Why Different Web Analytics Report Different Numbers?

By Daniel Scocco

questions and answersThis post is part of the Friday Q&A section. If you want to ask a question just send it via the contact form.

Mayur asks:

Is the number of pages viewed on my site equal to the number of banner ad impressions I should have? Because my Aw Stats shows a higher number than the ad networks I work with. Google Analytics shows even lower numbers. What is the deal?

Welcome to the world of web analytics!

As you already noticed, each web analytics software, ad network and traffic estimation service (e.g., Compete, Alexa, Quantcast) will report a different number.

Why is that the case? Because what represents a unique visitor and a page view to your website is subjective (i.e., each service and software has its own criteria for deciding when a visit and a page load happens). For example, some services and software count search bots as visitors, while others don’t.

Second, different services and analytics programs use different tracking methods (e.g., JavaScript, pixel tracking) so not all visitor will be tracked equally. For example, a visitor with JavaScript disabled will not be counted by a service using JavaScript tracking, while it will be a service using pixel tracking.

Given all these differences and nuances, here are some rules of thumb most webmasters and online marketers use:

1. Raw logs are useless

Most servers store raw logs, which are lists of all the accesses and page requests on your website. It’s possible to interpret those raw logs with special programs, creating graphic reports which will contain the number of visitors, page views and so on.

Those numbers are grossly overestimated, though, because all kinds of search bots and automated queries are counted together.

2. Webalizer and AW Stats overestimate

Webalizer and AW Stats are very popular web analytics programs, and that is because they are usually installed by default on cPanel (the control panel software on most hosting companies). Both of them tend to overestimate the number of visitors and page views your website receives, however, and such data should always be used with a grain of salt.

3. Ad networks underestimate a bit, but there is nothing you can do about it

The number of impressions you’ll see on most ad networks control panel usually is an underestimation of your total traffic, and that is because they won’t track people who can’t see ads or who block them on purpose.

There is nothing you can do about it though, and if you want to make money using ad networks you need to play under their rules. The alternative is to have your own banners embed with HTML code, in which case they would be seen by 100% of your visitors.

4. Google Analytics underestimates a bit, but it’s the industry standard

The numbers reported by Google Analytics also underestimate your traffic slightly, and that is because the software has very strict rules regarding what should be considered a visitors and a page view.

GA’s tracking is very reliable, though, and that is why it’s used as the industry standard. If you want to sell a website, for example, most serious buyers will ask for Google Analytics data before they make an offer.

Summing up: Go with Google Analytics if you want to get a sense of your “real” traffic.



Share

14 Responses to “Why Different Web Analytics Report Different Numbers?”

  • Angelica

    So what’s the solution? Well, I guess i will be getting an average between my analytics and webalizer.

    They are not that far apart in numbers.

  • Daniel Scocco

    The solution is to use Google Analytics.

    • Mir Imran Elahi

      Feedjit is showing me i got many traffic from stumbleupon.But Google Analytics is showing nothing.Why?

  • Mahesh

    Hi daniel what’s your opinion to use on histats or mint analytics for niche sites? my another questîon -is using google analytics on multiple niche sites from same account in any way devalue sites?

  • Toby

    Blimey so much information to take in, bit over my head – but interesting

  • Vivek Parmar

    What about Sitemeter??
    Google Analytics is used as an industry standard

  • G R Peacock

    I wrote a little article on this subject after seeing a, in-depth study reported from the Columbia University School of Journalism. It’s online along with a link to the original article at: http://www.measurementmedia.com/2010/09/24/publications/media-measurement-has-never-been-an-exact-science/.

    The differences between javascript analytics, including Google (GA), and the in-house server scripts is not trivial, it can be huge, sometimes more than a factor of 10 based on my dozen or so websites! It all depends on the speed of the server and a bunch of additional delay sources.

    Pure html pages have the least delays (if compact) and the closest results between AWStats (AWS) and GA – again, in my experience.

    Massaging the server side scripts like AWS is more satisfying to me, since they seperate the bots from the beauts. Funny thing, there used to be an even better system called Urchin Analytics, but Google bought them and it is not commonly available for much less than about $10K a pop!

    Still, I think AWS provides valuable information about what’s really happening on a website and is inherently more accurate than any external javascript reporting system, by design. But the ad buyers have been convinced GA is the best and they rule the money!

    Not long ago Google seemed to recognize the fact that its free Scripts posted at the bottom of a webpage were underreporting page views due to various delays. Their solution was to recommend loading the script into the page header instead of the older method of just before the literal end of a page (just before the tag). I do not believe Quantcast has changed yet.

    How hard can it be to count page views? Depends on who is counting. Perception Rules!

  • Eugene Farber

    Unfortunately you pretty much have to rely on Google Analytics because that is the number that anyone really cares about. But I think GA’s estimates are REALLY low.

    Google uses Java script to measure visits. Most of the time it is in the footer and the last thing to load. Some people may switch pages before the script even runs and they are not counted. We don’t really care much about that though, because they aren’t spending too much time on your site anyway.

    The bigger problem is Firefox add-ons like NoScript that are EXTREMELY popular. It has weekly downloads in the hundreds of thousands. This add-on will block the script from running.

    These visitors you do care about…but they will never be counted by Google Analytics.

  • Vladimir Radmilovic

    Neither is perfect, and some people choose to use both types of analytics (JavaScript and raw log files analysis), there are also hybrid solutions. Note that not Webalizer and AwStats are not only raw log analyzers. Our Web Log Storming, for example, is quite different than others and gives you lot of marketing options, conversion calculations, spider filtering, etc.

  • Leon

    You just saved me a load of money by not buying any of the software people tend to imply is better than Google stats.

  • BlogTipss

    Best way.. just use Google Analytics.

  • Latief

    I say that Google Analytics are the best. Neither is perfect but Google Analytics is the most accurate I think, but you have to choose Google Analytics, Google’s team and developers are the best so.. they know what they’re doing so you can count on them.

  • Bash Bosh

    I agree, Google Analytics is the industry standard today.

    Of course, there is no (and probably wouldn’t be ever) any 100% accurate website traffic analyzer.

    But I personally think that Google Analytics is the closest to that “perfect” value, and it is used by all experienced website owners. So, definitely go with this great and free tool!

  • Sergio Felix

    I have a WP blog, so I’m using WordPress Stats (plugin that claims to use Google Analytics without the ‘extras’) Google Analytics as well and AW Stats, mainly to get an overall between these three.

    None is even close to each other, but my still safest bet is definitely Google Analytics, WP Stats as they provide me real hits excluding my own and AW Stats I just use that to get the extras which I don’t in the other too.

    Logs, 404’s, time on site, etc.

Comments are closed.