Top of the League Table and best site overall was The Cheshire, with Bradford & Bingley in second place. At the bottom of the table was Abbey.

- Site with the lowest number of warnings [poor html] was Co-operative Bank [with 83]. AIB again had the highest number of warnings with more than 22,000 representing a significant reduction from the last report; second highest was Halifax with over 18,000, indicating that site quality is dropping.
- One site was error free, LloydsTSB.
- 54% of the sites tested passed basic metadata test for their front page.
- 51% of the sites passed basic speed tests for the front page, testing the site for a modem, ADSL and network based user – a great improvement.
- Site with fastest overall server response time was Leeds Holbeck, with an average page response time of 0.05 seconds. The site with the slowest response time was Abbey, with an average page response of over 2.45 seconds per page.
- Site with the fastest download speed was the Northern Rock; the slowest site was Birmingham Midshires.
- The site that contained the most errors was (again) the Coventry Building Society, with over 1,500 the next highest was Abbey with 129.

Accessibility testing
- Top site this month was Northern Bank
- 9 of the sites tested achieved greater than 90% compliance
- 7 sites scored between 50% and 90% compliance for the pages tested
- 17 sites scored less than 50%, with 10 of the 17 sites achieving less than 5% compliance
- 2 sites had no compliant pages, Britannia and Portman, and Abbey had less than 0.5% compliance.

Full Accessibility results are shown on page 7 of the report. The tests completed address a number of the requirements for mandatory priorities 1 only.

The results of accessibility testing currently do not affect the overall ratings or rankings for the League Tables contained in this report,. The range of tests [Web Accessibility Initiative WAI] that can be completed automatically are limited, 100% compliance with the automated tests, does not mean 100% compliance to the requirements.

The findings take into account the website size; e.g. a website with 10 pages and 1 error would score the same as a website with 100 pages and 10 errors – as both having the same page/error ratio of 10:1. Ratings are based on the many millions of website tests conducted.

Tests were confined to the top 250 pages of each website and were completed on the 25th March 2004 - a free copy of your own site summary is available on request. Our next report will be produced around the 28th May 2004.