FYI – This case study was my entry for a competition on the AffLift Forum.
You can view my original thread HERE

Update: The competition finished and my follow-along won first prize!

Hi all,

I’ve been doing push ads since the original Chrome update came out. 99% of my focus has been on mobile push as desktop didn’t perform well for the types of offers I generally promoted, namely cc submit sweeps. Since the whole COVID situation, I was thinking of the types of offers that could be popular and have been testing different verticals outside my comfort zone. This will be an interesting ZeroPark Push Notification Case Study as I haven’t seen many case studies/follow-alongs that focus on Desktop traffic. This particular offer has done quite well on a number of other push traffic sources and GEOs, I am focussing on getting the highest AOV and user-retention possible as to continue to get my payout increased overtime. I’m running completely whitehat with strict rules surrounding creatives, LPs, terminology and advertisement techniques.

Without further delay, let’s jump into this special case study. Shout out to ZeroPark and AffLift for the contest!

Traffic Source : Zeropark

Tracker : Voluum

Offers : CPS Utilities/Software

Payout : $45-$60

Budget : $20k+

Geo : T1 & T2

I launched this campaign on ZP yesterday and have been seeing some promise. I’ve been running this particular offer for almost a month now in 5 or 6 different GEOs and on many other traffic sources so I have a feel for LP and offer combination has the best performance, although I do still split-test on each new source.

Initial Optimization Rules

I developed my own private automated optimization platform which integrates with the majority of push sources I buy traffic from, so long as they have an API. ZeroPark is a great source for my optimization platform as I can rest assured the statistics my platform uses to optimize are 100% accurate due to the bridge between ZP and Voluum.

  1. IF {target} clicks > 75 & LP clicks == 0 THEN pause {target}
  2. IF {source} clicks > 130 & LP clicks == 0 THEN pause {source}
  3. IF {target} cost > $27 & conversions == 0 THEN pause {target} & mark “?” {target} (mark it so I can manually assess it and decide based on other KPIs whether to resume)
  4. IF {source} cost > $50 & conversions == 0 THEN pause {source} & mark “?” {source} “”
  5. IF {target} conversions >= 1 & {target} spent > $50 THEN adjust {target} CPC to 50% of {target} EPC

My reasoning for the above rules are simple. Since I don’t do direct-linking, if visitors from a specific target click my notification over 75 times but not a single one of them clicks-through, I class that as a relatively low-quality placement. Same for the source, although I give it a bit more leeway to allow lower-volume targets within the source to have a chance to register some click-throughs.

As for rules 3 and 4, since there is so much volume available in the GEOs I’m targeting – I’m using an aggressive strategy of pausing sources when they reach 55-70% spend of the conversion value. Those recently paused sources/targets marked with a “?” marker, I will analyse manually to see if they are worthy of resuming (higher than 3-4% CTR, 0% suspicious visits, more expensive CPC, etc.

Rule #5 is just for controlling the higher-spending targets to ensure they don’t get out of control with a high-CPC winning all the traffic.

Let’s see how these rules work when applied to Desktop campaigns as I don’t have much experience targeting this platform, as I mentioned my main focus was previously on CC submit sweeps on mobile.


Campaign 1 (T1 GEO):

Day 1 – 4th May:

  • Launched the campaign towards the end of my day, got a couple quick conversions and was able to identify 3 or 4 targets to blacklist right-away. I let it run with $200/daily budget set and woke up the next day (5th of May) to about breakeven. I also saw my optimizer was able to pause another few targets with 0% LP CTR.

Day 2 – 5th May:

  • Playing around with the bid as I noticed I was averaging only 15% win-rate. Ended the day negative but overall about breakeven which is pretty good considering I only launched it a day ago.

Day 3 – 6th May:

  • I narrowed it down to 1 creative that was outperforming the others (Creative 1) which has a $22.90 CPA, not bad at all. However I’ll be monitoring for any performance decreases and won’t hesitate to resume Creative 3 and Creative 4. I paused #4 before I had done a lot of blacklisting of bad targets/sources so it definitely has potential for lower CPA. I’ll have a think about it.
  • Also I started bidding much higher on my top converting target. $72 spend and $534 revenue! 642% ROI – not too shabby!! Let’s hope this target sticks around and is only converting for me so others don’t start bidding higher too. Regardless, at this performance I could easily bid up to $0.65 CPC and still be in profit.
  • Happy with performance overall. Additionally I looked through all the sources/targets my optimization rules have blocked so far and from my manual analysis I decided only 1 target was worth resuming, the rest didn’t show much promise. I prefer being aggressive, especially on sources like ZeroPark that have a huge amount of volume. Also having run probably hundreds of campaigns with ZP, I already have an idea which targets and sources to blacklist right away and the behaviour exhibited by sources that I will need to inevitably blacklist.
  • After my past experience with Desktop push campaigns being terrible, it seems if you pick the right offer, optimize your LPs for desktop devices. Interesting fact – I’m not using push collection scripts on any of my LPs for this campaign. Opt-in rates on desktop, at least in my experience, are upsettingly low – not worth the small drop in CTR that comes with having a push opt-in prompt on the landing pages. I need all the CTR I can get considering the strict regulations imposed on promoting this offer.
  • I’ll check back in tomorrow, hopefully I keep seeing green and a ROI increase wouldn’t go unmissed either! Push can be a strange beast, that’s for sure – consistency is not a word I’d associate with push traffic anymore. However, so far so good on this campaign! Let’s get this bread $$$$$$$

Day 4 – 7th May:

  • Not much to say here, didn’t make any real changes but it seems the competition against my bids is rising. Increased the bids on profitable Targets.

Day 5 – 8th May:

  • Much the same story as above. The campaign is feeling neglected despite my custom auto-optimisation rules constantly monitoring the Targets/Sources. Need to create some bid-increase rules for profitable Targets with decreasing winrate%.

Day 6 – 9th May:

  • I’ve identified the current best Target and have been incrementally increasing the bid each day to see how it affects my win-rate % on the Target. It’s currently sitting at $207 cost and $762.90 revenue. Noice!

Day 7 – 10th May:

  • So today I created a new LP using a fresh angle I haven’t seen before. I try to think of new angles and create new LPs/creatives at least weekly – especially if campaign performance isn’t great. I added the lander about halfway through the day today and it finished the day at 223.72% ROI. Not bad! We might have a new winner, let’s see how it goes tomorrow.
  • Additionally the top Target I mentioned in the Day 6 report above is still going strong. Considering creating a Target campaign specifically for that 1 target. Perhaps I’ll add the top 2-4 targets into the new Target campaign and blacklist them in this main campaign. Pictured below – Top 3 Targets overall (Check out #2 – $1.30 spend, $149.94 revenue!):
  • Also as a note, the first Target in the screenshot above is the one I posted yesterday. It’s sitting at 280% ROI overall now.
  • I admit my attention has been directed to my more profitable campaigns on other sources but I’m determined to get this campaign to its full potential. Let’s do it!

Day 8, 9 – 11th, 12th May:

  • Decided to ax the campaign altogether after Day 10. Whilst I know I could’ve probably made it profitable by putting in more work – considering I have the same campaign on other traffic sources with much higher volume and ROI – it’s a much better use of my time to focus on those sources instead of ZP for this particular campaign. As far as my plan for this ZeroPark follow-along, I am going to be diverting my focus towards my new ZP campaign outlined below!

Leave a Reply

Your email address will not be published.