Transitioning to Server-Side Data Collection: A Case Study with a Fortune 500 Grocery Retailer

In the ever-evolving landscape of digital marketing, privacy concerns and data control are at the forefront of many companies' strategies. This was especially true for a Fortune 500 grocery retailer we worked with, who was keen to move away from third-party data collection and take full control of their data streams. Our journey with them involved integrating Pinterest and shifting from client-side to server-side data collection. Here's how we did it and the results we achieved.

The Initial Setup

When we began the Pinterest integration for this retailer, they were already collecting server-side event data. The logical next step was to remove the client-side third-party mechanisms in favor of controlling the data stream. This approach not only allowed the retailer to manage how data was shipped but also minimized the number of assets loaded on each page, reducing both total page load times and potential security threats.

The Testing Process

Our testing process was straightforward but meticulous:

1. Development and Initial Testing: After some initial development, we deployed the solution to one of the retailer's smaller sites—a popular chain in the Denver area.
2. Parallel Integration: We set the server-side integration to run in parallel with the existing client-side integration. This allowed us to run tests and directly compare event volumes and match rates between both setups over the same periods.
3. Iterative Testing: We ran several tests, iterating until the results were satisfactory, aiming for parity or near parity with the client-side tags.

The server-side event stream included data from both the retailer's native phone applications and websites. For our testing, we filtered this event stream to focus solely on website traffic. We ensured the server-side events contained all the key information provided by the client-side tags.

Key Metrics and Results

Our primary metrics were:

• Event Volume Match: Measuring how many events could be tied to an end user on Pinterest.
• Conversion Attribution: Ensuring conversions were accurately tracked and reported for the retailer's partner connections.

We ran a two-day test over the last weekend in February, comparing event volumes and match rates by event name. The results were compelling:

1. Higher Event Volumes: The server-to-server integration consistently showed a much higher event volume—often more than double that of the client-side tag.
2. Increased Conversion Attribution: Conversion tracking was more accurate with the server-side setup, leading to more meaningful reports for partner connections.

Despite the higher raw event volumes, the match rates for server-side events were only slightly lower than those for client-side tags. This was significant because it indicated that the server-side tag was capturing more events from a broader set of users, not just the same users the client-side tag was tracking.

Conclusion

The transition to server-side data collection for this Fortune 500 grocery retailer proved highly effective. By taking control of their data streams and integrating directly with Pinterest, they saw significant improvements in event volume and conversion attribution. This approach not only enhanced their data accuracy and reporting but also aligned with a more privacy-focused digital environment.

For more details on how to stay ahead in the game and ensure your marketing efforts are effective and respectful of user privacy, watch our full video on the process and results:

Key Takeaways

• Server-side data collection provides more control and security.
• Higher event volumes and more accurate conversion tracking can be achieved.
• This method reduces page load times and potential security threats.

To learn more on how MetaRouter can help your organization, let's chat!