|
|
|
|
|
|
Traffic to websites :
A less pronounced seasonal downturn in July 2008 |
|
| Website traffic is seasonal and specifically experiences a habitual summer break. This study compares the seasonal effect for the month of July 2008 with the average effect for the months of July over the previous 6 years. |
| Perimeter: |
- Study conducted from July 2001 to July 2008
- Perimeter of 249,712 mobile websites
|
|
| Fewer vacations for websites? |
An analysis of website traffic allowed us to identify the existence of seasonal patterns in website audiences, with key moments over the course of the year, such as a summer downturn and a revival in January. Thanks to non-parametric deseasonalization methods, we were able to investigate these monthly seasonal effects.
The following graph shows the average recorded monthly seasonal effects for website traffic between 2002 and 2007. The months of January and March experience intense seasonal effects with peaks of +9% and +11% respectively in website traffic. While the month of August is most seriously affected by a downturn with 14% less website visits, followed by July.
|
|
Between 2002 and 2007, an average July is characterized by a downturn of -9% in website traffic. Now look at the figure for July 2008.
In order to characterize the seasonality of website traffic in July 2008, we compared the seasonal effect for the months of July from 2002 to 2007 to the same effect for July 2008.
This revealed that the seasonal effect for July 2008 is less pronounced than the average seasonal effect for the months of July from 2002 to 2007. It actually corresponds to a downturn of -7.7%, compared to an average seasonal effect of -8.7% for the months of July from 2002 to 2007. The seasonal effect for July 2008 is therefore less pronounced than for July 2007 (-8.8%) and especially for July 2006, which experienced a particularly severe downturn (-12.1%).
Is the trend in website traffic in the process of moving towards less pronounced summer downturns or in part shifting from July to August? It would be interesting to analyze the seasonal effects for the month of August in order to cast more light in this point.
Naturally, seasonal effects and shifts in them may vary greatly depending on the business sectors examined (e-commerce being extremely active in December/January for example). So, please don’t hesitate to contact us if the variation in this study of your market is of interest to you..
We will be back in XiTi Monitor in September with a study of the seasonal effect for the month of August 2008.
|
Methodology:
The registered audience for the websites audited by XiTi enabled us to detect, very early, the existence of seasonal effects, which investigative methods allowed us to examine.
In this study, the figures in which we are interested are those for website traffic, which represent changes in audience generated by a fixed number of websites. The data series successively integrates the daily rates of change, calculated on the basis of a constant perimeter for all the websites audited by XiTi. The rates of change thereby integrate into the calculation method the arrivals and departures for websites within the perimeter of XiTi.
Tools designed* for processing chronological series enabled us to break down this monthly series. The components are:
- The trend in the series, representing long-term change in the series.
- The seasonal component, representing sub-annual fluctuations, monthly in our case, which are repeated more or less regularly from year to year. It highlights phases of growth and recession.
It is this seasonal component which we examined during the course of this study.
*Non-parametric deseasonalization method.
|
|
|
|
|
|
|
|