Google Signals is one of the most consequential settings in GA4, and most teams have it enabled without fully understanding what it does to their data. The short version: Google Signals enables cross-device reporting and demographic data by linking user sessions to Google accounts. The catch: it also triggers data thresholding, which hides rows in your reports whenever the combination of dimensions would allow identifying individual users. If your reports are full of "(other)" rows and your demographic data is missing, Google Signals is almost certainly the cause.
This guide covers what Google Signals is, what it unlocks, what it breaks, how to diagnose whether thresholding is affecting your reports, and when to disable it.
What Google Signals Actually Does
When Google Signals is enabled, GA4 uses the Google account data of users who have opted into "Ads personalization" in their Google account settings. For those users, GA4 can:
- Link their sessions across devices (phone, laptop, tablet) into a single user journey
- Include them in demographic reports (age, gender) based on their Google account profile
- Populate the "Google Signals" reporting identity in the User Explorer
- Make those users available for remarketing audiences in Google Ads
For users who have NOT opted into Ads personalization — which is a significant and growing proportion — Google Signals does nothing. GA4 falls back to device ID or user ID for those users. This means the cross-device benefit of Signals is partial at best, depending on your audience.
The tradeoff is data thresholding: to protect user privacy, GA4 applies automatic thresholds to any report that could reveal information about individuals. When Google Signals is enabled, the threshold kicks in more aggressively. Small user counts, rare dimension combinations, or detailed demographic breakdowns all get suppressed — removed from the report entirely and lumped into the ambiguous "(other)" bucket.
What Is Data Thresholding?
Data thresholding is GA4's mechanism for preventing reports from revealing information about identifiable individuals. When a row in your report would represent fewer users than GA4's minimum threshold (Google does not publish the exact number, but it is commonly cited as 30-50 users), GA4 removes that row and adds its data to the "(other)" row instead.
Thresholding is different from data sampling. Sampling processes a subset of events to estimate results quickly. Thresholding removes specific rows to protect privacy. You can tell them apart: sampled reports show a shield icon in Explorations and the sampling percentage. Thresholded reports show no warning — rows simply disappear into "(other)" silently.
The presence of an "(other)" row in a standard report is the most reliable signal that thresholding is active. If you filter your Acquisition report by a specific campaign and see "(other)" consuming a significant share of sessions, thresholding is suppressing the rows with small user counts — usually new or niche segments that fall below the threshold for that dimension combination.
How to Tell if Google Signals Is Causing Your "(other)" Problem
The most direct test: go to Admin → Property → Reporting Identity. You will see three options: Blended (the default when Signals is on), Observed, and Device-based. If it is set to Blended, Google Signals is contributing to your reporting identity and thresholding may be active.
To confirm thresholding is the cause of "(other)" rows (versus low traffic or a filter issue):
- Go to the Acquisition → Traffic Acquisition report
- Add a secondary dimension — try "Device category" or "Country"
- Look at the "(other)" row. Note the session count in it
- Remove the secondary dimension. Does "(other)" shrink or disappear?
If "(other)" gets much smaller without the secondary dimension, that confirms dimension combinations are falling below the threshold — which is the classic thresholding signature. The primary dimension alone has enough users per row, but the combination creates rows too small for GA4 to report.
You can also check Admin → Property → Reporting Identity → switch from "Blended" to "Observed" and compare your reports before and after. If the "(other)" rows shrink significantly in Observed mode, Google Signals' cross-device tracking was creating the small user counts that triggered thresholding.
When to Keep Google Signals On
Google Signals is worth keeping enabled if all three of the following are true:
- You run Google Ads remarketing audiences. Signals-based audiences include cross-device users who opted into Ads personalization. Without Signals, you are remarketing to device-level segments only.
- Your traffic is high enough that thresholding doesn't materially affect reports. If you have 100,000+ sessions per month with reasonably uniform audience composition, thresholding will suppress only a tiny fraction of rows.
- You have reviewed your Consent Mode v2 configuration. Google Signals only applies to users who have consented to analytics storage and who have opted into Ads personalization. If your Consent Mode setup is correct, the scope of Signals is already limited to consenting users.
If you are primarily focused on analytics accuracy rather than advertising, the case for keeping Signals on is weak. The cross-device data it provides is partial, the demographic data is unreliable for small properties, and the thresholding it causes actively degrades report accuracy.
When to Disable Google Signals
Disable Google Signals if any of the following apply:
- You are seeing significant "(other)" rows in standard reports. Thresholding that makes your acquisition or page reports unreliable is worse than losing cross-device data you may not have been using anyway.
