WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: help with consluting services

for

From: Steve Green
Date: Apr 22, 2023 10:07AM

I agree with what Glen says, with the exception that you tend to get faster once you have tested a few pages because you are seeing the same coding techniques and non-conformances again and again. At least you should be - there's something wrong if you aren't.

Once we have determined that the coding is indeed consistent, we usually take a component-based approach to pricing. On most websites you can identify numerous components that are always the same wherever they appear, such as page headers, footers, accordions, date pickers, embedded videos etc. We work out time allowances for the components and test one example of each, taking into account any variants. Then when we are estimating the cost of testing individual pages, we exclude those components. The result is that it might take 45 minutes to test a page instead of 90 minutes.

This approach minimises the time and cost of fixed-price projects, but doesn't lend itself to a price-per-page model. That said, a price-per-page model is a hopelessly inaccurate way to try and estimate the cost. Fixed-price and time-and-materials are better.

If you don't have any experience of estimation, I would advise going for a T&M model and tell them you can probably test between 3 and 5 pages per day on average, depending on the size and complexity of the pages, the amount of unique content and how badly coded it is. If they insist on a fixed price per page, guess high (perhaps 3 pages per day) because you will be taking all the risk.

Steve Green
Managing Director
Test Partners Ltd