event tracking. What about the rendering ability of different browsers and connection speeds, how did you factor that in

  • Ravi

    Aman,
    Thanks for the f/b.

    well, you can create advance segments for different browsers or even see the page load time for different browsers for page(s) or group of pages. In this caste it appeared that the problem was wide spread and similar across all the browsers so we really didnt care going that path.

    HTH ?!

Previous post:

Next post: