It’s puzzling why so many otherwise intelligent people think the web site statistics purveyed by Alexa and Compete.com are reliable, when in reality they’re usually rough estimates which they derive from their own data sources, not actual metering from the web sites they list. If you’re the owner of one of the web sites whose performance is being low balled in this manner, you’re probably quite upset knowing the negative fallout this could produce.

Image by Dominik Gwarek

How rough are their numbers? I checked a web site I own on both these free services. I know from using Google Analytics that it gets about 35,000 unique visitors a month, 9,630 of which originate from the USA, yet the US unique visitor traffic figure was off by over 50 percent in Compete. Meanwhile in Alexa, which doesn’t give an estimate, the same web site ranked in the 770,000 range, while sites that have much less traffic than mine ranked in the 200,000 range. Testing Quantcast, which also only bothers with traffic originating in the USA, the uniques number was closer to the mark for the same web site (8,900) . A lucky guess?

What do they use to come up with these numbers? Alexa, for its part, has a help page that attempts to answer this question: “Alexa computes traffic rankings by analyzing the Web usage of millions of Alexa Toolbar users and data obtained from other, diverse traffic data sources.” Alexa toolbar users? How many of those millions actually use it? Furthermore, “diverse traffic data sources” is mightily vague. Perhaps they mean Alexa web site searches? Millions of Alexa toolbar users still can’t replace the accuracy of actual metering on the real web sites. Oddly, Alexa doesn’t offer any such device, which I’m sure millions of web site owners would be glad to add to their home page to set the record straight.

Alexa does utter this disclaimer on their site, to be fair (excerpted): “The traffic data are based on the set of toolbars that use Alexa data, which may not be a representative sample of the global Internet population. To the extent that our sample of users differs from the set of all Internet users, our traffic estimates may over- or under-estimate the actual traffic to any particular site…Sites with relatively low traffic will not be accurately ranked by Alexa…Generally, traffic rankings of 100,000 and above should be regarded as not reliable“. So, from the horse’s mouth: the stats for most sites in their database mean nothing! I suspect that the toolbar is a strong marketing tool for Alexa and they have less interest in accuracy.

Compete.com, on the other hand, takes the high road in its help page: “Compete has developed a unique methodology created by experts in the fields of mathematics, statistics and the data sciences to aggregate, transform, enhance and normalize data in order to estimate U.S. Internet traffic.” Then they go on to say that their numbers are arrived by monitoring the activity of two million members of the Compete community, which I believe to mean people who register on their web site, the vast majority of whom would have opted for the free plan, being that the next plan up costs 199 dollars a month! Looking over what services a free account provides, I fail to see how meaningful statistics can be extracted from their activities on the site.

As for tools to meter your own site’s traffic, that doesn’t appear to be an option with Compete until you buy the enterprise package, which entails having to call them for the cost (uh-oh!). Compete gives the appearance of being the luxury web site metrics provider. I conclude that unless you’re part of that exclusive club, their stats for your site are meaningless! I also found it enormously humorous that they describe their methodology as being superior to that of Google Analytics, going on and on about how deleted cookies and users with multiple computers can seriously alter unique visitor tracking. Yes, using actual data is sub-par as opposed to estimates!

Quantcast is the only one of the three to have estimates that have some semblance in reality, Their explanation for how they get their data: “We collect directly-measured data from the millions of web destinations controlled by our Quantified Publishers.” I don’t know what that means, but it sounds serious! The good thing about Quantcast is that they do give you free traffic monitoring for your site. You just paste a little bit of code in your site’s footer just as you would for Google Analytics.

Compete, Alexa, and Quantcast aren’t the only sources of free site metrics, but they’re the ones most frequently consulted. Of the three, Quantcast is the only one that gets a passing grade, for acting responsibly. Alexa and Compete, however, are knowingly peddling largely inaccurate data, and for this reason should be shunned. Note that Google has a free product: Google Trends. It currently doesn’t include smaller sites, but instead of inventing data for them, it just tells you they don’t have any. Google, the trustworthy giant of the web, will hopefully extend this service to you and I!

Related Articles:

• Which Traffic Stats and Auditing Tools Should you Use for your Web Site?