7
Mar
Should you rely on Alexa Rank estimates for CEE website traffic estimates?
  • News
  • 0 Shares
  • 472 Views
alexa-traffic

We sometimes hear professionals or companies state how important is Alexa Rank for a website. In fact, very often marketers use Alexa Rank to decide whether they should advertise on a website or not.

Alexa is a California based company that provides commercial web traffic data and analytics. It analyzes data provided by millions of users who have installed on their browser the Alexa Toolbar or use their data panel. These data are considered a sample of all internet users and the rankings are calculated using the following algorithm.

“Alexa’s Traffic Ranks are based on the traffic data provided by users in Alexa’s global data panel over a rolling 3 month period. Traffic Ranks are updated daily. A site’s ranking is based on a combined measure of Unique Visitors and Pageviews. Unique Visitors are determined by the number of unique Alexa users who visit a site on a given day. Pageviews are the total number of Alexa user URL requests for a site. However, multiple requests for the same URL on the same day by the same user are counted as a single Pageview. The site with the highest combination of unique visitors and pageviews is ranked #1. Additionally, we employ data normalization to correct for biases that may occur in our data.”

So the above approach sounds solid, but how accurate can the traffic estimates be? Well, it depends.

If the sample that Alexa is analyzing is big enough, and the website traffic being considered is visited by many users, then this approach should work. However, this is not the case for the majority of websites in certain domain industries.

A significant accuracy factor is the actual volume of the total traffic recorded for a given website. Alexa states that there are limits on their traffic estimates for website rankings beyond #100,000. Thus, rankings beyond #100,000 are not statistically meaningful. This means that, for example, the difference in traffic between a site ranked #1,000,000 and a site ranked #2,000,000 has low statistical significance, i.e., it is not reliable.

Sites ranked #100,000+ may be subject to large ranking swings due to the scarcity of the collected data. It is not unusual for such sites to decline or improve their rank by hundreds of thousands within a few days. This is happening due to the long tail effect. A small change in the number of visitors to a site can result in a huge improvement or decline in Alexa rank.

Long_Change-Visitors-vs.-Rank

As some of you know, ARGO-E manages many websites in the civil, environmental and construction (CEE) industry. So we decided to make an experiment: We selected four of the websites we maintain. Three of our monitored CEE websites had about the same traffic A, but are being ranked by Alexa between #1,700,000 and #12,800,000! Another website with three times the traffic (3*A) is also ranked near #1,700,000! Large ranking swings have been also observed to many of our websites, with more than a #1,000,000 drop or improvement over a few months period. This is not real!

So, unless your website receives hundreds of thousands or millions of visitors per month, the Alexa traffic estimates should not be accurate. It is not Alexa’s fault! It is the statistics!!

An additional factor to be considered is Alexa’s users diversity. Go around your office and see how many colleagues of yours have installed Alexa’s toolbar at home or in the office. Let us answer this for you: Most likely, none!  Considering that the average CEE professional is less likely to have installed Alexa toolbar on their browser or use the actual Alexa data panel, they will not be represented properly.

According to our own survey data and the Alexa rankings, it is safe to say that not many CEE professionals are using the Alexa tools. As an immediate result many well known and popular CEE related websites are not ranked properly within Alexa, while websites, from different industries, with less traffic are ranked significant higher. So although we love, as a company, to look at data, we know that bad data is worse than no data!


Comments are closed.