The market share of web browsers can be estimated based on the user-agent strings of visitors to popular websites. Statistics from safalra.com suggest a market share last month of 40% for Internet Explorer based browsers, and 52% for Gecko based browsers (such as Mozilla Firefox).
News (July 2009): Daily and hourly data from Net Applications
At the time of writing (July 2009), a coding error means that graphs of Net Applications’ daily and hourly market share data are available without needing a premium account. The following graphs show the daily market share of various web browsers:
- Microsoft Internet Explorer 8.0 daily (launched 19th March 2009)
- Microsoft Internet Explorer 7.0 daily
- Microsoft Internet Explorer 6.0 daily
- Mozilla Firefox 3.5 daily (launched 30th June 2009)
- Mozilla Firefox 3.0 daily
- Google Chrome 2.0 daily
- Google Chrome 1.0 daily
- Apple Safari 4.0 daily
- Apple Safari 3.1 daily
The following graphs show the hourly market share of various web browsers (note that the time scale shown under the graph is incorrect — the correct values are shown in the table under the graph):
- Microsoft Internet Explorer 8.0 hourly (released 19th March 2009)
- Microsoft Internet Explorer 7.0 hourly
- Microsoft Internet Explorer 6.0 hourly
- Mozilla Firefox 3.5 hourly (released 30th June 2009)
- Mozilla Firefox 3.0 hourly
- Google Chrome 2.0 hourly
- Google Chrome 1.0 hourly
- Apple Safari 4.0 hourly
- Apple Safari 3.1 hourly
Results (for January 2008)
18% of visits were from crawlers (mostly search engines) — these are not included in the statistics below.
| Market share | Browser |
|---|---|
| 52% | Gecko based |
| 40% | Internet Explorer based |
| 5.2% | KHTML based |
| 2.4% | Opera |
| 0.030% | NetFront |
| 0.010% | IBrowse |
| 0.0055% | Nokia (embedded) |
| 0.0039% | WebTV |
| 0.0037% | Blackberry (embedded) |
| 0.0014% | Lynx |
| 0.0011% | Links based |
| 0.00046% | NTTDoCoMo (embedded) |
| 0.00046% | w3m |
| 0.00023% | AWeb |
| 0.00023% | Dillo |
| 0.00023% | Tkhtml based |
Historical data
The following links link to plain text documents.
| 2008 | Jan | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2007 | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
| 2006 | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
| 2005 | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
| 2004 | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
| 2003 | Oct | Nov | Dec |
Notes
- All percentages are rounded to two significant figures
- ‘Internet Explorer based’ includes Internet Explorer and Avantbrowser
- ‘Gecko based’ includes Netscape, Mozilla, Phoenix, Firebird, Firefox, Galeon and Epiphany
- ‘KHTML based’ includes Konqueror, Safari and OmniWeb
- ‘Links based’ includes Links and ELinks
- ‘Tkhtml based’ includes hv3 and BrowseX
- Browsers for mobile devices and other embedded systemds are categorised by device maker unless the browser can be deduced from the user-agent string (for example, NetFront or Opera-Mini).
Why do your figures differ from those at [insert name of site]?
We’re often asked why the figures for Internet Explorer are so low, when other sites claim 80% market share for Internet Explorer. There are several reasons for this. Most sites quoting market share figures are very poor at identifying browsers from user-agent strings — it’s common to see them list ‘Netscape 5’ among their results, which doesn't even exist. The figures above distinguish between a large number of browsers, including many minor browsers that pretend to be Internet Explorer that most sites can’t identify. Demographic issues also have an effect — Safalra’s website tends to attract more technical readers (although the most popular pages are non-technical).