[phoenix-dev] More about conversion and website improvement

[Bjorn Freeman-Benson <bjorn.freeman-benson@xxxxxxxxxxx>]

Another good reason to have a numeric measurement (conversion) of the "goodness" of our website is that it then allows us to do A-B tests: we make a change, we run the two versions side by side (randomly dispatching readers to one or the other) and then we measure the change in conversion. Good science which rapidly leads to a better website.

We have enough traffic that we can run A-B tests in parallel or we could run them sequentially (a week of A, a week of B). Parallel has shown better results with mainstream websites so I recommend parallel but it does require a small bit of coding.

- Bjorn

--

[end of message]