|Website Ranking: Accurate Visitor Traffic Metrics|
|Articles - Opinion & Editorials|
|Written by Olin Coles|
|Thursday, 12 August 2010|
Website Ranking: Monitoring Accurate Visitor Traffic Metrics
Most experienced webmasters agree that constant web traffic is the key to any website's success, but sharing these often-secret visitor metrics to prove ranking position can be risky. In the computer hardware review industry, webmasters often jockey for position simply because it means getting support from the big-name manufacturers. Get enough visitors each month, and maybe Intel will answer your email request... maybe. With so much competition between websites, it's becomes vitally important that rank is reported correctly. There are various traffic estimate websites available to report statistics to the anonymous public, but while some use hard-gathered data there are others that offer numbers based on the patterns of selected participants. Web statistics are reported by three types of tools: on-site (server side), off-site (client side), and hybrid methods such as packet sniffing or tracking applications.
Traffic Monitoring Tools
Of the many web analytics tools available, few people can agree on their use or usefulness. For the webmaster it's all about purpose: are the statistics meant to impress others, or educate and inform? On-site tools such as Webalizer and AWStats are located server side with the website, and generally record traffic statistics without filtering search engine bots or non-browser network traffic. On-site tools usually report accurate traffic numbers, but only to the extent that anything and everything is considered valid traffic. As a result, some competitive webmasters resort to using server-side logging tools because they report inflated traffic statistics. Off-site tools may seem better by this perspective, but include their own set of drawbacks.
Then there are the others: guess-timate websites that gather from several sources, such as DNS server polling, cookie reading, and browser-installed tracking applications. The most popular of these are Alexa and Compete, which could also be considered the least accurate. Alexa has taken great strides to collect data from something other than their self-installed tracking tool, and over the past two years has started polling server logs for more clues to a websites traffic trends. Compete is among the least accurate traffic tools available, as it relies solely on a select few individuals participating in feedback, or as they word it "combining the online behaviors and attitudes from 2 million consumers across the United States". Of course, if your website is unfamiliar to one of these selected 'consumers' or you receive traffic from outside the US, then you're out of luck.
Finally, there are aggregator tools that combine data from several sources to factor an outcome. WebsiteGrader is a good example of an aggregator tool that offers unique perspective. There are also websites available that observe the trends at other analytics website to build their own ranking, such as TrafficEstimate and StatBrain. While none of these aggregator sites offer a better insight than directly-measured server or client-side tools, they offer yet another alternative view to a website's performance metrics. Use these tools in addition to, but not in place of, the other tools mentioned in this article.
Each of these options are not without a long list of potential flaws. Server-side tools like Webalizer and AWStats report data is not comparable to another server running the same software, and each software version records traffic data differently than the next. Making matters worse, reports statistics can be changed because they're delivered on an easily-edited HTML page. On-site monitoring tools also lacks validation, because any type of traffic that crosses the wire is counted as a visit. Client-side tools such as Google Analytics and Quantcast may be among the best available, but they're nowhere near perfect. The code that these tools depend on can be suppressed by browser privacy programs or ad blockers, which makes it difficult to accurately track proper bounce rate and visitor metrics. Although reports can (and should) be emailed directly from the provider's website, often times a doctored screen shot is substituted when traffic inflation is desired. Obvious problems exist with the tracking applications from Alexa, and especially Compete. To begin with, users can abuse these browser tools by running them on multiple computers and amplifying the effect by opening multiple tabs. Compete users, whoever these company-selected individuals are, might not be aware of your new website or care for the topic.
Still, even with these other factors taken into consideration, there are unknown contributors that can alter web metrics. I'll use Benchmark Reviews as a firsthand example: from March 2007 to mid-June 2010, this website was hosted in Seattle, WA. After several years of slow and steady growth, we relocated our website to another server farm. You're probably wondering what this has to do with web metrics, and here's the punch line: our visitor time on site metric reported by Google Analytics (GA) had been an abnormally high 25-28 minutes from day one, but when our website was cloned onto an identical server at the new datacenter this metric suddenly dropped to around 4 minutes. At first glance, the server transfer seemed to fix a longtime issue with our time on site metric. But that wasn't all... the number of daily visitors also immediately grew by 10%, yet average pageviews were cut in half and the bounce rate jumped from 40% to 70% (not good). Despite this odd behavior reported by Google Analytics, our traffic statistics remained consistent in Quantcast and Webalizer. Keeping in mind that our website file and server configurations had never changed, this event raises the question: how do you know which tool is correct?
It's not easy - because there are numerous factors impacting web metrics, and a webmaster can never truly know the precise number of unique visits or pageviews his website generates from real visitors. Some webmasters complain that tools such as Google Analytics or Quantcast can never be accurate because of user-interference or privacy tools, and they'd be correct; but they must also consider that all websites would then be discriminated against equally. Unfortunately, this also means that interested parties (such as product sponsors) may never have a truly accurate number and could see inconsistencies reported. This often leads to competitor speculation, and occasionally accusations based more on doubt and less on data.
This article covered some of the most popular software tools available, while revealing their major defects. On-side (server-side) tools such as Webalizer and AWStats are both good for private website monitoring on a single server, but shouldn't be used for comparing one site against another hosted on different servers. Google Analytics is a great (free) traffic monitoring tool, and although website statistics are comparable between servers the information is kept private to webmasters. For public access, Quantcast is a good tool for traffic comparison between websites that have installed the code. Thanks to browser privacy tools some traffic goes unaccounted for both Google Analytics and Quantcast, but this is something that effects all websites. Traffic aggregators such as Alexa and TrafficEstimate are fine for alternative comparison, but are highly inaccurate and swayed by outside influence. Finally, restricted community tools such as Compete are among the among the least accurate, as they only measure web surfing habits for a small group of privately selected individuals to form traffic metrics.
Benchmark Reviews has used Google Analytics from the beginning, and we'll continue to use this free tool, but our recent datacenter transfer revealed that GA is not impervious to unexplainable outside influence. We don't know what these factors could be, and we're uncertain if the statistics were either correct before of after the site transfer. But what we do know is that this industry is rife with ignorant webmasters, desperately willing to fill the gaps with hyperbole and conjecture, and it pays to educate yourself on the options. Hopefully this article has introduced you to some new tools and helped you learn the facts behind visitor traffic metrics. At the end of the day, information provided by any of these tools should only be used for researching your own website. If you must compare one site to another, be forewarned that even the most accurate tools will only offer a best estimate.