ROI 72% on iPhone 11

Traffic sources: ComeAds, PropellerAds
Affiliate network: ExpertMobi
Offer: (12794) [WAP+WEB] Get iPhone 11 / US
Flow: User signs up with their email > Conversion (Soi)
GEO: US
Campaign period: 16.09.19 — 21.09.19
Cost: $1056
Total revenue: $1815
Profit: $759
ROI: ~72%

CC submit sweepstakes were a real trend of the end of this summer. We got lucky to test them on time and had some nice results with FR sweeps. Things got harder by the middle of September, so we decided to test some classic SOI sweepstakes instead. We've been working with EU traffic for a pretty long time by that moment, so we wanted to try US sweepstakes.

We needed a new and fresh campaign. We've found a newsworthy event pretty quickly: it was a new iPhone 11 release date coming. We got our first iPhone 11 offers on September 11. We've been working with our US advertiser for a really long time, I do believe that many of you know who I'm talking about.

So, we decided to use pop traffic just as usual. We tested ComeAds and PropellerAds. We were to face some common problems: we needed to get rid of bot traffic, refine live traffic and find the best performing creatives.

In order to check traffic we used google captcha + mobile orientation set up on the landing page.

The algorithm was the following:
- If traffic did not pass captcha and mobile orientation, it was considered bot traffic.

There was also another interesting feature: even live traffic could sometimes fail verification, because in order to pass through the captcha, user needed to stay on the landing page for 2-3 seconds.

Here is the basic landing page:




After taking a little quiz user gets to the following page:





We've also tested this option:



The algorithm was the following:
- If there were clicks, then traffic was considered bot, because no human being would click onto this.

We started testing US offer with ComeAds traffic. We chose them because their traffic is clean as a rule, though volumes are not big. But this source is perfect to check if offers are live and perform well or not. We saw several leads pretty quickly even with low volumes of traffic. So we decided not to make black lists and just left traffic going in small volumes. Statistics looked like this:


We sent traffic from PropellerAds the first time on September 14, we spent $20 and got our first blacklists. ROI was 88%, which was a nice result already. We started to ban zones checking following parameters:

- we sent traffic with the cost of one lead and got no conversions = ban.
- CTR was lower than 1% = ban
- less than 70% of traffic passed through mobile and less than 30% of traffic passed through captcha = ban

On September 16 we sent traffic the same way. We sent it in a pretty fast pace. We noticed that once volumes fell, bid began to increase. Things started to get better soon. In the afternoon we found another source that turned out to be the best performing one, as you can see from the screenshots. To sum it up, we got 114 conversions and CR 1.33% on September 16.



We continued optimizing our campaign. We removed LPs that didn't really work, banned some more sources using the following criterion: it got banned if we received no conversions from traffic that we spent at least one lead cost at. The offer kept performing bringing us conversions.

On September 18 we realized that our top sources now looked different during traffic verification procedures:

- 90% of traffic passed through mobile orientation
- 50% of traffic passed through captcha

We received the peak number of leads (189) on September 21. After that we lost this source and we didn't reach similar numbers using any other source we had with this campaign.

We kept sending traffic to turn a profit for some time and then we paused traffic, because our main goals have already been reached by that moment. We tested US, collected some useful data on PropellerAds and found a working solution for Win Samsung/Win iPhone SOI sweepstakes.

September 19 turned our to be the most profitable day with 276.53% ROI.

Here is the stats from tracker: