Thanks, Andrew — super helpful to hear how you’ve handled it. The part that still trips me up a bit is user identification across domains and brands. In our case, a single user might interact with multiple domains (e.g., during onboarding) and also belong to different brands, though they share the same top-level user ID. I’m trying to figure out how best to stitch that together in Mixpanel so the journey feels coherent in analysis — not split or misattributed. We agreed to run a POC where we send events to different mixpanel projects and keep users separate although some of them might have accounts in domain1.com and domain2.com 🤞
Guidance needed on cross-domain tracking and multi-brand setup Hi everyone, I’d appreciate your help with a setup challenge we’re facing. Our company operates several brands, each with its own website (e.g., alpha.com, bravo.io, etc.). Some features—like our website builder—are shared across multiple brands and embedded within their respective sites. We also have scenarios where a user’s journey within a single brand can temporarily cross into a different domain. Mixpanel identifies users via domain-based cookies and device IDs, which leads to two challenges: 1. Single-brand journey across multiple domains Here’s a simplified user flow: Page visit on alpha.com >Page visit on alpha.com/page2 >Page visit on alpha.com/page3 >Page visit on bravo.io/page2 All of this is a continuous journey from the user’s perspective within alpha.com and the user is identified under alpha.com, but because of the domain switch, I assume Mixpanel treats it as a separate user or session. What would be the recommended way to track this as a unified journey? 2. Cross-brand feature analysis We have shared products, like our website builder, that are embedded into multiple brand domains. Ideally, the website builder team would like to analyze usage across all brands in one Mixpanel project. We can assign a global user ID across brands, but we’re unsure whether this would lead to identity issues and multiple other. The clean approach would be keeping them separated and analysing each feature in the context of the website they're embedded, but I can understand how usefule it would be to analyse a let's say a funnel conversion for the feature and then break it down by a different domain to identify differences Should we send all events into a single project or keep them separated per brand? What setup would you recommend for a clean, scalable solution? Is there a concept of Portfolio (like ion Amplitude) ? Thanks in advance for your guidance! Best, Luka
Question about merging anonymous events across browsers after identify() Hi all, I just ran a test with the Mixpanel JavaScript SDK and encountered a behavior I didn’t fully expect. Here’s what I did: 1. In Google Chrome, I opened the site and Mixpanel created an anonymous profile with a random device_id / distinct_id. 2. The user did 4–5 page views anonymously. 3. Then I called identify() with a known user ID (e.g., user_123), and continued browsing. 4. Mixpanel correctly merged the anonymous and identified events into one profile ✅ So far, all good. Then I tried this in Opera (completely different browser): 1. Again, Mixpanel created an anonymous user when I landed on the site. 2. I did a few anonymous events (e.g., page views) - PROBLEM 3. Then I logged in and called identify("user_123") Now the issue: The anonymous events in Opera stayed under a separate profile and were not merged. Expected: Every time an anonymous user browse around in any browser/device as soon as they are identified all those anonymous events should be assigned to their profile Question: Why doesn’t Mixpanel merge the anonymous events from the second browser session to the existing profile once identify() is called with the same user ID? Is this expected behavior? Thanks in advance!
