The Order ID is a unique identifier for a transaction (like a purchase or payment). Cometly uses it to prevent duplicate events and ensure accurate revenue tracking.
Why This Matters
If you send duplicate events without proper deduplication, your revenue and conversions will be inflated.
Using Order ID correctly ensures:
Clean reporting
Accurate attribution
No duplicate purchases or payments
What Is Order ID?
The Order ID is the unique ID assigned to a transaction by your platform.
Examples:
Shopify → Order ID
Stripe → Charge ID
CRM / Checkout → Internal order or transaction ID
When you send events with an Order ID, Cometly uses it to identify whether the event is new or a duplicate.
How Deduplication Works
When an event is received, Cometly checks:
“Have I already seen this event name + Order ID for this customer?”
If YES:
The event is treated as a duplicate
It is ignored completely
No updates are made
If NO:
The event is recorded normally
Important Behavior
First-in wins
The first event received is the one that counts
Later duplicates are dropped
No updates happen
Amounts, fields, or properties are NOT updated
The original record stays unchanged
Shopify Exception (Upsells)
There is one exception for the direct Shopify integration, (this does not apply if you are using Webhooks in your Shopify store) this applies only to the direct Shopify integration:
If a duplicate purchase event comes in with a higher amount
Cometly will update the original event amount
This handles post-purchase upsells automatically. No extra setup is required.
Deduplication Priority Order
If multiple identifiers are present, Cometly uses this priority:
1. Idempotency Key (Highest Priority)
Same key → duplicate
Different key → treated as separate events (even with same Order ID)
Use this when you need full control
2. Order ID
Same Order ID + Event Name → duplicate
This is the default and most common method
3. Order Name
Used as a fallback (e.g.
#1042)Only applies if Order ID is not present
4. Time Window (Lowest Priority)
Used only if no identifiers exist:
Purchases → 1-day window
Other events → 1-week window
⚠️ Important:
Deduplication Only Works Within the Same Event Type. Order ID does NOT deduplicate across different event names.
Example
If you send:
Purchase→ Order ID: ABC123ClosedWon→ Order ID: ABC123Upsell→ Order ID: ABC123
👉 These are treated as 3 separate conversions
Even though the Order ID is the same.
What This Means
If your integration sends multiple event types for the same transaction:
Revenue will be counted multiple times
Conversions will be inflated
Common Use Cases
Use Case 1: Simple Purchase Tracking
Send one Purchase event per transaction
Include Order ID
✅ Result:
Duplicate webhooks are ignored
Clean and accurate revenue
Use Case 2: Track Total Order Value (Booked Revenue)
Send
Purchasewith the full order amount
Best for:
Ecommerce
Tracking total order value at checkout
⚠️ Do NOT also send:
ClosedWonUpsell
Unless you want separate metrics
Use Case 3: Track Revenue from Deals (Sales Teams)
Send
ClosedWonwhen a deal is finalized
Best for:
B2B / Sales-led companies
CRM-based pipelines
⚠️ Do NOT also send:
Purchase
Unless you intentionally want:
Sales pipeline revenue AND purchase revenue tracked separately
Use Case 4: Track Additional Revenue (Upsells)
Send
Upsellevents when additional revenue is added after the initial purchase
Best for:
Post-purchase offers
Add-ons or upgrades
⚠️ Important:
This is a separate event type, not a duplicate
It will count as additional revenue
Use Case 5: Force Separate Events (Advanced)
If you need multiple events with the same Order ID to be treated as separate:
Add different Idempotency Keys
✅ Result:
Events are treated as separate transactions
💡 Pro Tip
Pick one primary event type per business metric:
Want ecommerce revenue → use
PurchaseWant sales pipeline revenue → use
ClosedWonWant add-ons → use
Upsell
Do not mix event types for the same transaction unless you intentionally want multiple conversion records.
The Key Takeaway
Order ID prevents duplicates within the same event type
First event wins, duplicates are ignored
Different event names = separate conversions
Choose your event strategy carefully to avoid inflated data