WatchMouse Site Performance Index July 2007
The WatchMouse Site Performance Index (SPI) represents the relative web site performance of sites.
The SPI research is performed in 28 business sectors
and 5 countries. The main web site of a set of representative organisations from the sector Publishing and printing in Sweden is tracked.
Organisations in this sector: Publishing and printing
- Albert Bonniers Forlag
- Allduplo
- Allers förlag
- Arkitektkopia
- B. Wahlströms
- Bonnier AB
- Bra Böcker
- Color Print
- Liber
- Norstedts Förlag
- Nässjötryckeriet
- Studentlitteratur AB
- Åbergs Tryckeri
SPI Trend Publishing and printing Sweden July 2007
This graph shows the monthly trend of the average Site Performance Index for the websites in the Publishing and printing sector in Sweden, for July 2007. A higher bar indicates a slower website response.
Organisations in this sector: Albert Bonniers Forlag, Allduplo, Allers förlag, Arkitektkopia, B. Wahlströms, Bonnier AB, Bra Böcker, Color Print, Liber, Norstedts Förlag, Nässjötryckeriet, Studentlitteratur AB, Åbergs Tryckeri
Availability Publishing and printing Sweden July 2007
This graph shows the monthly trend of the availability for the websites in the Publishing and printing sector in Sweden, for July 2007. The yellow line is a 2-day moving average.
Availability
Availability (moving average)
Organisations in this sector: Albert Bonniers Forlag, Allduplo, Allers förlag, Arkitektkopia, B. Wahlströms, Bonnier AB, Bra Böcker, Color Print, Liber, Norstedts Förlag, Nässjötryckeriet, Studentlitteratur AB, Åbergs Tryckeri
Introducing the WatchMouse Site Performance Index
The WatchMouse Site Performance Index (SPI) represents the relative web site performance of sites.
The SPI research is performed in 28 business sectors
and 5 countries.
For each sector in each country, the main web sites of a set of representative organisations are tracked.
In this research, the performance in terms of: load time, speed and availability are measured. These figures are then used in a formula that represents the users' perceptions of waiting time, and compares these with the results of competitors in the same SPI category.
Why is the WatchMouse SPI important?
A poorly performing website has a very negative impact on customer experience. Customers tend to rate a site badly when it loads slowly, and worse still, if it is unavailable.
The WatchMouse Site Performance Index (definition) site gives you a quick overview of trends
in relative performance (more) and availability (more) across sectors and countries.
WatchMouse Site Performance Index: Definition
The WatchMouse Site Performance Index (SPI) quantifies the user perception of speed and availability of a site in a single number. The SPI is computed by adding up the time needed to load the main page, adding a penalty for each failed request. The higher the number, the lower the customer experience of the site. Sites with an SPI of 1000 perform well, whereas an index of 1500 or higher is an indication of a seriously negative user experience.
The WatchMouse Site Performance Index was developed by Dr Peter van Eijk and his team, based on many years of practical experience and research into the performance and availability of computer systems. The current version of the SPI is computed as follows.
One of the 35 test stations of the WatchMouse worldwide network retrieves the homepage of the site at a fixed frequency, without images, frames, etc. This HTML page is expected to download within 8 seconds, without any errors. If this time is exceeded, or if an error occurs, this is verified by one of the other test stations. This check is called the "2nd opinion check". If this check also reports an error, this is counted as "not available".
All 2nd opinion checks, and all successful checks, are taken into account. All the connect and download times of all successful checks are added together. For each failed (2nd opinion) check, a 10 second penalty is added. The SPI is the sum of these, expressed as an average over 1000 checks.
About performance: the four seconds rule
Even if a web site is up and running, still not all visitors may experience a good performance.
The longer it takes for pages to load, the lower the chance that users will have the patience to view them.
Some years ago 8 seconds for a page to load was considered acceptable, but as people get more experienced users, their tolerance diminishes.
Frustrated visitors translate to lost sales opportunities, for example.
Research by Jupiter Research, supports this view, and it appears that right now four seconds is a more sensible target for a web site to load.
About availability: 99% uptime is not good
An uptime of 99.99% or higher is extremely desirable for high-profile companies and those that achieve high turnovers through their sites.
Many companies stipulate an uptime of 99.9% as a minimum requirement in the Service Level Agreement with their IT department or provider,
even though this allows for 8 hours of downtime on a yearly basis.
An uptime below 99% (four full days in one year) is usually regarded as unacceptable.
