Discover XiTi 7.5

Web analytics
News | Contact |
Home > Internet users behaviour > 09/25/2006   

Forecast until the end of January 2007: slowing growth

Forecasts of the traffic to internet sites until January 2007: a slow down in the growth of web site audiences recorded since 2004 may be confirmed.
  • Study carried out from 01/01/03 to 08/21/2006
  • The entire XiTi perimeter
On average, a given website will record almost 2.5 times more traffic in January 2007 than in January 2003. In a previous study, we noted a slow-down in web audience growth since 2004. This phenomenon seems to taking further hold: an adjusted model showing the recorded trend is shown below.


On average, website audiences grew by 50% in a year and a half starting in early 2003, and doubled in two and a half years. This growth has been slowing gradually since 2004, and the trend has been felt even more in 2006:

  • Between the beginning of 2005 and the beginning of 2006, the model's audience trend shows an increase of 27% (175 à 222).

  • If the model conditions continue to hold, the increase in the audience trend between the beginning of 2006 and the beginning of 2007 will be only 9% (222 à 242).

A number of parameters might explain this phenomenon: has home Internet access reached saturation point? have web users reached an upper threshold for visits in a given period?…

Methodology :

The recorded visits to the sites we audit have enabled us to calculate weekly rates of evolution. For each weekly rate, only the sites that were present in consecutive weeks are considered. This maintains a constant, maximum pool of sites. We assume that the bias arising from the weekly period used is negligible.

These rates make it possible to obtain the evolution in visit numbers for a constant number of sites. The influence of sites leaving XiTi and new sites joining is thus accounted for in the model.

The evolution in the numbers of visits recorded is corrected with the evolution in the number of sites audited by XiTi. Remember that no institute is able to list all the sites that exist on the Internet. We have chosen to represent the evolution in average audience in a closed system, revealing the evolution of visit potential.

With regard to the prediction, a mobile average for each annual period removes seasonal effects on the audience. The corrected audience was then adjusted with a parametric model.

In this study, the basis of 100 is the value of the adjusted model in week 1. This enables the use of direct predictions of the trend up to January 2007.

© 1997-2019 AT Internet/XiTi - All rights reserved  | FR  | EN  | ES
deg_gauche deg_droit