- You do not run Google Ads remarketing. If there are no audiences, there is no benefit to compensate for the thresholding cost.
- You operate in a jurisdiction where cross-device tracking of consenting users raises compliance concerns. Signals relies on Google's own user identification — it is worth reviewing with counsel whether enabling it creates any obligations beyond what your standard analytics setup requires.
- Your reporting identity audit (Admin → Reporting Identity) shows demographic data is consistently empty. This means your users are not opted into Ads personalization in large enough numbers for Signals to contribute meaningful data anyway — all cost, no benefit.
To disable: Admin → Property → Reporting Identity → switch to "Observed" (which uses your own user ID or device ID, not the Signals layer). Or go to Admin → Data Collection → Google Signals data collection → toggle off. The Reporting Identity path is the more surgical option — it disables Signals for reporting without affecting data collection.
The Reporting Identity Hierarchy
GA4's Reporting Identity setting controls how users are identified across sessions in reports. The three options are:
- Blended: Uses User ID if present → Google Signals if available → device ID as fallback. This is the thresholding-prone option.
- Observed: Uses User ID if present → device ID as fallback. Google Signals is not used. Thresholding applies only when User ID creates small user counts, which is less common. This is the recommended setting for most analytics-first properties.
- Device-based: Uses device ID only. No User ID, no Signals. This matches the old Universal Analytics user definition — one device = one user. Thresholding is minimal. Used when you want maximum compatibility with UA-era benchmarks or when cross-device user counts would confuse stakeholders.
The Reporting Identity setting affects your reports and the User Explorer but does not affect data collection. Changing it does not delete any data — it only changes how existing data is aggregated for display.
Google Signals and User-Level Reporting
One of Google Signals' more useful features is the User Explorer report (Reports → User → User Explorer). When Signals is enabled and a user is signed into a Google account, the User Explorer can show their cross-device journey — sessions from their phone, laptop, and tablet stitched together under a single user identifier. This is useful for understanding high-value user journeys before a conversion.
Without Signals, User Explorer still works but shows device-level journeys only. A user who first visited on mobile and later converted on desktop appears as two separate users. Whether that limitation matters depends on how much of your conversion flow spans devices.
Important: Google Signals data in User Explorer is subject to the same thresholding rules as standard reports. Users in small segments may not appear at all. And User Explorer is sampled for large properties, compounding the accuracy limitations.
FAQ
Does disabling Google Signals affect my historical data?
No. Changing the Reporting Identity setting does not modify collected data. It changes how GA4 aggregates and displays data in reports going forward. Your historical event data is unchanged — GA4 will simply re-aggregate it using the new identity method the next time you view reports.
I disabled Google Signals but still see "(other)" rows. Why?
Thresholding can also occur without Signals in some situations — typically when a secondary dimension creates user counts below the threshold. Check Admin → Reporting Identity and confirm it is set to "Observed" or "Device-based." If thresholding persists, try removing secondary dimensions or broadening your date range to increase the per-row user counts above the threshold.
Can I have Google Signals for Ads but not for GA4 reports?
Yes, with some nuance. You can set Reporting Identity to "Observed" (which removes Signals from your report aggregation) while still having Google Signals data collection enabled for audience building. Signals will still power your remarketing audiences in Google Ads, but it will not contribute cross-device stitching to your GA4 reports, which reduces thresholding. Go to Admin → Reporting Identity → Observed, while leaving Admin → Data Collection → Google Signals → On.
Does Google Signals work if I have Consent Mode v2 enabled?
Yes, but only for users who consent to analytics_storage. Signals requires both: the user must have opted into Ads personalization in their Google account AND your site must have collected their data with consent granted. For users who decline your banner, GA4 receives only modeled data for them — Signals doesn't apply to modeled data. In practice, on sites with proper Consent Mode v2 and typical EU opt-in rates (often 50-70%), the pool of Signals-eligible users is smaller than the total consenting pool, since not all consenting users have Ads personalization enabled in their Google accounts.
Free GA4 audit
NiceLookingData's automated GA4 audit checks your Reporting Identity setting, diagnoses whether Google Signals is likely causing data thresholding, and surfaces the reporting gaps this creates — alongside 60 other checks across data accuracy, compliance, and measurement quality. Free to run.
Run the GA4 audit free →Analytics consultant turned founder. After years running the same GA4 and GTM audits across client engagements, Ludde built the audit into a product — so the pattern-matching takes a minute, not a meeting. More about Ludde →
Run a free GA4 audit.
Connect your Google Analytics 4 property. Our auditor runs 61 checks and gives you an instant health score with a plain-English action plan.
