The question "what should I test?" comes just above the question "What should I test first?" when it comes to website optimisation.

Depending on your progress through the optimisation journey, you will inevitably need to think about how you prioritise all the tests you have thought about doing on your website or app.  What we need to do is remove the emotional attachment to tests created by internal stakeholders.

What we can do is setup rules for ourselves surrounding the tests we have in the backlog.  You can create as many rules as you feel necessary and each one gets a point if it's met or not by the test.

For example...

Main Goal of test - Does it support the company's KPIs?  Yes = 1, no = 0.  No doesn't have to be a bad thing at all, but the test could be driving sign-ups to a newsletter rather than sales of a product.

Location of test on the page - Does the test appear above or below the fold?  Above = 1, below = 0.  Again, below the fold isn't a bad thing and it'll completely depend on how each sites customers act, but something above the fold is more than likely going to have a bigger impact that something below.

Affected customers - How many customers are included in this test?  Above 90% = 1, below 90% = 0.  If the test includes all customers across all segments then it's more likely going to have a larger effect that one that only affects 10%.  However, this will depend on how much the test moves into personalisation as this may be exactly what we're after.

There are other rules you can set for yourself but they would need to become more company specific as they'll work better for each company.

Once tests have been scored on something like Google Sheets or Excel then the test order has been created with a little sum function and sorting.  All you would need are columns for; the test code, the test title, the test description, each of the rules, and a total.

Have you got a better idea than this, let us know below and let us help the industry really push optimisation.