.png)
We're excited to share our latest success story: the growth of Paysend across the US and Europe. Their team was helpful and collaborative in setting up the right strategy to acquire new high-quality users, who would become valuable customers.
.png)
The Challenge
Paysend’s challenge was acquiring not just new customers, but users who’d send money via their app. These are high-quality users, who generate money and add value to their clients. Paysend's mission was a North America- and Western Europe-targeted user acquisition campaign for iOS.
To meet these objectives, we agreed to optimize and run the campaigns towards post-install activity. Focusing on post-install activities requires a huge amount of data, more than is needed for scaling cost per install (CPI) campaigns.
The solution
Together with the Paysend team and based on their data and experience with other acquisition channels, we tested a variety of model setups and from the results Paysend ultimately wanted to focus on the following post install activity KPI:
- Get more than 30% of newly registered users to complete an in-app money transfer
.png)
Our first step is to understand what is valuable for our client. To do so, we need to understand the user journey: this is key in all fintech products, where friction during onboarding can lead to massive loss of potential clients.
By tracking as many data points as possible, from the interaction with the ad to the most valuable in-app events, our algorithms are able to identify correlation and patterns useful to optimize the app campaigns.
Especially for fintech apps, we don’t stop at the first transaction or money transfer: tracking key subsequent in-app actions allows our A.I. to maximise user LTV. From a technical perspective, this means a data-rich environment, with many in-app events tracked and shared.
Down-the-funnel events such as registration or in-app purchase can be difficult to target. To do so, our A.I. generated hundreds of installs in the first phase buying ad placements within the multiple ad exchanges and SSPs we are integrated with.
Once this initial phase was completed our machine learning algorithms started the optimization process, which focused on three main steps:
- CPA optimization: following a learning period, our algorithms modeled a correlation between the account registration and money transfer events - it therefore optimized the app campaigns towards registrations;
- Contextual signals analysis: amongst the many contextual dimensions analyzed, operating system turned out to be the most relevant, with most recent versions driving better results;
- State Targeting: for our North America campaigns, our algorithms identified top performing locations in terms of installs and registrations. Our platform then automatically redistributed traffic volumes to these locations, which led to an exponential growth.
The Results
Targeting a post-install event earlier in the user funnel allowed our technology to scale both installs and registrations in the first weeks of the campaign.
During the first phase, aimed at generating installs and understanding the correlation between in-app events, we increased the app install rate by 200%. Then, once our A.I.-powered optimizations kicked in, we grew towards Paysend’s KPI and exceeded their registration-to-money transfer rate goal across all targeted geos.
Not only this, but in North America we also succeeded in generating an eCPI of 10% cheaper than that at the beginning of the campaign, as well as driving the registration rate up by 176% within this first month.
.png)
.png)
.png)
