Google Analytics Server Events
Report confirmed outcomes to Google Analytics when browser tracking cannot be trusted.
What it adds
Server-side delivery of completed conversions to Google Analytics, joined to the browser session and deduplicated against it.
What your agent is told to do
5
What your agent is told to do
5-
1
Identify the outcomes the browser cannot see or reports unreliably — a payment confirmed by a webhook, a subscription that renewed, a signup completed after a redirect, anything blocked by an ad blocker — and send those from the server.
-
2
Capture the analytics client identifier from the browser when the visit begins, store it against the session or the record, and send it with the server event so the conversion joins the same session and campaign rather than appearing as a new anonymous visit.
-
3
Generate a stable event identifier at the moment the outcome happens and use it for both the browser and the server send, so the provider can collapse the pair into one event. Assign each event type a primary origin and treat the other as the fallback.
-
4
Keep the measurement credential in server configuration only, never in a browser bundle, and never send an email address, a name, a payment detail, a health or financial attribute, or anything else the provider's policies prohibit. Send an opaque internal identifier if a user must be identified.
-
5
Queue every send through the app's existing background processing so the user's transaction commits and returns without waiting on the provider. This entry owns server-side delivery only; if the app has a separate brief for browser tracking, that one owns page views and in-page interactions.
Edge cases it handles
8
Edge cases it handles
8- Without the browser's client identifier the server event lands as a brand-new session with no campaign attribution, which makes the conversion useless for measuring acquisition. Capture and store it early, and record what happens when it is absent.
- The same conversion sent from both the browser and the server double-counts unless both carry the same stable event identifier. Generate that identifier once, at the outcome, not independently on each side.
- Session and campaign context is only legitimately available for a visit the app actually observed. Do not fabricate a source, and do not reuse a stale identifier from a previous visitor on a shared device.
- Sending prohibited values — email addresses, names, payment details, or sensitive category attributes — breaches the provider's terms and can have the property's data deleted. Enforce an allowed parameter list rather than passing the record through.
- Server sends must not sit on the request path. Queue them, so a slow or unreachable provider cannot delay a checkout or leave a transaction half-finished.
- Events accepted at the endpoint can still be discarded downstream for a malformed or unknown parameter, often with no error returned. Validate against the configured event and parameter schema before sending, and verify in the provider's own debug view.
- Events arriving long after the fact may fall outside the provider's accepted time window and be dropped. Send promptly, and set expectations for anything replayed from a backlog.
- Rate limits and outages need backoff with jitter and a bounded queue. Failing telemetry must never fail or retry the underlying business transaction.
Definition of done
8
Definition of done
8- Outcomes the browser cannot reliably observe are reported from the server.
- Server events carry the stored client identifier and join the originating session and campaign where one was observed.
- Browser and server sends of the same conversion share one stable event identifier and are counted once.
- No prohibited or personally identifying value is sent, enforced by an allowed parameter list.
- Sends are queued off the request path and never delay or fail the user's transaction.
- Credentials exist only in server configuration, and outages result in backoff and a bounded queue rather than lost transactions.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
Data Table Cell Primitives
Data Table Cell Primitives
Render every kind of table cell the same way, wherever the table happens to be.
What it does
A shared set of cell renderers for numbers, dates, people, statuses, links, and row actions, used by every table in the app.
How it works
- 1 Catalogue the cell types the app's tables already render and reduce them to a small set: number and currency, date and relative time, person or avatar, status or badge, link or identifier, and an actions cell.
- 2 Separate each cell's underlying value from its presentation, so sorting, filtering, grouping, and export operate on the raw value while the user sees the formatted one.
- 3 Align and format numerically consistent cells the same way everywhere — figures right-aligned with tabular figures and a fixed precision per column, dates in the user's locale and time zone, currency with its code where more than one is possible.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/data-table-cell-primitives
Segment Event Forwarding
Segment Event Forwarding
Send one clean stream of product events to Segment and let it fan out downstream.
What it does
A documented event schema and a single forwarding path emitting track, identify, group, and page calls with consistent identity.
How it works
- 1 Write the event schema before writing any code: the exact set of event names, the properties each carries with their types, and the traits attached to a user and to an account. Every downstream tool inherits this schema, so a name chosen carelessly is expensive to change later.
- 2 Establish one identity model and apply it everywhere. Use a stable internal user identifier that never changes, associate the anonymous identifier from the first visit with it at signup, and attach the account or workspace so downstream tools can roll events up by customer.
- 3 Route every event through one internal emitter rather than calling the provider from feature code. That emitter validates the event against the schema, drops or flags anything unrecognised, and is the single place where identity and default properties are attached.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/segment-event-forwarding
UI Event Naming Schema
UI Event Naming Schema
Give every tracked interaction a name that still means the same thing in a year.
What it does
A documented vocabulary and structure for event names and properties, applied across the app's existing tracking.
How it works
- 1 Audit the events the app already sends and list them. The duplicates, the near-synonyms, and the ones nobody can explain are the reason the schema is needed, and they are also the migration list.
- 2 Fix a structure of object, action, and context: what was acted on, what happened to it, and where. Fix a closed set of actions — viewed, clicked, opened, submitted, selected, failed — and require every new event to reuse one of them.
- 3 Keep variable detail in properties, never in the name. A record identifier, a plan name, or a workspace slug baked into an event name produces thousands of unaggregatable one-off events.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/ui-event-naming-schema
How it works
-
1
Copy the link
Grab the Markdown instruction URL for this feature.
-
2
Give it to your AI
Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
-
3
It inspects, then implements
Your agent reads your existing app first, then adds the feature to fit it.
Works with your stack
These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.
Need it tighter than that? Customize the feature and tell it exactly what you're running.