As someone actively earning through traffic arbitrage, I'm always on the lookout for new tools to improve my results. Recently, I decided to research new ad networks, CPA networks, and other tools for arbitrage, and ended up testing a combination of three tools.
Below, I'll share my case study.
In their catalog, I chose the VotTak offer from the mobile app vertical. It turned out to be a good offer with decent rates. I also liked that it had two payment models: for installs and for watching videos in the app. This allowed me to choose the best model depending on the traffic.
I immediately disabled the platforms that were significantly underperforming (ROI below -20%). I decided to leave the platforms with small losses (-10-20% ROI) alone for now, thinking they might balance out later with pre-landing page tests.
I repeated the same actions for other GEOs.
At some point, optimizing everything in one stream became inconvenient, so I decided to separate the selected platforms into a different path named Only VT. The screenshot below shows the general flow and the selected platforms.
The ROI fluctuated daily, but overall, we were positive. This is crucial to consider when evaluating results, as turning off a seemingly underperforming platform too quickly can be a mistake. My recommendation is to evaluate based on larger data sets. We’ll increase ROI through pre-landing pages and further platform optimization.
Then I tested landing pages. By the time I wrote this article, some pre-landing pages were still being tested, but a few clear leaders with over 50% ROI compared to the initial pre-landing page with 4% ROI had emerged.
As a result, I expect further ROI growth for positive platforms and for negative ROI platforms to break even within reasonable limits.
Fraud levels vary across platforms, and generally, up to 20% is normal. As shown in the screenshot, we didn’t exceed this fraud threshold. Kaminari also has a suspicious user block, where the system can flag users as fraud based on internal settings. For example, users with various proxies, anti-detect browsers, AdBlock, and other settings that can negatively impact traffic assessment.
A plus of anti-fraud systems is that you can start your optimization with them. You run 50-100 clicks on each publisher and decide what to do with them, thus saving significant testing budgets. This can be done simultaneously or afterwards.
Below, I'll share my case study.
Common Challenges
Almost every arbitrager faces the problem of low conversion rates and high advertising costs, which negatively impact ROI. I aimed to find a solution that would help me attract better quality traffic, monetize it effectively, and minimize fraud risks. Here’s what I achieved.CPA Network
I spent a lot of time analyzing different CPA networks, comparing offers, rates, and conditions. For this test, I chose OfferGate. They’ve been in the market for six years, have plenty of offers, and I found no negative reviews about payments.In their catalog, I chose the VotTak offer from the mobile app vertical. It turned out to be a good offer with decent rates. I also liked that it had two payment models: for installs and for watching videos in the app. This allowed me to choose the best model depending on the traffic.
Ad Network
After researching many ad networks, I chose the well-known and highly-rated network Kadam. I decided not to focus on a specific GEO but launched ads worldwide to assess the potential. I chose the popunder ad format, as it’s cheaper than in-app push.Process
After the first few days of running ads and analyzing the statistics, clear winners emerged: IQ, NL, TR, AE, IR, and a few other GEOs. IQ stood out the most because it provided the highest traffic volume and was relatively easy to optimize.I immediately disabled the platforms that were significantly underperforming (ROI below -20%). I decided to leave the platforms with small losses (-10-20% ROI) alone for now, thinking they might balance out later with pre-landing page tests.
I repeated the same actions for other GEOs.
At some point, optimizing everything in one stream became inconvenient, so I decided to separate the selected platforms into a different path named Only VT. The screenshot below shows the general flow and the selected platforms.
The ROI fluctuated daily, but overall, we were positive. This is crucial to consider when evaluating results, as turning off a seemingly underperforming platform too quickly can be a mistake. My recommendation is to evaluate based on larger data sets. We’ll increase ROI through pre-landing pages and further platform optimization.
Then I tested landing pages. By the time I wrote this article, some pre-landing pages were still being tested, but a few clear leaders with over 50% ROI compared to the initial pre-landing page with 4% ROI had emerged.
As a result, I expect further ROI growth for positive platforms and for negative ROI platforms to break even within reasonable limits.
Fraud Protection
Despite the efforts of ad networks to combat fraud, I still encountered it in my campaigns. The issue lies more with the platforms within the networks. Poor-quality traffic sources harm the networks as much as they do us. Here, I decided to use the demo version of the Kaminari tool, which provided protection and saved my budget on traffic acquisition. Using anti-fraud products is an often overlooked but critical aspect of success.Fraud levels vary across platforms, and generally, up to 20% is normal. As shown in the screenshot, we didn’t exceed this fraud threshold. Kaminari also has a suspicious user block, where the system can flag users as fraud based on internal settings. For example, users with various proxies, anti-detect browsers, AdBlock, and other settings that can negatively impact traffic assessment.
A plus of anti-fraud systems is that you can start your optimization with them. You run 50-100 clicks on each publisher and decide what to do with them, thus saving significant testing budgets. This can be done simultaneously or afterwards.
