Leave a comment
Get the GH Bookmarklet

Ask GH

High Tempo testing is great and at the beginning we can have lots of A/B ideas. The problem is when generating over 2000 A/Bs, finding more high quality A/Bs become a hard thing to do. Benchmarking competitors and similar websites and reading articles, etc. can be very repetitive and hard to find good new ideas to test. I would like to know what are you guys doing to keep a big and high quality backlog after a long time of high tempo testing. I'm trying to create a good framework/methodology that will always work when backlog is weak or short.

  • AA

    Anuj Adhiya

    almost 2 years ago #

    This is a good question.

    We have >1500 ideas in our ideas backlog - and I've generated 1/3rd of them so I know exactly what you're talking about.

    In my view, this is a matter of focus and setting the right objectives (more about this in @danihart's post here: https://blog.growthhackers.com/growth-hacking-chaos-becomes-clear-with-roadmap-objectives-f59375091e39)

    Your objective should be a reflection of your hypothesis for your highest leverage opportunity at the moment. Once you have an objective, that automatically filters out tests that don't apply.
    Now, when you're talking about growth experiments, there's two types of tests to consider:
    a. discovery type tests (ie you're trying something new for the first time) and
    b. optimization type tests (you've run at least one experiment testing the same thing before and are now iterating on it)

    Logic dictates that every discovery type test will lead to learnings that will tell you whether there's enough of a signal to make it worth doubling down with further optimization type tests or not.

    If you do have a signal, the learnings themselves should give you multiple ideas for optimizations that you can try.
    By definition these will be of higher quality then other tests because you're trying to make that initial signal stronger.
    If you do not have a signal, the learnings should give you ideas for other potential areas that you can run discovery type tests in.

    Such a process where objectives narrow your focus to discovery tests related to that objective, which further inform optimization tests should keep your backlog full of high quality ideas you can test.

Join over 70,000 growth pros from companies like Uber, Pinterest & Twitter

Get Weekly Top Posts
High five! You’re in.
SHARE
14
14