There are many tools available in the market that allow you to do A/B or Multivariate Testing on your webpages. The solutions offered are either hosted solutions or stand-alone products. The hosted solutions are more popular, because it is as simple as pasting a script on all the pages of your website. Almost all the solutions are priced based on the Page Views served during the test, and most of them offer a few days of trial for you to evaluate the capabilities.
Here are some of the popular tools
VertSter offers A/B Testing as well as Multivariate Testing solutions. Using Vertster's A/B testing module, you can have the rotation of elements in either a sequential fashion or in a variable rate. In sequential rotation, one element is served to every other visitor. In Variable Rate Split, the client can decide on what percentage of users should be served an element being tested. Vertster also has a solution for Multivariate Testing which is based on the famous Taguchi algorithm.
Offermatica offers a hosted solution for A/B Split Testing as well as MultiVariate Testing. Their solution splits visitors into two segments randomly and displays one of the two desired elements that is being tested. In Multivariate Testing, Offermatica's solution allows you to test only a subset of all the possible combinations of elements. The outcome of the rest of the combinations are predicted based on their statistical model. At the end of the test period, the solution gives the winning combination of elements. It also notes a combination that it predicts to be a potential winner, with the leverage of it's statistical model.
PageLab comes from the SiteLab stable, that has all the tools required right from website building to optimization. The PageLab software provides solution only for A/B testing and not Multivariate Testing. The highlight of this software is that after establishing sufficient data, the PageLab software automatically determines the page-element that is resulting in maximum conversions. PageLab then starts serving that element to both the visitor segments, thereby taking the decision by itself.