Testing website layouts takes time.
For any testing platform to learn what works best on a website, it needs to be able to gather enough information to make statistically relevent decisions.
Similar to an election or public opinion pool where you might have heard the term “margin of error”, Ezoic needs to be able to test a layout on enough users before it can be confident in deciding what works and doesn’t work for your website.
Factors affecting the speed of optimization
1. The size of your site effects the speed of testing optimization. This is because the system needs 7,000 visits for any particular layout test to become 'a result'. This is for the system to know that the data or result is good (i.e. reliable, and not just a few random clicks or unduly influenced by a few users).
For the system to get through enough data - it needs a few page layouts to run concurrently, for each screen size and each get approx 7k visits.
You can see why a site with 20k visits a month will take a few months to improve. Our current minimum site size is 30k visits a month.
2. Testing dilution. We know how challenging it is to be patient and wait for results - especially if it feels like you are temporarily not making more money. This can happen when the majority of the layouts being tested are generating incomes lower than your original. This can result in a dip (for a short time) in income, but doesn't last long.
We know the methodology works and we also know that we will improve the results a lot over time. Usually it takes approximately 12 weeks before you can really start to see the full effects of the technology.
3. Almost without fail, the single most important factor on whether a site succeeds or fails with Ezoic is, how much time is given over to testing.
Our platform has been hardwired to speed things up as much as possible, including variation clustering (using data from 'similar sites', etc), but at the end of the day we are limited by good old math and the fact that we need to see 7,000 - 9,000 visitors on a variation before it can make a decision.
We have thousands of layouts available for desktop, mobile and tablet (and each new test is adjusted/modified over time). We are gathering user experience and income data from millions of visitors.
Quite often it's just a matter of time and traffic before the system will find layouts for each screen size that will out-perform your old site.