Client Background:

Our client is a performance marketing company that specializes in assisting clients in achieving impressions and click goals. Acting as a third-party campaign manager, their primary focus lies in driving conversions in the app marketplace to boost app installs. However, the challenge they face is a low conversion rate, largely due to fraudulent clicks and bot-generated traffic, hindering their ability to meet client goals and risking throttling or shutdowns from marketplaces.

Objective:

This company's objective is to enhance their click-to-conversion ratio (CVR) through the implementation of machine learning. This initiative is part of a 12-week program, including a discovery phase and an eight-week model accelerator proof-of-concept (POC). The goal is to improve CVR by at least 1.8 times from the current baseline.

Approach:

The project is divided into two phases: Discovery and Model Accelerator. Early results from the model accelerator phase show promising outcomes, with a rudimentary model achieving a 3X lift on CVR while maintaining 90% of click traffic. This sets the stage for a comprehensive solution to significantly improve This companies''s click-to-conversion ratios.

Preliminary Results:

The early success of the model accelerator phase is indicative of the potential for substantial improvements in click-to-conversion ratios. By achieving a 3X lift on CVR while retaining a high percentage of click traffic, the team is confident in their ability to positively impact the client's campaign performance.

Return on Investment (ROI) and Future State:

The successful implementation of the machine learning solution not only addresses the immediate challenge but also positions the client for a tangible return on investment. In this case, this company realized ROI in three different ways:  

1. Efficiency. More accurately detecting which clicks are likely to convert means they can send less click traffic overall. This means they can improve their margins, either by selling the same service at the same price while reducing their costs or by increasing their prices while maintaining/decreasing their costs
2. Scale. If/when this system is more fully automated, that means that campaign managers would be able to manage more campaigns, enabling them to scale more effectively.
3. Valuation. Given that they’re PE-backed, having a successful AI solution inside their product suite driving a core aspect of their business, they are able to position themselves competitively and drive a much higher valuation multiple.

Implementation and Scalability:

Over the past 6 years of developing & delivering custom AI strategies and solutions for our clients, KUNGFU.AI has perfected a deployment methodology which ensures scalability and flexibility. Leveraging the concept of microservices, KUNGFU.AI deploys AI capabilities via containerized Docker images and APIs such as REST. This methodology is a lightweight framework which allows us to be platform agnostic and our capabilities to be easily scaled.

Challenges and Strategic Considerations:

Balancing the total number of conversions with the conversion ratio is a key challenge. The strategy involves educating the client on the nuanced use of the model to make informed decisions, optimizing performance without risking over-filtering traffic.

Conclusion:

The collaboration between the AI team and this client is poised not only to address performance marketing challenges but also to drive substantial business growth. With the potential for improved click-to-conversion ratios and a clear return on investment, the client is positioned to offer enhanced services to clients, showcasing the impact of leveraging machine learning in the performance marketing landscape.

‍