Split testing, like most other things, is more about consistency than a burst of effort.
In theory there exists a Perfect Page. There is a “perfect set of copy” that you can use, a perfect headline, perfect imagery that converts best given your products.
In practice, you’ll never get there. And even if you did, you wouldn’t know it. But the pursuit of that page is what will increase your conversion rate and AOV above anything else.
Consistent split-testing is important for two reasons:
- Your traffic changes throughout the year. If you get 3x the site visitors, it’s extremely unlikely that your “type” of visitors just increased by 3x. It’s much more likely the type of visitor and what they respond best to will change
It informs your learnings. We don’t split-test in a vacuum. We take what we learn and use that feedback to prepare new split-tests
In pursuit of that “Perfect Page,” consistent split-testing is the thing that will get you further than anything else.
Ok so enough theory.
Here’s a framework for how often you should be running tests:
- Pick a cadence you know you can absolutely follow. Don’t do something for a week then stop. I PROMISE that consistent testing is worth it if you dedicate yourself to it
- Focus on preparing 4 tests per month. You don’t necessarily have to run them all, just prepare the tests
- You can prepare these tests all in one sitting/at the beginning of the month/weekly, etc. Usually it’s best to stagger your ideation because, again, you’ll learn from the results you get
- Once test briefs are prepared, you can then have your designer and developer execute the tests
- If you have a team/contractors helping you, it’ll actually get easier because you have the social pressure/deadlines to do it. Give a 4 tests per month quota to devs and designers and they’ll start bugging YOU for test ideas.
- Follow our Inside Out framework if you don’t know what to test.
Remember, tests need to get a sufficient amount of traffic to make a data based decision. It does no good to spend hours preparing a test for a page that gets 10 sessions a week.
Tests also need to run for 7 days to call it as a loser and 14 days to call a winner (at least).
So trying to run 5 tests a week probably isn’t going to work unless you have huge amounts of traffic.
Now look, I know that every person tells you that THEIR thing is the thing you should be spending time on. If you followed everyone’s biased best practices you’d be flossing your teeth 5x a day and deep-cleaning your oven for 90 minutes a day.
That being said, you can run a very fruitful CRO program in a few hours per week. It’s also one of the few times in marketing that you can do work once and reap the benefit for years to come. Imagine trying that with ads… make one ad and it brings in customers for years (I wish).
CRO is a worthwhile endeavor because once you get a conversion rate increase, you get it in perpetuity. It’s like passive incremental revenue.
Remember, tests need to get a sufficient amount of traffic to make a data based decision. It does no good to spend hours preparing a test for a page that gets 10 sessions a week.
Tests also need to run for 7 days to call it as a loser and 14 days to call a winner (at least).
So trying to run 5 tests a week probably isn’t going to work unless you have huge amounts of traffic.
Now look, I know that every person tells you that THEIR thing is the thing you should be spending time on. If you followed everyone’s biased best practices you’d be flossing your teeth 5x a day and deep-cleaning your oven for 90 minutes a day.
That being said, you can run a very fruitful CRO program in a few hours per week. It’s also one of the few times in marketing that you can do work once and reap the benefit for years to come. Imagine trying that with ads… make one ad and it brings in customers for years (I wish).
CRO is a worthwhile endeavor because once you get a conversion rate increase, you get it in perpetuity. It’s like passive incremental revenue.