Referral Tracking and Program Management

We started by collaborating closely with the client to fully understand their vision and requirements. This involved detailed discussions and reviews of the provided design.

Our team conducted a thorough analysis of the required functionalities, focusing on the integration of the tracking snippet and the multi-level referral system.

We developed a lightweight tracking snippet intended for embedding in third-party websites. The snippet is loaded asynchronously so it never blocks the host page's render pipeline. On load, it reads the referral token from the URL query string, writes it to a first-party cookie scoped to the host domain, and listens for a registration completion event. When the event fires, it sends a POST to the RevEngine collection endpoint via the Fetch API with a no-cors mode flag, keeping the request compatible with strict CSP headers on the host site. The payload includes the referral token, the host domain, the registration timestamp, and the user's email hash - no plaintext PII leaves the snippet without explicit host-side consent configuration. The final minified snippet weighed in under 4 KB, well within the threshold that would trigger performance objections from site owners.

The commission hierarchy was modeled in PostgreSQL as a self-referencing adjacency structure: each user record carries a referred_by foreign key pointing to the user who invited them. Commission calculations walk the tree upward from the converting user, applying the percentage configured for each program level. To prevent infinite loops from data corruption (a user whose referral chain accidentally points back to themselves), the commission calculation function includes a maximum depth guard and a cycle-detection pass that checks the full ancestor chain before writing any commission record. Programs support up to three levels - primary, secondary, and tertiary - with independent commission rates per level, all configurable by the admin without a code change.

The tracking snippet required a dedicated QA matrix that went beyond standard browser testing. Our QA specialist tested the snippet across sites built on WordPress, Webflow, Squarespace, and raw HTML/CSS stacks, covering both the happy path (snippet loads, referral token present, registration fires correctly) and the failure modes (snippet loads but token missing, registration event fires before snippet is ready, CSP blocks the outbound POST). For each failure mode, the snippet degrades gracefully rather than throwing an uncaught exception that would surface in the host site's console.

Based on testing feedback and client reviews, we continuously refined the platform to enhance its reliability and performance.

The platform was successfully launched within the agreed timeline, fully functional and aligned with the client’s vision.

We provided post-launch support to address any issues and ensure the platform’s smooth operation.

The platform was delivered on time, fully functional and meeting all of the client’s requirements. The tracking snippet was validated across all target site stacks before launch, and the first live referral conversions were confirmed within the first week of operation.

Before RevEngine, referral partners had to email the client’s team to get reports on their clicks, conversions, and commission balances. After launch, every partner had a self-serve dashboard showing their referral link performance, conversion history, and pending payout status in real time. The volume of inbound reporting requests to the client’s support team dropped to near zero after the first billing cycle.

The combination of server-side token validation and cookie-based session bridging produced referral attribution that held up across the multi-step registration flows on the client’s partner sites, including flows where users returned to complete registration 24 to 48 hours after the initial click.

The platform’s PostgreSQL-backed architecture and Azure hosting allow the client to onboard additional partner sites and referral programs without changes to the core codebase. New programs are configured through the admin UI, and the tracking snippet is the same file regardless of which program a site is participating in - the program association is determined server-side from the referral token at collection time.

As RevEngine continues to grow, plans are in place to integrate more advanced analytics features, providing deeper insights into referral performance.

The platform is designed to scale efficiently, supporting increasing numbers of users and referral programs without compromising performance.