• http://www.analyticspros.com Caleb Whitmore


    You raise some good points, however I feel the need to call out caution around using the GA-supplied “mobile” tag. 

    That tag will produce much higher traffic counts because it will report *all user agents*.  In testing I’ve done we saw about 3x the traffic in terms of visits and 2x the volume of pageviews with the “mobile” tag vs. the standard JavaScript tag.  That’s because the mobile tag uses a completely different means of visitor and session tracking as well as being server-side and thus vulnerable to data from bots.  

    I’d be curious to see data for the tests you reference and it would be illuminating to your readers to know more about that, because I fear many will rush off an implement the mobile tag on their mobile sites when in fact it should only be implemented on sites that are purely dedicated to WAP devices that are unable to execute JavaScript.

    Readers *should* be worried about accuracy, but those issues can be largely resolved by using the latest version of the GA JavaScript tag and placing the async code in the of the page.  Placement of the JS code, non-async especially, lower in the page can produce 10% to 30% data loss easily, and more for mobile due to latency issues on mobile networks.