Push, like other formats, has been "buried" more than once, although traffic volumes and CPCs have not declined, the competition remains high, and new push networks regularly emerge. One of these relatively fresh networks, I decided to test almost from scratch since I hadn't worked with pushes for a long time and had no working connections at that time. The last time I ran pushes was about 2-3 years ago, and back then they converted well on almost everything. According to the case studies, the top verticals now are dating, sweeps, gambling, and crypto, so we'll choose from this list. The idea was to spend $300 on an unfamiliar source and understand how this format works and how to work with it in 2023.
At the beginning, I chose 3 push networks and created accounts, but the first two didn't respond right away, while Rollerads immediately connected me with a Russian-speaking manager. I checked their traffic volumes and average CPCs by country and decided not to spread myself too thin across multiple sources. I immediately clarified with the manager which verticals perform best specifically for them because each network has different sets of publishers and rules. There might be a bias towards more gray niches, like adult. The freshness of the database can also affect conversions in different verticals, so it's always better to stay in touch with the manager as they have up-to-date information on current trends. I was told that crypto, sweeps, software, and a bit of dating are currently on top. It doesn't make sense to test crypto with $300, so I chose sweeps because I'm currently working with a partner network on sweeps smartlinks on Facebook, and they have offers for tier-3 countries. This will allow me to acquire more traffic with my modest budget, test several offers/GEOS, and gather a larger amount of data. I selected several tier-3 Asian countries where there is a lot of cheap traffic: Thailand, Indonesia, Vietnam, and Malaysia. I'm going to the spy service to see the latest approaches and find texts in the required languages. I didn't want to pay for a subscription, so I found a free spy service for pushes - https://spy.house/ - great, RollerAds is included in their network list, and that's where I'll be running my campaigns.
At the beginning, I chose 3 push networks and created accounts, but the first two didn't respond right away, while Rollerads immediately connected me with a Russian-speaking manager. I checked their traffic volumes and average CPCs by country and decided not to spread myself too thin across multiple sources. I immediately clarified with the manager which verticals perform best specifically for them because each network has different sets of publishers and rules. There might be a bias towards more gray niches, like adult. The freshness of the database can also affect conversions in different verticals, so it's always better to stay in touch with the manager as they have up-to-date information on current trends. I was told that crypto, sweeps, software, and a bit of dating are currently on top. It doesn't make sense to test crypto with $300, so I chose sweeps because I'm currently working with a partner network on sweeps smartlinks on Facebook, and they have offers for tier-3 countries. This will allow me to acquire more traffic with my modest budget, test several offers/GEOS, and gather a larger amount of data. I selected several tier-3 Asian countries where there is a lot of cheap traffic: Thailand, Indonesia, Vietnam, and Malaysia. I'm going to the spy service to see the latest approaches and find texts in the required languages. I didn't want to pay for a subscription, so I found a free spy service for pushes - https://spy.house/ - great, RollerAds is included in their network list, and that's where I'll be running my campaigns.
As we can see, the top-performing creatives for sweeps are the latest models of iPhones, Samsungs, PlayStations, and coupons from local supermarkets. Since I'm planning to run a broad campaign using a smartlink, where multiple offers will be in rotation, I'm starting with creatives that have general phrases like "You've won!" or "Claim your prize!" These will work for most sweepstakes offers. I'm selecting texts for each GEO, and from them, I'll create push notification texts:
I have the texts and images, the smartlink is ready, it's time to create the campaigns. In the campaign settings of Rollerads, I can choose between standard or premium traffic. I'll start with premium and test standard later. The rest is fairly typical: selecting the GEO, setting the bid, limits, and choosing subscription age ranges (1-3 days and 4-7 days). For devices, I'll target mobile only, Android, all versions. I'm not touching browsers, carriers, feeds, or CPA optimization for now. With these settings, I created four campaigns for each GEO, with eight creatives in each campaign, based on the recommended bids. Here are some examples of push notifications:
The approval process took up to 30 minutes, and the campaigns were launched. I double-checked all the links, and the traffic started flowing smoothly. Since it's too early to optimize based on the initial clicks, I'm waiting for enough data to accumulate. The first conversions have started, indicating that everything is working fine. For the test, I set campaign limits to $20 everywhere. On the next day, with all the push impressions, a total of $102 was spent, and optimization opportunities are becoming apparent. I added a few placements and feeds to the blacklist, identified the most successful push combinations in terms of images and triggering phrases. However, the most significant conversion insights came from the connection method and specific countries:
It's evident that mobile traffic has almost twice the conversion rate, which is a common trend with push traffic. In terms of countries, two leaders stand out with negative ROIs of -17% and -19%: Indonesia and Thailand. Therefore, the main optimization points will be as follows:
I also wanted to take a look inside the smartlink and find out which offers in Indonesia and Thailand yielded the best results on my traffic. This information could provide insights for improving the creatives. Not all affiliate networks provide this information, for some reason, as if they themselves don't know which offers are running. However, the manager provided me with the current list, revealing that in Indonesia, it's a local supermarket gift card offer, while in Thailand, it's a sweepstakes offer for the latest iPhone. Now armed with this knowledge, I added more relevant creatives to the successful ones, incorporating logos and brand mentions to increase conversions.
- Disable traffic from stationary connection methods.
- Disable Vietnam and Malaysia.
- Disable underperforming ads and scale up successful ones.
I also wanted to take a look inside the smartlink and find out which offers in Indonesia and Thailand yielded the best results on my traffic. This information could provide insights for improving the creatives. Not all affiliate networks provide this information, for some reason, as if they themselves don't know which offers are running. However, the manager provided me with the current list, revealing that in Indonesia, it's a local supermarket gift card offer, while in Thailand, it's a sweepstakes offer for the latest iPhone. Now armed with this knowledge, I added more relevant creatives to the successful ones, incorporating logos and brand mentions to increase conversions.
With the new campaign settings, we immediately started seeing positive results. I deliberately didn't increase the bids to maintain a manageable pace and ensure timely monitoring of the blacklist and control over the process. The traffic volume decreased significantly as I disabled Wi-Fi connections in the campaigns, but the conversion rate (CR) increased almost twofold. Over the course of three days, the return on investment (ROI) with the new settings reached 50%, which is quite impressive for push campaigns.
The overall ROI of the campaign, considering the initial launch, turned out to be lower, at 22%.
Now there is a working combination that will bring in profit without taking up too much time. Since there is no hold in this affiliate network, I immediately withdrew the earnings and reinvested in the traffic source. From now on, it's a matter of monitoring the traffic, adding new placements to the blacklist, and periodically refreshing the creatives.