Lorelle

Is Google Analytics Slow or Not?

Web Analytics, February 4th, 2009 by Lorelle

In Is Google Analytics Slooowing Down Your Blog? by Easton Ellsworth on the , he asks a very important question:

Do you think Google Analytics is worth the trouble? Why or why not?

The first question we usually get when talking about is “will it slow down my site.”

Personally, I find this fascinating, as people seem to have no fear of adding WordPress Plugins, Google Gadgets, and JavaScript, AJAX, and all kinds of bells and whistles to their site, but add a statistics package and instantly they ask about load times and impacts on the user experience. Clearly, there is some experiential history here worth exploring.

A quick search for reports on load times across the web and found a long history of people blaming the statistics program for dragging down their site loading times, with a lot of articles disputing those blames.

Many claim the Google Analytics slow down is a myth. SwamiSEO found very little page loading slow down using Google Analytics, and discussions on Stackoverflow also reported there was little difference. The Royal Pingdom report featured on the Blog Herald admitted that while the numbers weren’t significant, they were measurable across the three months of testing.

A quick glance into Google Analytics loading time history and mythology found it used to be slow. Many reports date back to when Google Analytics was using the urchin.js script, which Google replaced with the new ga.js script at the end of 2007. They recommend replacing the slower script with new faster one. AskApache has an article on how to speed up the urchin.js script loading times if you wish to continue with the older version.

The Blog Herald article referenced the slow load times for Google Analytics study by Royal Pingdom on Google Analytics script load times in Europe compared to North America, which were admittedly small, but the responses of the readers in the article comments shows a continued believe that Google Analytics is a resource hog and drag on their system, in spite of research testing reporting otherwise around the world.

There are many ways of testing your own page load times. Truwex Online Web Page Check Tool for Google Analytics checks Google Analytics specifically, and Pingdom Tools and Pagetest are just a few of the many page load time checking services available for free and a fee.

Last year, we did an informal evaluation of loading speeds for Woopra against the competition, and recent updates to servers and code have made Woopra even faster. We’ll be doing more load time studies and testing soon as the new version is released, but the feedback so far continues to be that Woopra puts little if any significant load on your site and users’ experience.

Woopra is seeing more than 50 million page views a day across more than 50,000 sites using our live web analytics program. This means we’re collecting 1,000,000,000 pieces of raw statistical data each day – and Woopra is still in beta testing development! That’s a lot of data streaming across servers, so page loading and speed issues are important to us.

5 Responses to “Is Google Analytics Slow or Not?”

  1. Sarsura says:

    i think woopra js is faster than google`s analytic js.. and i like the improvements of woopra and looking forward to the new desktop client version of woopra.

  2. Ferodynamics says:

    Slow or not, how many kilobytes are we talking about? I’m not that concerned about your tracking mechanism because all you can really do is look at some basic environment info and make it look pretty.

    In terms of bandwidth, interaction, I’m more fascinated, and afraid of, the “chat”, because I know a little Javascript chat window could slow down my site, especially once you start adding cute stuff like smileys, colors, etc. I’m not *that* concerned about people on dialup, (they are still a significant % of traffic) I’m more of a believer in speed for the sake of speed, because a fast site gets more pageviews, more comments, more everything, that’s just how I roll and it works for me.

    If *anything* is loaded onto my website, it better be damn important. Those people that pile on plugins and such, as you pointed out, they don’t know what they’re doing. To those people: if you have traffic, over the long run, just one little image will end up costing you hundreds of dollars, not really in terms of bandwidth, but lost attention (ad visibility) and causing you to upgrade your server before you really need to, as well as increased load time which affects you in various tangential ways. There’s a reason my CTR is high and with only two ads per page.

    This isn’t penny pinching either, every kilobyte makes a huge difference! It’s not a joke when it’s money out of my pocket!

    On top of that, did you know search engines weigh page-load time? Yeah, so a few kilobytes here and there could end up costing you big.

    As far as Analytics, I let Google see what’s going on because I want to build trust, even if they’re already collecting everything they want to know via ad units, because they send me traffic and pay my bills.

  3. mercurythread says:

    From my point of view I still run both packages at the same time and I’ve not found any problems with load time – every so often at periods of highload the Google Analytics file does take a little longer to load – but as its at the end of the code its not a biggie.

    I’ve found far more load time problems with the plugins for wordpress – and some of my dodgy PHP coding than I ever have with these scripts. Long may it continue.

  4. It is a fact that Google asking a person to place their code before the closing body tag means that the page will not finish loading until the Google script does. For our sites this is a problem as we often use javascripts that execute only after the page is loaded. For many websites, they will not see any change as they do not have pages that rely on changes made after the page loads to look correct. I think therein may lie the disparity in numbers. I can tell you this with confidence, when I remove the code the javascripts executed after the page is loaded execute very quickly. I am looking at this point at either reverting to just using the image that they provide in case of “noscript” or putting the google analytics into a function and executing the function and body onload. I hope the latter will work.

  5. This brightend my day so can I brighten yours?

    If everything seems to be going right, you obviously don’t know what the hell is going on. :)

Leave a Reply