kapa.ai

There are a few documented reasons why $referring_domain (or $initial_referring_domain) may increasingly appear as (not set) in your reports:

$referring_domain is only collected when document.referrer returns a value in the browser. It will be omitted if:

In those cases, the property is not sent at all, which Mixpanel displays as (not set). [Default Properties]

Over time, more browsers and security settings prevent referrer data from being passed. This is a known trend — browsers increasingly strip or block referrer headers, which means more events arrive without a referring domain. This would explain a year-over-year increase in (not set) values. [Traffic Attribution]

If any portion of your tracking has shifted to server-side over time, $referring_domain is not automatically collected — it must be parsed manually from the Referer request header. If this isn't done, the property will be missing. [Server-Side Best Practices]

If your share of direct or organic traffic has grown YoY, you'd naturally see more (not set) for UTM and referrer properties, since those users arrive without a referrer. [Traffic Attribution]

To mitigate this, for server-side implementations you can manually extract and forward the referrer:

from urllib.parse import urlparse

if "Referer" in request.headers:
    properties.update({
        "$referrer": request.headers["Referer"],
        "$referring_domain": urlparse(request.headers["Referer"]).hostname
    })

[Server-Side Attribution]

For client-side tracking, this is handled automatically by the JavaScript SDK, but you cannot recover referrer data that browsers choose not to send.

Since it works in Chrome Incognito mode (where extensions are disabled by default), this strongly points to a Chrome extension as the culprit — even though you've already disabled ad blockers.

Here are some steps to identify the specific extension causing the issue:

The knowledge sources don't go further into specific extension-by-extension troubleshooting for the Mixpanel UI itself, so if you're unable to identify the offending extension, I'd still recommend contacting Mixpanel Support with this new information (that Incognito works) — it will help them narrow down the cause much faster.

Thank you for ruling those out. Unfortunately, the knowledge sources I have access to don't contain specific information about the "Something went wrong" error in the Mixpanel UI itself (as opposed to tracking/data ingestion issues), and they don't cover Chrome-specific browser compatibility troubleshooting beyond ad blockers and CORS.

Since the issue is isolated to Chrome and the common causes have been eliminated, I'd recommend:

The knowledge sources I have don't provide enough information to give you a more specific answer for this particular UI error in Chrome. Mixpanel Support would be the best next step here.

The knowledge sources don't contain specific information about the "Something went wrong" error message you're seeing in Chrome. However, based on what's documented, here are some Chrome-specific troubleshooting steps worth trying:

Since the issue is specific to Chrome and not Edge, a browser extension or Chrome-specific setting is the most likely cause. If none of these steps resolve the issue, I'd recommend reaching out to Mixpanel Support directly, as this error ("Something went wrong... we have been notified") suggests a client-side issue that their team may be able to investigate further with your account details.

Based on the knowledge sources available, Mixpanel does provide real-time analysis. Specifically, it is mentioned that Mixpanel provides "real-time, unsampled analysis." [Mixpanel vs GA4]

Additionally, when events are tracked via the SDKs or the /track API endpoint, they are ingested in real-time into your project. [Track Events]

However, there are some nuances worth noting from the knowledge sources:

For a direct comparison, GA4 has a minimum wait time of 4–8 hours to process data, whereas Mixpanel does not have this delay. [Mixpanel vs GA4]

Here are the rate limits for Mixpanel's Query API:

If you exceed these limits, you'll receive a 429 response code. [Rate Limits]

To reduce rate-limiting errors, Mixpanel recommends:

Note: If you're using the Mixpanel MCP Server, the rate limit is different — a maximum of 600 MCP requests per hour per user. [Mixpanel MCP Server]

Yes, your organization plan is likely the reason. The Attribution feature (including the "Attributed by..." breakdown) is only available to users on an Enterprise or Growth plan. [Attribution feature]

If you're on the Free plan, you won't have access to this computed property.

Here's what you can do:

If you are already on a Growth or Enterprise plan and still can't see the "Attributed by..." option, here are a few other things to check:

Sorry Novita O., I don't have permission to access files in this workspace. Please ask a workspace admin to reinstall the kapa.ai bot to grant the required permission.

This is a nuanced situation. Let me address both parts of your question:

When you break down a funnel by cohorts, Mixpanel counts users in each cohort segment they qualify for. This means a user who belongs to both cohorts will appear in both segments — there is no built-in mechanism to force a user to be counted in only one cohort when breaking down by cohorts. [Funnels Breakdowns]

The knowledge sources do not explicitly describe a way to deduplicate users across overlapping cohort segments in a funnel breakdown.

For attribution, you can use the "Attributed by…" breakdown in Funnels, which attributes credit for the funnel to touchpoints along the user journey using the first step of the funnel. You can select First Touch as your attribution model, which gives 100% credit to the first touchpoint within the attribution lookback window. [Attributed by Breakdown; Attribution Models]

To set this up:

⚠️ Note: Attribution is available on Enterprise or Growth plans only. [Attribution]

If your goal is to attribute users to a single channel/touchpoint, the "Attributed by…" breakdown with First Touch is the right tool. If you need mutually exclusive cohort segments, you would need to redefine your cohorts so they don't overlap before applying the breakdown.

Yes, Mixpanel has several character and size limits on properties (attributes):