How much is enough data to make any decisions about anything?

[petebowen]

You can use standard stats calcs for this, but my approach is to supplement those by asking if a few more clicks/conversions/whatever would change the way I feel about the result.

For example

If I had 10 clicks and 1 conversion – a 10% conversion rate and I ask what would happen if I got 1 more conversion – raising the conversion rate to 20%. A 20% conversion rate feels different to a 10% conversion rate.

But, if I had 1000 clicks and 100 conversions – the same 10% conversion rate, and I got 1 more conversion it’d change the conversion rate from 10% to 10.1%. I wouldn’t feel any different about 10% vs 10.1% conversion rate.

I found this article on the subject useful: [https://www.searchenginejournal.com/interpret-ppc-data-correctly/](https://www.searchenginejournal.com/interpret-ppc-data-correctly/)

[Jodiefoatersucks]

I make my changes based off of weekly performances. However if a holiday is within that week I have to factor that into my data

[incrementality]

Sounds like you’re trying to get to a level of confidence that what you’re seeing isn’t due to chance, which forms the basis of hypothesis testing.

I would recommend getting into the basics of A/B testing and understanding how volume and effect sizes interplay between the level of confidence one has in the results.