To analyze your website's conversion rates, you can analyze the access log. You can also conduct a User Test.
To conduct your User Test, you need people to do the testing, but you don't need to gather hundreds of participants like big corporations do. Four or five people are enough to precisely identify any problems with your site. Recruiting four or five people shouldn't cost you an arm and a leg. Just choose people similar to your target audience. Also, rather than testing with everybody at once, if you conduct the test with one person at a time, you don't have to spend extra money setting up a test facility and other things. That way, testing is cheaper.
Conducting the Test
When you test, don't just tell people to do general tasks like "browse the site." Instead, give them specific tasks, i.e. order products (without actual money transactions), find a product with feature X, etc. During the test, participants usually ask questions like "How can I add items to the shopping cart?", but you shouldn't answer them. One of the reasons for the test is to find out if people can place orders without your help. Also, pay close attention to how they browse your site. How many times they click the mouse, facial expressions... these are reflections of your site's usability. Also, telling participants to verbalize their feeling, i.e. frustrations, would be much better than trying to read their minds.
Instead of testing only once, repeating the test (test, revise site, re-test) will make your site have higher performance. There is no rule on how many times you should repeat the test. You should decide based on the results. When reviewing the results, don't be distracted by comments based on people's preferences, i.e. liking pink more than red. Focus on the flaws affect conversion rates, i.e. confusing checkout process. Also, don't focus on small details; if you can sell, other things don't matter.
© January, 2011