Scott Miller at Vertster weighs up the pluses and minuses of running A/B split tests on your home page. In summary, he breaks his reasons into the following categories:

Why home page testing is a good idea:

• There is a lot of traffic.
• It can be influential.
• The home page gets a mix of traffic types.

Why homepage testing is a bad idea:

• The traffic is highly variable over time.
• There are more than just prospects hitting your homepage.
• Determining the goal can be a challenge.
• Any change you make will affect your search engine rankings.

Read on to learn more...

Ever wonder how your home page's Google PageRank score has changed over the past couple years in comparison to your competitors? There's a way to go back in time and check, assuming you've been listed in the Google Directory for a while. Just use the Wayback Machine and dig up archived versions of your category page in the Google Directory. For example, if your site is listed under Shopping: General Merchandise: Major Retailers, simply type into the Wayback Machine the URL "directory.google.com/ Top/Shopping/General_Merchandise/Major_Retailers/" (view the archives of this page here). This isn't just useful for checking the actual PageRank scores. Since the sites listed on the Google Directory are sorted by PageRank score, you can also review your movement relative to the other sites in your category for smaller incremental PageRank changes. For example, did your site climb to the 5th position in that category even though your integer score didn't change?

Don't rely on the Internet Archive to capture these Google Directory pages with any sort of regularity. Going into the future, you should ideally be monitoring and recording the PageRank scores of you and your competitors on a regular basis.