Home > Paid Traffic Sources > Adult Traffic (NSFW)

Statistical Significance in Testing on Adult (5)


07-19-2016 05:15 PM #1 sfmedia (Member)
Statistical Significance in Testing on Adult

Hi guys,

I'm running adult dating offers on TrafficJunky. The great part about these low payout/high converting offers is that you can get testing results quickly.

How many conversions do you typically look for before determining a winner? Sometimes I find myself comparing ads that have 3 conversions vs 2 conversions and 1 conversion can happen via random chance so this probably isn't ideal.

Thanks!


07-19-2016 08:39 PM #2 casual_dating_offers (Member)

Hey,
What geos are you promoting? mob or web?


07-28-2016 12:05 PM #3 iMonetizeIt (Senior Member)

Usually the stats can show true information about offers, when you get from 50 to 100 conversions on each GEO. Less is not too informative.


07-28-2016 12:50 PM #4 crysper (Member)

Judging by gut like most affiliates do is a recipe for failure, especially if you don't have a lot of experience on this.

I recommend using a Bayesian calculator, here's one free https://www.peakconversion.com/2012/...al-calculator/

or the paid one from AffKit(mine) http://www.affkit.com/tools/when-to-cut-landers

or an optimization tool that does this automatically(check my signature).


07-28-2016 03:53 PM #5 matuloo (Legendary Moderator)

You are right, with low amounts of conversions all it takes is a random lead here and there and the results change completely.

In adult, its very important to run with several funnels head to head for a longer period of time and only cut some of them when they really fall behind significantly OR one of them really takes the lead.

Getting something like 11 to 8 to 7 leads in a day is not enough for example, the next day can look totally different - which is the case with Adult dating more often than with any other vertical I ran so far, the day-to-day differences can be huge. Thats why it's important to run more than one banner/LP/offer in adult.

To determine if you have a winning offer or funnel, I would recommend to try to run for a full week so you can see how the cvr changes on weekends, during night etc ... you don't have to do this all the time, but you need to go through this process a few times to figure out how it works.


Home > Paid Traffic Sources > Adult Traffic (NSFW)