According to the training, what are the best options for testing a hypothesis on a site that has low traffic, under 3,000 visits?
Testing your hypothesis is not very important unless you are a large enterprise website with millions of visitors.
Run split tests even if you can’t hit statistical significance because it will still give you an answer of which variation will perform better over the long-term.
First focus on driving traffic to the site in the “audience” step of the website hierarchy. Then you can look at pre- and post-implementation behavior using cohort and/or event reporting along with multiple qualitative sources to build confidence that the changes are having the intended impact on user behavior.
Set-up split tests and let them run for a few months until you have a big enough sample size to hit statistical significance.