[CASE STUDY] Testing PropellerAds CPA Goal 2.0 (40% ROI) (11)
[CASE STUDY] Testing PropellerAds CPA Goal 2.0 (40% ROI)
Hello STMers,
New case study. It's a small one this time, but some members might still find a lot of value in it.
If you advertise on PropellerAds, feel free to share your experience with CPA Goal 2.0 bidding model.
Original Blog Post: https://www.affiliatecase.com/propellerads-case-study/
Campaign Breakdown
Offer Vertical: Sweepstakes
Payout: €12.00
GEO: Portugal
Affiliate Network: Clickdealer
Traffic Network: PropellerAds
Traffic Type: POP/PPV
Introduction
Most of the affiliates marketers should be familiar with PropellerAds. They are a Cyprus based POP/Push traffic source platform that have been at the top of the industry for quite some time.
Early last year they have introduced a new bidding price model option that has been getting refined since then, prompting me to give it a shot.
Campaign
Let's first discuss what CPA Goal 2.0 is and how it differs from other bidding models.
This definition is taken straight from PropellerAds FAQ: "CPA goal 2.0 is an auto-optimization technology performed according to the required price per conversion. The algorithm is able to predict what traffic slices are most likely to drive quality conversions at the desired price."
Essentially, instead of you manually setting your bid price per thousand impression, the system will automatically decide how much to bid per placement in order to get the cost per action (normally set to just below the offer payout) .
You can read more about CPA Goal 2.0 here.
The campaign was a Run Of Network campaign, without any zone restrictions. I really wanted to test this to see the full power of what this price bidding model is capable of.
Offer and Landing Page
It just happened that my Affiliate Manager at ClickDealer just sent me a list of highest converted CC-Submit Sweepstake offers over the past week and there was a GEO on that list that I haven't had any other campaigns in. It was a standard "Win Samsung Phone" Sweepstake offer that showed a promising EPC compared to other offers in Portuguese GEO.
Traffic Source and Tracker Settings
Tracker Settings
Since I knew that the offer is working and that landing page style has been profitable in other GEOs on that particular traffic source, I've decided to only run a single offer/landing page combo for this test. This is not something I would recommend to beginners, at least until you have a very good idea of what works and what doesn't on your select traffic source.
Cost is set to CPC with "Auto" checked off.
Please beware that after some other tests with the CPA Goal 2.0 bidding model, the cost had to be adjusted manually for the first few days as there was a huge discrepancy in some cases of what the tracker shows and what the traffic source actually spent. I'm assuming this is due to the system's testing period, where it's trying to figure out and optimize correct pricing.
Traffic Source Settings
I've decided to set a rather conservative CPA of $8 (Roughly 60% of the offer payout), which is definitely something that should be tested further. (Higher percentage payouts vs lower). Majority of the rest of the settings were left as default.
As you can see there is no way to select a frequency cap (how many times your ad will be displayed per user), as system will automatically optimize varying frequency per different zones based on the conversion rate.
Results
To my excitement, the conversions quickly started rolling resulting in almost 500% ROI just in the first few hours of the campaign. Until the next day when the conversion dropped into the negative. As you can see from the chart below, the campaign performed all over the place, on day to day basis, but in the end resulted in being profitable.
Conclusion
It looked like upon launch, campaigns with the CPA Goal 2.0 bidding model seem to really takes off with conversions rolling fast and furious, but then eventually die off towards the end of the day or a few days after. Which is not to say that one can simply launch campaigns every few days to ensure high performance while collecting those profitable zones. Once enough zones is obtained, it might be worthwhile launching a whitelist CPA Goal 2.0 campaign that could potentially generate even more profits, and perhaps not die off as fast.
Nice!
What was the "conversion type" you selected on propeller ads?
Sent from my YAL-L21 using STM Forums mobile app
Nice case study, thanks for sharing.
Seems like pops are not dead as people used to think 
I've been looking into using CPA (and understanding it), so it's nice to see some information and a great case study. Very interesting! Thanks
Nica case study.
Was the same with the old Smart CPA campaigns and can happen on other platforms where you run on CPA as well.
First you get a boost of good traffic to check if the campagn is worth it and then the real testing starts.
Usually I see 2 different behaviours for CPA campaigns.
They either start strong and slow down continually so that they basically flatline after few days.
Or they start strong and have some (sometimes pretty hefty) performance swings the next days but keep going good overall and settle in good performance after some time.
The latter ones can run for some time pretty good.
Whenever I tested CPA G2 campaigns with WL or a good BL I had the problem to have too less volume.
Then once I asked my AM if the low volume is rather a problem of too low bid or because I started with a BL and he said a lower bid doesn´t have such impact, it´s because I started with a BL.
Since then I didn´t mess around with lists on CPA goal campaigns anymore and just run them RON.
About the bids, I mostly test at least 2 different bids.
The first is about 70%-80% payout, the other one is about 50% payout.
I don´t know why but I had several campaigns where the lower bid performed way better than the higher bid, in terms of volume as well as CR.
Interesting dicusssion.
From these campaigns I'm then fishing out the good placements and starting them as a separate WL campaign.
CC submit offer set as a SOI conversion with low payout (discovered this by mistake as I initially started it with SOI) and it works really well. But might depend from case to case.
Originally Posted by s3ks3k
Originally Posted by erikgyepes
Seems like pops are not dead as people used to think
Originally Posted by affiliatecase
Before the changes I mostly chose "Custom" conversion type because it happened that I had better results with it than chosing the correct type.
Originally Posted by twinaxe
I also think about running a small comparison test.
Same campaign and then 1 x Single Opt-In (1 step), 1 x Double Opt-In (2 steps) and 1 x Complicated conversion.
Curious to see if it will result in diffderent performance or maybe different traffic that you receive.
Thank you for such an interesting case study. By the way, we are buying stories about advertising campaigns with PropellerAds traffic. And we are ready to pay A LOT for good case studies!
All details are here: https://stmforum.com/forum/showthrea...l=1#post398150
Home > Affiliate Marketing Forum >