Discover XiTi 7.5

Web analytics
News | Contact |
Home > Internet users behaviour > 02/21/2007   

Websites in 2006: slowdown in growth

Web sites visits have doubled in 4 years but the slowdown in the growth of the average web site traffic, noted in a preceding survey, is confirmed.
  • Study carried out from 12/30/2002 to 01/31/2007
  • Perimeter of 186,768 Websites
Visits to Websites doubled between January 2003 and January 2007…

It should be understood first and foremost that our study presents growth in the average number of visits to Websites and not growth in total consumption of the Web. As we saw in our previous analysis, it confirms a slowdown in the growth of Website audiences.

Despite this recent slowdown, Website visits have nevertheless doubled in four years.

Looking at the same range of sites, we observe that there were 198 visits in January 2007, compared to 100 visits recorded in January 2003:


Let us look at evolutions year by year in order to obtain a better analysis of the slowdown…

but this growth slowed down in 2006, particularly in the second half of the year.

The audience for Websites is seasonal, with peaks and troughs throughout the year: usage falls off in the summer, starts up again in September and grows through the autumn before dropping back again in December, then picking up pace again in January.

Up to 2005, although the audience for Websites in the XiTi range went through troughs at certain times of the year, it continued to grow overall: the number of visitors never fell back to the level of the previous year. This was not the case in 2006: at the end of December 2006, the audience level was below that of December 2005.

The following assumption may be made from this observation. More and more Internet equipment is being purchased and more and more people are browsing the Internet, which suggests that the overall number of Web visits is increasing at the same rate. In this case, this slowdown in the growth of Website audiences may be explained by an even more marked growth in the number of pages available on the Internet than in the number of people using the Internet. Faced with a wider choice, Internet users may well be spread across a larger number of sites. This confirms the increasingly important role that visitor loyalty plays in Website development.

The above graph shows the growth in audiences over the 4 years of the study (superimposed). The trend for the 2003 series started from a base of 100 on Monday 30 December 2002. For the other series, the starting value is the value reached at the end of the previous series. The graph allows us to pinpoint any fallback in audience at a particular time of the year compared to previous years (overlapping curves)


Is the seasonal phenomenon the same size every year?

From the graph below, we can see that, in 2006, the January peak is less pronounced than in the two previous years and that the growth in audience from June 2006 is still proportionally lower overall than the growth shown for the same months in the 3 previous years, apart from August 2003 (a month of heat wave), which shows the most pronounced fallback for an August in the last 4 years. The December 2006 trough is significant compared to December 2005 in particular.

The graph below shows the growth in audiences over the 4 years of the study in comparison with each other: the 4 series start from a base of 100 on the 1st Monday before 1st January each year. The graph allows us to compare the effects of seasonal peaks and troughs between the 4 years.


How will Website activity develop over 2007? Is this trend towards a slowdown in growth temporary, or will it become more noticeable?

Join us shortly on XiTi Monitor for our first audience forecasts up to June 2007.

Methodology :

The indicator used for this study – measuring the number of visits recorded by a constant group of sites – shows the growth in Website activity.

In the absence of official data concerning the number and composition of existing Websites, it is not possible to present an indicator that reflects growth in total Web consumption by Internet users. We have therefore chosen to work on a constant range of sites.

We have chosen to monitor audience growth in visits to a constant range of sites, to which we successively apply the actual daily growth rates calculated for all the sites audited by XiTi. Growth rates are those for the XiTi range, which are constantly corrected in terms of site withdrawals and additions.

We have changed the analysis start-out base compared to our previous audience studies (month of January instead of the 1st week in January), so the growth rate cannot be directly compared. Every year, week numbers are subject to a shift in the calendar, which is not the case with monthly analysis.

We use two moving averages in this study. Their role is to smooth out a series of data and reflect a general trend. For measurement in time t, the moving averages used require series measurement in time t, along with before and after measurements. For Graph 1, the data series is monthly and the moving average over 13 terms measures an annual trend. To measure the moving average at month m, we need months m + 1 to m + 6, which explains the absence of a moving average at the end of the series.

Similarly, in Graphs 2 and 3, the data series is daily and the moving average over 7 terms measures a weekly trend. To measure the moving average at day d, we need days d + 1 to d + 3. The moving average for the last three days of the period is therefore not provided.

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