How to Prioritise Your Website Experiments by ROI

If you’ve already got a wealth of evidence-based data, before long you’ll have a spreadsheet full of tantalisingly profitable feedback from your web visitors.

The question is, which of the findings is going to improve your experimentation, conversion rate optimisation and ultimately your sales, and which will be a lot of time and effort for comparably little reward?

A surprisingly easy exercise will tell you:

Go through the data and group similar responses together.
Count the responses for each group.

– The frequency of the response gives you an idea of the improvement in conversion you can expect to see if you fix the problem.

– Allocate the group a number from 1 to 10 based on the volume. 1 being, not many. 10 being the most.

Next, index the groups by the technical difficulty / cost of development. 1 being slow and/or expensive. 10 being quick and/or cheap.
Multiply the two figures and divide by 10.
Sort by your new combined profitability and difficulty index. You now have a road map for tests based on their potential return on investment. Simply execute the highest ROI tests first.