During the last quarter of 2008 I've been freelancing for a Latin American Website Incubator. In my research I've found that at some point I needed to estimate traffic (number of unique monthly visitors) in each of the studied countries in order to have at least an approximation of the revenue income from Advertising. This is how my search started...
In the first month of research, I couldn't find what I was looking for: a simple tool or website where if you pasted the URL you'd get as a result an estimate of monthly visits/users or page views. The only websites that I found does this kind of estimation are trafficestimate.com and compete.com. I didn't try compete's services, some of which are not free but I did try trafficestimate.com. After extensive comparison of its results and real data that I managed to get, I have to say that the error margin of this site is huge. The variability of the estimates (compared to real data) that this website has in its results is so big that I cannot recommend it for traffic estimations or traffic forecasting.
Even though I couldn't find exactly what I was looking for to accurate predict traffic, during the search process I found some interesting tools that can be useful:
• Alexa.com: This is the oldest and most comprehensive webtraffic statistics website available for free use. Even though it's not statistically accurate (Its sampling method is not representative) is the closest we have to an overall traffic ranking system. It tends to work pretty well with websites with a "Traffic Rank" higher than 100.000.
• Ranking.com: Provides information in a very similar way than Alexa.com does: Traffic Rankings, Number of users per Million, trend graphs, etc.
Given that there wasn't such a tool, I decided to create one myself and I developed an Excel Model to predict Traffic for any site. This model had to be based on free and publicly available information so anybody could use it. What I came up with is a statistical function (the average of a linear function with a 2nd degree polinomic function) that predicts the number of users based on Alexa's reach. This model, links real stats (past real number of users of selected websites) to estimate users of any website, based on its reach. In some cases, because Alexa's information is not accurate always, the "Reach" has to be "transformed" into a relevant number before being plugged into the Excel function.
Although it still needs a lot of work, so far the average estimation error is around 7%, which is much better than the other tools found online.
To find more about my Excel Model to forecast website monthly visits, visit my post Website Traffic Estimation Excel Model
Visit my blog MBA Internet Marketing Manager
Article Source: http://EzineArticles.com/?expert=Pedro_Neira
No comments:
Post a Comment