Test: Propeller Ads Auto Optimization (5)
I run few CPA Goal campaigns on Propeller Ads where I wanted to test if it could bring good additional volume (and profits) when I also run them on CPC with Propeller Ads Auto Optimization feature to block unprofitable zones.
The idea was that campaigns with Auto Optimization enabled are in the end normal CPC campaigns just that zones get automatically blocked when they exceed a certain spend without good performance so that they could open some nice traffic pockets that are not active in the CPA Goal campaigns.
Usually I use TheOptimizer for it but when I can add such automation directly in the trafficsource it could probably make the setup easier and faster.
For the tests I created 6 CPC campaigns, 3 Push campaigns with High/Medium/Low activity and 3 InPage campaigns with High/Medium/Low activity.
I used the same funnel with a good converting offer for all 6 campaigns and set the desired CPA to $0.30.
For all campaigns I used a reasonable CPC bid and the campaigns started all RON with no BL or WL.
A picture is worth a thousand words so before I bore you now with a long story I will just show you some stats right away.
It´s the first 4 pages = first 200 zones sorted by impressions.
Campaign 1 (Push, High activity):
Campaign 2 (InPage, High activity):
Campaign 3 (Push, Medium activity):
Campaign 4 (InPage, Medium activity):
Campaign 5 (Push, Low activity):
Campaign 6 (InPage, Low activity):
What do we see here?
On first sight we can see that in 4 campaigns by far most of the zones are blacklisted.
"This zone is excluded from your traffic sources. The targeted traffic segments were ineffective for the chosen CPA Goal."
The InPage campaign with Medium activity shows the same result on page 1 and 2, on page 3 there are some more active zones and on page 4 even more although there are still many excluded zones as well.
Only the Push campaign with Low activity targeting shows by far mostly active zones and only few zones are excluded.
Now let´s have a closer look at the stats.
I was pretty surprised to see that so many zones were excluded because the desired CPA for all campaigns is $0.30 but by far most of the zones didn´t even spend $0.30 yet they still got excluded.
I would have expected rather the opposite, that many zones spend more than the desired CPA and not less but it seems for the Auto Optimization the CPA is only one factor to consider.
Then I noticed that it seems to make a difference if the zones have conversions or not.
Apart from one zone all zones with conversions were more than the desired CPA in loss.
On the other hand the CPA of all active zones with conversions is lower than the desired CPA so it´s pretty obvious that for converting zones the CPA is the main factor to exclude them.
For zones without any conversions it´s not that clear.
When I think about reasons why these zones get blacklisted the following things come to my mind:
- They exceeded the desired CPA and didn´t convert
- The CTR is too low
- The win rate is too low
The desired CPA can´t be the main reason because many zones got excluded without even spending $0.30.
It also can´t be the win rate alone, when it´s too low it should rather result in less volume and not necessarily in a block so the most logical and definitely an understandable reason would be that the CTR is too low so that the trafficsource doesn´t make enough money from the zone in the affected campaign.
More likely is that it´s a combination of win rate and CTR.
But then there are zones that got excluded although they have higher CTR and higher win rate as well as lower losses compared to other zones that are still active.
In the end I can´t really see what exactly triggers the Auto Optimization to exclude zones without conversions but what I can clearly see is that there are many zones excluded that I would give some more time.
A good thing is that in all 6 test campaigns the algo didn´t let zones overspend much, in by far most cases it blocked zones after latest 1 1/2x desired CPA in loss, no matter if the zones converted or not.
Nonetheless I better stick with TheOptimizer for such things because it not only gives me more options for optimization but there I can also decide myself when to block zones.
In my opinion it would be better when the Auto Optimization would only blocks zones that really spent at least the desired CPA instead of blocking zones by other algos as well, the way it works now is more like a restricted version of CPA Goal.
It also blocks zones automatically but in contrast to CPA Goal campaigns it doesn´t do any other optimization so your campaigns can´t really grow with it.
What´s even worse is that the Auto Optimization also excludes very good zones and cuts lots of potential from your campaigns.
How do I know?
As I said in the beginning, I run the funnel that I used for the Auto Optimization test also on CPA Goal and I use also a bid of $0.30 CPA there.
Although all CPA Goal campaigns were profitable most of them were still pretty small so let´s skip them.
The Push campaigns with High and Medium activity targeting however were really good.
When we compare now the results of the two CPA Goal campaigns with their Auto Optimization counterparts we can see quite a difference.
Here are some stats from the Push campaign with High activity
And here are stats from the Push campaign with Medium activity
As you can see in some cases the very same zones that were excluded by Auto Optimization generated several hundred conversions on CPA Goal.
The most extreme difference is where a zone generated 876 conversions on CPA Goal and only 1 conversion with Auto Optimization with exactly same targeting and same target CPA.
876 vs 1
Conclusion: I never used Auto Optimization before but it seemed like a nice little helper.
After my tests I have to say that I probably won´t use the Auto Optimization feature in future again.
P.S. I know that the post is long because of the images but in my opinion it´s was the best way to show what I wanted to show 
Nice! Thank you for the info! Short but very sweet. Goes to show that not all optimization features are helpful.
Amy
For me most surprising was the massive difference then between Auto Optimization and CPA Goal, didn´t expect it.
Little update: The Auto Optimization experiment inspired me to run another test.
Stay tuned 
Originally Posted by twinaxe
Stay tuned
Inspiring - and looking forward!
Amy
Home > General > Affiliate Marketing Forum