Annotations

Annotations are notes attached to dates in your property’s timeline. They help the AI understand why metrics changed and provide context for better analysis.

Why Annotations Matter

When you ask “Why did traffic drop last week?”, the AI has two options:

Without annotations: Speculate based on patterns (“Perhaps there was a technical issue…”)

With annotations: Provide context (“According to your annotation, the site was down for maintenance on March 15, which explains the 4-hour traffic gap…”)

Annotations turn guesswork into informed analysis.

Types of Annotations

Manual Annotations

Annotations you create for your property:

Marketing events: Campaign launches, email sends, promotions
Site changes: Redesigns, new features, page updates
Technical events: Deployments, outages, migrations
Business events: Product launches, pricing changes
External factors: Holidays, industry events, PR coverage

Global Annotations

Annotations from feeds that affect all properties:

Algorithm updates: Google Core Updates, Spam Updates, etc.
Platform outages: Google Search incidents, GA4 issues
Industry events: Major announcements affecting many sites

Global annotations are provided by AI Data Stream and automatically available to all properties.

Creating Annotations

Adding a Manual Annotation

Go to your property’s Annotations tab
Click Add Annotation
Choose a Category (required - every team starts with a set of categories to pick from)
Under Date Type, choose Single Date or Range (for multi-day events) and set the date and time
Choose the Visibility: Shared or Private
Enter a Title and, optionally, a Description
Optional: enter an Affected Page URL (full URL, relative path or wildcard like /blog/*; leave blank for site-wide)
Click Create

Annotation Fields

Field	Description
Category	Type of annotation (required, used for filtering)
Date Type	Single Date or Range, with date and time
Visibility	Private (you only) or Shared (team)
Title	What happened (required, keep concise)
Description	More detail (optional)
Affected Page URL	One URL, relative path or wildcard pattern (optional)

Writing Good Notes

Too vague:

Site update

Better:

Launched new checkout flow - simplified from 4 steps to 2

Best:

Checkout redesign launched: 4 steps → 2, removed account requirement, added guest checkout

Include what changed and why it matters for analysis.

Annotation Categories

Categories help you organize and filter annotations. Each category has a name, color, and optional icon.

Built-in Categories

Global annotations use two built-in categories of their own:

Algorithm Update - Google Core Updates, Spam updates, ranking changes
System Outage - Platform downtime, API issues, hosting problems

These appear in the category filter under Global Categories. You can’t choose them for your own annotations.

Custom Categories

Every team starts with five categories, created in the team owner’s language:

Marketing Campaign
Site Change
Content Update
Product Launch
Bug Fix

Only the team owner can add, edit or delete categories. To add one:

Go to Team Settings → Annotation Categories (or Annotations → Manage Categories)
Click Add Category
Enter a name
Choose an icon and color
Click Create Category

Global Annotation Feeds

AI Data Stream maintains global feeds that automatically add annotations to your properties.

Available Feeds

Algorithm Updates

Source: Search Engine Land, official Google announcements
Includes: Core Updates, Spam Updates, Product Reviews, etc.
Updates: Checked daily for newly confirmed updates

Google Search Status

Source: Google Search Status Dashboard
Includes: Crawling issues, indexing problems, ranking issues
Updates: Checked every 6 hours

Admin Annotations

Source: Events added by the AI Data Stream team

Managing Global Feeds

Each property can enable or disable global feeds:

Go to your property’s Settings page
Find the Annotation Settings section
Toggle Show Global Annotations to show/hide global annotations
Enable or disable individual feeds under Annotation Feeds
Click Save Changes

You might disable algorithm update feeds for a property that doesn’t depend on organic search.

Historical Algorithm Updates

The algorithm feed includes updates dating back to 2018:

August 2025 Spam Update
June 2025 Core Update
March 2025 Core Update
December 2024 Core Update
And 50+ more historical updates

This means you can analyze historical data with algorithm context already in place.

Using Annotations in Analysis

Automatic Checking

All prompt presets instruct the AI to check annotations before analyzing time-based data. You don’t need to mention annotations—the AI does this automatically.

You ask:

Why did organic traffic drop in March?

AI checks:

Annotations for March
Global algorithm updates in March
Search Console data
GA4 data

AI responds:

Organic traffic dropped 23% in March. Looking at your annotations, I see the March 2025 Core Update rolled out March 5-12. Your traffic decline aligns with this update period…

Referencing Annotations in Questions

You can also explicitly reference annotations:

“What annotations do I have for last quarter?”
“Check if any algorithm updates affected my traffic in December”
“Factor in the site migration when analyzing Q1 performance”

Syncing Annotations to GA4

If you have Google Analytics 4 connected, you can sync individual annotations to GA4.

How It Works

When creating or editing an annotation, you’ll see an Also create in Google Analytics switch if GA4 is connected. Enable it to push that annotation to GA4’s annotation feature. This keeps your GA4 property in sync if your team also works directly in Google Analytics.

Limitations

GA4 annotations have a 150-character limit (if your description is longer, you’re asked for a GA4 Short Description)
Only dates are synced, not times
Requires GA4 connection with write access
Deletions don’t sync: If you delete an annotation in AI Data Stream that was synced to GA4, you’ll need to delete it manually in Google Analytics (the API doesn’t support deletion)

You can also pull annotations from GA4 into AI Data Stream.

Annotation Best Practices

Annotate Consistently

Create a habit:

Log marketing campaigns as you launch them
Note site changes during deployments
Record known issues when they occur

Don’t rely on memory—annotate in real-time.

Be Specific About Dates

Use exact dates when known
Use date ranges for rollouts (“Launched March 5, completed March 7”)
Note time zones if relevant

Include Affected Pages

For page-specific changes, note which pages:

“/blog/*” for all blog changes
“/products/widget” for specific product changes
Leave blank for site-wide changes

Keep Notes Scannable

The AI and your team will read these notes. Write clearly:

Lead with what changed
Add brief context if needed
Skip obvious details

Review Regularly

Periodically review annotations:

Archive old annotations that are no longer relevant
Update ongoing campaigns
Add any missed events

Managing Annotations

Viewing Annotations

Go to Annotations in your property to see:

Timeline view of all annotations
Filter by category
Filter by date range
Search notes

Editing Annotations

Find the annotation in the list
Click the edit icon
Make your changes in the form
Click Update

Archiving Annotations

Archiving removes annotations from AI context while preserving the record:

Find the annotation in the list
Click the delete/archive icon
Confirm

Archived annotations are excluded from AI analysis but can be restored later.

Restoring Archived Annotations

Go to Annotations
Click the Archived button (shown once you have archived annotations) to see archived annotations
Find the annotation to restore
Click Restore

The annotation returns to active status and will be included in AI analysis again. Click View Active to return to your active annotations.

Deleting Archived Annotations Permanently

Open the Archived view
Click the delete icon (Delete permanently) on the annotation
Confirm with Delete permanently

This removes the annotation for good; it can’t be restored. Only the annotation’s creator or someone with Admin permission on the property can do this. An annotation that was also created in Google Analytics stays in GA4 and has to be deleted there.