Discover XiTi 7.5

Web analytics
News | Contact |
Home > Internet users behaviour > 06/14/2006   

Internet audience since 2003

How has the traffic of Internet sites changed since 2003? The survey measures the increase in traffic exactly, and highlights the progressively less significant effects of seasonality.
  • Study conducted from 01/01/03 to 05/07/06
  • The entire XiTi perimeter
Visits multiplied by over 2.5 in 3 years!

Having stagnated until August 2003, the number of visitors then grew steadily to reach a visit score in 2006 that is more than two and a half times as high as the figure three years ago.
This progression can be correlated with the increase in ownership of Internet equipment, and is a good reflection of the growing role of the Internet in today's lifestyles.


This first curve gives us a global trend in the medium term, but what of the evolution year by year?

Seasonal behaviour

Regardless of the year in question, the web audience varies according to well-defined periods. It is relatively stable until the summer, when it starts to fall, with a marked drop in August (and July, in 2005). Internet activity picks up again in late August/early September, continuing to grow throughout the autumn before another decrease at the end of the year (remaining higher than at the beginning of the year, however). After this pause, the audience takes off once more in the first weeks of January, exceeding scores a year before.

Thus, although the audience dips at certain times of the year, its overall level is growing constantly: the number of visits never falls back to the level of the previous year. This is true for the average across all sectors of the Internet: behaviours actually vary quite significantly from one sector to another…

These seasonal effects are marked by two particular times in the year when web users probably have other priorities, such as summer holidays or preparing for the end-of-year festivities…


Does this seasonal phenomenon have the same importance from year to year?

The graph below demonstrates that seasonal effects tend to become gradually less significant. It is also interesting to note that over the 3 years of the study the annual variation in audience decreased, representing a slow-down in the growth in visits.


Thus visits are increasing through the years, but their growth is slowing a little.
What about the duration of these visits?

Average weekly duration per visit: a drop of 10% in 3 years, or visits 30 seconds shorter than in early 2003.

While the Internet audience continues to grow, web users are spending less and less time on each visit. Several hypotheses could explain these increasingly brief visits: the spread of broadband, more user-friendly sites, visitors who are keener to cut to the chase…

Since early 2005, the trend is actually slightly upwards, but still falls short of 2003's average visit duration. In fact, between week 10 of 2003 and week 12 of 2005, the average length of visits fell by 13.3% (42 seconds), growing by 3.9% (10 seconds) between week 17 of 2005 and week 11 of 2006.

It will be interesting to see how this aspect of web users' behaviour develops in 2006…


For a more precise analysis of your own market, this study can be focused on a particular activity sector (for example, the study has already been carried out for the recruitment and video games sectors).

Do not hesitate to contact us for more detailed information.


For each week of the study, the evolution in the number of visits is based on visits generated by all the sites audited by XiTi in two consecutive weeks. These rates of evolution made it possible to obtain the evolution in visit numbers for a constant number of sites scaled to a basis of 100. The influence of sites leaving XiTi and new sites joining is thus accounted for in the study. The study interprets the evolution in the numbers of visits recorded, corrected with the evolution in the number of sites audited by XiTi.

For the study of annual seasonality in visit numbers, the periods extend from the fourth week of January to the third week of the following January. The decision was taken not to start with the first week in January, firstly because this would not have emphasised the strong growth at the beginning of the year. Secondly, given the weekly breakdown, this significant evolution is recorded either between the last week of December and the first week of January, or between the first two weeks of January. On the other hand, there is very little change between weeks 3 and 4, which guided our choice.

The average visit duration per week is the figure recorded for all the sites audited by XiTi in all sectors. The floating average used is a floating average over 4 months (17 terms), preserving linear trends and minimising the variance in the irregular component.

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