HI Everyone,

For one of our client, we use mainly one LP.

Since it’s creation 20 months ago we have been testing a lot and have reached the variant BI ((number 61th if i'm correct). The thing is that it’s becoming hard to improve or even understand what’s really happening with the testing. We usually have 5 to 7 variants running with 2 or 3 being quite old and having a lot more data and leads for them (more than 12 000 leads for the first one). What make it difficult for me is the way the graph works by showing the 7 rolling days and the difficulties to compare really old variants with new ones.

I was wondering how far do you go in A/B ou smart testing ?

Do you reset the stats of the LP when you are that far ? (i only do it when we launch the LP).

Do you still use A/B testing over Smart testing ? I feel like i would be more in control and have an easier way to improve with A/B but i’m surely wrong.