Help Needed

Risto

Risto

— 2 min read
Help Needed

Tackling the Challenge of Cross-Domain User IDs

Introduction

One of the challenges in user identification across multiple domains is the limitation imposed by local-only cookies and IDs. In Tracardi, this issue becomes apparent when a user transitions between domains, as we currently have limited options for passing their profile ID. To address this, I’ve embarked on a research project aimed at retaining the profile ID across domains using a probabilistic approach.

However, there’s a hurdle: I don’t have sufficient traffic to test edge cases thoroughly. This blog post outlines my approach and how you can help support this initiative.

The Research Approach

The solution I’m exploring involves collecting the following minimal data points:

  • Browser Fingerprint (app_id): A unique identifier for the browser.
  • Web Page ID (web_id): A randomly generated identifier for the page.
  • IP Address: To help identify user sessions across domains.
  • Browsing Time: The time a user spends on the page.
  • Visited Page: The current page URL.
  • Possible Out-links: Links the user might click to leave the domain.

By combining these data points, the goal is to develop a probabilistic algorithm that can reliably detect when a user transitions between domains.

Request for Support

To test this proof of concept effectively, We need access to a webpages with significant traffic, where a small JavaScript snippet could be placed. Here’s the script:

<script src="https://ck.tracardi.org/script/ck.min.js?v=3"
        id="crosskey"
        data-api-endpoint="https://ck.tracardi.org"
        data-allowed-domains="blog.tracardi.com,tracardi.app,github.io"></script>

How It Works

This script collects the following dataset, which is strictly limited to proof-of-concept testing and does not include any personal information:

const payload = {  
    app_id: appId,  
    web_id: webId,  
    domain: domainName,  
    url: currentUrl,  
    paths: outlinks,  
    ip: userIp  
};

Configuration Instructions

To configure the script for your use case, define the domains you’d like to track for cross-domain traffic. For example:

data-allowed-domains="blog.tracardi.com,tracardi.app"

This will enable the script to detect and analyze traffic between the specified domains.

Why Your Support Matters

By adding this script to high-traffic domains, we can gather more data to validate the algorithm’s ability to detect cross-domain traffic. The more diverse the data, the better we can refine the approach and ensure its reliability.

If you have a high-traffic webpage and are open to collaborating, I’d greatly appreciate your support in adding this script.

Closing Thoughts

Cross-domain user ID tracking is a complex challenge, but with collective effort and collaboration, we can make significant strides in solving it. If you’re interested in helping or have questions, feel free to reach out.

Your input and support are invaluable in making this research successful!

Read more